> the pedestrian suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path. Our technology immediately detected the individual as soon as they began to emerge from behind the stopped vehicle. The Waymo Driver braked hard, reducing speed from approximately 17 mph to under 6 mph before contact was made.
> Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene.
> Following the event, we voluntarily contacted the National Highway Traffic Safety Administration (NHTSA) that same day.
I honestly cannot imagine a better outcome or handling of the situation.
> Waymo said in its blog post that its “peer-reviewed model” shows a “fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph.”
It's likely that a fully-attentive human driver would have done worse. With a distracted driver (a huge portion of human drivers) it could've been catastrophic.
You're omitting the context provided by the article. This wasn't just a random scenario. Not only was this by an elementary school, but during school drop off hours, with both children and doubled parked cars in the vicinity. If somebody doesn't know what double parking is - it's when cars parallel park beside one another, implicitly on the road, making it difficult to see what's beyond them.
So you are around young children with visibility significantly impaired because of double parking. I'd love to see video of the incident because driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do, because a kid popping out from behind one of those cars is not only unsurprising but completely expected.
Another reason you also slow way down in this scenario is one of those cars suddenly swinging open their door which, again, would not be particularly surprising in this sort of context.
That's my thinking as well. Taken in some abstract scenario, all those steps seems very reasonable, and in that abstract scenario we can even say it would do better than an average human would. But that is missing the overall context that this was an elementary school during drop-off hours. That's when you crawl at 3 mph expecting kids to jump behind any car, and not going at 17mph.
> But that is missing the overall context that this was an elementary school during drop-off hours. That's when you crawl at 3 mph expecting kids to jump behind any car, and not going at 17mph.
Indeed. Sure the car knows the limit, it knows it is a school zone, it can precisely track people within the reach of its sensors (but not behind blockages it can't see through).
But it is missing the human understanding of the situation. Does it know that tiny humans behave far more erratically then the big ones? Obvious to us humans, but does the car take that into account? Does it consider that in such a situation, it is likely that a kid that its sensors can't possibly detect has a high probability to suddenly dart out from behind an obstacle? Again obvious to us humans because we understand kids, but does the car know?
Urm, ime people frequently drive significantly over the speed limit in all these places, at all times of the day.
Blows my mind how you guys confidently state this with authority as if that's the normal behavior, when the reality is that it probably should be - but isn't actually.
So you're confidently stating it's not the normal behaviour... Can you tell me what the average speed is for human drivers outside elementary schools at drop-off times?
But if you’re plan on building a fleet of cars operating all over the country or the world, do you want to model them after the careful driver, who has awareness about the situation (school, drop off/pickup hours, etc) or say “what the heck, some drivers are not paying attention so neither will my robots, it’s fine”
"I'd love to see video of the incident because driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do"
I am not sure what your definition of typical is. The reason we have lower speed zones in school districts at specific times is because humans typically drive fast. The reason police officers frequently target these areas and write a plethora of tickets is because humans typically ignore speed limits.
Your claim seems to be that a human would drive much slower than the posted speed limit considering the conditions, but the laws and the court room suggest otherwise.
There is probably a multimodal distribution of behaviors for people in these situations. Any robotaxi ought to behave in the more cautious mode no? I drive at near idle speeds in these situations.
While I completely agree with your premise, the software can be improved. It can be programmed to drop down to 5 miles an hour or less depending on street size, pedestrian proximity, school zone, etc.
If only the same could be said for the other parents in the school zone. I’ve seen people roar by in similar scenarios at 30+ miles an hour.
Driving is based so much off of feel so my numbers may be off, but in the scenario you are talking about 5mph seems reasonable, 10mph already seems like to much.
The want to be E but really armchair engineer in me for this context says there's far too little Engineering safety of the situation.
That school should not be on a busy roadway at all, it should also not have a child dropoff area anywhere near one but instead, ideally, a slow loop where the parents do drop off children, and then proceed forward in a safe direction away from the school in a flow.
It's funny because now you're sounding like you're blaming the school/the city for the situation.
Things are what they are. Driving situations are never perfect and that's why we adapt. The Waymo was speeding in a school zone. Did a dangerously fast overtake of a double parked car. It's engineering safety failure over engineering safety failure from Waymo's part, on nobody else.
Source? The article doesn't list a speed limit, but highways.dot.gov suggests to me that the speed limit would be 25mph in the school zone, in which case the waymo was going significantly under the speed limit.
If you drive in Sweden you will occasionally come up to a form of speed reduction strategy that may seem counterintuitive. They all add to make driving harder and feel more dangerous in order to force attention and lower speed.
One is to merge opposite directional roads into a single lane, forcing drivers to cooperate and take turn to pass it, one car at a time.
For a combined car and pedestrian road (max speed of 7km/h) near where I live, they intentionally added large obfuscating objects on the road that limited visibility and harder to navigate. This forces drivers to drive very slow, even when alone on the road, as they can't see if a car or person may be behind the next object.
In an other road they added several tight S curves in a row, where if you drive anything faster than 20km/h you will fail the turns and drive onto the artificial constructed curbs.
In other roads they put a sign in the middle of two way roads while at the same time drastically limiting the width to the curb, forcing drivers to slow down in order to center the car in the lane and squeeze through.
In each of those is that a human driver with human fear of crashing will cause drivers to pay extra attention and slow down.
In Bulgaria we have a similar speed reduction strategy but we are a bit ahead of Sweden:
We use medium-radius but very deep potholes. If you lose attention for even a split second, you are forced to a full stop to change a tire.
Near schools it gets more "advanced": they put parked cars on both sides of the road, and the holes positioned so you can't bypass them. For example, two tire-sized holes on both sides of the road right next to the parked cars. You have to come to a complete stop, then slowly descend into the hole with the front wheels, climb back out, and repeat the process for the rear wheels. Occasionally, even though we (technically) have sidewalks, they are covered in mud or grass or bushes, so pedestrians are forced to walk in the middle of the road. This further reduces driving speed to walking pace and increases safety in our cities. Road markings are missing almost everywhere and they put contradicting road signs so drivers are not only forced to cooperate but also to read each other minds.
That’s genius but one has to ask: how much does it cost to maintain these speed restricting features?
In the UK, the cost of owning a car is high yet our potholes, while frequent, are small enough to survive. Thus being more of an annoyance rather than a speed restriction.
It's fairly common at least in the Netherlands, Germany, and Switzerland too. In Switzerland they also place street parking spots on alternating sides on narrow streets, which also makes you more attentive and lower your speed.
I've heard that that is why roundabouts are safer than their alternatives: counterintuitively, they're safer because they're less safe, forcing the user to pay more attention as a result.
Roundabouts are safer. They're safer because they prevent everybody from speeding through the intersection. And, even in case of an accident, no head-on collisions happen in a roundabout.
They're safer specifically for vehicles, as they convert many conflicts that would be t-bones (worst for passengers) into getting rear-ended (maximum crumple zone on both vehicles).
Roundabouts are worse for land use though, which impacts walkability, and the safety story for pedestrians and bike users with them is decidedly not great as well.
>Roundabouts are worse for land use though, which impacts walkability, and the safety story for pedestrians and bike users with them is decidedly not great as well.
They're much safer for pedestrians than intersections. You're only crossing and dealing with traffic coming from one direction, stopping at a median, and then crossing further over.
Unlike trying to navigate a crosswalk where you have to play guessing games as to which direction some vehicle is going to come at you from while ignoring the lights (people do the stupidest things, and roundabouts are a physical barrier that prevents a bunch of that)
In Waterloo Region I used to cycle through multiple intersections that were "upgraded" some years ago from conventional stoplights to roundabouts and imo it was a huge downgrade to my sense of safety. I went from having a clear right of way (hand signal, cross in the crosswalk) to feeling completely invisible to cars, essentially dashing across the road in the gaps in traffic as if I was jaywalking.
I could handle it as an adult just walking my bike but it would be a nightmare for someone pushing a stroller or dependent on a mobility device.
Especially bad when crossings are like 30cm from the roundabout. Some are better with at least one car's length between the two.
Otherwise you either risk getting run over by a car exiting the roundabout without seeing you; or getting run over by the car that stopped, but was rear-ended by another inside the roundabout.
> and the safety story for pedestrians and bike users with them is decidedly not great as well.
The what now? Seriously, what in the world are you talking about? Roundabouts are heaven. They physically force drivers to slow down when approaching or leaving them, creating a safe environment for pedestrians and cyclists.
For example, there's no such thing as "running a red light at full speed" at a roundabout, no speeding up to "make the light", etc.
For cyclists specifically, they're amazing, because they eliminate the deadly left-turns. Every turn is a right turn, which is super safe.
As a cyclist, I'm not a big fan of roundabouts, because I'm always worried I'll get hit on the side by a car entering/leaving the roundabout whenever I don't take the first exit, mostly because I feel like I have less visibility in the direction from which the car might come from, compared to a standard crossing.
Though I've never been in an accident either on a crossing or roundabout, so I can't really judge how true my impression is.
One-lane-roundabouts are very safe. I lived in Hannover (Germany) in the 80s and 90s, they had 2 or 3 lanes in the roundabouts. There were large signs that counted the accidents (200+/year) to raise awareness and during the trade fairs (anybody remembers Cebit?) the number of accidents peaked. Today they are all a lot safer because of a lot of traffic lights.
I thought that the idea of roundabouts was that they lead to slightly more accidents than before, but they are of much lower severity than before (the 90 degree intersections they replace).
I would say it depend on where you are. City driving is generally not a great experience and its not that uncommon to see a speed bump before almost every crossing, to the point where you get surprised if there isn't one. That said, as long you don't leave the designated main roads that goes through the city areas it is not that bad. They do demand a lot of attention.
Street parking has mostly been turned into exclusive residential parking, so parking houses are often the only choice. As a result they are quite expensive, and you got to walk to the destination.
Parking and access is much better in the country side, and the highways are fairly good and similar to those found in the west Europe. It not as straight or wide as authobahn, but not as much traffic either.
why not just put in speedbumps if all you're trying to do is slow people down? Are you sure this was the purpose of these designs? sounds a little too freakonomics to me.
They also are rather expensive to maintain, because the leading ledge gets many repeated stresses. And in nordic climates like Sweden there is a snow plow in the winter to remove snow - those occasionally snag on the speed bump - which tends to chip of big chunk, triggering rapid wear.
Many roads in London have parked cars on either side so only one can get through - instead of people cooperating you have people fighting, speeding as fast as they can to get through before someone else appears, or race on-coming cars to a gap in the parked cars etc. So when they should be doing 30mph, they are more likely doing 40-45. Especially with EVs you have near-instant power to quickly accelerate to get to a gap first etc.
And putting obstacles in the road so you cant see if someone is there? That sounds really dangerous and exactly the sort of thing that caused the accident in the story here.
Yes. They have made steady progress over the previous decades to the point where they can now have years with zero road fatalities.
> And putting obstacles in the road so you cant see if someone is there? That sounds really dangerous and exactly the sort of thing that caused the accident in the story here.
Counterintuitive perhaps, but it's what works. Humans adjust their behaviour to the level of perceived risk, the single most important thing is to make driving feel as dangerous as it is.
I think the humans in London at least do not adjust their behaviour for the perceived risk!
From experience they will adjust their behaviour to reduce their total travel time as much as possible (i.e. speed to "make up" for lost time waiting etc) and/or "win" against other drivers.
I guess it is a cultural thing. But I cannot agree that making it harder to see people in the road is going to make anything safer. Even a robot fucking taxi with lidar and instant reaction times hit a kid because they were obscured by something.
> I think the humans in London at least do not adjust their behaviour for the perceived risk!
Sure they do, all humans do. Nobody wants to get hurt and nobody wants to hurt anyone else.
(Yes there are few exceptions, people with mental disorders that I'm not qualified to diagnose; but vast majority of normal humans don't.)
Humans are extremely good at moderating behavior to perceived risk, thank evolution for that.
(This is what self-driving cars lack; machines have no fear of preservation)
The key part is perceived though. This is why building the road to match the level of true risk works so well. No need for artificial speed limits or policing, if people perceive the risk is what it truly is, people adjust instictively.
This is why it is terrible to build wide 4 lane avenues right next to schools for example.
There are always going to be outlier events. If for every one person who still manages to get hit—at slow, easily-survivable speeds—you prevent five others from being killed, it’s a pretty obvious choice.
Does physics work? If it does, then these physical obstacles work too. Go ahead, try to drive faster than 10mph through a roadway narrowed so much it's barely wider than your car, with curbs. And yeah, I'm describing a place in London.
It's a runaway process of prioritizing safety over convenience -- and it's wrecking their road base just before self-driving cars would allow them to have both.
I was wondering how much convenience is worth one kid's life. This thread reminded me of some interesting terms like "value of statistical life." It appears that all those annoying low speed limits and purposeful obstructions in residential areas really do save lives.
> An evaluation of 20 mph zones in the UK demonstrated that the zones were effective both in reducing traffic speed and in reducing RTIs. In particular child pedestrian injuries were reduced by 70 per cent from 1.24 per year in each area before to 0.37 per year after the zones were introduced
20mph residential is pretty close to standard. Note the Waymo car was going slower than that. That's far from the 5mph GP was reacting too, or the super tight curves.
What an American framing. My convenience at the cost of your eventual safety. I guess this is why we also have toddler death machines with 5-foot grills that we call “full size” vehicles.
> If you've ever driven more than 5 miles an hour, you risked hurting someone for your convenience.
Taken literally, that's clearly not true.
For example you can easily drive 150mph in the flat desert where there is nothing for a hundred miles and you can see many miles ahead. You have zero risk of hurting anyone else unless they somehow teleport in front of you.
But driving 5mph in tight street full of elementary school kids running around can be extremely dangerous.
You’re egocentric instead of system-centric. Life has risks, but risk is to be managed, not accepted blindly with disregard of available options. A systemic approach to minimizing risk of injury on roads looks exactly like inconvenience to the individual.
In many civilized countries and locales, even bringing up the word “convenience” in the context of road safety would be considered tasteless. Maybe a phrase like “excessively obstructive” or other euphemisms would be used, but the word “inconvenient” regarding safety measures that would e.g. help prevent the death of toddlers today would be appalling.
There’s this techbro utopia mindset leaking through as well, just like it does for climate change topics, that pragmatic solutions that work for us today are deprioritized because some incredible technology is right around the corner. This is also distinctly American, specifically Silicon Valley, culture.
Gosh, no, the self-driving cars will be forced to drive at safe speeds in pedestrian corridors as opposed to voluntarily driving at safe speeds in pedestrian corridors. How awful.
It is, but it's laughable to suggest it's happening anywhere. Our world is dominated by cars. You likely can't see it precisely because it's so normalized.
“Just before” … this would mean all cars would be required to be self driving and that they’re forced to adhere to the set speed limits. You think this is just around the corner? In a country like Sweden with a lot of snow? Let’s talk about that this when we’re actually close to hitting 100% of self driving cars on the road.
And it’s not “runaway”, it’s exactly the right prioritisation. I’d encourage you to spend some time on Not Just Bikes and the say whether you’d like to live in a Nordic or an American neighbourhood. The Nordic style is also about convenience because car centric infrastructure makes a lot of things less accessible and convenient.
Those things all sound easy to remove in some hypothetical future where there are enough and safe enough self driving cars to have both. Makes sense to design for human driven cars for now though.
> It's likely that a fully-attentive human driver would have done worse.
We'd have to see video of the full scene to have a better judgement, but I wouldn't call it likely.
The car reacted quickly once it saw the child. Is that enough?
But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
If that's the scenario, there is a real probability that a child might appear, so I'm going to be over-slowing way down pre-emptively even thought I haven't seen anyone, just in case.
The car only slows down after seeing someone. The car can react faster that I can after seeing someone, but as a human I can pre-react much earlier based on the big picture, which is much better.
As someone who lives on a residential street right by a primary school in the UK, the majority of drivers are going over 20mph even at the peak time when there are children everywhere.
While in theory human drivers should be situationally aware of the higher risks of children being around, the reality is that the majority will be in their own bubble of being late to drop their kid off and searching for the first free spot they can find.
the human driver would usually drive more closely to the centerline of such a residential road. If the road is clear ahead i'd drive almost over the centerline of the road having enough clearance between my path and the parked cars for any such "jumper" to be visible long enough for me to react. If there is an opposite traffic i get back strictly into my lane and slow down much more if the parked cars are close and they block sidewalk view, etc.
The autonomous cars have really got more aggressive recently as i mentioned before:
I have a similar school drop-off, and can confirm that the cars are typically going around 17-20mph around the school when they're moving. Also that yes, human drivers usually do stay much closer to the centerline.
However, Waymo was recently cleared to operate in my city, and I actually saw one in the drop-off line about a week ago. I pulled out right in front of it after dropping my kid off. And it was following the line of cars near the centerline of the road. Honestly its behavior was basically indistinguishable from a human other than being slightly more polite and letting me pull out after I put my blinker on.
> the human driver would usually drive more closely to the centerline of such a residential road
I certainly do this. But asserting that most humans would usually do this? Have you ever actually seen humans drive cars? This is absolutely not what they do. On top of that, they run stop signs, routinely miss pedestrians in blind spots, respond to texts on their phone, or scroll around on their display to find the next song they want to put on.
I vividly recall a shot within a commercial, in which a driver was shown in slow motion, chucking his coffee into the passenger foot well in order to have two hands on the wheel for an emergency. I don’t remember what was about to happen to the car or the world around it. I’m pretty sure that a collision occurred.
Have you been in a waymo? It knows when there are pedestrians around (it can often see over the top of parked cars) and it is very cautious when there are people near the road and it frequently slows down.
I have no idea what happened here but in my experience of taking waymos in SF, they are very cautious and I'd struggle to imagine them speeding through an area with lots of pedestrians milling around. The fact that it was going 17mph at the time makes me think it was already in "caution mode". Sounds like this was something of a "worst case" scenario and another meter or 2 and it would have stopped in time.
I think with humans, even if the driver is 100% paying attention and eyes were looking in exactly the right place where the child emerged at the right time, there is still reaction times - both in cognition but also physically moving the leg to press the pedal. I suspect that a waymo will out-react a human basically 100% of the time, and apply full braking force within a few 10s of milliseconds and well before a human has even begun to move their leg.
You can watch the screen and see what it can detect, and it is impressive. On a dark road at night in Santa Monica it was able to identify that there were two pedestrians at the end of the next block on the sidewalk obscured by a row of parked cars and covered by a canopy of overgrown vegetation. There is absolutely no way any human would have been able to spot them at this distance in these conditions. You really can "feel" it paying 100% attention at all times in all directions.
Yep I've often noticed this as well - it has many many times detected humans that I can't even see (and I like to sit in the front), especially at night.
Sometimes it would detect something and I think "huh? Must be a false positive?" but sure enough it turns out that there really was someone standing behind a tree or just barely visible around a corner etc.
Sure none of those have run out in front of us, but the fact it is spotting them and tracking their movement before I am even aware they're there is impressive and reassuring.
> I suspect that a waymo will out-react a human basically 100% of the time, and apply full braking force within a few 10s of milliseconds and well before a human has even begun to move their leg.
Correct. Human reaction time is at its very best ~250ms. And that's when you're hyper-focused on reacting to a specific stimuli and actively try to respond to it as fast as possible.
During normal driving, a focused driver will react on the order of 1s. However, that's assuming actively paying attention to the road ahead. If you were to say, be checking your mirrors or looking around for any other reason this can easily get into multiple seconds. If you're say, playing on your phone (consider how many drivers do this), forget it.
A machine however is 100% focused 100% of the time and is not subject to our poor reaction times. It can brake in <100ms every time.
On the other hand, a software fault could make it run into an obstacle that'd be obvious to a human at full speed.
Your opinion of "most humans" is vastly overinflated. The median human driver would be going 5 over the speed limit, on their cell phone, and paying fuck all attention. Humans never drive as slow as 17 mph, even in the context of being directly in front of schools with visible children.
True, but it seems fair to evaluate Waymo against the median American driver. If they expand to whatever other countries you're thinking of, then it will be fair recalibrate accordingly.
> But most humans would have been aware of the big picture scenario much earlier.
I wouldn't call it likely. Sure, there are definitely human drivers who are better than Waymo, but IME they're few and far between. Much more common to be distracted or careless.
When walking along a busy street facing traffic, I like to play a game of "who's using a phone?" I sometimes score in excess of 50% of drivers texting or otherwise manipulating a phone instead of actually driving.
It's amazing how much nonsense we let slide with human drivers, and then get uptight about with anything else. You see the same attitude with bicycles. Cars run stop signs and red lights all day long and nobody bats an eye, but a cyclist does it and suddenly they're a menace.
I don't think it makes sense to lump some drivers better than waymo and worse than waymo. A human brain automatically thinks of all the scenarios, where Waymo has pre-programmed ones (and some NN based ones). So it's scenarios by scenario.
Consider this scenario:
5 kids are walking on the sidewalk while you're driving past them. But suddenly a large dumpster is blocking your view of them just as you pass. You saw them before the dumpster, but not after your car and the dumpster completely blocks the view.
Does a human brain carry some worry that they suddenly decide to run and try to cross the street after the dumpster?
Does Waymo carry that worry or just continue to drive at the exact same speed.
Again, it's not like every driver will think about this, but many drivers will (even the bad ones).
> A human brain automatically thinks of all the scenarios
I don't think this is true. There are infinitely many scenarios in a complex situation like a road with traffic, cars parked, pedestrians about, weather, etc. My brain might be able to quickly assess a handful, but certainly not all.
There aren't infinitely many scenarios to consider, but even if that's a figure of speech, there aren't thousands or even hundreds.
If there's ten kids nearby, that's basically ten path scenarios, and that might be reduced if you have great visibility into some of them.
> My brain might be able to quickly assess a handful, but certainly not all.
What would you do if you can't assess all of them? Just keep driving same speed?
If the situation is too overwhelming you'll almost certainly back off, I know I would. If I'm approaching that school block and there's like 50 small kids running around in all directions, I have no idea what's going on and who is going where, so I'm going to just stop entirely until I can make some sense of it.
It should be trivial for Waymo to implement a "drive carefully near schools" feature, and if really spicy "drive REALLY carefully near schools at these times" feature.
Safe driving starts with speed, lowering speed and informing the passengers seems like a no-brainer.
> here aren't infinitely many scenarios to consider, but even if that's a figure of speech, there aren't thousands or even hundreds.
There are a very, very large number of scenarios. Every single possible different state the robot can perceive, and every possible near future they can be projected to.
Ten kids is not 10 path scenarios. Every kid could do a vast number of different things, and each additional kid raises the number of joint states to another power.
This is trivially true. The game that makes driving possible for humans and robots is that all these scenarios are not equally likely.
But even with that insight, it’s not easy. Consider a simple case of three cars about to arrive at an all-way stop. Tiny differences in their acceleration - potentially smaller differences than the robot can measure - will result in a different ordering of cars taking turns through the intersection.
> Ten kids is not 10 path scenarios. Every kid could do a vast number of different things, and each additional kid raises the number of joint states to another power.
This is the difference between computing and humans. The car will attempt to compute all possible path scenarios because it has no instict, and it might not be possible to compute everything in real time so it might fail.
But the human will easily deal with the situation.
Try running through a sports field in an elementary school during lunch, full of unpredictable kids running around. Can you make it from one side to the other without crashing into a whole bunch of kids? Of course you can. You didn't need to try to compute an exponential number of scenarios, you just do it easily. The human brain is pretty amazing.
In fact no computer approach attempts to compute all possible path scenarios since we know that’s not tractable.
And current practical approaches are mostly end to end (or nearly) ML systems that do not compute a lot of alternative paths, and they work in approximately constant time independent of the scenario.
You strongly imply that computers can’t drive, but you could have written that in a Waymo.
It was a figure of speech, but I think you're undercounting. When you consider interactions between all the things, even with just a handful of variables (and I think there are many more than a handful) you get a huge number of scenarios.
This is the classical ‘Frame Problem” of AI. How do you consider, even if only to reject, infinite scenarios in finite time? Humans and other animals don’t seem to suffer from it.
God I wish I re-read my statement, I was more focused on Humans think of an unlimited number of scenarios - not necessarily all. A computer will only think of pre-programmed ones.
The computer isn't pre-programmed though. These computers are trained similar to how human brains are (though obviously brains are still vastly, vastly, vastly superior to computers for tasks like this).
> A human brain automatically thinks of all the scenarios, ...
Patently, obviously false. A human brain will automatically think of SOME scenarios. For instance, if a collision seems imminent, and the driver is holding a cup of coffee, these ideas are likely to occur to the driver:
IF I GRAB THE STEERING WHEEL AND BRAKE HARD, I MIGHT NOT HIT THAT PEDESTRIAN IN FRONT OF ME.
IF I DON'T CONTINUE HOLDING THE COFFEE CAREFULLY, I MIGHT GET SCALDED.
THIS SONG ON MY RADIO IS REALLY ROCKING!
IF I YANK MY WHEEL TO THE LEFT, I MIGHT HIT A CAR INSTEAD OF A HUMAN.
IF I BRAKE HARD OR SWERVE AT ANY TIME IN TRAFFIC, I CAN CAUSE AN ACCIDENT.
Experiments with callosal patients (who have damaged the connective bridge between the halves of their brains) demonstrate that this is a realistic picture of how the brain makes decisions. It offers up a set of possible actions, and attempts to choose the optimal one and discard all others.
A computer program would do likewise, EXCEPT it won't care about the coffee cup nor the radio (remove two bad choices from consideration).
It still has one bad choice (do nothing), but the SNR is much improved.
I'm not being hyperbolic; self-preservation (focusing on keeping that coffee in my hand) is a vital factor in decision-making for a human.
> ...where Waymo has pre-programmed ones (and some NN based ones).
Yes. And as time goes on, more and better-refined scenarios will be added to its programming. Eventually, it's reasonable to believe the car software will constantly reassess how many humans are within HUMAN_RUN_DISTANCE + CAR_TRAVEL_DISTANCE in the next block, and begin tracking any that in an unsafe margin. No human on Earth does that, continually, without fail.
> Does a human brain carry some worry that they suddenly decide to run and try to cross the street after the dumpster? Does Waymo carry that worry or just continue to drive at the exact same speed.
You continue to imply that Waymo cannot ever improve on its current programming. Does it currently consider this situation? Probably not. Will it? Probably.
God I wish I re-read my statement, I was more focused on Humans think of an unlimited number of scenarios - not necessarily all. A computer will only think of pre-programmed ones.
For what it's worth, that kind of lumping of drivers is more-or-less one of the metrics Waymo is using to self-evaluate. Perfect safety when multi-ton vehicles share space with sub-300-pound humans is impossible. But they ultimately seek to do better than humans in all contexts.
According to the article the car was traveling at 17 miles an hour before it began braking. Presumably this was in a 25 mph school zone, so it seems the Waymo was already doing exactly what you describe - slowing down preemptively.
This is close to a particular peeve I have. Occasionally I see signs on the street that say "Slow Down". I'm not talking about the electronic ones connected to radar detectors. Just metal and paint.
Here's my problem. If you follow the instructions on the sign, it still says to slow down. There's no threshold for slow enough. No matter how slow you're going, the sign says "Slow Down". So once you become ensnared in the visual cone of this sign, you'll be forced to sit stationary for all eternity.
But maybe there's a loop-hole. It doesn't say how fast you must decelerate. So if you come into the zone going fast enough, and decelerate slowly enough, you can make it past the sign with some remaining non-zero momentum.
You know, I've never been diagnosed on the spectrum, but I have some of the tendencies. lol.
Obviously a static sign is not aware of your current state, so it's message can only be interpreted as relevant to your likely state... i.e. the posted speed limit.
Usually the reason is the "slow down" portion is very small, and it's confusing to shift down the actual speed limit for a 200 foot stretch of road then increase it again.
Much better to be specific than a vague "slow down". There's a road near me with two tight turns a couple blocks apart. One advises 25mph and the other advises 10mph.
Except we do that all the time in school zones... normally 35+, but from 7am-9am and again from 2pm-4pm the limit drops to 25mph (which is still to fast if the kids are actually crossing the street or walking alongside en masse).
Think of it like they're saying "my children play on this street and my neighbors walk here. Please think about that when you decide how fast to go here."
A lot of clickbait headlines have the same problem. "You're using too much washing powder!"
Everyone's replying to you as if you truly don't understand the sign's intention but I'm sure you do. It's just annoying to be doing everything right and the signs and headlines are still telling you you're wrong.
There was a driving safety safety ad campaign here: "Drive to the conditions. If they change, reduce your speed." You can imagine how slow we'd all be going if the weather kept changing.
Yes. You have understood precisely the spirit in which I intended it.
In advertising: "Treat yourself. You deserve it!"
Me: What if someone who didn't deserve it heard this message. How can you possibly know what I deserve? Do all people deserve to treat themselves? Is the notion of deserving or treating really so vacuous?
There's a mental health awareness campaign going on around here at the moment with all the generic messages like that. "You're doing great" is completely devalued by the sign giving the same message to everyone, and the best one says something like "Don't push yourself too hard. If you want to rest, rest." Wondering if I can tell my boss the sign told me it's okay not to get any work done.
Humans are supposed to deal with this kind of ambiguity. Actually, that's one of our nicest abilities.
I hate when people pretend to be smarter than everyone else by pointing this kind of utterance and insisting that someone, somehow, will parse those statements in the most literal and stupid manner.
Then there are the ignorant misanthropes that can't waste a chance to repeat their reductionist speculations about human cognition. Just like the idiot Elon Musk that wasted billions in irrecoverably fucked self-driving system based on computer-version because he underestimated the human visual cortex.
My driving test was so thorough that I had to parallel park between two entirely fictional cars. There was certainly no consideration of eccentric signage.
I apologize if I gave the impression that I did not understand how to put them into context. Although I don't think my driving lessons ever mentioned it.
A 25mph school zone? That seems fast. 15mph would be more the norm, which is in line with the 17mph the car believed itself to be traveling.
FYI, unless you are a commerical truck, a cop, or a racer, your speedometer will read slightly fast, sometimes as much as 5 to 10%. This is normal practice for cars as it limits manufacturer liability. You can check this using independant gps, ie not an in-dash unit. (Just imagine the court cases if a speedo read slower than the actual speed and you can understand why this started.)
I mostly see 25 mph for school zones, though I'm in NC. Checking California, it sounds like 25 is standard there as well.[0] Some will drop to 15, but 25 is the norm as far as I can find.
Also, a different wheel diameter than the speedometer was calibrated with and you will have a larger difference between actual velocity and speedometer reading. The odometer will also not record actual distance traveled.
It depends. I had a honda motorcycle where the speedo was 10ish % fast (not unussual on bikes due to tire shape) but the odo was accutrate. Same sensor, but the computer just counted wheel rotations slightly differently for each use.
Virtually all speedos read fast. The federal standards have a fairly high margin for being allowed to read high, and a zero margin for reading low. Thus speedos are more or less universally calibrated to read at least 5% high.
It does seem fast to me -- school zones are 20 mph in Seattle, at least when children are present. But Google suggests 25 is the norm in Santa Monica, where the incident occurred.
To put it another way. If an autonomous vehicle has a reaction time of 0.3 seconds, the stopping distance from 17 mph is about the same as a fully alert human driver (1 second reaction time) driving 10.33 mph.
I have a hard time believing that a human driver would be as slow as this Waymo, or even slower. I drive my kid to school where it's posted 20mph and there are cameras (with plenty of warnings about the presence of said cameras) and witness a constant string of flashes from the camera nailing people for speeding through there.
Hardcoded limits are problematic because they completely lack context.
On that very same road with a 20mph limit, 40mph might be completely safe or 3mph might be extremely negligently dangerous. It all depends on what is going on in the area.
A small child jumped out in front of it, which is about the worst case scenario you can have... and the kid was fine. So it sounds like it was slow enough?
In this situation, the car was already driving under the legal speed required for a school zone (25mph when children are present) [edit: some comments in the post suggest there is a 15mph sign, which is sometimes posted; to me, driving 17mph in a 15mph zone is acceptable).
I think any fair evaluation of this (once the data was available) would conclude that Waymo was taking reasonable precautions.
That's exactly part of the problem. If it is programmed to be over-cautious and go 17 in a 25 zone, that feels like it is safe. Is it?
It takes human judgment of the entire big picture to say meaningfully whether that is too slow or too fast. Taking the speed limit literally is too rigid, something a computer would do.
Need to take into account the flow of the kids (all walking in line vs. milling around going in all directions), their age (younger ones are a lot more likely to randomly run off in an unsafe direction), what are they doing (e.g. just walking, vs. maybe holding a ball that might bounce and make them run off after it), their clustering and so on.
Driving past a high school with groups of kids chatting on the sidewalk, sure 20mph is safe enough. Driving past an elementary school with a mass of kids with toys moving in different directions on the same sidewalk, 17mph is too fast.
And if I'm watching some smaller kids disappear behind a visual obstruction that makes me nervous they might pop up ahead of it on the street, I slow down to a crawl until I can clearly see that won't happen.
None of this context is encoded in the "25mph when children are present" sign, but for most humans it is quite normal context to consider.
But would be great to see video of the Waymo scene to see if any of these factors was present.
Curiously enough Google could have access to how fast humans usually drive through that street.. if they record people's Google Maps trips, they can show the court that "Look, 80% of Google Maps users drive through here at 30 mph!".
Google might even know how many drivers aren't obeying the speed limit or slow-rolling through stop signs. I wonder if they already have partnerships with law enforcement to detect areas where the traffic law is more ignored than others.
I've read studies saying that most drivers don't brake at max effort, even to avoid a collision. This may be at least one of the reasons that Waymo predicted that an attentive human would likely have been going faster than their car at the moment of impact. I've got a good idea of my fun-car's braking performance, because I drive it hard sometimes, but after reading that I started practicing a bit with my wife's car on the school run, and... Yeah: it's got a lot more braking power than I realized. (Don't worry, I brake hard on a long straight exit ramp, when no one's behind me, a fast slow-down is perfectly safe, and the kiddo loves it.) I've now got an intuitive feel for where the ABS will kick in, and exactly what kind of stopping distance I have to work with, which makes me feel like a safer driver.
Second, going off my experience of hundreds and hundreds of ride-share rides, and maybe thirty Waymo journeys, I'd call the best 10-15% of humans better drivers than Waymo. Like, they're looking further up the road to predict which lane to be in, based on, say, that bus two blocks away. They also drive faster than Waymos do, without a perceptual decrease in safety. (I realize "perceptual" is doing some work in that sentence!) That's the type of defensive and anticipatory urban driver I try to be, so I notice when it's done well. Waymo, though, is flat-out better, in every way, than the vast majority of the ride-share drivers I see. I'm at the point where I'll choose a Waymo any time it'll go where I'm headed. This story reinforces that choice for me.
> I've read studies saying that most drivers don't brake at max effort, even to avoid a collision.
Ha! It is unbelievable how difficult it is to make someone brake hard. You'd think it's the easiest thing possible in the age of ABS - just press hard as you can.
I have a lot of experience on this, I used to teach car control both to teens and adults. One of the frequent exercises was seemingly very simple: Drive at Xmph until this spot, then brake at maximum power.
The vast majority of people can't do it on the first or second try, they'll just meekly press on the brake like they're coasting to a stop. After more coaching that hard means hard, they start to get it, but it takes many many tries.
The reason attentive humans don't equal the Waymo here is reaction time. When a thing happens the human takes a moment to process what it means, and choose a reaction. It's not, by our standards, a long time but it's way longer than it takes the Waymo.
Going early means you slow early, which means you also take longer to reach the child, but you're braking for all of that extra time, so you're slowing down even more.
> But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
Waymos do this and have for years. They know where the people are around them and will take precautionary action based on that.
That video is nearly 7 years old at this point and they've gotten much, much better since then.
If you think a fully-attentive human driver would have done better, I think you're kidding yourself.
I know you didn't make this point, but if anyone think the average LA driver would have done better than this I've got a bridge to sell you and that's really what matters more. (I say that as someone who used to live like half a mile from where this happened)
It would be nice to see the video (although maybe there are some privacy issues, it is at a school after all).
Anyway, from the article,
> According to the NHTSA, the accident occurred “within two blocks” of the elementary school “during normal school drop off hours.” The safety regulator said “there were other children, a crossing guard, and several double-parked vehicles in the vicinity.”
So I mean, it is hard to speculate. Probably Waymo was being reasonably prudent. But we should note that this description isn’t incompatible with being literally in an area where the kids are leaving their parents’ cars (the presence of “several double parked cars brings this to mind). If that’s the case, it might make sense to consider an even-safer mode for active student unloading areas. This seems like the sort of social context that humans might have and cars might be missing.
But things speculation. It would be good to see a video.
In principle, attentive drivers, who have either somehow come independently to the appropriate understanding or have been trained in how to react to hazards ahead...
... could in some circumstances know that there's a likelihood that a child will emerge suddenly and reduce their speed in anticipation where circumstances allow.
Note that: If you cut speed but other drivers can't see why they may overtake, even unsafely, because you are a nuisance to them. Slowing in anticipation that a child will run out from behind the SUV, only for a car behind you to accelerate around you and smack straight into the child at even higher speed, is not the desired outcome even though you didn't hurt anybody...
And yes, we'd need to see the video to know. It's like that Sully scenario. In a prepared test skilled pilots were indeed able to divert and land, but Sully wasn't prepared for a test. You're trained to expect engine failure in an aeroplane - it will happen sometimes so you must assume that, but for a jet liner you don't anticipate losing both engines, that doesn't happen. There's "Obviously that child is going in the road" and "Where the fuck did they come from?" and a lot in between and we're unlikely to ever know for sure.
There's a bus stop right behind my house. I routinely hear the driver honking and yelling at people who ignore when the stop sign is extended (which is a misdemeanor in my state). So forgive me for not assuming a human would have done better.
I live in an area where there are pedestrians stepping into the street without looking, all over the place, and you can drive / cycle without hitting them but have to slow down appropriately if you have to go near something that you can't see behind. Like you say it would be interesting to see the video.
> But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
Waymos constantly track pedestrians nearby, you can see it on the status screen if you ride in one. So it would be both better able to find pedestrians and react as soon as one was on a collision course. They have a bit more visibility than humans do due to the sensor placement, so they also can see things that aren't that visible to a person inside the car, not to mention being constantly aware of all 360 degrees.
While I suppose that in theory, a sufficiently paranoid human might outdo the robot, it looks to me like it's already well above the median here.
Do they speculate about things like “we’re near a school zone, kids are unloading, there might be a kid I’ve never seen behind that SUV?” (I’m legitimately asking I’ve never been in a Waymo).
Asking whether an entity has modeled and evaluated a specific situation, using that evaluation to inform its decisions, is not about subjective experience.
If you're asking whether their training data includes situations like this, and whether their trained model/other pieces of runtime that drive the car include that feature as part of their model, the answer is yes. But not in the way a normal human driver would think about it; many of the details of its decision making process are based on large statistical collections, rather than "I'm in a school zone and need to anticipate children may be obscured and run out into traffic." There are many places where the car needs to take caution without knowing specifically it's within 50 feet of a school zone.
>The car can react faster that I can after seeing someone
and that can potentially allow internal planning algorithm to choose more risky and aggressive trajectories/behavior, etc. say to reach target destination faster and thus deliver higher satisfaction to the passengers.
Anecdote, but I live next to an elementary school and also on a route frequented by Waymos. Human drivers routinely cruise down the 25mph roads at 40+ and blow stop signs, even during school intake and release. Waymo vehicles always seem a lot more cautious.
When thinking about these things you have to factor in the prior probability that a driver is fully attentive, not just assume they are.
If you’ve ever been in a Waymo you quickly realize their field of view is pretty good. You often see the vehicle sensing small pets and children that are occluded to a passenger or driver. For this reason and my experience with humans near aforementioned school, I doubt a human would out perform the Waymo in this particular incident and it’s debatable they even have more context to inform their decisions.
All that said, despite having many hours in a Waymo, it’s not at all clear to me how they factor in sidewalk context. You get the sense that pedestrians movement vectors are accounted for near intersections, but I can’t say I’ve experienced something like a slow down when throngs of people are about.
Precisely. Environmental context is not considered in Waymo's "peer-reviewed model" (I encourage reflexive commenters to first read it: https://waymo.com/safety/collision-avoidance-benchmarking), only basic driver behavior and traffic signal timings.
Note the weaselly "immediately detected the individual as soon as they began to emerge" in the puff piece from Waymo Comms. No indication that they intend to account for environmental context going forward.
If they already do this, why isn't it factored in the model?
Not OP but I interpret that as they are focusing exclusively on what happened after the car saw the kid.
And I completely agree that from that instant forward, the car did everything correctly.
But if I was the accident investigator for this, I would be far more interested in what happened in the 30 seconds before the car saw the kid.
Was the kid visible earlier and then disappear behind an obstruction? Or did the kid arrive from the side and was never earlier visible? These are the more important questions.
Possibly, but Waymos have recently been much more aggressive about blowing through situations where human drivers can (and generally do) slow down. As a motorcyclist, I've had some close calls with Waymos driving on the wrong side of the road recently, and I had a Waymo cut in front of my car at a one-way stop (t intersection) recently when it had been tangled up with a Rivian trying to turn into the narrow street it was coming out of. I had to ABS brake to avoid an accident.
Most human drivers (not all) know to nose out carefully rather than to gun it in that situation.
So, while I'm very supportive of where Waymo is trying to go for transport, we should be constructively critical and not just assume that humans would have been in the same situation if driving defensively.
Certainly, I'm not against constructive criticism of Waymo. I just think it's important to consider the counterfactual. You're right too that an especially prudent human driver may have avoided the scenario altogether, and Waymo should strive to be that defensive.
> I'm not against constructive criticism of Waymo.
I feel like you have to say this out loud because many people in these discussions don't share this view. Billion dollar corporate experiments conducted in public are sacrosanct for some reason.
> I just think it's important to consider the counterfactual
More than 50% of roadway fatalities involve drugs or alcohol. If you want to spend your efforts improving safety _anywhere_ it's right here. Self driving cars do not stand a chance of improving outcomes as much as sensible policy does. Europe leads the US here by a wide margin.
> I feel like you have to say this out loud because many people in these discussions don't share this view. Billion dollar corporate experiments conducted in public are sacrosanct for some reason.
Yes, and I find it annoying that some people do seem to think Waymo should never be criticized. That said, we already have an astounding amount of data, and that data clearly shows that the experiment is successful in reducing crashes. Waymos are absolutely, without question already making streets safer than if humans were driving those cars.
> If you want to spend your efforts improving safety _anywhere_ it's right here.
We can and should do both. And as your comment seems to imply but does not explicitly state, we should also improve road design to be safer, which Europe absolutely kicks America's ass on.
> and that data clearly shows that the experiment is successful in reducing crashes
I disagree. You need way more data, like orders of magnitude more. There are trillions of miles driven in the US every year. Those miles often include driving in inclement weather which is something Waymo hasn't even scraped the surface of yet.
> without question
There are _tons_ of questions. This is not even a simple problem. I cannot understand this prerogative. It's far too eager or hopeful.
> We can and should do both
Well Google is operating Waymo and "we" control road policy. One of these things we can act on today and the other relies on huge amounts of investments paying off in scenarios that haven't even been tested successfully yet. I see an environment forming where we ignore the hard problems and pray these corporate overlords solve the problem on their own. It's madness.
> You need way more data, like orders of magnitude more. There are trillions of miles driven in the US every year.
Absurd, reductive, and non-empirical. Waymos crash and cause injury/fatality far less frequently than human drivers, full stop. You are simply out of your mind if you believe otherwise, and you should re-evaluate the data.
> Those miles often include driving in inclement weather which is something Waymo hasn't even scraped the surface of yet.
Yes. No one is claiming that Waymos are better drivers than humans in inclement weather, because they don't operate in those conditions. That does not mean Waymos are not able to outperform human drivers in the conditions in which they do operate.
> I see an environment forming where we ignore the hard problems and pray these corporate overlords solve the problem on their own. It's madness.
What's madness is your attitude that Waymos' track record does not show they are effective are reducing crashes. And again, working on policy does not prevent us from also improving technology as you seem to believe it does.
>data clearly shows that the experiment is successful in reducing crashes.
That's fine. But crashes are relatively rare and what matters is accountability. Will Waymo be accountable for hitting this kid the way a human would? Or will they fight in court to somehow blame the pedestrian? Those are my big concerns when it comes to self driving vehicles, and history with tech suggests that they love playing hot potato instead of being held accountable.
And yes, better walkable infrastructure is a win for all. The minor concern I have is the notion that self driving is perfect and we end up creating even more car centric infrastructure. I'm not sure who to blame on that one.
I hope so too. I'll be keeping a close eye on how they handle this, though. My benefit of the doubt for tech was already long drained, and is especially critical for safety critical industries.
> More than 50% of roadway fatalities involve drugs or alcohol. If you want to spend your efforts improving safety _anywhere_ it's right here. Self driving cars do not stand a chance of improving outcomes as much as sensible policy does. Europe leads the US here by a wide margin.
Could you spell out exactly what "sensible" policy changes you were thinking of? Driving under the influence of drugs and/or alcohol is already illegal in every state. Are you advocating for drastically more severe enforcement, regardless of which race the person driving is, or what it does to the national prison population? Or perhaps for "improved transit access", which is a nice idea, but will take many decades to make a real difference?
>Driving under the influence of drugs and/or alcohol is already illegal in every state.
FWIW, your first OWI in Wisconsin, with no aggravating factors, is a civil offense, not a crime, and in most states it is rare to do any time or completely lose your license for the first offense. I'm not sure exactly what OP is getting at, but DUI/OWI limits and enforcement are pretty lax in the US compared to other countries. Our standard .08 BAC limit is a lot higher than many other countries.
That's true, but note that getting much more severe on enforcement and punishment for DUI/OWI will result in an even higher prison population, more serious life consequences for poor and minorities, etc, when the US is constantly getting trashed for how bad those things are already.
To be a bit snarkier, and not directed at you, but I wish these supposedly superior Europeans would tell us what they actually want us to do. Should we enforce OWI laws more strictly, or lower the prison population? We can't do both!
I suspect you could step up enforcement in ways that don’t involve prison time simply by taking away people’s licenses, and then having a fast feedback loop to catch people driving without a license.
In addition to what the sibling said regarding the impracticality of not driving in most of the US, which I completely agree with, I'd also ask exactly what you want to do with your "fast feedback loop to catch people driving without a license". What do you do with the people who drive anyways because not driving is so impractical and get caught?
We already took their license, we can't double-take it to show we really mean it. Fining them seems a bit rough when they need to drive to get to the job to make the money to pay those fines. Or we're right back to jail time and an even higher prison population.
> I'd also ask exactly what you want to do with your "fast feedback loop to catch people driving without a license".
Unless the vehicle is stolen, seize and impound the vehicle. If the driver is the owner, auction it off and give them back the proceeds, minus costs.
I feel like I'm living in some different world where drunk driving is a-okay when I face these types of objections to actually enforcing the rules around it.
Taking away licenses is a bad way to enforce driving rules because so many people have to be able to drive or their life collapses. The problems of aggressive license revocation are similar to the problems of aggressive prison time.
I get where you're coming from, but it's pretty hard to be sympathetic given the crimes we're talking about and the impact they have on others.
Like that would sound nuts if we applied it to other things - e.g. "take away the professional license of a mid-career pilot/surgeon/schoolteacher/engineer because he was drinking on the job and his life collapses".
Various people can't drive because of e.g. visual impairments, age, poverty, etc. - I find it an ugly juxtaposition to be asserting that we must allow people with DUIs to drive because otherwise their lives would "collapse" to the same point as those other people who can't drive.
> Like that would sound nuts if we applied it to other things - e.g. "take away the professional license of a mid-career pilot/surgeon/schoolteacher/engineer because he was drinking on the job and his life collapses".
The analogy is closer to "take away their ability to get any job" and then it sounds even more harsh.
> Various people can't drive because of e.g. visual impairments, age, poverty, etc. - I find it an ugly juxtaposition to be asserting that we must allow people with DUIs to drive because otherwise their lives would "collapse" to the same point as those other people who can't drive.
If you can't see well enough to drive, then life was unfair to you, and you can often get help with transportation that isn't available to someone that violated the law. For age, if you're young then your parents are supposed to care for you, if you're too old to drive you're supposed to have figured out your retirement by now. For poverty, you kinda still need a car no matter what, that's just how the US is set up in most areas. And it's not ugly to make the comparison to extreme poverty, to say that kicking someone down to that level is a very severe punishment.
> must allow
I wasn't saying what we should do, just that turning up the aggressiveness has serious unwanted consequences.
> The analogy is closer to "take away their ability to get any job" and then it sounds even more harsh.
If you take away the license of a pilot mid-career, they may be able to pivot to something else, but have a huge sunk cost of education and seniority where they ground out poor pay/schedules and then never made it to the part of the career with better pay. For a substantial segment of them, the career impact would be comparable to taking away the ability to drive from a random person.
> For poverty, you kinda still need a car no matter what, that's just how the US is set up in most areas.
You really don't. If you don't already live somewhere with public transit, you'll probably have to move. You'll have to make some sacrifices. But it's workable, I lived without a car and relied on city busses for all my transportation for several years. (And while I wouldn't necessarily recommend it, prior to that, I lived in a small town of ~4k people without transit service. I walked everywhere, and took the inter-city bus when I needed to leave the town.)
Absolutely, I can tell you right now that many human drivers are probably safer than the Waymo, because they would have slowed down even more and/or stayed further from the parked cars outside a school; they might have even seen the kid earlier in e.g. a reflection than the Waymo could see.
It seems it was driving pretty slow (17MPH) and they do tend to put in a pretty big gap to the right side when they can.
There are kinds of human sensing that are better when humans are maximally attentive (seeing through windows/reflections). But there's also the seeing-in-all-directions, radar, superhuman reaction time, etc, on the side of the Waymo.
17MPH is way to fast, depending on the details. I do not think the article gives the details to know if it was a reasonable speed to be going or not, enough details to know it might be to fast, like proximity to a school and children present, yes.
I think my problem is that it reacted after seeing the child step out from behind the SUV.
An excellent driver would have already seen that possible scenario and would have already slowed to 10 MPH or less to begin with.
(It's how I taught my daughter's to drive "defensively"—look for "red flags" and be prepared for the worst-case scenario. SUV near a school and I cannot see behind it? Red flag—slow the fuck down.)
At least it was already slowed down to 17 mph to start. Remember that viral video of some Australian in a pickup ragdolling a girl across the road? Most every comment is "well he was going the speed limit no fault for him!" No asshole, you hit someone. It's your fault. He got zero charges and the girl was seriously injured.
Just about absolute. Fall off bridge onto car, I guess not. Olympic sprinter dashes out from car intentionally trying to be hit? Guess not either. Clothed mostly in black on a rainy night on a freeway? Not either.
But you hit a kid in daytime? It's your fault. Period.
It’s possible a driver turns a corner (not wearing sunglasses) and suddenly the sun briefly blinds them, while a kid darts into the street.
I’ve seen kids (and ADULTS!) walk on the side of the street at night in all black or very very dark clothing. It’s extra amusing when they happen to be black (are they trying to get themselves killed?) It’s not the drivers fault if they genuinely can’t see a camouflaged person. I’ve had numerous close calls like this on rural and suburban roads and I think i’m a cautious driver. Make sure you are visible at night.
Or if a kid is riding a bicycle down a hill and flies into the middle of an intersection (dumb? brakes failed? etc). very possible to accidentally mow down the child.
HOWEVER, i do agree that 95% of the time it’s the drivers fault if they hit a kid. Poor awareness and speed are the biggest factors. It is certainly not 100% of the time the drivers fault though. That’s absurd. You really misunderstand how dumb some pedestrians (and parents) are.
But….it’s all besides the point. A child that doesn’t understand the importance of cross walks and looking both ways is too young to be walking alone, period. Yes even if they’re “right”. Being right isn’t helpful if you’re dead.
> You seem to be implying that there are no circumstances in which a vehicle can hit a pedestrian and the driver not be at fault... which is absurd.
It is absurd, but that doesn't mean that the attitude can't be useful!
In teaching my teenager to drive, I drilled into him the fact that, in every accident, regardless of who is "at fault", there is almost always something that the other party could have done to mitigate it. I gave him plenty of situations as examples...
You're going down a street that has kids on the sidewalk? You better be prepared to have one of those kids come out in front of the car while rough-housing, playing, whatever.
You had right of way? Maybe you did, but did you even look at the opposing traffic to see if it was safe to proceed or did you just look at the traffic light?
I've driven, thus far in my life, roughly 600000km (maybe more) with 2x non-trivial accidents, both ruled not my fault. In hindsight, I could have avoided both of them (I was young and not so self-aware).
I'm paranoid when driving, and my stats are much much better than Waymo's (have never injured anyone - even my 2x accidents only had me injured), even though I drive in all sorts of conditions, and on all sorts of roads (many rural, some without markings).
Most people don't drive like this though (although their accident rate is still better than Waymo's).
No it's not. The same principle applies to rules of right of way on the water. Technically the 32 foot sailboat has right of way over a triple-E because the triple-E uses mechanical propulsion.
You have a responsibility to be cautious in heavy equipment no matter what the signage on the road says, and that includes keeping a speed at which you can stop safely if a person suddenly steps onto the road in situations where people are around. If you are driving past a busy bar in downtown, a drunk person might step out and you have a responsibility to assume that might happen. If you have to go slower sometimes, tough.
I don't think that's a great analogy since a sailboat's right-of-way isn't unlimited and it can certainly be found at fault for a collision with a triple-E container ship - especially given maritime law uses the comparative fault system where fault is shared between parties.
For instance, a sailboat must alter course if a collision can't be avoided by the give-way vessel alone:
Rule 17(b):
> When, from any cause, the vessel required to keep her course and speed finds herself so close that collision cannot be avoided by the action of the give-way vessel alone, she shall take such action as will best aid to
avoid collision.
So if you sail your boat into a container ship and it tries to give way, but doesn't have the ability to do so quickly enough to prevent a collision, you're violating the rules if you don't also alter course as well.
Plus, if we're going to connect this to a pedestrian, if a sailboat suddenly cut in front of a container ship with zero concern for its limited maneuverability/ability to stop, the sailboat would also violate Rule 2 by neglecting precaution required by seamen and failing to consider the limitations of the vessels involved.
And if a pedestrian jumps from a bridge to land right in front of you? or how about a passenger jumps of out the car next to you? still going to stand on your absolute?
As an aside, because it would not be germane to automotive safety…
In the Coast Guard Auxiliary “Sailing and Seamanship” class that I attended, targeting would-be sailboat skippers, we were told the USS Ranger nuclear-powered aircraft carrier had the right-of-way.
You mean the Aussie one where the guy was going an appropriate speed for the area and when the cops arrived the parents and their neighbors LIED TO THE POLICE and said he was hooning down the road at excess speed and hit the kid? And that he was only saved from prison by having a dash cam that proved the lies to be lies? That one?
That logic is utter bs, if someone jumps out when you're travelling at an appropriate speed and you do your best to stop then that's all that can be done. Otherwise by your logic the only safe speed is 0.
It's not the drivers fault when they hit a kid who darts out in front of them and they had no time to react and weren't doing anything illegal like speeding.
I don't see how that's feasible without introducing a lot of friction.
Near my house, almost the entire trip from the freeway to my house is via a single lane with parked cars on the side. I would have to drive 10 MPH the entire way (speed limit is 25, so 2.5x as long).
Why can't we add friction to save lives? Automobiles are the single leading cause of death for children in the USA! We're not talking about something uncommon.
Remove the free parking if that's making the road unsafe. Or drive 10 mph. Done.
But you most likely don't have that entire road be full of little kids in the sidewalk all the way. If you did, then yes probably 10mph or less would be wise.
That's why the speed limit is 25 (lower when children are present in some areas) and not 35 or 40 etc. It's not reasonable to expect people to drive at 40% of the posted speed limit the entire way. We're also not talking about zero visibility (e.g. heavy fog). We're talking about blind spots behind parked cars, which in dense areas of a city is a large part of the city. If we think as a society in those situations the safe speed is 10 mph, then the speed limit should be 10mph.
It absolutely is reasonable to expect people to drive below the limit. Speed limits are just that, maximum upper bounds of how fast you legally can go, not "recommended" speeds or minimum speeds, nor are they necessarily "safe" speeds. It's just a legal upper bounds.
It is a drivers responsibility to drive for the conditions. If conditions are calling for driving 40% slower, then that's what you do and suck it up.
If too many roads have conditions that require that, take that up with your municipality to fix the situation. Or, even better, advocate for better public transit and trains, and designing cities to move people, not move cars.
I mean, you are putting your finger right on the answer: the whole car thing doesn't work or make sense, and trying to make autonomous vehicles solve the unsolvable is never going to succeed.
Car culture in the US is toxic, and a lot of accidents and fatalities are a result of how poorly designed our infrastructure is. We design for cars, not for people (just one more lane bro, will totally fix traffic. Nevermind that a train can move double the capacity of that entire line of traffic).
Cars are the wrong solution, particularly in urban areas. A self driving car is still a car, and comes along with all the same problems that cars cause.
Aye, and to always look for feet under and by the front wheel of vehicles like that.
Stopped buses similarly, people get off the bus, whip around the front of them and straight into the streets, so many times I’ve spotted someone’s feet under the front before they come around and into the street.
Not to take away from Waymo here, agree with thread sentiment that they seem to have acted exemplary
You can spot someone's feet under the width of a bus when they're on the opposite side of the bus and you're sitting in a vehicle at a much higher position on the opposite side that the bus is on? That's physically impossible.
In normal (traditional?) European city cars, yes, I look for feet or shadows or other signs that there is a person in the other side. In SUVs this is largely impossible but then sometimes you can see heads or backpacks.
Or you look for reflections in the cars parked around it. This is what I was taught as “defensive“ driving.
I think you're missing something though, which I've observed from reading these comments - HN commenters aren't ordinary humans, they're super-humans with cosmic powers of awareness, visibility, reactions and judgement.
>reacted after seeing the child step out from behind the SUV.
Lmao most drivers I see on the roads aren't even capable of slowing down for a pedestrian crossing when the view of the second half of the crossing is blocked by traffic (ie they cannot see if someone is about to step out, especially a child).
Yes and no. Tons of situations where this is simply not possible, whole traffic goes full allowed speed next to row of parked cars. If somebody unexpectedly pops up distracted, its a tragedy guaranteed regardless of driver's skills and experience.
In low traffic of course it can be different. But its unrealistic to expect anybody to drive in expectation that behind every single car passed there may be a child jumping right in front of the car. That can be easily thousands of cars, every day, whole life. Impossible.
We don't read about 99.9% of the cases where even semi decent driver can handle it safely, but rare cases make the news.
I slow down considerably near parked cars. And I try to slow down much earlier approaching intersections where there are parked cars blocking my view of cross walk entries. I need to be able to come to full stop earlier than intersection if there happens to be a pedestrian there.
I kind of drive that way. I slow down, move as far away in my lane from the parked cars as possible. It's certainly what I would expect from a machine that would claim to be as good as the best human driver.
This is generally the problem with self-driving cars, at least in my experience (Tesla FSD).
They don't look far enough ahead to anticipate what might happen and already put themselves in a position to prepare for that possibility. I'm not sure they benefit from accumulated knowledge? (Maybe Waymo does, that's an interesting question.) I.e., I know that my son's elementary school is around the corner so as I turn I'm already anticipating the school zone (that starts a block away) rather than only detecting it once I've made the turn.
Evidence of this? I own a Tesla (HW4, latest FSD) as well as have taken many Waymo rides, and have found both to react well to unpredictable situations (i.e. a car unexpectedly turning in front of you), far more quickly than I would expect most human drivers to react.
This certainly may have been true of older Teslas with HW3 and older FSD builds (I had one, and yes you couldn't trust it).
So... the thing is it's really not. They're behind on schedule, having just launched in public. But FSD has been showing capabilities in regular use (highway navigation, unprotected left turns in traffic, non-geofenced operation areas based solely on road markings) that Waymo hasn't even tried to deploy yet.
It's much more of a competition than I suspect a lot of people realize.
Yes I agree, but why 10mph? Why not 5mph? or 2mph? You'll still hit them if they step out right in front of you and you don't have time to react.
Obviously the distances are different at that speed, but if the person steps out so close that you cannot react in time, you're fucked at any speed.
10mph will do serious damage still, so please for the sake of the children please slow yourself and your daughter's driving down to 0.5mph where there are pedestrians or parked cars.
But seriously I think you'd be more safe to both slow down and also to put more space between the parked cars and your car so that you are not scooting along with a 30cm of clearance - move out and leave lots of space so there is more space for sight-lines for both you and pedestrians.
multiple children in my area have died due to being hit by distracted drivers driving near schools. One incident resulted in 2 children being dragged 60 yards. Here's a snippet from an article about the death I was referencing:
> The woman told police she was “eating yogurt” before she turned onto the road and that she was late for an appointment. She said she handed her phone to her son and asked him to make a call “but could not remember if she had held it so face recognition could … open the phone,” according to the probable cause statement.
> The police investigation found that she was traveling 50 mph in a 40 mph zone when she hit the boys. She told police she didn’t realize she had hit anything until she saw the boys in her rearview mirror.
The Waymo report is being generous in comparing to a fully-attentive driver. I'm a bit annoyed at the headline choice here (from OP and the original journalist) as it is fully burying the lede.
I usually take extra care when going through a school zone, especially when I see some obstruction ('behind a tall SUV', was the waymo overtaking?), and overtaking is something I would probably never do (and should be banned in school zones by road signs).
This is a context that humans automatically have and consider. I'm sure Waymo engineers can mark spots on the map where the car needs to drive very conservatively.
> especially when I see some obstruction ('behind a tall SUV', was the waymo overtaking?)
Yep. Driving safe isn't just about paying attention to what you can see, but also paying attention to what you can't see. Being always vigilant and aware of things like "I can't see behind that truck."
Honestly I don't think sensor-first approaches are cut out to tackle this; it probably requires something more akin to AGI, to allow inferring possible risks from incomplete or absent data.
I appreciate your sensible driving, but here in the UK, roads outside schools are complete mayhem at dropping off/picking up times. Speeding, overtaking, wild manoeuvres to turn round etc.
When reading the article, my first thought was that only going at 17mph was due to it being a robotaxi whereas UK drivers tend to be strongly opposed to 20mph speed limits outside schools.
Most US states cap speed limits around schools at 15mph when children are present. There may also be blinking lights above these signs during times that will be likely.
I'm not sure how much of that Waymo's cars take into account, as the law technically takes into account line of sight things that a person could see but Waymo's sensors might not, such as children present on a sidewalk.
> Most US states cap speed limits around schools at 15mph when children are present.
Are you sure? The ones I've seen have usually been 20 or 25mph.
Looking on Image Search (https://www.google.com/search?q=school+zone+speed+limit+sign) and limiting just to the ones that are photos of real signs by the side of the road, the first 10 are: 25, 30, 25, 20, 35, 15, 20, 55, 20, 20. So only one of these was 15.
School pick up and drop off traffic is just about the worst drivers anywhere. Like visibly worse than a bunch of "probably a little drunk" people leaving a sports stadium. It's like everyone reverts to "sixteen year old on first day behind the wheel" behavior. It's baffling. And there's always one token dad picking up his kid on a motorcycle or in a box truck or something that they all clutch their pearls at.
This exact scenario happened with my dad 50 years ago when a little girl ran out to the street from between some parked cars. It's an extremely difficult scenario to avoid an accident in.
> Does Waymo have the same object permanence and trajectory prediction (combined) to that of a human?
On this note specifically ive actually been impressed, ie when driving down Oak st in SF (fast road, tightly parked cars) I've often observed it slow if someone on a scooter on the sidewalk turns to look toward oncoming traffic (as if to start riding), or to slow passing parked box trucks (which block vision of potential pedestrians)
>It's likely that a fully-attentive human driver would have done worse.
Maybe. Depends on the position of the sun and shadows, I'm teaching my kids how to drive now and showing them that shadows can reveal human activity that is otherwise hidden by vehicles. I wonder if Waymo or other self-driving picks up on that.
So you say, with hindsight. Regardless, what you would do is irrelevant. The question is: what would the ~average driver do? If you tell me the average driver goes <=14mph in that situation, I'm gonna look at you funny.
A human driver in a school zone during morning drop off would be scanning the sidewalks and paying attention to children that disappear behind a double parked suv or car in the first place, no?
As described by the nhtsa brief:
"within two blocks of a Santa Monica, CA elementary school during normal school drop off hours; that there were other children, a crossing guard, and several double-parked vehicles in the vicinity"
The "that there were other children, a crossing guard, and several double-parked vehicles in the vicinity" means that waymo is driving recklessly by obeying the speed limit here (assuming it was 20mph) in a way that many humans would not.
It's possible, but likely is a heavy assertion. It's also possible a human driver would have been more aware of children being present on the sidewalk and would have approached more cautiously given obstructed views.
Please please remember that any data from Waymo will inherently support their position and can not be taken at face value. They have significant investment in making this look more favorable for them. They have billions of dollars riding on the appearance of being safe.
I wonder if that is a "fully attentive human drive who drove exactly the same as the Waymo up until the point the child appeared"?
Personally, I slow down and get extra cautious when I know I am near a place where lots of kids are and sight lines are poor. Even if the area is signed for 20 I might only be doing 14 to begin with, and also driving more towards the center of the road if possible with traffic.
I do the same, and try to actively anticipate and avoid situations like this. Sadly, in my experience most drivers instead fixate on getting to their destination as fast as possible.
You clearly don't spend much time around a school measuring the speed of cars. Head on down and see for yourself how often or not a human driver goes >17mph in such a situation.
> It's likely that a fully-attentive human driver would have done worse.
> a huge portion of human drivers
What are you basing any of these blind assertions off of? They are not at all born out by the massive amounts of data we have surrounding driving in the US. Of course Waymo is going to sell you a self-serving line but here on Hacker News you should absolutely challenge that. In particular because it's very far out of line with real world data provided by the government.
Your data that shows 30% of fatal crashes involve alcohol, not to mention any of the other factors I named? Seems like your data supports my conclusion!
Again, I welcome you to point to data that contradicts my claims, but it seems you are unable
What data completely disagrees with them and what does it disagree with them about?
The "Persons Killed, by Highest Driver Blood Alcohol Concentration (BAC) in the Crash"[1] report shows that in 2023, 30% of fatal crashes involved at least one driver with a BAC > 0.08 (the legal limit), and 36% involved a BAC > 0.01.
Interesting that "Non-motorist" fatalities have dropped dramatically for everyone under the age of 21, but increased for everyone between 21 and 74.[2] Those are raw numbers, so it'd be even more interesting to display them as a ratio of the group's size. Are less children being killed by drivers because there are less children generally? Changes in parents' habits? Backup cameras?
Waymo is intentionally leaving out the following details:
- Their "peer-reviewed model" compares Waymo vehicles against only "Level 0" vehicles. However even my decade-old vehicle is considered "Level 1" because it has an automated emergency braking system. No doubt my Subaru's camera-based EBS performs worse than Waymo's, still it's not being included in their "peer-reviewed model." That comparison is intentionally comparing Waymo performance against the oldest vehicles on the road -- not the majority of cars sold currently.
- This incident happened during school dropoff. There was a double-parked SUV that occluded the view of the student. This crash was the fault of that double-parked driver. But why was the uncrewed Waymo driving at 17 mph to begin with? Do they not have enough situational awareness to slow the f*ck down around dropoff time immediately near an elementary school?
Automotive sensor/control packages are very useful and will be even more useful over time -- but Waymo is intentionally making their current offering look comparatively better than it actually is.
Emergency braking in non-camera/non-LIDAR cars requires a significant radar signal which you're only going to get from another vehicle (and even then it's noisy and tends to produce frustrating false positives, leading to later-than-you-want stops). It very likely won't detect a child or a dog, I'm not aware of a single instance of an EBS claiming to have done so in practice (and kids and dogs get hit every day!).
It depends on the situation, and we need more data/video. But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast, and the Waymo should have been driving more conservatively.
> But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast, and the Waymo should have been driving more conservatively.
UK driving theory test has a part called Hazard Perception: not reacting on children milling around would be considered a fail.
Many states in the US have the Basic Speed Law, e.g. California:
> No person shall drive a vehicle upon a highway at a speed greater than is reasonable or prudent having due regard for weather, visibility, the traffic on, and the surface and width of, the highway, and in no event at a speed which endangers the safety of persons or property.
The speed limit isn't supposed to be a carte blanche to drive at that speed no matter what; the basic speed law is supposed to "win." In practice, enforcement is a lot more clear cut at the posted speed limit and officers don't want to write tickets that are hard to argue in court.
That law seems more likely to assign blame to drivers if they hit someone. So practically it's not enforced but in accidents it becomes a justification for assigning fault.
I mean yeah. If you were traveling at some speed and caused damage to persons or property, that's reasonable, but refutable, evidence that you were traveling at a speed that endangered persons or property.
And at the same time, if you were traveling at some speed and no damage was caused, it's harder to say that persons or property were endangered.
Exactly. That’s why I’ve always said the driving is a truly AGI requiring activity. It’s not just about sensors and speed limits and feedback loops. It’s about having a true understanding for everything that’s happening around you:
Having an understanding for the density and make up of an obstacle that blew in front of you, because it was just a cardboard box. Seeing how it tumbles lightly through the wind, and forming a complete model of its mass and structure in your mind instantaneously. Recognizing that that flimsy fragment though large will do no damage and doesn’t justify a swerve.
Getting in the mind of a car in front of you, by seeing subtle hints of where the driver is looking down, and recognizing that they’re not fully paying attention. Seeing them sort of inch over because you can tell they want to change lanes, but they’re not quite there yet.
Or in this case, perhaps hearing the sounds of children playing, recognizing that it’s 3:20 PM, and that school is out, other cars, double parked as you mentioned, all screaming instantly to a human driver to be extremely cautious and kids could be jumping out from anywhere.
How many human drivers do you think would pass the bar you're setting?
IMO, the bar should be that the technology is a significant improvement over the average performance of human drivers (which I don't think is that hard), not necessarily perfect.
> How many human drivers do you think would pass the bar you're setting?
How many humans drivers would pass it, and what proportion of the time? Even the best drivers do not constantly maintain peak vigilance, because they are human.
> IMO, the bar should be that the technology is a significant improvement over the average performance of human drivers (which I don't think is that hard), not necessarily perfect.
In practice, this isn't reasonable, because "hey we're slightly better than a population that includes the drunks, the inattentive, and the infirm" is not going to win public trust. And, of course, a system that is barely better than average humans might worsen safety, if it ends up replacing driving by those who would normally drive especially safe.
I think "better than the average performance of a 75th or 90th percentile human driver" might be a good way to look at things.
It's going to be a weird thing, because odds are the distribution of accidents that do happen won't look much like human ones. It will have superhuman saves (like that scooter one), but it will also crash in situations that we can't really picture humans doing.
I'm reminded of airbags; even first generation airbags made things much safer overall, but they occasionally decapitated a short person or child in a 5MPH parking lot fender bender. This was hard for the public to stomach, and if it's your kid who is internally decapitated by the airbag in a small accident, I don't think you'll really accept "it's safer on average to have an airbag!"
The parent comment said the bar should be "significant improvement" over the average performance of human drivers.
Then you said, "this isn't reasonable", and the bar shouldn't be "slightly better" or "barely better". It should be at least better than the 75th percentile driver.
It sounds like you either misread the parent comment or you're phrasing your response as disagreement despite proposing roughly the same thing as the parent comment.
All depends on what you read as "significant improvement".
A 20% lower fatal crash rate compared to the average might be a significant improvement-- from a public health standpoint, this is huge if you could reduce traffic deaths by 20%.
But if you don't get the worst drivers to replace their driving with autonomous, that "20% less than average" might actually make things worse. That's my point. The bar has to be pretty dang high to be sure that you will actually make things better.
> In practice, this isn't reasonable, because "hey we're slightly better than a population that includes the drunks, the inattentive, and the infirm" is not going to win public trust.
Sadly, you're right, but as rational people, we can acknowledge that it should. I care about reducing injuries and deaths, and the %tile of human performance needed for that is probably something like 30%ile. It's definitely well below 75%ile.
> > And, of course, a system that is barely better than average humans might worsen safety, if it ends up replacing driving by those who would normally drive especially safe.
It's only if you get the habitually drunk (a group that is overall impoverished), the very old, etc, to ride Waymo that you reap this benefit. And they're probably not early adopters.
Uber and Lyft were supported by police departments because they reduced drunk driving. Drunk driving isn't just impoverished alcoholics. People go to bars and parts and get drunk all the time.
You also solve for people texting (or otherwise using their phones) while driving, which is pretty common among young, tech-adopting people.
> Drunk driving isn't just impoverished alcoholics. People go to bars and parts and get drunk all the time
Yes, but the drivers who are 5th percentile drivers who cause a huge share of the most severe accidents are "special" in various ways. Most of them are probably not autonomy early adopters.
The guy who decided to drive on the wrong side of a double yellow on a windy mountain road and hit our family car in a probable suicide attempt was not going to replace that trip with Waymo.
> But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast
Hey, I'd agree with this-- and it's worth noting that 17^2 - 5^2 > 16^2, so even 1MPH slower would likely have resulted in no contact in this scenario.
But, I'd say the majority of the time it's OK to pass an elementary school at 20-25MPH. Anything carries a certain level of risk, of course. So we really need to know more about the situation to judge the Waymo's speed. I will say that generally Waymo seems to be on the conservative end in the scenarios I've seen.
(My back of napkin math says an attentive human driver going at 12MPH would hit the pedestrian at the same speed if what we've been told is accurate).
Swedish schools still have students who walk there. I live near one and there are very few cars that exceed 20km/h during rush hours. Anything faster is reckless even if the max over here is 30 km/h (19 mph).
The schools I'm thinking of have sidewalks with some degree of protection/offset from street, and the crossings are protected by human crossing guards during times when students are going to schools. The posted limits are "25 (MPH) When Children Are Present" and traffic generally moves at 20MPH during most of those times.
There are definitely times and situation where the right speed is 7MPH and even that feels "fast", though, too.
> Hey, I'd agree with this-- and it's worth noting that 17^2 - 5^2 > 16^2, so even 1MPH slower would likely have resulted in no contact in this scenario.
Only with instant reaction time and linear deceleration.
Neither of those are the case. It takes time for even a Waymo to recognize a dangerous situation and apply the brake and deceleration of vehicles is not actually linear.
> It takes time for even a Waymo to recognize a dangerous situation
Reaction time makes the math even better here. You travel v1 * reaction_time no matter what, before entering the deceleration regime. So if v1 gets smaller, you get to spend a greater proportion of time in the deceleration regime.
> linear deceleration.
After reaction time, stopping distance is pretty close to n^2. There's weird effects at high speed (contribution from drag) and at very low speed, but they have pretty modest contributions.
I think the strategy is a lot more nuanced than that.
In any case, with zero reaction time, linear deceleration time to stop is proportional to velocity squared. With reaction time, the linear deceleration time is that plus the velocity times the reaction time.
so the two cases we're comparing are 17 * r + (17^2 - 5^2) vs. 16 * r + (16^2), or 17 * r + 264 vs 16 * r + 256. As long as reaction time isn't negative, a vehicle that could slow to 5MPH starting at 17MPH could slow to 0MPH starting at 16MPH.
(There are weird things that happen at <2.5MPH reducing deceleration to sublinear, but the car moves only a few inches at these speeds during a panic stop).
Edit: Not 'allowed' but people do it constantly. Regular drivers, delivery drivers, city workers, construction trucks, etc. There may be laws but very little enforcement.
I bet we'll the the SUV mania in the future as something crazy, like smoking in a plane or using lead for gasoline. Irrational large size cars that people get because everyone it's afraid of another SUV hitting them in a sedan. The tragedy of the commons.
What I find a bit confusing is that no one is putting any blame on the kid. I did the same thing as a kid, except it was a school bus instead of SUV, and that was a fucking stupid thing to do (I remember starting to run over the street, and the next thing is that I am in the hospital bed), even though I had been told to always cross the street from behind the bus, not in front of it.
Of course the kid is at fault. But everyone knows that kids do stupid and reckless things, which is why drivers are generally expected to take more care around schools and similar institutions. If robotaxis are not able to do that, then the results will be easy to predict
The best reaction from Waymo would have been to start to lobby against letting those monster-trucks park on streets near schools. They are killing so many children, I'm flabbergasted they are still allowed outside of worksites.
From a "my opinion" standpoint, yes, I would love to see this.
From a tactical PR standpoint, it would be a disaster. Muh big truuuucks is like a third rail because Americans are car obsessed as a culture. They already hit a kid, best to save some energy for the next battle.
Besides if Waymo wins (in general) private car ownership will decrease which is a win regardless. And maybe Waymo can slowly decrease the size of their fleet to ease up the pressure on this insane car size arms race.
AV’s with enough sensing are generally quite good at stopping quickly. It is usually the behavior prior to the critical encounter that has room for improvement.
The question will be whether 17 mph was a reasonably cautious speed for this specific scenario. Many school zones have 15 mph limits and when there are kids about people may go even slower. At the same time, the general rule in CA for school zone is 25 mph. Clearly the car had some level of caution which is good.
It does sound like a good outcome for automation. Though I suppose an investigation into the matter would arguably have to look at whether a competent human driver would be driving at 17mph (27km/h) under those circumstances to begin with, rather than just comparing the relative reaction speeds, taking the hazardous situation for granted.
What I would like to see is a full-scale vehicle simulator where humans are tested against virtual scenarios that faithfully recreate autonomous driving accidents to see how "most people" would have acted in the minutes leading up to the event as well as the accident itself
Indeed, 15 or 25 mph (24 or 40 km/h) are the speed limits in school zones (when in effect) in CA, for reference. But depending on the general movement and density and category of pedestrians around the road it could be practically reckless to drive that fast (or slow).
If my experience driving through a school zone on my way to work is anything to go off of, I rarely see people actually respecting it. 17 mph would be a major improvement over what I'm used to seeing.
>Though I suppose an investigation into the matter would arguably have to look at whether a competent human driver would be driving at 17mph (27km/h) under those circumstances to begin with, rather than just comparing the relative reaction speeds, taking the hazardous situation for granted.
Sure but also throw in whether that driver is staring at their phone, distracting by something else, etc. I have been a skeptic of all this stuff for a while but riding in a Waymo in heavy fog changed my mind when questioning how well I or another driver would've done at that time of day and with those conditions.
For me it would be interesting to know if 17 mi/h was a reasonable speed to be driving in this environment under these conditions to begin with. In my school zones that's already close to the maximum speed allowed. What was the weather, were there cars parked which would make a defensive driver slow down even more?
The autonomous vehicle should know what it can't know, like children coming out from behind obstructions. Humans have this intuitive sense. Apparently autonomous systems do not, and do not drive carefully, or slower, or give more space, in those situations. Does it know that it's in a school zone? (Hopefully.) Does it know that school is starting or getting out? (Probably not.) Should it? (Absolutely yes.)
This is the fault of the software and company implementing it.
Some do, some of the time. I'm always surprised by how much credence other people give to the idea that humans aren't on average very bad at things, including perception.
I'm not sure what your point is here since this was better than many humans. Was it better than all humans? No. But there also isn't a single human who's better than all humans.
So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
That’s why they purchase goods and services (from others) and then cry about things they don’t and probably never will understand.
And why they can be ignored and just fed some slop to feel better.
I could lie but that’s the cold truth.
Edit: I'm not sure if the repliers are being dense (highly likely), or you just skipped over context (you can click the "context" link if you're new here)
> So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
That is the general public sentiment I was referring to.
So if they were 100% self-sufficient and understood everything they'd be smart enough to interpret a child being hit at 6 mph as progress? Fun how "general public" is always a "they" vs "you".
Your comment sounds like subconsciously you're trying to come off as stronger than the general public, which begs the question: Why? Why do you need to prove your strength over the populace?
are you being dense on purpose, or you just don't understand how context works? hint, check out the "context" link
look at what I was replying to. if you still don't get it, then yeah I'm just proving my point and you can keep crying about it.
> So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
That is the general public sentiment I was referring to.
The fact that you go around asking dumb questions in bad faith to people is enough for me, last time I engage with you.
Immediately hitting the brakes when a child suddenly appears in front of you, instead of waiting 500ms like a human, and thereby hitting the child at a speed of 6 instead of 14 is a success.
What else to you expect them to do, only run on grade–separated areas where children can't access? Blare sirens so children get scared away from roads? Shouldn't human–driven cars do the same thing then?
I don't know the implementation details, but success would be not hitting pedestrians. You have some interesting ideas on how to achieve that but there might be other ways, I don't know.
>I don't know the implementation details, but success would be not hitting pedestrians.
So by that logic, if we cured cancer but the treatment came with terrible side effects it wouldn't be considered a "success"? Does everything have to perfect to be a success?
The limit is 20 MPH in Washington state, in California the default is 25 MPH, but is going to 20 MPH soon and can be further lowered to 15 MPH with special considerations.
The real killer here is the crazy American on street parking, which limits visibility of both pedestrians and oncoming vehicles. Every school should be a no street parking zone. But parents are going to whine they can't load and unload their kids close to the school.
On street parking is so ingrained into the American lifestyle that any change to the status quo is impossible. Cars have more rights on public property than people. Every suburban neighborhood has conflicts over people's imagined "ownership" of the street parking in front of their house. People rarely use their garages to store their car since they can just leave it on the street. There are often laws that prevent people from other neighborhoods from using the public street to park. New roads are paved as wide as possible to allow both street parking and a double-parked car to not impede traffic. And we've started building homes without any kind of parking that force people to use the street.
Europe is much better at this than we are. Even when you have on street parking, they make sure there are clearances around cross walks and places where there are lots of pedestrians. Most US cities don't even care, even a supposedly pedestrian friendly one like Seattle.
Lol tell that to my city. They removed the cross walks and declared the zone a "shared zone" where pedestrians have right of way.
Result? No more safe places to cross, drivers are not stopping for pedestrians when no cross walk. They added parking zones right up to the old cross walks that pedestrians still use (since it were the safest places) where vans are regularly parked and obscure the entirety one side of the road. Even outside of these shared zones, there are lots lots lots of places where parking space is obscuring the crosswalk, where huge vans park right on it, even though legally you have to be 5+ meters away. Never seen a cop give a ticket for that in my life.
One day someone will get killed right there, I'm sure of it, and it'll be mainly the city's fault.
Impossible is probably the wrong word. But where I live, a superficially "progressive" area, many of these traffic calming, road diet, etc. measures are met with regular opposition.
If it had no parking, then the parents would be parked somewhere else and loading and unloading their kids there, and then that would need to be a no-parking zone too.
I guess you could keep doing that until kids just walk to and from school?
Our local school has them unload a block away unless they are handicapped. A kid isn't going to die walking a block. But its pointless because they still allow residential on street parking around the school, and my son has to use a crosswalk where cars routinely park so close to, I had to tell him that the traffic (pretty heavy) on the road wouldn't see him easily, and he should always ease his way into a crosswalk and not assume he would be easily seen.
In the UK we have a great big yellow zig-zag road marking that extends 2/3rds the width of an average car across the road. It means "this is a school, take your car and fuck off". You find it around school gates, to a distance of a few car lengths either side of the gate, and sometimes all along the road beside a school.
It doesn't stop all on street parking beside the school, but it cuts it down a noticeable amount.
For a school near me, the road is no parking during pick up/drop off times. It even changes to one way traffic. The no parking windows is similar to alternate street sweeping days. There are signs posted that indicate the times.
Same for my tiny town. Stopping on the road is 100% not allowed, and parking isn't allowed there either. The school has its own parking area to park and pick up/drop off kids, and cars in there creep at 2 or 3 MPH.
how about 10-15 mph if directly adjacent to a school, especially during the bands before and after school stars or ends. route away from schools whenever feasible.
A vehicle can't go 10 to 15 miles per hour at the same time. If it was going 15, then it should have been going 10. Or driving further away from occluded spaces. And again, routing away from schools.
The simple fact is that it hit a child and even though it wasn't a serious issue due to their safety policies, there's still room for improvement in these technologies.
And since it's a robot, and not a human, you can actually make changes and have them stick. For example, routing away from schools during certain hours.
This isn't Apollo 13 with a successful failure. A driverless car hit a human that just happened to be a kid. Doesn't matter if a human would have as well, the super safe driverless car hit a kid. Nothing else matters. Driverless car failed.
Tesla report ids from SGO-2021-01_Incident_Reports_ADAS.csv with no or unknown airbag deployment status: 13781-13330, 13781-13319, 13781-13299, 13781-13208, 13781-8843, 13781-13149, 13781-13103, 13781-13070, 13781-13052... and more
Being transparent about such incidents is also what stops them from potentially becoming a business/industry-killing failures. They're doing the right thing here, but they also surely realize how much worse it would be if they tried to deny or downplay it.
Was it unpredictable? They drove past a blind corner (parked SUV) in a school zone. I'm constantly slowing down in these situations as I expect someone might run out at any second. Waymo seemed to default to the view that if it can't see anyone then nobody is there.
> I honestly cannot imagine a better outcome or handling of the situation.
It's the "best outcome" if you're trying to go as fast as possible without breaking any laws or ending up liable for any damage.
German perspective, but if I told people I've been going 30km/h next to a school with poor visibility as children are dropped off around me, I would be met with contempt for that kind of behavior. I'd also at least face some partial civil liability if I hit anyone.
There's certainly better handling of the situation possible, it's just that US traffic laws and attitudes around driving do not encourage it.
I suspect many human drivers would've driven slower, law or no law.
I look for shadows underneath stationary vehicles. I might also notice pedestrians "vanishing". I have a rather larger "context" than any robot effort.
However, I am just one example of human. My experience of never managing to run someone over is just an anecdote ... so far. The population of humans as a whole manages to run each other over rather regularly.
A pretty cheap instant human sensor might be Bluetooth/BLE noting phones/devices in near range. Pop a sensor in each wing mirror and on the top and bottom. The thing would need some processing power but probably nothing that the built in Android dash screen couldn't handle.
There are lots more sensors that car manufacturers are trying to avoid for cost reasons, that would make a car way better at understanding the context of the world around it.
I gather that Tesla insist on optical (cameras) only and won't do LIDAR. My EV has four cameras and I find it quite hard to see what is going on when it is pissing down with rain, in the same way I do if I don't clean my specs.
I'll just remind anyone reading: they're under no obligation to tell the unvarnished truth on their blog.
Even if the NHTSA eventually points out significant failures, getting this report out now has painted a picture of Waymo only having an accident a human would have handled worse.
It would be wise to wait and see if the NHTSA agree. Would a driver have driven at 17mph in this sort of traffic or would they have viewed it as a situation where hidden infant pedestrians are likely to step out?
I'm picturing a 10 second clip showing a child with a green box drawn around them, and position of gas and brake, updating with superhuman reactions.
That would be the best possible marketing that any of these self driving companies could hope for, and Waymo probably now has such a video sitting somewhere.
It depends on the video. What they should do is arrange for the video to get leaked and let the Internet courts argue about it, and then based on the Internet verdict, come out and claim it's real and they fired somebody for leaking it, or it's AI generated.
Love him or hate him, releasing the video is something I can see Elon doing because assuming a human driver would have done worse, it speaks for itself. Release a web video game where the child sometimes jumps out in front of the car, and see how fast humans respond like the "land Starship" game. Assuming humans would do worse, that is. If the child was clearly visible through the car or some how else avoidable by humans, then I'd be hiding the video too.
I honestly think that Waymo's reaction was spot on. I drop off and pick up my kid from school every day. The parking lots can be a bit of a messy wild west. My biggest concern is the size of cars especially those huge SUV or pickup trucks that have big covers on the back. You can't see anything incoming unless you stick your head out.
When I was a boy, I ran into the street from between two parked cars. I did not notice the car coming, but he noticed me popping out from nowhere, and screeched to a stop.
I saw a girl dart out between to parked cars on a strode. She was less lucky. The car did slam on their breaks. I have no idea what speed it was ultimately going when they hit the girl. It wasn't enough to send her flying but it was enough to knock her over hard. The dad, was sitting in his front yard and had her up and in his car and I'm guessing rushed to the hospital.
Those kind of neighborhoods where the outer houses face the fast large roads I think are less common now but lots of them left over from the 50+ years ago.
I once rounded a blind curve on a non-residential street only to find a man on a bicycle pulling a trailer with his baby in it, stopped in the middle of the road. I stopped and yelled at him, which surprised him.
Whatever our human laws and morality says about right and wrong and fault, the laws of physics usually judges the car a winner when it hits somebody.
Placing yourself somewhere where pedestrians are not expected (non-residental road) mostly hidden from oncoming traffic for an extended period is putting yourself in undue risk.
You don't always have a choice about where you are momentarily and anybody turning a blind corner has an obligation to immediately reduce their speed (prior to turning the corner!) to where they can safely come to a stop without endangering others. That's drivers education 101. Right after 'don't text while driving', 'don't drink while driving' and 'slow down when there are pedestrians, bicycles and other fragile road users around'.
If the person got up and walked away I'm not sure what damage you'd be doing by reasonably removing your car from blocking others while waiting for police.
Waymo driver? The vehicles are autonomous. I otherwise applaud Waymo's response, and I hope they are as cooperative as they say they will be. However, referring to the autonomous vehicle as having a driver is a dangerous way to phrase it. It's not passive voice, per se, but it has the same effect of obscuring responsibility. Waymo should say we, Waymo LLC, subsidiary of Alphabet, Inc., braked hard...
Importantly, Waymo takes full ownership for something they write positively: Our technology immediately detected the individual.... But Waymo weasels out of taking responsibility for something they write about negatively.
the "Waymo Driver" is how they refer to the self-driving platform (hardware and software). They've been pretty consistent with that branding, so it's not surprising that they used it here.
> Importantly, Waymo takes full ownership for something they write positively [...] But Waymo weasels out of taking responsibility for something they write about negatively
Pretty standard for corporate Public Relations writing, unfortunately.
In fact I would call that “superhuman” behavior across the board.
The vast vast vast majority of human drivers would not have been able to accomplish that braking procedure that quickly, and then would not have been able to manage the follow up so quickly.
I have watched other parent drivers in the car pick up line at public schools for the last 16 years and people are absolutely trash at navigating that whole process and parents drive so poorly it’s absurd. At least half parents I see on their phones while literally feet away from hitting some kid.
Human drivers are smart enough to slow down when around a school where kids are being dropped of. This piece of software wasn't. Clearly not superhuman.
How do you know how quickly the software braked? A blog post by a company selling a product is not credible material. We need independent sources.
> The vast vast vast majority of human drivers ... would not have been able to manage the follow up so quickly
You are saying the "vast vast vast majority of human drivers" wouldn't pull over after hitting a child?
I remember similar blind faith in and unlimited advocacy for anything Tesla and Musk said, and look how that has turned out. These are serious issues for the people in our communities, not a sporting event with sides.
Yeah. I'm a stickler for accountability falling on drivers, but this really can be an impossible scenario to avoid. I've hit someone on my bike in the exact same circumstance - I was in the bike lane between the parked cars and moving traffic, and someone stepped out between parked vehicles without looking. I had nowhere to swerve, so squeezed my brakes, but could not come to a complete stop. Fortunately, I was going slow enough that no one was injured or even knocked over, but I'm convinced that was the best I could have done in that scenario.
The road design there was the real problem, combined with the size and shape of modern vehicles that impede visibility.
Building on my own experience I think you have to own that if you crash with someone you made a mistake. I do agree that car and road design for bicycles(?) makes it almost impossible to move around if you do not risk things like that.
This is the classic Suddenly Revealed Pedestrian test case, which afaik, most NCAP (like EuroNCAP, Japan NCAP) have as part of their standard testing protocols.
Having performed this exact test on 3 dozen vehicles (L2/L3/L4) for several AV companies in the Bay Area [1], I would say that Waymo's response, per their blog post [2] has been textbook compliance. (I'm not defending their performance... just their response to the collision). This test / protocol is hard for any driver (including human driven vehicles), let alone ADAS/L3/L4 vehicles, for various reasons, including: pedestrian occlusion, late ped detection, late braking, slick roads, not enough braking, etc. etc.
Having said all that, full collision avoidance would have been best outcome, which, in this case, it wasn't. Wherever the legal fault may lie -- and there will be big debate here -- Waymo will still have to accept some responsibility, given how aggressively they are rolling out their commercial services.
This only puts more onus on their team to demonstrate a far higher standard of driving than human drivers. Sorry, that's just the way societal acceptance is. We expect more from our robots than from our fellow humans.
Waymo’s performance, once the pedestrian was revealed, sounds pretty good. But is 17mph a safe speed at an active school dropoff area? I admit that I don’t think I ever personally pay attention to the speedometer in such a place, but 17mph seems excessive even for an ordinary parking lot.
I wonder whether Waymo’s model notices that small children are present or likely to be present and that it should leave extra margin for error.
(My general impression observing Waymo vehicles is that they’ve gone from being obnoxiously cautious to often rather aggressive.)
> But is 17mph a safe speed at an active school dropoff area?
Now you're asking interesting questions... Technically, in CA, the speed limit in school zones are 25 mph (which local authorities can change to 15 mph, as needed). In this case, that would be something the investigation would check, of course. But regardless of that, 17 mph per se is not a very fast speed (my gut check: turning around intersections at > 10-11 mph feels fast, but going straight at 15-20 mph doesnt feel fast; YMMV). But more generally, in the presence of child VRUs (vulnerable road users), it is prudent to drive slowly just because of the randomness factor (children being the most unaware of critters). Did the Waymo see the kids around in the area? If so, how many and where? and how/where were they running/moving to? All of that is investigation data...
My 2c is that Waymo already took all of that into account and concluded that 17 mph was indeed a good speed to move at...
...which leads to your observation below:
> (My general impression observing Waymo vehicles is that they’ve gone from being obnoxiously cautious to often rather aggressive.)
Yes, I have indeed made that same observation. The Waymos of 2 years ago were very cautious; now they seem much more assertive, even a bit aggressive (though that would be tough to define). That is a driving policy decision (cautious vs assertive vs aggressive).
One could argue if indeed 17 mph was the "right" decision. My gut feel is Waymo will argue that (but likely they might make the driving policy more cautious esp in presence of VRUs, and child VRUs particularly)
> Technically, in CA, the speed limit in school zones are 25 mph
Legally a speed limit is a 'limit' on speed, not a suggested or safe speed. So it's never valid to argue legally that you were driving under the limit, the standard is that you slow down or give more room for places like a school drop-off while kids are being dropped off or picked up.
In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
> In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
That's a difficult question to answer, and the devil really is in the details, as you may have guessed. What I can say that Waymo is, by far, the most prolific publisher of research on AV safety on public roads. (yes, those are my qualifiers...)
Here's their main stash [1] but notably, three papers talk about comparison of Waymo's rider-only (i.e. no safety driver) performance vis-a-vis human driver, at 7.1 million miles [2], 25 million miles [3], 56 million miles [4]. Waymo has also been a big contributor to various AV safety standards as one would expect (FWIW, I was also a contributor to 3 of the standards... the process is sausage-making at its finest, tbh).
I haven't read thru all their papers, but some notable ones talk about the difficulty of comparing AV vs human drivers [5], and various research on characterising uncertainty / risk of collision, comparing AVs to non-impaired, eyes-on human driver [6]
As one may expect, at least one of the challenges is that human-driven collisions are almost always very _lagging indicators_ of safety (i.e. collision happened: lost property, lost limbs, lost lives, etc.)
So, net-net, Waymo still has a VERY LONG WAY to go (obviously) to demonstrate better than human driving behavior, but they are showing that their AVs are better-than-humans on certain high-risk (potential) collisions.
As somebody remarked, the last 1% takes 90% of time/effort. That's where we are...
There are many cases when the owner wouldn’t be liable as well, like if the victim was performing an illegal act like attacking the owner or dog, or trespassing. If a child isn’t following the law or being supervised by a parent, some consequences are inevitable and the driver isn’t instantly liable. For example, if a student jumps in front of a car in an attempted suicide, it would be very hard for a driver to avoid that in certain situations.
In a technical sense, maybe, but it's all going to be about optics. They have a responsibility to handle the situation well even if it's not their fault, and the public will hold them accountable for what they deem the involvement was, which may not be the actual scenario.
> In a technical sense, maybe, but it's all going to be about optics.
Indeed, it is, and that is exactly why Waymo will have to accept some responsibility. I can bet that internally Waymo's PR and Legal teams are working overtime to coordinate the details with NHTSA. We, the general public, may or may not know the details at all, if ever.
However, Waymo's technical teams (Safety, etc) will also be working overtime to figure out what they could have done better.
As I mentioned, this is a standard test, and Waymo likely has 1000s of variations of this test in their simulation platforms; they will sweep across all possible parameters to make this test tighter, including the MER (minimum expected response from the AV) and perhaps raise the bar on MER (e.g. brake at max deceleration in some cases, trading off comfort metrics in those cases; etc.) and calculate the effects on local traffic (e.g. "did we endanger the rear vehicles by braking too hard? If so, by how much??" etc). All these are expected actions which the general public will never know (except, perhaps via some technical papers).
Regardless, the PR effects of this collision do not look good, especially as Waymo is expanding their service to other cities (Miami just announced; London by EOY2026). This PR coverage has potential to do more damage to the company than the actual physical damage to the poor traumatized kid and his family. THAT is the responsibility only the company will pay for.
To be sure, my intuition tells me this is not the last such collision. Expect to see some more, by other companies, as they commercialize their own services. It's a matter of statistics.
The performance of a human is inherently limited by biology, and the road rules are written with this in mind. Machines don't have this inherent limitation, so the rules for machines should be much stronger.
I think there is an argument for incentivising the technology to be pushed to its absolute limits by making the machine 100% liable. It's not to say the accident rate has to be zero in practice, but it has to be so low that any remaining accidents can be economically covered by insurance.
At least in the interim, wouldn’t doing what you propose cause more deaths if robot drivers are less harmful than humans, but the rules require stronger than that? (I can see the point in making rules stronger as better options become available, but by that logic, shouldn't we already be moving towards requiring robots and outlawing human drivers if it's safer?)
> In your experience, where do we find a credible source of info? Do we need to wait for the government's investigation to finish?
Most likely, yes, the NHTSA investigation will be credible source of info for this case. HOWEVER, Waymo will likely fight it tooth-and-nail from letting it be public. They will likely cite "proprietary algorithms / design", etc. to protect it from being released publicly. So, net-net, I dunno... Will have to wait and see :shrug.gif:
But meanwhile, personally I would read reports from experts like Phil Koopman [1] and Missy Cummings [2] to see their take.
> Remember Tesla's blog posts?
You, Sir, cite two companies that are diametrically opposite on the safety spectrum, as far as good behavior is concerned. Admittedly, one would have less confidence in Waymo's own public postings about this (and I'd be mighty surprised if they actually made public their investigation data, which would be a welcome and an pioneering move).
On the other hand, the other company you mentioned, the less said the better.
There is already widespread discussion on LinkedIn about this thread... but usefully, here [1] is the NHTSA's Office of Defects Investigation report. Nothing much new there, tbh.
As I did suspect, legal scholars are already calling for "voluntary disclosure" from Waymo re: its annotated videos of the collision [2]. FWIW, my skepticism about Waymo actually releasing it remains...
“The event occurred when the pedestrian suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path. Our technology immediately detected the individual as soon as they began to emerge from behind the stopped vehicle. The Waymo Driver braked hard, reducing speed from approximately 17 mph to under six mph before contact was made,” a statement from Waymo explains.
1. I often see signs in such areas that flash when people exceed the limit. I’d urge you to pull over and see how often humans drive above the limit.
2. I’d urge you to also pull over and watch for how many drivers are not consistently looking at the road, such as using their phones, looking down at climate/entertainment/vehicle controls, looking at a passenger, etc
Meanwhile the news does not report the other ~7,000 children per year injured as pedestrians in traffic crashes in the US.
I think the overall picture is a pretty fantastic outcome -- even a single event is a newsworthy moment _because it's so rare_ .
> The NHTSA’s Office of Defects Investigation is investigating “whether the Waymo AV exercised appropriate caution given, among other things, its proximity to the elementary school during drop off hours, and the presence of young pedestrians and other potential vulnerable road users.”
Meanwhile in my area of the world parents are busy, stressed, and on their phones, and pressing the accelerator hard because they're time pressured and feel like that will make up for the 5 minutes late they are on a 15 minute drive... The truth is this technology is, as far as i can tell, superior to humans in a high number of situations if only for a lack of emotionality (and inability to text and drive / drink and drive)... but for some reason the world wants to keep nit picking it.
A story, my grandpa drove for longer than he should have. Yes him losing his license would have been the optimal case. But, pragmatically that didn't happen... him being in and using a Waymo (or Cruise, RIP) car would have been a marginal improvement on the situation.
Err, that is not the desirable statistic you seem to think it is. American drivers average ~3 trillion miles per year [1]. That means ~7000 child pedestrian injurys per year [2] would be ~1 per 430 million miles. Waymo has done on the order of 100-200 million miles autonomously. So this would be ~2-4x more injurys than the human average.
However, the child pedestrian injury rate is only a official estimate (it is possible it may be undercounting relative to highly scrutinized Waymo vehicle-miles) and is a whole US average (it might not be a comparable operational domain), but absent more precise and better information, we should default to the calculation of 2-4x the rate.
I suspect that highway miles heavily skew this statistic. There's naturally far fewer pedestrians on highways (lower numerator), people travel longer distances on highways (higher denominator), and Waymo vehicles didn't drive on highways until recently. If you look only at non-highway miles, you'll get a much more accurate comparison.
Then you or Waymo can meet the burden of proof and present that more precise and better information. There is little reason to assume against safety at this point in time except as a intellectual exercise for how more accurate information could be found.
Until then, it is only prudent to defer snap judgements, but increase caution, insist on rigor and transparency, and demand more accurate information.
> we should default to the calculation of 2-4x the rate.
No we should not. We should accept that we don't have any statistically meaningful number at all, since we only have a single incident.
Let's assume we roll a standard die once and it shows a six. Statistically, we only expect a six in one sixth of the cases. But we already got one on a single roll! Concluding Waymo vehicles hit 2 to 4 times as many children as human drivers is like concluding the die in the example is six times as likely to show a six as a fair die.
More data would certainly be better, but it's not as bad as you suggest -- the large number of miles driven till first incident does tell us something statistically meaningful about the incident rate per mile driven. If we view the data as a large sample of miles driven, each with some observed number of incidents, then what we have is "merely" an extremely skewed distribution. I can confidently say that, if you pick any sane family of distributions to model this, then after fitting just this "single" data point, the model will report that P(MTTF < one hundredth of the observed number of miles driven so far) is negligible. This would hold even if there were zero incidents so far.
Sure, but we shouldn't stretch the analogy too far. Die rolls are discrete events, while miles driven are continuous. We expect the number of sixes we get to follow a binomial distribution, while we expect the number of accidents to follow a Poisson distribution. Either way, trying to guess the mean value of the distribution after a single incident of the event will never give you a statistically meaningful lower bound, only an upper bound.
The Poisson distribution is well approximated by the binomial distribution when n is high and p is low, which is exactly the case here. Despite the high variance in the sample mean, we can still make high-confidence statements about what range of incident rates are likely -- basically, dramatically higher rates are extremely unlikely. (Not sure, but I think it will turn out that confidence in statements about the true incident rate being lower than observed will be much lower.)
Would this Waymo incident be counted as an injury? Sounds like the victim was relatively unharmed? Presumably there are human-driver incidents like this where a car hits a child at low speeds, with effectively no injuries, but is never recorded as such?
People's standards for when they're willing to cede control over their lives both as the passenger and the pedestrian in the situation to a machine are higher than a human.
And for not totally irrational reasons like machine follows programming and does not fear death, or with 100% certainty machine has bugs which will eventually end up killing someone for a really stupid reason—and nobody wants that to be them. Then there's just the general https://xkcd.com/2030/ problem of people rightfully not trusting technology because we are really bad at it, and our systems are set up in such a way that once you reach critical mass of money consequences become other people's problem.
Washington banned automatic subway train operation for 15 years after one incident that wasn't the computer's fault, and they still make a human sit in the cab. That's the bar. In that light it's hard not to see these cars as playing fast and loose with people's safety by comparison.
>People's standards for when they're willing to cede control over their lives both as the passenger and the pedestrian in the situation to a machine are higher than a human.
Are they? It is now clear that Tesla FSD is much worse than a human driver and yet there has been basically no attempt by anyone in government to stop them.
FSD is already better than at least one class of drivers. If FSD is engaged and the driver passes out, FSD will pull over to the side of the road and stop. And before we leap to conclusions that it only helps in the case of drunk drivers who shouldn't be driving in the first place (which, they shouldn't be), random strokes and seizures happen to people all the time.
We should all think twice before taking a company PR statement completely at face value and praising them for slowing down faster than their own internal "model" says a human driver would. Companies are heavily interested in protecting their bottom line and in a situation like this probably had 5-10 people carefully craft every single word of the statement for maximum damage control.
Surprised at how many comments here seem eager to praise Waymo based off their PR statement. Sure it sounds great if you read that the Waymo slowed down faster than a human. But would a human truly have hit the child here? Two blocks from a school with tons of kids, crossing guards, double parked cars, etc? The same Waymo that is under investigation for passing school busses illegally? It may have been entirely avoidable for the average human in this situation, but the robotaxi had a blind spot that it couldn't reason around and drove negligently.
Maybe the robotaxi did prevent some harm by braking with superhuman speed. But I am personally unconvinced it was a completely unavoidable freak accident type of situation without seeing more evidence than a blog post by a company with a heavily vested interest in the situation. I have anecdotally seen Waymo in my area drive poorly in various situations, and I'm sure I'm not the only one.
There's the classic "humans are bad drivers" but I don't think that is an excuse to not look critically into robotaxi accidents. A human driver who hit a child next to a school would have a personal responsibility and might face real jail time or at the least be put on trial and investigated. Who at Waymo will face similar consequences or risk for the same outcome?
It's going to sound batshit insane what I say - the problem is, if we don't praise company PR, the other side will use this as an excuse to push even harder regulations, not allow them in newer cities, slow down the adoption rate, while factually ignoring that this is just a safer method of transport. I wish I was not a bootlicker, but I really want robotaxis to be available everywhere in the world at some point, and such issues should not slow them down IF it's better, and especially, not worse than humans on average.
I understand it sounds stupid, but there was huge push back for introducing Uber to the cities I lived in. And obviously this is even bigger change. However, if a private company is willing to foot the bill, go above and beyond to prove its usefulness and safety, I will be repping for it.
Do you know anyone who works at Waymo? The cynicism is silly. Just because some people at some companies behave horribly, it doesn't mean all or even most do.
Look at Waymo's history in the space, meet some of the people working there, then make a decision.
You don't have to think anyone is behaving horribly to acknowledge that a company's PR department will tend to put out the version of the story that makes them look best.
The a human would do it better people are hilarious. Given how many times I have been hit by human drives on my bike and watched others get creamed by a cars. One time in Boulder at a flashing cross walk a person ran right through it and the biker they creamed got stuck in the roof rack.
For real, I am convinced these are people who never walk or bike, at least around cities like Santa Monica. I am an everyday urban walker and I have to constantly be on alert not to be hit, even when I'm behaving predictably and with the right of way.
Yeah I have to wonder if any of the "humans would do it better" people actually have children and have dropped them off in a school zone. Drivers are on their phones rolling through school zones at 25-30 during pickup/dropoff hours all the fucking time.
I was just dropping my kids off at their elementary school in Santa Monica, but not at Grant Elementary where this happened.
While it's third-hand, word on the local parent chat is that the parent dropped their kid off on the opposite side of the street from Grant. Even though there was a crossing guard, the kid ran behind a car an ran right out in to the street.
If those rumors are correct, I'll say the kid's/family's fault. That said, I think autonomous vehicles should probably go extra-slowly near schools, especially during pickup and dropoff.
At some point children are capable of pursuing Darwin Awards. Parents may enable this, but ultimately if one’s child does something stupid contrary to one’s guidance and restrictions, they may end up with a Darwin for it. Two hundred years ago the child mortality rate was half, as in you lost one child per two, and most of those were not the fault of the child or parents. Society for quite some years has been pushing that down, to the point that a near-death involving a neglectful parent and a witless child is apparently (?) newsworthy — but the number of deaths will never reach zero, whether humans or robots or empty plains and blue skies. There will always be a Veruca Salt throwing themselves into the furnace no matter how many safety processes we impose onto roads, cars, drivers, and/or robots.
If you want to see an end to this nonsensical behavior by parents, pressure your local city into having strict traffic enforcement and ticketing during school hours at every local school, so that the parent networks can’t share news with each other of which school is being ‘harassed’ today. Give license points to vehicles that drop a child across the street, issue parking tickets to double parkers, and boot vehicles whose drivers refuse to move when asked. Demand they do this for the children, to protect them from the robots, if you like.
But.
It’ll protect them much more from the humans than from the robots, and after a few thousand rockets are issued to parents behaving badly, you’ll find that the true threat to children’s safety on school roads is children’s parents — just as the schools have known for decades. And that’s not a war you’ll win arguing against robots. (It’s a war you’ll win arguing against child-killing urban roadway design, though!)
The US commercial aviation industry did not get to its excellent safety record by simply shrugging and accepting a “no-fault accident”.
There are always systemic factors that can be improved, for example working on street design to separate dangerous cars from children, or transportation policy by shifting transportation to buses, bikes, and walking where the consequences of mistakes are significantly reduced.
Cars are the #2 killer of children in the US, and it’s largely because of attitudes like this that ignore the extreme harm that is caused by preventable “accidents”
"No fault" does not mean "no cause" and air crash investigations always focus on causes, not fault. When you understand causes, you can think about how to prevent them happening again.
A human driver travelling at the same speed would have hit that child at exactly 17 mph, before their brain even registered that child was there. If that driver would also have been driving a large SUV that child would have been pushed on the ground and ran over, so probably a fatality. And also functionally nobody would have given a shit apart from some lame finger pointing at (probably) the kid’s parents.
And it is not the child’s or their parents’ fault either:
Once you accept elementary school aged children exist, you have to accept they will sometimes run out like this. Children just don’t have the same impulse control as adults. And honestly even for adults stepping out a bit from behind an obstacle in the path of a car is an easy mistake to make. Don’t forget that for children an SUV is well above head height so it isn’t even possible for them to totally avoid stepping out a bit before looking. (And I don’t think stepping out vs. running out changes the outcome a lot)
This is why low speed limits around schools exist.
So I would say the Waymo did pretty well here, it travelled at a speed where it was still able to avoid not only a fatality but also major injury.
> A human driver travelling at the same speed would have hit that child at exactly 17 mph, before their brain even registered that child was there.
Not sure where this is coming from, and it's directly contradicted by the article:
> Waymo said in its blog post that its “peer-reviewed model” shows a “fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph.” The company did not release a specific analysis of this crash.
No, Waymo’s quote supports the grandparent comment - it was about a “fully attentive human driver” - unless you are arguing that human drivers are consistently “fully attentive”?
Fair enough, so then how fast would a semi-attentive driver stop?
The comment I originally replied to makes the claim a human's brain wouldn't have even responded fast enough to register the child was there. That's going WAY further than how Waymo is claiming a human would have responded.
I don't see how that's a more reasonable assumption that a human driver actually being "fully attentive", and I'm not sure Waymo's definition of that term is the same as what you're using.
> And it is not the child’s or their parents’ fault either: Once you accept elementary school aged children exist, you have to accept they will sometimes run out like this. Children just don’t have the same impulse control as adults.
I get what you are trying to say and I definitely agree in spirit, but I tell my kid (now 9) "it doesn't matter if it isn't your fault, you'll still get hurt or be dead." I spent a lot of time teaching him how to cross the street safely before I let him do it on his own, not to trust cars to do the right thing, not to trust them to see you, not to trust some idiot to not park right next to cross walk in a huge van that cars have no chance of seeing over.
If only we had a Dutch culture of pedistrian and road safety here.
Vehicle design also plays a role: passenger cars have to meet pedestrian collision standards. Trucks don't. The silly butch grilles on SUVs and pickups are deadly. This is more of an argument for not seeing transportation as a fashion or lifestyle statement. Those truck designs are about vanity and gender affirming care. It's easier to make rational choices when it's a business that's worried about liability making those choices.
Cheers to cities pedestrianizing school streets even in busy capitals (e.g. Paris). Cars have no place near school entrances. Fix your urbanism and public transportation.
Yes, kids in developed countries have the autonomy to go to school by themselves from a very young age, provided the correct mindset and a safe environment. That's a combination of:
* high-trust society: commuting alone or in a small group is the norm, soccer moms a rare exception,
* safe, separated lanes for biking/walking when that's an option.
you're exactly right. the fixation on human vs AV error rates completely misses the point. even if we achieve 'perfect' AVs, mixing heavy machinery with children guarantees conflict. physics dictate cars can't stop instantly. the only solution is removing cars, not better drivers.
most commenters here are ignoring the structural incentives. the long term threat of waymo isn't safety, its the enclosure of public infrastructure. these companies are building a permission structure to lobby personal vehicles and public transit off the road.
transportation demand is inelastic. if we allow a transition where mobility is captured by private platforms, the consumer loses all leverage. the endgame is the american healthcare model: capture the market, kill alternatives, and extract max rent because the user has no choice. we need dense urban cores and mass transit, not a dependency on rent seeking oligopolies
If the speed limit was 15 mph, and the Waymo vehicle was traveling at 17 mph before braking, why do you believe the Waymo vehicle would honor a 12 mph speed limit? It didn't honor the 15 mph limit.
> If the speed limit was 15 mph, and the Waymo vehicle was traveling at 17 mph before braking, why do you believe the Waymo vehicle would honor a 12 mph speed limit?
+/- 2 mph is acceptable speedometer and other error. (15 mph doesn’t mean never exceed under any legal inteprerstion I know.)
It’s reasonable to say Waymo would reduce speed in a 12 versus 15 in a way most humans would not.
Ignored by some, not all humans. I absolutely drive extra slowly and cautiously when driving past an elementary school during drop off and pick up precisely because kids do dumb stuff like this. Others do too, though not everyone of course, incredibly.
The great thing about doing things like driving the speed limit in school zones is you get to witness other drivers drive even worse, like passing you in a no passing zone in front of the school, because they can't bear to drive slow for three blocks.
We are responsible for the consequences of our actions. The speed limit is almost irrelevant; drive slowly enough so you don't hit anyone - especially in a school zone.
> We are responsible for the consequences of our actions.
We're not though. Drivers are allowed to kill as many people as they like as long as they're apologetic and weren't drinking; at most they pay a small fine.
So the waymo was speeding! All the dumbasses on here defending waymo when it was going 17 > 15.
Oh also, that video says "kid ran out from a double parked suv". Can you imagine being dumb enough to drive over the speed limit around a double parked SUV in a school zone?
The 15 mph speed limit starts on the block the school is on. The article says the Waymo was within two blocks of the school, so it's possible they were in a 25 mph zone.
> Can you imagine being dumb enough to drive over the speed limit around a double parked SUV in a school zone?
Can you imagine being dumb enough to think that exceeding a one size fits all number on a sign by <10% is the main failing here?
As if 2mph would have fundamentally changed this. Pfft.
A double parked car, in an area with chock full street parking (hence the double park) and "something" that's a magnet for pedestrians, and probably a bunch of pedestrians should be a "severe caution" situation for any driver who "gets it". You shouldn't need a sign to tell you that this is a particular zone and that warrants a particular magic number.
The proper reaction to a given set of indicators that indicate hazards depends on the situation. If this were easy to put in a formula Waymo would have and we wouldn't be discussing this accident because it wouldn't have happened.
Are you comparing robot drivers to the existing alternative? Next time you see one of those blinking speed displays, I’d urge you to pull over and see how fast many human drivers go, and watch for what percent of them aren’t consistently even looking at the road ahead.
So let me get this straight, the car should have been going less than the speed limit, but the fact that it was going a hair over the speed limit is the problem?
The car clearly failed to identify that this was a situation it needed to be going slower. The fact that it was going 17 instead of 15 is basically irrelevant here except as fodder for moral posturing. If the car is incapable of identifying those situations no amount of "muh magic number on sign" is going to fix it. You'll just have the same exact accident again in a 20 school zone.
If the car is going slower than the speed limit in this scenario, it is difficult to tell over the internet if that speed was appropriate. If the car is going over the speed limit, it is obviously inappropriate.
My read is the Waymo was operating normally in a situation a human would operate with heightened awareness and caution. This is not the win they think it is.
Personally in LA I had a Waymo try to take a right as I was driving straight down the street. It almost T-boned me and then honked at me. I don’t know if there has been a change to the algorithm lately to make them more aggressive but it was pretty jarring to see it mess up that badly
In recent weeks I've found myself driving in downtown SF congestion more than usual, and observed Waymos doing totally absurd things on multiple occasions.
The main saving grace is they all occurred at low enough speeds that the consequences were little more than frustrating/delaying for everyone present - pedestrians and drivers alike, as nobody knew what to expect next.
They are very far from perfect drivers. And what's especially problematic is the nature of their mistakes seem totally bizarre vs. the kinds of mistakes human drivers make.
I'm curious as to what kind of control stack Waymo uses for their vehicles. Obviously their perception stack has to be based off of trained models, but I'm curious if their controllers have any formal guarantees under certain conditions, and if the child walking out was within that formal set of parameters (e.g. velocity, distance to obstacle) or if it violated that, making their control stack switch to some other "panic" controller.
This will continue to be the debate—whether human performance would have exceeded that of the autonomous system.
From a purely stats pov, in situations where the confusion matrix is very asymmetric in terms of what we care about (false negatives are extra bad), you generally want multiple uncorrelated mechanisms, and simply require that only one flips before deciding to stop. All would have to fail simultaneously to not brake, which becomes vanishingly unlikely (p^n) with multiple mechanisms assuming uncorrelated errors. This is why I love the concept of Lidar and optical together.
The true self-driving trolley problem. How many rear-end collisions and riders' annoyance caused by phantom braking a manufacturer (or a society) is going to tolerate to save one child per N million miles?
Uncorrelated approach improves sensitivity at the cost of specificity. Early sensor fusion might improve both (maybe at the cost of somewhat lesser sensitivity).
Yeah, if a human made the same mistakes as the Waymo driving too fast near the school, then they would have hurt the kid much worse than the Waymo did.
So if we're going to have cars drive irresponsibly fast near schools, it's better that they be piloted by robots.
Maybe we should not only replace the unsafe humans with robots, but also have the robots drive in a safe manner near schools rather than replicating the unsafe human behavior?
One argument for the robots is that they can be programmed to drive safer, while humans cant.
But that depends on reliability, especially in unforseen (and untrained-upon) circumstances. We'll have to see how they do, but they have been doing better than expected
Depends on the school zone. The tech school near me is in a 50 zone and they don't even turn on the "20 when flashing" signs because if you're gonna walk there, you're gonna come in via residential side streets in the back and the school itself is way back off the road. The other school near me is downtown and you wouldn't be able to go 17 even if you wanted to.
Kinetic energy is a bad metric. Acceleration is what splats people.
Jumping out of a plane wearing a parachute vs jumping off a building without one.
But acceleration is hard to calculate without knowing time or distance (assuming it's even linear) and you don't get that exponent over velocity yielding you a big number that's great for heartstring grabbing and appealing to emotion hence why nobody ever uses it.
Absent more precise information, this is a statistical negative mark for Waymo putting their child pedestrian injury rate at ~2-4x higher than the US human average.
US human drivers average ~3.3 trillion miles per year [1]. US human drivers cause ~7,000 child pedestrian injurys per year [2]. That amounts to a average of 1 child pedestrian injury per ~470 million miles. Waymo has done ~100-200 million fully autonomous miles [3][4]. That means they average 1 child pedestrian injury per ~100-200 million miles. That is a injury rate ~2-4x higher than the human average.
However, the child pedestrian injury rate is only a official estimate (possible undercounting relative to highly scrutinized Waymo miles) and is a whole US average (operational domain might not be comparable, though this could easily swing either way), but absent more precise and better information, we should default to the calculated 2-4x higher injury rate; it is up to Waymo to robustly demonstrate otherwise.
Furthermore, Waymo has published reasonably robust claims arguing they achieve ~90% crash reduction [5] in total. The most likely new hypotheses in light of this crash are:
A. Their systems are not actually robustly 10x better than human drivers. Waymos claims are incorrect or non-comparable.
B. There are child-specific risk factors that humans account for that Waymo does not that cause a 20-40x differential risk around children relative to normal Waymo driving.
C. This is a fluke child pedestrian injury. Time will tell. Given their relatively robustly claimed 90% crash reduction, it is likely prudent to allow further operation in general, though possibly not in certain contexts.
I don't think this comparison is meaningful given the sample size of 1 and the differences between the between your datasets. The standard error margins from the small sample size alone are so large that you could not reasonably claim humans are safer (95% CI for Waymo is about 1 per 20 million miles to 1 per 8 billion miles). Then there are the dataset differences:
1. The NHTSA data is based on police-reported crash data, which reports far fewer injuries than the CDC reports based on ED visits. The child in this case appeared mostly unharmed and situations like this would likely not be counted in the NHTSA data.
2. Waymo taxis operate primarily in densely populated urban environments while human driver milage includes highways and rural roads where you're much less likely to collide with pedestrians per mile driven.
Waymo's 90% crash reduction claim is at least an apples-to-apples comparison.
I don't think I'd want to take much from such a statistical result yet. A sample size of 1 accident just isn't enough information to get a real rate from, not that I want to see more collisions with children. Though this is also muddied by the fact that Waymo will most likely adjust their software to make this less likely, and we won't know exactly how or how many miles each version has. I'd also like to see the data for human incidents over just the temperate suburban areas like Waymo operates in.
> child pedestrian injury rate at ~2-4x higher than the US human average.
If this incident had happened with a human driven vehicle would it even have been reported?
I don't know exactly what a 6mph collision looks like but I think it's likely the child had nothing more than some bruises and if a human has done it they would have just said sorry, made sure they were ok, and left
Do we even know that the child was injured? All I've seen anyone officially claim is that the Waymo made contact, the kid fell over, then stood up and walked to the side of the road. Assuming the Waymo was still braking hard, 6mph means it was about 1/4s and about 30cm from reaching a full stop, so it could be a very minor incident we're talking about here.
I'm not aware of any statistics for how often children come into contact with human-driven cars.
The Waymo driver tech is impressive. That said an experienced driver might have recognized the pattern where a stopped big vehicle occludes a part of the road leading to such situation, and might have stopped or slowed down almost to a halt before passing. The Waymo driver reacts faster but is not able to predict such scenarios by filling the gaps, simulating the world to inform decisions. Chapeau to Waymo anyways
There have been many instances of Waymo preventing a collision by predicting pedestrians emerging from occlusion. This isn’t new information at all for them. Some accidents are simply physically impossible to prevent. I don’t know for sure if this one was one of those, but I’m fairly confident it couldn’t have been from prediction failure.
See past examples:
https://youtube.com/watch?v=hubWIuuz-e4 — first save is a child emerging from a parked car. Notice how Waymo slows down preemptively before the child starts moving.
That’s probably how they do it, which is again very clever stuff, chapeau. But they do it like that b/c they can’t really predict the world around them fast enough. It might be possible in the future with AI World Models though
What do you mean “fast enough”? You can’t predict something that doesn’t exist i.e. not visible to the sensors. A Waymo wouldn’t move at all if it assumed people would always jump out of nowhere.
Even if you detect “fast enough”, there are physical limits for braking and coming to a stop.
I think this definitely an improvement to consider, but when comparing I think that big number, i.e. statistics are the only thing that matters. Some human could detect the pattern and come to full halt another human driver could be speeding while texting
> The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post.
The issue is that I don’t trust a private company word. You can’t even trust the president of the USA government nowadays… release the video footage or get lost.
> any other car been there, probably including Tesla
Cheap shots. If this was Tesla there would be live media coverage across every news outlet around the world and congressmen racing to start investigation.
Look at any thread where Tesla is mentioned and how many waymo simps are mansplaning lidar.
The only question I have is whether the speed it was going was situationally appropriate and whether we’d expect a human to be considered reckless under the same circumstances. 17mph sounds pretty slow but it really depends on context.
Slowing the adoption of much-safer-than-humans robotaxis, for whatever reason, has a price measured in lives. If you think that the principle you've just stated is worth all those additional dead people, okay; but you should at least be aware of the price.
Failure to acknowledge the existence of tradeoffs tends to lead to people making really lousy trades, in the same way that running around with your eyes closed tends to result in running into walls and tripping over unseen furniture.
You may not have any way of knowing but the rest of society has developed all sorts of systems of knowing. "Scientific method", "Bayesian reasoning", etc. or start with the Greek philosophy classics.
Waymo is not a machine, it is a corporation, and corporations can, in fact be held accountable for decisions (and, perhaps more to the point, for defects in goods they manufacture, sell, distribute, and/or use to provide services.)
What? No? The main selling point is eliminating costs for a human driver (by enabling people to safely do other things from their car, like answering emails or doomscrolling, or via robotaxis).
Sure, but the companies building them are just shoving billions of dollars into their ears so they don't have to answer "who's responsible when it kills someone?"
The question of responsibility, while philosophically interesting wrt. increasingly autonomous machines, is not going to be an issue in practice. We'll end up dealing with it like we always do with multi-party responsibility in complex systems: regulators setting safety standards and outlining types and structures of liabiliy, contracts shifting the liability around, and lots and lots of insurance.
In fact, if you substitute "company providing self-driving solution (integrated software + hardware)" for "company renting out commercial drivers" (or machine operators), then self-driving cars already fit well into existing legal framework. The way I see it, the only change self-driving cars introduce here is that there is no individual operator we could blame for the accident, no specific human we could heavily fine or jail, and then feel good about ourselves because we've issued retributive justice and everything is whole now. Everything else has already long been worked out.
> They have to be, as a machine can not be held accountable for a decision
This logic applies equally to all cars, which are machines. Waymo has its decision makers one more step removed than human drivers. But it’s not a good axiom to base any theory of liability on.
In traditional vehicles, liability is structurally centered on the human driver, and product liability exists but is difficult to assert absent clear or systemic defects.
Autonomous driving systems disrupt this by directly assuming the driving function, forcing liability upstream (where it's significantly more difficult to navigate).
So now it's driver vs injured party. Self-driving makes it trillion dollar company vs injured party. Night and day difference.
That sucks, and I love to hate on "self driving" cars. But it wasn't speeding to start with (assuming speed limit in the school zone was 20 or 25), braked as much as possible, and the company took over all the things a human driver would have been expected to do in the same situation. Could have been a lot worse, probably wouldn't have been any better with a human driver (just going to ignore as no-signal Waymo's models that say an attentive human driver would have been worse). It's "fine". In this situation, cars period are the problem, not "self driving" cars.
It's hard to imagine how any driver could have reacted better in this situation.
The argument that questions "would a human be driving 17mph in a school zone" feels absurd to the point of being potentially disingenuous. I've walked and driven through many school zones before, and human drivers routinely drive above 17mph (in some cases, over the typical 20mph or 25mph legal limit). It feels like in deconstructing some of these incidences, critics imagine a hypothetical scenario in which they are driving a car and its their only job to avoid a specific accident that they know will happen in advance, rather than facing the reality of what human drivers are actually like on the road.
When I was a kid (age 12, or so), I got hit by a truck while crossing the road on my bike.
In that particular instance, I was cited myself -- after the fact, at the hospital -- and eventually went before a judge. In that hearing, it was established that I was guilty of failing to yield at an intersection.
(That was a rather long time ago and I don't remember the nature of the punishment that resulted. It may have been as little as a stern talking-to by the judge.)
A person who hits a child, or anyone, in America, with no resulting injury, stands a roughly 0% chance of facing a judge in consequence. Part of Waymo's research is to show that even injury accidents are rarely reported to the police.
Are you thinking of civil liability or criminal liability?
Waymo is liable in a civil sense and pays whatever monetary amount is negotiated or awarded.
For a criminal case, some kind of willful negligence would have to be shown. That can pierce corporate veils. But as a result Waymo is being extremely careful to follow the law and establish processes which shield their employees from negligence claims.
Waymo is going to make sure they are never criminally liable for anything, and even if they were, a criminal case against a corporation just ends up being a modest fine.
It's interesting how polarized this comments section is. Lots of people claiming a human driver would definitely have been driving slower. Lots of people claiming statistics show that human drivers do worse in this scenario aggregate. Of course neither side presenting convincing evidence.
I tried to find some data: In a 20 mph school zone, ~70-80% of drivers drove 26 mph or faster, regardless of whether there were flashing lights, an always-20 zone, or a 20-when-schoolchildren-present zone. [1]
Although note that ~70-80% of drivers drove 6 or more mph over the speed limit in a school zone, while it seems like the claims in some of these comments are that a human driver would drive at less than 17 mph out of an abundance of caution.
> Normal person would drive carefully around blind spots.
I can't tell if you are using sarcasm here or are serious. I guess it depends on your definition of normal person (obviously not average, but an idealized driver maybe?).
> Waymo said its robotaxi struck the child at six miles per hour, after braking “hard” from around 17 miles per hour. The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post. Waymo said its vehicle “immediately detected the individual as soon as they began to emerge from behind the stopped vehicle.”
As this is based on detection of the child, what happens on Halloween when kids are all over the place and do not necessarily look like kids?
These systems don't discriminate on whether the object is a child. If an object enters the path of the vehicle, the lidar should spot it immediately and the car should brake.
Sorry, I was being oblique. Humans kill other humans with cars every day. They kill even more on Halloween. Let's start addressing that problem before worrying whether Waymos might someday decide it's OK to drive through ghosts.
Autonomous vehicles won't be perfect. They'll surely make different mistakes from the ones humans currently make. People will die who wouldn't have died at the hands of human drivers. But the overall number of mistakes will be smaller.
Suppose you could wave your magic wand and have a The Purge-style situation where AVs had a perfect safety record 364 days of the year, but for some reason had a tricky bug that caused them to run over tiny Spidermen and princesses on Halloween. The number of fatalities in the US would drop from 40,000 annually to 40. Would you wave that wand?
> In October 2025, a Waymo autonomous robotaxi struck and killed KitKat, a well-known bodega cat at Randa's Market in San Francisco's Mission District, sparking debates over self-driving car safety
It's a child now. All I wanna ask - what should happen, so they stop killing pets and people?
The real but contentious answer is to change our street and urban design. You can only do so much to make a giant metal machine safe for children and small animals to be struck by. Reducing the frequency of cars and pedestrians occupying the same space will go further than trying to engineer the equivalent of a pool that is impossible to drown in.
Do you think that a company that operates autonomous vehicles will support legislation that makes it easier and safer to move around on foot without getting hit by a car? Or will they lobby for car-centric urban design, like many many companies before them?
I know submissions are not meant to contain modifications to article titles, but would it be so bad to have added "at 6mph" and/or "minor injuries" to the title?
I'm a big fan of Waymo and have enjoyed my Waymo rides. And I don't think Waymno did anything "bad" here.
> The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post. Waymo said its vehicle “immediately detected the individual as soon as they began to emerge from behind the stopped vehicle.”
BUT! As a human driver, I avoid driving near the schools when school's letting out. There's a high school on my way home and kids saunter and jaywalk across the street, and they're all 'too cool' to press the button that turns on the blinking crosswalk. So I go a block out of my way to bypass the whole school area when I'm heading home that way.
Waymos should use the same rationale. If you can avoid going past a school zone when kids are likely to be there, do it!
> Waymos should use the same rationale. If you can avoid going past a school zone when kids are likely to be there, do it!
I can see that, prioritize obstacle predictability over transit time. A school zone at certain times of day is very unpredictable with respect to obstacles but a more car congested area would be easier to navigate but slower. Same goes for residential areas during Halloween.
OK
Its like this!, If I hit a child in a school district, I loose my licence for many years, and if I dont or cant show remourse, it could be longer, I pay fines, etc
Therefor waymo, must have it's algorythm terminated, ie: totaly destroyed, all the hardware smashed, and they never get to try again with any derivitive of this technology, as there is no reasonable, understandable path towards repentence and rehabilitation, it is litteraly a monster running over children.
or was it carrying an ICE team, then nevermind.
>To put this in perspective, our peer-reviewed model shows that a fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph. This significant reduction in impact speed and severity is a demonstration of the material safety benefit of the Waymo Driver.
The statistically relevant question is: How many human drivers have hit children near elementary schools, since Waymo's last accident?
If Waymo has fewer accidents where a pedestrian is hit than humans do, Waymo is safer. Period.
A lot of people are conjecturing how safe a human is in certain complicated scenarios (pedestrian emerging from behind a bus, driver holds cup of coffee, the sun is in their eyes, blah blah blah). These scenarios are distractions from the actual facts.
This is wrong, although something quite like it is right.
Imagine that there are only 10 Waymo journeys per year, and every year one of them hits a child near an elementary school, while there are 1000000 non-Waymo journeys per year, and every year two of them hit children near elementary schools. In this scenario Waymo has half as many accidents but is clearly much more dangerous.
Here in the real world, obviously the figures aren't anywhere near so extreme, but it's still the case that the great majority of cars on the road are not Waymos, so after counting how many human drivers have had similar accidents you need to scale that figure in proportion to the ratio of human to Waymo car-miles.
(Also, you need to consider the severity of the accidents. That comparison probably favours Waymo; at any rate, they're arguing that it does in this case, that a human driver in the same situation would have hit the child at a much higher and hence more damaging speed.)
Who is liable when FSD is used? In Waymo's case, they own and operate the vehicle so obviously they are fully liable.
But in a human driver with FSD on, are they liable if FSD fails? My understanding is yes, they are. Tesla doesn't want that liability. And to me this helps explain why FSD adoption is difficult. I don't want to hand control over to a probabilistic system that might fail but I would be at fault. In other words, I trust my own driving more than the FSD (I could be right or wrong, but I think most people will feel the same way).
I believe Mercedes is the only consumer car manufacturer that is advertising an SAE Level 3 system. My understanding is that L3 is where the manufacturer says you can take your attention off the road while the system is active, so they're assuming liability.
That's pretty hyperbolic. At less than 20 mph, car vs pedestrial is unlikely to result in death. IIHS says [1] in an article about other things:
> As far as fatalities were concerned, pedestrians struck at 20 mph had only a 1% chance of dying from their injuries
Certainly, being struck at 6 mph rather than 17 mph is likely to result in a much better outcome for the pedestrian. And that should not be minimized; although it is valuable to consider the situation (when we have sufficient information) and validate Waymo's suggestion that the average human driver would also have struck the pedestrian and at greater speed. That may or may not be accurate, given the context of a busy school dropoff situation... many human drivers are extra cautious in that context and may not have reached that speed; depending on the end to end route, some human drivers would have avoided the street with the school all together based on the time, etc. It's certainly seems like a good result for the premise, child unexpectedly appears from between large parked vehicles, but maybe there should have been an expectation.
Plenty. Have you ever driven on a freeway at rush hour? Have you driven in a pickup/dropoff line at a school or an airport? You may or may not want to go 100, but when there's a vehicle in front of you going 20mph or less, you're kind of stuck.
Yes, I have seen how people struggle to drive at low speed and can only do start/stop, accelerating to a way too high speed only to slam on brakes, repeating endlessly.
But really, did you seriously read my post as meaning people literally can't go slower than 20 so just plough into whatever is in the way? I'm obviously talking about an open road situation. Hardly any human drivers go under 20 by choice.
For me, the policy question I want answered is if this was a human driver we would have a clear person to sue for liability and damages. For a computer, who is ultimately responsible in a situation where suing for compensation happens? Is it the company? An officer in the company? This creates a situation where a company can afford to bury litigants in costs to even sue, whereas a private driver would lean on their insurance.
So you're worried that instead of facing off against an insurance agency, the plantiff would be facing off against a private company? Doesn't seem like a huge difference to me
Is there actually any difference? I'd have though that the self-driving car would need to be insured to be allowed on the road, so in both cases you're going up against the insurance company rather than the actual owner.
Waymo hits you -> you seek relief from Waymo's insurance company. Waymo's insurance premium go up. Waymo can weather a LOT of that. Business is still good. Thus, poor financial feedback loop. No real skin in the game.
John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
NOW ... add criminal fault due to driving decision or state of vehicle ... John goes to jail. Waymo? Still making money in the large. I'd like to see more skin in their game.
> John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
John probably (at least where I live) does not have insurance, maybe I could sue him, but he has no assets to speak of (especially if he is living out of his car), so I'm just going to pay a bunch of legal fees for nothing. He doesn't car, because he has no skin in the game. The state doesn't care, they aren't going to throw him in jail or even take away his license (if he has one), they aren't going to even impound his car.
Honestly, I'd much rather be hit by a Waymo than John.
I see. Thank you for sharing. Insurance here is mandatory here for all motorists.
If you are hit by an underinsured driver, the government steps in and additional underinsured motorist protection (e.g. hit by an out of province/country motorist) is available to all and not expensive.
Jail time for an at-fault driver here is very uncommon but can be applied if serious injury or death results from a driver's conduct. This is quite conceivable with humans or AI, IMO. Who will face jail time as a human driver would in the same scenario?
Hit and run, leaving the scene, is also a criminal offence with potential jail time that a human motorist faces. You would hope this is unlikely with AI, but if it happens a small percentage of the time, who at Waymo faces jail as a human driver would?
I'm talking about edge cases here, not the usual fender bender. But this thread was about policy/regs and that needs to consider crazy edge cases before there are tens of millions of AI drivers on the road.
Insurance here is also mandatory for all motorists. Doesn't matter if the rules aren't actually enforced.
Waymo has deep pockets, so everyone is going to try and sue them, even if they don't have a legitimate grievance. Where I live, the city/state would totally milk each incident from a BigCo for all it was worth. "Hit and run" by a drunk waymo? The state is just salivating thinking about the possibility.
I don't agree with you that BigCorp doesn't have any skin in the game. They are basically playing the game in a bikini.
> Insurance here is mandatory here for all motorists.
You do know that insurance being mandatory doesn't stop people from driving without insurance, right?
> If you are hit by an underinsured driver, the government steps in and additional underinsured motorist protection (e.g. hit by an out of province/country motorist) is available to all and not expensive.
Jolly good for you.
If I don't carry underinsured coverage, and someone totals my car or injures me with theirs, I'm basically fucked.
>John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
Ah great, so there's a lower chance of that specific John Smith hitting me again in the future!
The general deterrence effect we observe in society is that punishment of one person has an effect on others who observe it, making them more cautious and less likely to offend.
No, the fact is that the child sustained minor injuries. And, fact: no human driver made the decision to drive a vehicle in that exact position and velocity. Imagining a human-driven vehicle in the same place is certainly valid, but your imagination is not fact. I imagine that the kid would be better off if no vehicle was there. But that's not a fact, that's an interpretation -- perhaps the kid would have ended up dead under an entirely different tire if they hadn't been hit by the waymo!
Disagree, most human drivers would notice they are near an elementary school with kids coming/going, crossing guard present, and been driving very carefully near blocked sight lines.
Better reporting would have asked real people the name of the elementary school, so we could see some pictures of the area. The link to NHTSA didn't point to the investigation, but it's under https://www.nhtsa.gov/search-safety-issues
"NHTSA is aware that the incident occurred within two blocks of a Santa Monica, CA elementary school during normal school drop off hours; that there were other children, a crossing guard, and several double-parked vehicles in the vicinity; and that the child ran across the street from behind a double parked SUV towards the school and was struck by the Waymo AV. Waymo reported that the child sustained minor injuries."
We're getting into hypotheticals but i will say in general i much much prefer being around Waymos/Zooxs/etc. than humans when riding a bicycle.
We're impatient emotional creatures. Sometimes when I'm on a bike the bike lane merges onto the road for a stretch, no choice but to take up a lane. I've had people accelerate behind me and screech the tyres, stopping just short of my back wheel in a threatening manner which they then did repeatedly as i ride the short distance in the lane before the bike lane re-opens.
To say "human drivers would notice they are near an elementary school" completely disregards the fuckwits that are out there on the road today. It disregards human nature. We've all seen people do shit like i describe above. It also disregards that every time i see an automated taxi it seems to drive on the cautious side already.
Give me the unemotional, infinite patience, drives very much on the cautious side automatic taxi over humans any day.
Couldn’t be anymore callous and clinical. This press release alone makes me want not to use their service.
* “Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene,” Waymo wrote in the post.*
The 'next' comment after yours is "Alternate headline: Waymo saves child's life" and the 'prev' comment is "A human driver would most likely have killed this child. That's what should be on the ledger." - would either of those be less 'callous and clinical'?
Other accident reports I've seen (NTSB, etc) often seem to take a similar approach - is it a bad thing?
Or what kind of language wouldn't make you 'want not to use their service'?
Wow this is why I feel comfortable in a Waymo. Accidents are inevitable and some point and this handling was well-rehearsed and highly ethical. Amazing company
That would be one hell of a convoluted route to avoid school zones. I wonder if it would even be possible for a large majority of routes, especially in residential areas.
It might not be possible for a lot of places — I don’t really know.
But I know when I drive, if it’s a route I’m familiar with, I’ll personally avoid school zones for this very reason: higher risk of catastrophe. But also it’s annoying to have to slow down so much.
Maybe this personal decision doesn’t really scale to all situations, but I’m surprised Waymo doesn’t attempt this. (Maybe they do and in this specific scenario it just wasn’t feasible)
Most people prefer the shortest ride. Circling around school zones would be the opposite of that. Rides are charged based on distance, so maybe this would interest Waymo, but one of the big complaints about taxi drivers was how drivers would "take them for a ride" to increase the fare.
Seems like a solvable problem: make it clear on the app/interior car screens that a school zone is being avoided — I think most riders will understand this.
You also have to drive much more slowly in a school zone than you do on other routes, so depending on the detour, it may not even be that much longer of a drive.
At worst, maybe Waymo eats the cost difference involved in choosing a more expensive route. This certainly hits the bottom line, but there’s certainly also a business and reputational cost from “child hit by Waymo in school zone” in the headlines.
And before the argument "Self driving is acceptable so long as the accident/risk is lower than with human drivers" can I please get that out of the way: No it's not. Self driving needs to be orders of magnitude safer for us to acknowledge it. If they're merely as safe or slightly safer than humans we will never accept it. Becase humans have a "skin in the game". If you drive drunk, at least you're likely to be in the accident, or have personal liability. We accept the risks with humans because those humans accept risk. Self driving abstracts the legal risk, and removes the physical risk.
I'm willing to accept robotaxis, and accidents in robotaxis, but there needs to be some solid figures showing they are way _way_ safer than human drivers.
I think those figures are already starting to accumulate. Incidents like this are rare enough that they are news worthy. Almost every minor incident involving Waymo, Tesla's FSD, and similar solutions gets a lot of press. This was a major incident with a happy end. Those are quite rare. The lethal ones even rarer.
As for more data, there is a chicken egg problem. A phased roll out of waymo over several years has revealed many potential issues but is also remarkable in the low number of incidents with fatalities. The benefit of a gradual approach is that it builds confidence over time.
Tesla has some ways to go here. Though arguably, with many hundreds of thousands of paying users, if it was really unsafe, there would be some numbers on that. Normal statistics in the US are measured in ~17 deaths per 100K drivers per year. 40K+ fatalities overall. FSD for all its faults and failings isn't killing dozens of people per years. Nor is Waymo. It's a bit of an apples and oranges comparison of course. But the bar for safety is pretty low as soon as you include human drivers.
Liability weighs higher for companies than safety. It's fine to them if people die, as long as they aren't liable. That's why the status quo is tolerated. Normalized for amounts of miles driven with and without autonomous, there's very little doubt that autonomous driving is already much safer. We can get more data at the price of more deaths by simply dragging out the testing phase.
Perfect is the enemy of good here. We can wait another few years (times ~40K deaths) or maybe allow technology to start lowering the amount of traffic deaths. Every year we wait means more deaths. Waiting here literally costs lives.
> ~17 deaths per 100K drivers per year. 40K+ fatalities overall.
I also think one needs to remember those are _abysmal_ numbers, so while the current discourse is US centric (because that's where the companies and their testing is) I don't think it can be representative for the risks of driving in general.
Naturally, robotaxis will benefit from better infra outside the US (e.g. better separation of pedestrians) but it'll also have to clear a higher safety bar e.g. of fewer drunk drivers.
Also fun to calculate how this compounds over say 40 years. You get to about 1 in 150 drivers being involved in some kind of deathly accident. People are really bad at numbers and assessing risk.
I am not sure. Self-driving is complex and involves the behavior of other, non-automated actors. This is not like a compression algorithm where things are easily testable and verifiable. If Waymos start behaving extra-oddly in school zones, it may lead to other accidents where drivers attempt to go around the "broken" Waymo and crash into it, other pedestrians, or other vehicles.
I know Tesla FSD is its own thing, but crowdsourced results show that FSD updates often increase the amount of disengagements (errors):
And we haven't reached the point where people start walking straight into the paths of cars, either obliviously or defiantly. https://www.youtube.com/shorts/nVEDebSuEUs
There are already anecdotes of people aggressively jaywalking in front of a Waymo because they know it will stop, and people driving more aggressively around Waymos because it will always defer to them.
> I'm willing to accept robotaxis, and accidents in robotaxis, but there needs to be some solid figures showing they are way _way_ safer than human drivers.
If waymo is to be believed, they hit the kid at 6mph and estimated that a human driver at full attention would have hit the kid at 14 mph. The waymo was traveling 17mph. The situation of "kid running out between cars" will likley never be solved either, because even with sub nanosecond reaction time, the car's mass and tire's traction physically caps how fast a change in velocity can happen.
I don't think we will ever see the video, as any contact is overall viewed negatively by the general public, but for non-hyperbolic types it would probably be pretty impressive.
That doesn't mean it can't be solved. Don't drive faster than you can see. If you're driving 6 feet from a parked car, you can go slow enough to stop assuming a worst case of a sprinter waiting to leap out at every moment.
But with waymos it would be possible. Mark those streets as "extremely slow" and never go there unless you are dropping someone off. (The computer has more patience than human drivers.)
If that's too annoying then bad parking by school areas so the situation doesn't happen.
I don't know if you've been to some cities or neighborhoods but almost every street has on-street parking in many of them.
And why would you make Waymo's go slower than human drivers, when it's the human drivers with worse reaction times? I had interpreted the suggestion as applying to all drivers.
Oh I have no problem believing that this particular situation would have been handled better by a human. I just want hard figures saying that (say) this happens 100x more rarely with robotaxis than human drivers.
> The situation of "kid running out between cars" will likley never be solved
Nuanced disagree (i agree with your physics), in that an element of the issue is design. Kids running out between cars _on streets that stack building --> yard --> sidewalk --> parked cars --> driving cars.
One simple change could be adding a chain link fence / boundary between parked cars and driving cars, increasing the visibility and time.
there's still an inlet and outlet (kinda like hotel pickup/drop off loops). It's not absolutely perfect, but it constrains the space of where kids can dart from every parked car to 2 places.
Also the point isn't the specifics, the point is that the current design is not optimal, it's just the incumbent.
Ok, that's not really a simple change anymore, because you need more space for that. Unless it's really just a drop off queue, but then it's not parked cars, since a parked car blocks the queue.
We would really need to see the site to have an idea of the constraints, Santa Monica has some places where additional roadway can be accomodated and some places where that's not really an option.
In high parking contention areas, I think there's enough latent demand for parking that you wouldn't observe fewer parked cars until reduce demand by a much greater amount.
>We accept the risks with humans because those humans accept risk.
It seems very strange to defend a system that is drastically less safe because when an accident happens, at least a human will be "liable". Does a human suffering consequences (paying a fine? losing their license? going to jail?) make an injury/death more acceptable, if it wouldn't have happened with a Waymo driver in the first place?
I think a very good reason to want to know who's liable is because Google has not exactly shown itself to enthusiastically accept responsibility for harm it causes, and there is no guarantee Waymo will continue to be safe in the future.
In fact, I could see Google working on a highly complex algorithm to figure out cost savings from reducing safety and balancing that against the cost of spending more on marketing and lobbyists. We will have zero leverage to do anything if Waymo gradually becomes more and more dangerous.
> Wherever I'm going, I'll be there to apply the formula. I'll keep the secret intact.
It's simple arithmetic.
It's a story problem.
If a new car built by my company leaves Chicago traveling west at 60 miles per hour, and the rear differential locks up, and the car crashes and burns with everyone trapped inside, does my company initiate a recall?
> You take the population of vehicles in the field (A) and multiple it by the probable rate of failure (B), then multiply the result by the average cost of an out-of-court settlement (C).
A times B times C equals X. This is what it will cost if we don't initiate a recall.
If X is greater than the cost of a recall, we recall the cars and no one gets hurt.
If X is less than the cost of a recall, then we don't recall.
Orders of magnitude? Something like 100 people die on the road in the US each day. If self-driving tech could save 10 lives per day, that’s wouldn’t be good enough?
"It depends". If 50 people die and 50 people go to jail, vs. 40 people die and their families are left wondering if someone will take responsibility? Then that's not immediately standing out as an improvement just because fewer died. We can do better I think. The problem is simply one of responsibility.
If the current situation was every day 40 people die but blame is rarely assigned, would you recommend a change where an additional 10 people are going to die but someone will be held responsible for those deaths?
People don't usually go to jail. Unless the driver is drunk or there's some other level of provable criminal negligence (or someone actively trying to kill people by e.g. driving into a crowd of protesters they disagree with), it's just chalked up as an accident.
Apart from a minority of car related deaths resulting in jail time, what kind of person wants many more people to die just so they can point at someone to blame for it? At what point are such people the ones to blame for so many deaths themselves?
In such situations it’s useful to put yourself in a hypothetical situation. Rules: you can’t pick who you will be: one of the dead or alive. It will be assigned randomly.
The driver cuts in front of one person on an e-bike so fast they can’t react and hit them. Then after being hit they step on the accelerator and go over the sidewalk on the other side of the road killing a 4 year old. No charges filed.
This driver will be back on the street right away.
Have you been in a self driving car? There are some quite annoying hiccups, but they are already very safe. I would say safer than the average driver. Defensive driving is the norm. I can think of many times where the car has avoided other dangerous drivers or oblivious pedestrians before I realized why it was taking action.
But, human drivers often face very little accountability. Even drunk and reckless drivers are often let off with a slap on the wrist. Even killing someone results in minimal consequences.
There is a very strong bias here. Everyone has to drive (in most of America), and people tend to see themselves in the driver. Revoking a license often means someone can’t get to work.
> Self driving needs to be orders of magnitude safer for us to acknowledge it. If they're merely as safe or slightly safer than humans we will never accept it
It’s already accepted. It’s already here. And Waymo is the safest in the set—we’re accepting objectively less-safe systems, too.
If you drive a car, you have a responsibility to do it safely. The fact that I am usually better than the bottom 50% of drivers, or that I am better than a drunk driver does not mean that when I hit someone it's less bad. A car is a giant weapon. If you drive the weapon, you need to do it safely. Most people these days are incredibly inconsiderate - probably because there's little economic value in being considerate. The fact that lots of drivers suck doesn't mean that waymo gets a pass.
Waymos have definitely become more aggressive as they've been successful. They drive the speed limit down my local street. I see them and I think wtf that's too fast. It's one thing when there are no cars around. But if you've got cars or people around, the appropriate speed changes. Let's audit waymo. They certainly have an aggressiveness setting. Let's see the data on how it's changing. Let's see how safety buffers have decreased as they've changed the aggressiveness setting.
The real solution? Get rid of cars. Self-driving individually owned vehicles were always the wrong solution. Public transit and shared infra is always the right choice.
> The fact that lots of drivers suck doesn't mean that waymo gets a pass.
But that fact does mean that we should encourage alternatives that reduce fatalities, and that not doing so results in fatalities that did not need to occur.
> The real solution? Get rid of cars.
I also support initiatives to improve public transit, etc. However, I don't think "get rid of cars" is a realistic idea to the general public right now, so let's encourage all of the things that improve things - robot drivers if they kill people less often than humans, public transit, etc. - let's not put off changes that will save lives on the hope that humanity will "get rid of cars" any time soon. Or when do you think humanity will "get rid of cars"?
From the Waymo blog...
> the pedestrian suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path. Our technology immediately detected the individual as soon as they began to emerge from behind the stopped vehicle. The Waymo Driver braked hard, reducing speed from approximately 17 mph to under 6 mph before contact was made.
> Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene.
> Following the event, we voluntarily contacted the National Highway Traffic Safety Administration (NHTSA) that same day.
I honestly cannot imagine a better outcome or handling of the situation.
Yup. And to add
> Waymo said in its blog post that its “peer-reviewed model” shows a “fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph.”
It's likely that a fully-attentive human driver would have done worse. With a distracted driver (a huge portion of human drivers) it could've been catastrophic.
You're omitting the context provided by the article. This wasn't just a random scenario. Not only was this by an elementary school, but during school drop off hours, with both children and doubled parked cars in the vicinity. If somebody doesn't know what double parking is - it's when cars parallel park beside one another, implicitly on the road, making it difficult to see what's beyond them.
So you are around young children with visibility significantly impaired because of double parking. I'd love to see video of the incident because driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do, because a kid popping out from behind one of those cars is not only unsurprising but completely expected.
Another reason you also slow way down in this scenario is one of those cars suddenly swinging open their door which, again, would not be particularly surprising in this sort of context.
That's my thinking as well. Taken in some abstract scenario, all those steps seems very reasonable, and in that abstract scenario we can even say it would do better than an average human would. But that is missing the overall context that this was an elementary school during drop-off hours. That's when you crawl at 3 mph expecting kids to jump behind any car, and not going at 17mph.
> But that is missing the overall context that this was an elementary school during drop-off hours. That's when you crawl at 3 mph expecting kids to jump behind any car, and not going at 17mph.
Indeed. Sure the car knows the limit, it knows it is a school zone, it can precisely track people within the reach of its sensors (but not behind blockages it can't see through).
But it is missing the human understanding of the situation. Does it know that tiny humans behave far more erratically then the big ones? Obvious to us humans, but does the car take that into account? Does it consider that in such a situation, it is likely that a kid that its sensors can't possibly detect has a high probability to suddenly dart out from behind an obstacle? Again obvious to us humans because we understand kids, but does the car know?
Urm, ime people frequently drive significantly over the speed limit in all these places, at all times of the day.
Blows my mind how you guys confidently state this with authority as if that's the normal behavior, when the reality is that it probably should be - but isn't actually.
So you're confidently stating it's not the normal behaviour... Can you tell me what the average speed is for human drivers outside elementary schools at drop-off times?
But if you’re plan on building a fleet of cars operating all over the country or the world, do you want to model them after the careful driver, who has awareness about the situation (school, drop off/pickup hours, etc) or say “what the heck, some drivers are not paying attention so neither will my robots, it’s fine”
"I'd love to see video of the incident because driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do"
I am not sure what your definition of typical is. The reason we have lower speed zones in school districts at specific times is because humans typically drive fast. The reason police officers frequently target these areas and write a plethora of tickets is because humans typically ignore speed limits.
Your claim seems to be that a human would drive much slower than the posted speed limit considering the conditions, but the laws and the court room suggest otherwise.
There is probably a multimodal distribution of behaviors for people in these situations. Any robotaxi ought to behave in the more cautious mode no? I drive at near idle speeds in these situations.
>driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do
Unfortunately, a vast overestimation of human danger recognition. Or empathy, unsure
anecdotally a couple of kids were hit by cars every year at my very small school of 300 people. pedestrians and drivers both acting recklessly
While I completely agree with your premise, the software can be improved. It can be programmed to drop down to 5 miles an hour or less depending on street size, pedestrian proximity, school zone, etc.
If only the same could be said for the other parents in the school zone. I’ve seen people roar by in similar scenarios at 30+ miles an hour.
Driving is based so much off of feel so my numbers may be off, but in the scenario you are talking about 5mph seems reasonable, 10mph already seems like to much.
The want to be E but really armchair engineer in me for this context says there's far too little Engineering safety of the situation.
That school should not be on a busy roadway at all, it should also not have a child dropoff area anywhere near one but instead, ideally, a slow loop where the parents do drop off children, and then proceed forward in a safe direction away from the school in a flow.
It's funny because now you're sounding like you're blaming the school/the city for the situation.
Things are what they are. Driving situations are never perfect and that's why we adapt. The Waymo was speeding in a school zone. Did a dangerously fast overtake of a double parked car. It's engineering safety failure over engineering safety failure from Waymo's part, on nobody else.
> The Waymo was speeding in a school zone
Source? The article doesn't list a speed limit, but highways.dot.gov suggests to me that the speed limit would be 25mph in the school zone, in which case the waymo was going significantly under the speed limit.
It is 15mph at this school with kids present. So percentage wise kind of high, but in absolute terms not much.
30kph is the posted speed limit in school zones here, precisely because it's an appropriately slow speed for the context.
> not something a human would typically do
lol I'm guessing you don't have kids. This is hilarious.
The full text says: "because driving 17mph (27kph for metric types) in this context is reckless and not something human would typically do"
While i dont have kids, i guess you dont either. Because usually kids dont drive cars, atleast i didnt when i was in elementary school.
If you drive in Sweden you will occasionally come up to a form of speed reduction strategy that may seem counterintuitive. They all add to make driving harder and feel more dangerous in order to force attention and lower speed.
One is to merge opposite directional roads into a single lane, forcing drivers to cooperate and take turn to pass it, one car at a time.
For a combined car and pedestrian road (max speed of 7km/h) near where I live, they intentionally added large obfuscating objects on the road that limited visibility and harder to navigate. This forces drivers to drive very slow, even when alone on the road, as they can't see if a car or person may be behind the next object.
In an other road they added several tight S curves in a row, where if you drive anything faster than 20km/h you will fail the turns and drive onto the artificial constructed curbs.
In other roads they put a sign in the middle of two way roads while at the same time drastically limiting the width to the curb, forcing drivers to slow down in order to center the car in the lane and squeeze through.
In each of those is that a human driver with human fear of crashing will cause drivers to pay extra attention and slow down.
In Bulgaria we have a similar speed reduction strategy but we are a bit ahead of Sweden: We use medium-radius but very deep potholes. If you lose attention for even a split second, you are forced to a full stop to change a tire. Near schools it gets more "advanced": they put parked cars on both sides of the road, and the holes positioned so you can't bypass them. For example, two tire-sized holes on both sides of the road right next to the parked cars. You have to come to a complete stop, then slowly descend into the hole with the front wheels, climb back out, and repeat the process for the rear wheels. Occasionally, even though we (technically) have sidewalks, they are covered in mud or grass or bushes, so pedestrians are forced to walk in the middle of the road. This further reduces driving speed to walking pace and increases safety in our cities. Road markings are missing almost everywhere and they put contradicting road signs so drivers are not only forced to cooperate but also to read each other minds.
Same in India! We go one better, we let people drive in the opposite lane as well!
That’s genius but one has to ask: how much does it cost to maintain these speed restricting features?
In the UK, the cost of owning a car is high yet our potholes, while frequent, are small enough to survive. Thus being more of an annoyance rather than a speed restriction.
That's the best part, the holes maintain themselves. Heck, they even appear without a site survey or paperwork.
It's fairly common at least in the Netherlands, Germany, and Switzerland too. In Switzerland they also place street parking spots on alternating sides on narrow streets, which also makes you more attentive and lower your speed.
I've heard that that is why roundabouts are safer than their alternatives: counterintuitively, they're safer because they're less safe, forcing the user to pay more attention as a result.
>they're safer because they're less safe
Roundabouts are safer. They're safer because they prevent everybody from speeding through the intersection. And, even in case of an accident, no head-on collisions happen in a roundabout.
They're safer specifically for vehicles, as they convert many conflicts that would be t-bones (worst for passengers) into getting rear-ended (maximum crumple zone on both vehicles).
Roundabouts are worse for land use though, which impacts walkability, and the safety story for pedestrians and bike users with them is decidedly not great as well.
>Roundabouts are worse for land use though, which impacts walkability, and the safety story for pedestrians and bike users with them is decidedly not great as well.
They're much safer for pedestrians than intersections. You're only crossing and dealing with traffic coming from one direction, stopping at a median, and then crossing further over.
Unlike trying to navigate a crosswalk where you have to play guessing games as to which direction some vehicle is going to come at you from while ignoring the lights (people do the stupidest things, and roundabouts are a physical barrier that prevents a bunch of that)
In Waterloo Region I used to cycle through multiple intersections that were "upgraded" some years ago from conventional stoplights to roundabouts and imo it was a huge downgrade to my sense of safety. I went from having a clear right of way (hand signal, cross in the crosswalk) to feeling completely invisible to cars, essentially dashing across the road in the gaps in traffic as if I was jaywalking.
I could handle it as an adult just walking my bike but it would be a nightmare for someone pushing a stroller or dependent on a mobility device.
To an extent… drivers tend to accelerate when leaving roundabouts which can make crossing difficult for pedestrians
Especially bad when crossings are like 30cm from the roundabout. Some are better with at least one car's length between the two.
Otherwise you either risk getting run over by a car exiting the roundabout without seeing you; or getting run over by the car that stopped, but was rear-ended by another inside the roundabout.
>from one direction, stopping at a median, and then crossing further over.
This assumes a median, which is not present at most smaller roundabouts in the US.
> and the safety story for pedestrians and bike users with them is decidedly not great as well.
The what now? Seriously, what in the world are you talking about? Roundabouts are heaven. They physically force drivers to slow down when approaching or leaving them, creating a safe environment for pedestrians and cyclists.
For example, there's no such thing as "running a red light at full speed" at a roundabout, no speeding up to "make the light", etc.
For cyclists specifically, they're amazing, because they eliminate the deadly left-turns. Every turn is a right turn, which is super safe.
As a cyclist, I'm not a big fan of roundabouts, because I'm always worried I'll get hit on the side by a car entering/leaving the roundabout whenever I don't take the first exit, mostly because I feel like I have less visibility in the direction from which the car might come from, compared to a standard crossing.
Though I've never been in an accident either on a crossing or roundabout, so I can't really judge how true my impression is.
One-lane-roundabouts are very safe. I lived in Hannover (Germany) in the 80s and 90s, they had 2 or 3 lanes in the roundabouts. There were large signs that counted the accidents (200+/year) to raise awareness and during the trade fairs (anybody remembers Cebit?) the number of accidents peaked. Today they are all a lot safer because of a lot of traffic lights.
I thought that the idea of roundabouts was that they lead to slightly more accidents than before, but they are of much lower severity than before (the 90 degree intersections they replace).
Same with driving in the winter. Anecdotally I always observe more accidents when the roads are clear.
I recently visited a friend that lives in Sweden (couple hours south of Stockholm). Something he said while I visited stuck with me:
"Sweden hates cars."
There must be a happy medium somewhere in between.
I would say it depend on where you are. City driving is generally not a great experience and its not that uncommon to see a speed bump before almost every crossing, to the point where you get surprised if there isn't one. That said, as long you don't leave the designated main roads that goes through the city areas it is not that bad. They do demand a lot of attention.
Street parking has mostly been turned into exclusive residential parking, so parking houses are often the only choice. As a result they are quite expensive, and you got to walk to the destination.
Parking and access is much better in the country side, and the highways are fairly good and similar to those found in the west Europe. It not as straight or wide as authobahn, but not as much traffic either.
It's true, Sweden isn't quite bike and pedestrian friendly enough yet, but they'll get that balance someday!
> There must be a happy medium somewhere in between.
"""
- Gavin, you know our shameful history of worker suicides. Since the renovation? Not a single one.
- Not even one? Ok. But there's gotta be like, a middle ground here...
"""
(https://youtu.be/EyyIrpf68SM?t=57)
why not just put in speedbumps if all you're trying to do is slow people down? Are you sure this was the purpose of these designs? sounds a little too freakonomics to me.
Speed bumps suck for both the driver and passangers of the car and generate road noise.
They also are rather expensive to maintain, because the leading ledge gets many repeated stresses. And in nordic climates like Sweden there is a snow plow in the winter to remove snow - those occasionally snag on the speed bump - which tends to chip of big chunk, triggering rapid wear.
Does it actually work though?
Many roads in London have parked cars on either side so only one can get through - instead of people cooperating you have people fighting, speeding as fast as they can to get through before someone else appears, or race on-coming cars to a gap in the parked cars etc. So when they should be doing 30mph, they are more likely doing 40-45. Especially with EVs you have near-instant power to quickly accelerate to get to a gap first etc.
And putting obstacles in the road so you cant see if someone is there? That sounds really dangerous and exactly the sort of thing that caused the accident in the story here.
Madness.
> Does it actually work though?
Yes. They have made steady progress over the previous decades to the point where they can now have years with zero road fatalities.
> And putting obstacles in the road so you cant see if someone is there? That sounds really dangerous and exactly the sort of thing that caused the accident in the story here.
Counterintuitive perhaps, but it's what works. Humans adjust their behaviour to the level of perceived risk, the single most important thing is to make driving feel as dangerous as it is.
I think the humans in London at least do not adjust their behaviour for the perceived risk!
From experience they will adjust their behaviour to reduce their total travel time as much as possible (i.e. speed to "make up" for lost time waiting etc) and/or "win" against other drivers.
I guess it is a cultural thing. But I cannot agree that making it harder to see people in the road is going to make anything safer. Even a robot fucking taxi with lidar and instant reaction times hit a kid because they were obscured by something.
> I think the humans in London at least do not adjust their behaviour for the perceived risk!
The evidence is that they do though. E.g. the Exhibition Road remodelling (removing curbs/signs/etc.) has been a great success and effectively reduced vehicle speeds, e.g. https://www.rbkc.gov.uk/sites/default/files/media/documents/...
> I think the humans in London at least do not adjust their behaviour for the perceived risk!
Sure they do, all humans do. Nobody wants to get hurt and nobody wants to hurt anyone else.
(Yes there are few exceptions, people with mental disorders that I'm not qualified to diagnose; but vast majority of normal humans don't.)
Humans are extremely good at moderating behavior to perceived risk, thank evolution for that.
(This is what self-driving cars lack; machines have no fear of preservation)
The key part is perceived though. This is why building the road to match the level of true risk works so well. No need for artificial speed limits or policing, if people perceive the risk is what it truly is, people adjust instictively.
This is why it is terrible to build wide 4 lane avenues right next to schools for example.
There are always going to be outlier events. If for every one person who still manages to get hit—at slow, easily-survivable speeds—you prevent five others from being killed, it’s a pretty obvious choice.
Does physics work? If it does, then these physical obstacles work too. Go ahead, try to drive faster than 10mph through a roadway narrowed so much it's barely wider than your car, with curbs. And yeah, I'm describing a place in London.
It's a runaway process of prioritizing safety over convenience -- and it's wrecking their road base just before self-driving cars would allow them to have both.
I was wondering how much convenience is worth one kid's life. This thread reminded me of some interesting terms like "value of statistical life." It appears that all those annoying low speed limits and purposeful obstructions in residential areas really do save lives.
> An evaluation of 20 mph zones in the UK demonstrated that the zones were effective both in reducing traffic speed and in reducing RTIs. In particular child pedestrian injuries were reduced by 70 per cent from 1.24 per year in each area before to 0.37 per year after the zones were introduced
https://www.rospa.com/siteassets/images/road-safety/road-saf...
The "Vision Zero" program was started in Sweden, and is becoming more widely adopted.
https://en.wikipedia.org/wiki/Vision_Zero
20mph residential is pretty close to standard. Note the Waymo car was going slower than that. That's far from the 5mph GP was reacting too, or the super tight curves.
What an American framing. My convenience at the cost of your eventual safety. I guess this is why we also have toddler death machines with 5-foot grills that we call “full size” vehicles.
If you've ever driven more than 5 miles an hour, you risked hurting someone for your convenience.
Acknowledging life has risk tradeoffs doesn't make you an American, but denying it can make you a self-righteous jerk.
> If you've ever driven more than 5 miles an hour, you risked hurting someone for your convenience.
Taken literally, that's clearly not true.
For example you can easily drive 150mph in the flat desert where there is nothing for a hundred miles and you can see many miles ahead. You have zero risk of hurting anyone else unless they somehow teleport in front of you.
But driving 5mph in tight street full of elementary school kids running around can be extremely dangerous.
It's all about context.
You’re egocentric instead of system-centric. Life has risks, but risk is to be managed, not accepted blindly with disregard of available options. A systemic approach to minimizing risk of injury on roads looks exactly like inconvenience to the individual.
In many civilized countries and locales, even bringing up the word “convenience” in the context of road safety would be considered tasteless. Maybe a phrase like “excessively obstructive” or other euphemisms would be used, but the word “inconvenient” regarding safety measures that would e.g. help prevent the death of toddlers today would be appalling.
There’s this techbro utopia mindset leaking through as well, just like it does for climate change topics, that pragmatic solutions that work for us today are deprioritized because some incredible technology is right around the corner. This is also distinctly American, specifically Silicon Valley, culture.
Gosh, no, the self-driving cars will be forced to drive at safe speeds in pedestrian corridors as opposed to voluntarily driving at safe speeds in pedestrian corridors. How awful.
> prioritizing safety over convenience
this sounds like exactly the right tradeoff, especially since these decisions actually increase convenience for those not in cars
Of course it sound right, because you cut off the word "runaway".
It is possible to go too far in either direction.
It is, but it's laughable to suggest it's happening anywhere. Our world is dominated by cars. You likely can't see it precisely because it's so normalized.
“Just before” … this would mean all cars would be required to be self driving and that they’re forced to adhere to the set speed limits. You think this is just around the corner? In a country like Sweden with a lot of snow? Let’s talk about that this when we’re actually close to hitting 100% of self driving cars on the road.
And it’s not “runaway”, it’s exactly the right prioritisation. I’d encourage you to spend some time on Not Just Bikes and the say whether you’d like to live in a Nordic or an American neighbourhood. The Nordic style is also about convenience because car centric infrastructure makes a lot of things less accessible and convenient.
Those things all sound easy to remove in some hypothetical future where there are enough and safe enough self driving cars to have both. Makes sense to design for human driven cars for now though.
If they're actually self-driving they should be able to drive around the obstacles just as well or better than human.
> It's likely that a fully-attentive human driver would have done worse.
We'd have to see video of the full scene to have a better judgement, but I wouldn't call it likely.
The car reacted quickly once it saw the child. Is that enough?
But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
If that's the scenario, there is a real probability that a child might appear, so I'm going to be over-slowing way down pre-emptively even thought I haven't seen anyone, just in case.
The car only slows down after seeing someone. The car can react faster that I can after seeing someone, but as a human I can pre-react much earlier based on the big picture, which is much better.
As someone who lives on a residential street right by a primary school in the UK, the majority of drivers are going over 20mph even at the peak time when there are children everywhere.
While in theory human drivers should be situationally aware of the higher risks of children being around, the reality is that the majority will be in their own bubble of being late to drop their kid off and searching for the first free spot they can find.
the human driver would usually drive more closely to the centerline of such a residential road. If the road is clear ahead i'd drive almost over the centerline of the road having enough clearance between my path and the parked cars for any such "jumper" to be visible long enough for me to react. If there is an opposite traffic i get back strictly into my lane and slow down much more if the parked cars are close and they block sidewalk view, etc.
The autonomous cars have really got more aggressive recently as i mentioned before:
https://news.ycombinator.com/item?id=46199294
Also Waymo handling road visibility issue:
https://news.ycombinator.com/item?id=39298290
I'd really like to see the video of the incident.
I have a similar school drop-off, and can confirm that the cars are typically going around 17-20mph around the school when they're moving. Also that yes, human drivers usually do stay much closer to the centerline.
However, Waymo was recently cleared to operate in my city, and I actually saw one in the drop-off line about a week ago. I pulled out right in front of it after dropping my kid off. And it was following the line of cars near the centerline of the road. Honestly its behavior was basically indistinguishable from a human other than being slightly more polite and letting me pull out after I put my blinker on.
> the human driver would usually drive more closely to the centerline of such a residential road
I certainly do this. But asserting that most humans would usually do this? Have you ever actually seen humans drive cars? This is absolutely not what they do. On top of that, they run stop signs, routinely miss pedestrians in blind spots, respond to texts on their phone, or scroll around on their display to find the next song they want to put on.
I vividly recall a shot within a commercial, in which a driver was shown in slow motion, chucking his coffee into the passenger foot well in order to have two hands on the wheel for an emergency. I don’t remember what was about to happen to the car or the world around it. I’m pretty sure that a collision occurred.
Have you been in a waymo? It knows when there are pedestrians around (it can often see over the top of parked cars) and it is very cautious when there are people near the road and it frequently slows down.
I have no idea what happened here but in my experience of taking waymos in SF, they are very cautious and I'd struggle to imagine them speeding through an area with lots of pedestrians milling around. The fact that it was going 17mph at the time makes me think it was already in "caution mode". Sounds like this was something of a "worst case" scenario and another meter or 2 and it would have stopped in time.
I think with humans, even if the driver is 100% paying attention and eyes were looking in exactly the right place where the child emerged at the right time, there is still reaction times - both in cognition but also physically moving the leg to press the pedal. I suspect that a waymo will out-react a human basically 100% of the time, and apply full braking force within a few 10s of milliseconds and well before a human has even begun to move their leg.
You can watch the screen and see what it can detect, and it is impressive. On a dark road at night in Santa Monica it was able to identify that there were two pedestrians at the end of the next block on the sidewalk obscured by a row of parked cars and covered by a canopy of overgrown vegetation. There is absolutely no way any human would have been able to spot them at this distance in these conditions. You really can "feel" it paying 100% attention at all times in all directions.
Yep I've often noticed this as well - it has many many times detected humans that I can't even see (and I like to sit in the front), especially at night.
Sometimes it would detect something and I think "huh? Must be a false positive?" but sure enough it turns out that there really was someone standing behind a tree or just barely visible around a corner etc.
Sure none of those have run out in front of us, but the fact it is spotting them and tracking their movement before I am even aware they're there is impressive and reassuring.
> I suspect that a waymo will out-react a human basically 100% of the time, and apply full braking force within a few 10s of milliseconds and well before a human has even begun to move their leg.
Correct. Human reaction time is at its very best ~250ms. And that's when you're hyper-focused on reacting to a specific stimuli and actively try to respond to it as fast as possible.
During normal driving, a focused driver will react on the order of 1s. However, that's assuming actively paying attention to the road ahead. If you were to say, be checking your mirrors or looking around for any other reason this can easily get into multiple seconds. If you're say, playing on your phone (consider how many drivers do this), forget it.
A machine however is 100% focused 100% of the time and is not subject to our poor reaction times. It can brake in <100ms every time.
On the other hand, a software fault could make it run into an obstacle that'd be obvious to a human at full speed.
Your opinion of "most humans" is vastly overinflated. The median human driver would be going 5 over the speed limit, on their cell phone, and paying fuck all attention. Humans never drive as slow as 17 mph, even in the context of being directly in front of schools with visible children.
You're describing the median driver in America or India. This is not universal.
True, but it seems fair to evaluate Waymo against the median American driver. If they expand to whatever other countries you're thinking of, then it will be fair recalibrate accordingly.
For context, Santa Monica is in America or India.
> But most humans would have been aware of the big picture scenario much earlier.
I wouldn't call it likely. Sure, there are definitely human drivers who are better than Waymo, but IME they're few and far between. Much more common to be distracted or careless.
When walking along a busy street facing traffic, I like to play a game of "who's using a phone?" I sometimes score in excess of 50% of drivers texting or otherwise manipulating a phone instead of actually driving.
It's amazing how much nonsense we let slide with human drivers, and then get uptight about with anything else. You see the same attitude with bicycles. Cars run stop signs and red lights all day long and nobody bats an eye, but a cyclist does it and suddenly they're a menace.
I don't think it makes sense to lump some drivers better than waymo and worse than waymo. A human brain automatically thinks of all the scenarios, where Waymo has pre-programmed ones (and some NN based ones). So it's scenarios by scenario.
Consider this scenario:
5 kids are walking on the sidewalk while you're driving past them. But suddenly a large dumpster is blocking your view of them just as you pass. You saw them before the dumpster, but not after your car and the dumpster completely blocks the view.
Does a human brain carry some worry that they suddenly decide to run and try to cross the street after the dumpster? Does Waymo carry that worry or just continue to drive at the exact same speed.
Again, it's not like every driver will think about this, but many drivers will (even the bad ones).
> A human brain automatically thinks of all the scenarios
I don't think this is true. There are infinitely many scenarios in a complex situation like a road with traffic, cars parked, pedestrians about, weather, etc. My brain might be able to quickly assess a handful, but certainly not all.
> like a road with traffic, cars parked, pedestrians about, weather
Not all of those need to be done "quickly". That's where LLMs fail
You note the weather when you leave. You understand the traffic five minutes ahead. You recognize pedestrians far ahead of time.
Computers can process a lot in fractions of a second. Humans can recognize context over many minutes.
The Waymo may have done better in the fraction of a second, but humans can avoid being in that situation to begin with.
Computers can take all of those things into account as well
Can, but don't.
It doesn't seem like self driving cars take into account the icy conditions of roads for one simple example.
There aren't infinitely many scenarios to consider, but even if that's a figure of speech, there aren't thousands or even hundreds.
If there's ten kids nearby, that's basically ten path scenarios, and that might be reduced if you have great visibility into some of them.
> My brain might be able to quickly assess a handful, but certainly not all.
What would you do if you can't assess all of them? Just keep driving same speed?
If the situation is too overwhelming you'll almost certainly back off, I know I would. If I'm approaching that school block and there's like 50 small kids running around in all directions, I have no idea what's going on and who is going where, so I'm going to just stop entirely until I can make some sense of it.
It should be trivial for Waymo to implement a "drive carefully near schools" feature, and if really spicy "drive REALLY carefully near schools at these times" feature.
Safe driving starts with speed, lowering speed and informing the passengers seems like a no-brainer.
Feels like bitter lesson fodder to special case things like this
> here aren't infinitely many scenarios to consider, but even if that's a figure of speech, there aren't thousands or even hundreds.
There are a very, very large number of scenarios. Every single possible different state the robot can perceive, and every possible near future they can be projected to.
Ten kids is not 10 path scenarios. Every kid could do a vast number of different things, and each additional kid raises the number of joint states to another power.
This is trivially true. The game that makes driving possible for humans and robots is that all these scenarios are not equally likely.
But even with that insight, it’s not easy. Consider a simple case of three cars about to arrive at an all-way stop. Tiny differences in their acceleration - potentially smaller differences than the robot can measure - will result in a different ordering of cars taking turns through the intersection.
It’s a really interesting problem.
> Ten kids is not 10 path scenarios. Every kid could do a vast number of different things, and each additional kid raises the number of joint states to another power.
This is the difference between computing and humans. The car will attempt to compute all possible path scenarios because it has no instict, and it might not be possible to compute everything in real time so it might fail.
But the human will easily deal with the situation.
Try running through a sports field in an elementary school during lunch, full of unpredictable kids running around. Can you make it from one side to the other without crashing into a whole bunch of kids? Of course you can. You didn't need to try to compute an exponential number of scenarios, you just do it easily. The human brain is pretty amazing.
In fact no computer approach attempts to compute all possible path scenarios since we know that’s not tractable.
And current practical approaches are mostly end to end (or nearly) ML systems that do not compute a lot of alternative paths, and they work in approximately constant time independent of the scenario.
You strongly imply that computers can’t drive, but you could have written that in a Waymo.
It was a figure of speech, but I think you're undercounting. When you consider interactions between all the things, even with just a handful of variables (and I think there are many more than a handful) you get a huge number of scenarios.
This is the classical ‘Frame Problem” of AI. How do you consider, even if only to reject, infinite scenarios in finite time? Humans and other animals don’t seem to suffer from it.
God I wish I re-read my statement, I was more focused on Humans think of an unlimited number of scenarios - not necessarily all. A computer will only think of pre-programmed ones.
The computer isn't pre-programmed though. These computers are trained similar to how human brains are (though obviously brains are still vastly, vastly, vastly superior to computers for tasks like this).
You are vastly overestimating most drivers. Most drivers aren't even looking out the window the majority of their time driving.
> A human brain automatically thinks of all the scenarios, ...
Patently, obviously false. A human brain will automatically think of SOME scenarios. For instance, if a collision seems imminent, and the driver is holding a cup of coffee, these ideas are likely to occur to the driver:
IF I GRAB THE STEERING WHEEL AND BRAKE HARD, I MIGHT NOT HIT THAT PEDESTRIAN IN FRONT OF ME.
IF I DON'T CONTINUE HOLDING THE COFFEE CAREFULLY, I MIGHT GET SCALDED.
THIS SONG ON MY RADIO IS REALLY ROCKING!
IF I YANK MY WHEEL TO THE LEFT, I MIGHT HIT A CAR INSTEAD OF A HUMAN.
IF I BRAKE HARD OR SWERVE AT ANY TIME IN TRAFFIC, I CAN CAUSE AN ACCIDENT.
Experiments with callosal patients (who have damaged the connective bridge between the halves of their brains) demonstrate that this is a realistic picture of how the brain makes decisions. It offers up a set of possible actions, and attempts to choose the optimal one and discard all others.
A computer program would do likewise, EXCEPT it won't care about the coffee cup nor the radio (remove two bad choices from consideration).
It still has one bad choice (do nothing), but the SNR is much improved.
I'm not being hyperbolic; self-preservation (focusing on keeping that coffee in my hand) is a vital factor in decision-making for a human.
> ...where Waymo has pre-programmed ones (and some NN based ones).
Yes. And as time goes on, more and better-refined scenarios will be added to its programming. Eventually, it's reasonable to believe the car software will constantly reassess how many humans are within HUMAN_RUN_DISTANCE + CAR_TRAVEL_DISTANCE in the next block, and begin tracking any that in an unsafe margin. No human on Earth does that, continually, without fail.
> Does a human brain carry some worry that they suddenly decide to run and try to cross the street after the dumpster? Does Waymo carry that worry or just continue to drive at the exact same speed.
You continue to imply that Waymo cannot ever improve on its current programming. Does it currently consider this situation? Probably not. Will it? Probably.
God I wish I re-read my statement, I was more focused on Humans think of an unlimited number of scenarios - not necessarily all. A computer will only think of pre-programmed ones.
For what it's worth, that kind of lumping of drivers is more-or-less one of the metrics Waymo is using to self-evaluate. Perfect safety when multi-ton vehicles share space with sub-300-pound humans is impossible. But they ultimately seek to do better than humans in all contexts.
According to the article the car was traveling at 17 miles an hour before it began braking. Presumably this was in a 25 mph school zone, so it seems the Waymo was already doing exactly what you describe - slowing down preemptively.
This is close to a particular peeve I have. Occasionally I see signs on the street that say "Slow Down". I'm not talking about the electronic ones connected to radar detectors. Just metal and paint.
Here's my problem. If you follow the instructions on the sign, it still says to slow down. There's no threshold for slow enough. No matter how slow you're going, the sign says "Slow Down". So once you become ensnared in the visual cone of this sign, you'll be forced to sit stationary for all eternity.
But maybe there's a loop-hole. It doesn't say how fast you must decelerate. So if you come into the zone going fast enough, and decelerate slowly enough, you can make it past the sign with some remaining non-zero momentum.
You know, I've never been diagnosed on the spectrum, but I have some of the tendencies. lol.
Obviously a static sign is not aware of your current state, so it's message can only be interpreted as relevant to your likely state... i.e. the posted speed limit.
If you should slow down relative to the posted speed limit why not change the speed limit to reflect that directly?
Usually the reason is the "slow down" portion is very small, and it's confusing to shift down the actual speed limit for a 200 foot stretch of road then increase it again.
There are signs for that. Advisory speeds that don't change the actual limit. https://wisconsindot.gov/PublishingImages/doing-bus/local-go...
Much better to be specific than a vague "slow down". There's a road near me with two tight turns a couple blocks apart. One advises 25mph and the other advises 10mph.
FWIW, it seems less confusing to me than longer speed limits, but with "Slow Down".
Except we do that all the time in school zones... normally 35+, but from 7am-9am and again from 2pm-4pm the limit drops to 25mph (which is still to fast if the kids are actually crossing the street or walking alongside en masse).
Think of it like they're saying "my children play on this street and my neighbors walk here. Please think about that when you decide how fast to go here."
A lot of clickbait headlines have the same problem. "You're using too much washing powder!"
Everyone's replying to you as if you truly don't understand the sign's intention but I'm sure you do. It's just annoying to be doing everything right and the signs and headlines are still telling you you're wrong.
There was a driving safety safety ad campaign here: "Drive to the conditions. If they change, reduce your speed." You can imagine how slow we'd all be going if the weather kept changing.
We might have OCPD.
Yes. You have understood precisely the spirit in which I intended it.
In advertising: "Treat yourself. You deserve it!"
Me: What if someone who didn't deserve it heard this message. How can you possibly know what I deserve? Do all people deserve to treat themselves? Is the notion of deserving or treating really so vacuous?
Normies: jfc
There's a mental health awareness campaign going on around here at the moment with all the generic messages like that. "You're doing great" is completely devalued by the sign giving the same message to everyone, and the best one says something like "Don't push yourself too hard. If you want to rest, rest." Wondering if I can tell my boss the sign told me it's okay not to get any work done.
Humans are supposed to deal with this kind of ambiguity. Actually, that's one of our nicest abilities.
I hate when people pretend to be smarter than everyone else by pointing this kind of utterance and insisting that someone, somehow, will parse those statements in the most literal and stupid manner.
Then there are the ignorant misanthropes that can't waste a chance to repeat their reductionist speculations about human cognition. Just like the idiot Elon Musk that wasted billions in irrecoverably fucked self-driving system based on computer-version because he underestimated the human visual cortex.
Fucking annoying midwits.
Sorry I made you mad. I wasn't trying to seem smarter than everyone. Maybe dumber.
A very kind response to such a grating lack of self-awareness.
You learn how to put those signs into context during your driving lessons, and fail your test if you don't apply that correctly.
My driving test was so thorough that I had to parallel park between two entirely fictional cars. There was certainly no consideration of eccentric signage.
I apologize if I gave the impression that I did not understand how to put them into context. Although I don't think my driving lessons ever mentioned it.
This is idle XKCD-style musing.
Think of the sign as a flag, not an instruction.
25 mph is an upper limit, not a minimum.
A 25mph school zone? That seems fast. 15mph would be more the norm, which is in line with the 17mph the car believed itself to be traveling.
FYI, unless you are a commerical truck, a cop, or a racer, your speedometer will read slightly fast, sometimes as much as 5 to 10%. This is normal practice for cars as it limits manufacturer liability. You can check this using independant gps, ie not an in-dash unit. (Just imagine the court cases if a speedo read slower than the actual speed and you can understand why this started.)
I mostly see 25 mph for school zones, though I'm in NC. Checking California, it sounds like 25 is standard there as well.[0] Some will drop to 15, but 25 is the norm as far as I can find.
[0] https://www.dmv.ca.gov/portal/handbook/california-driver-han...
I've lived all over California and I agree that 25mph is the norm here.
Edit: However, elsewhere in the thread someone linked this Streetview image that shoes that this particular school zone is 15mph: https://maps.app.goo.gl/7PcB2zskuKyYB56W8?g_st=ac
25 mph is typical non-school-zone residential around here, with school zones always slower.
Also, a different wheel diameter than the speedometer was calibrated with and you will have a larger difference between actual velocity and speedometer reading. The odometer will also not record actual distance traveled.
It depends. I had a honda motorcycle where the speedo was 10ish % fast (not unussual on bikes due to tire shape) but the odo was accutrate. Same sensor, but the computer just counted wheel rotations slightly differently for each use.
Virtually all speedos read fast. The federal standards have a fairly high margin for being allowed to read high, and a zero margin for reading low. Thus speedos are more or less universally calibrated to read at least 5% high.
It does seem fast to me -- school zones are 20 mph in Seattle, at least when children are present. But Google suggests 25 is the norm in Santa Monica, where the incident occurred.
In Encinitas, California, that sign would have no more than 20 MPH. In adjacent Carlsbad, I believe 25 is normal.
It was going 17 mph. That is rather slow.
To put it another way. If an autonomous vehicle has a reaction time of 0.3 seconds, the stopping distance from 17 mph is about the same as a fully alert human driver (1 second reaction time) driving 10.33 mph.
>It was going 17 mph. That is rather slow.
There's a case to be made that it wasn't slow enough.
I have a hard time believing that a human driver would be as slow as this Waymo, or even slower. I drive my kid to school where it's posted 20mph and there are cameras (with plenty of warnings about the presence of said cameras) and witness a constant string of flashes from the camera nailing people for speeding through there.
Hardcoded limits are problematic because they completely lack context.
On that very same road with a 20mph limit, 40mph might be completely safe or 3mph might be extremely negligently dangerous. It all depends on what is going on in the area.
A small child jumped out in front of it, which is about the worst case scenario you can have... and the kid was fine. So it sounds like it was slow enough?
Considering the car hit the child at only 6 mph and the kid just got up and brushed themselves off, it was plenty slow enough.
Nobody was injured.
Maybe. That level of safetyism seems pretty unreasonable when humans are 100x worse and still allowed on the road.
In this situation, the car was already driving under the legal speed required for a school zone (25mph when children are present) [edit: some comments in the post suggest there is a 15mph sign, which is sometimes posted; to me, driving 17mph in a 15mph zone is acceptable).
I think any fair evaluation of this (once the data was available) would conclude that Waymo was taking reasonable precautions.
> was already driving under the legal speed
That's exactly part of the problem. If it is programmed to be over-cautious and go 17 in a 25 zone, that feels like it is safe. Is it?
It takes human judgment of the entire big picture to say meaningfully whether that is too slow or too fast. Taking the speed limit literally is too rigid, something a computer would do.
Need to take into account the flow of the kids (all walking in line vs. milling around going in all directions), their age (younger ones are a lot more likely to randomly run off in an unsafe direction), what are they doing (e.g. just walking, vs. maybe holding a ball that might bounce and make them run off after it), their clustering and so on.
Driving past a high school with groups of kids chatting on the sidewalk, sure 20mph is safe enough. Driving past an elementary school with a mass of kids with toys moving in different directions on the same sidewalk, 17mph is too fast.
And if I'm watching some smaller kids disappear behind a visual obstruction that makes me nervous they might pop up ahead of it on the street, I slow down to a crawl until I can clearly see that won't happen.
None of this context is encoded in the "25mph when children are present" sign, but for most humans it is quite normal context to consider.
But would be great to see video of the Waymo scene to see if any of these factors was present.
Curiously enough Google could have access to how fast humans usually drive through that street.. if they record people's Google Maps trips, they can show the court that "Look, 80% of Google Maps users drive through here at 30 mph!".
Waymo itself has this as well. They record their drives after all which means they know the speed of the vehicles around them.
They even wrote a blog post about it:
https://waymo.com/blog/2023/07/past-the-limit-studying-how-o...
Google might even know how many drivers aren't obeying the speed limit or slow-rolling through stop signs. I wonder if they already have partnerships with law enforcement to detect areas where the traffic law is more ignored than others.
Two things:
I've read studies saying that most drivers don't brake at max effort, even to avoid a collision. This may be at least one of the reasons that Waymo predicted that an attentive human would likely have been going faster than their car at the moment of impact. I've got a good idea of my fun-car's braking performance, because I drive it hard sometimes, but after reading that I started practicing a bit with my wife's car on the school run, and... Yeah: it's got a lot more braking power than I realized. (Don't worry, I brake hard on a long straight exit ramp, when no one's behind me, a fast slow-down is perfectly safe, and the kiddo loves it.) I've now got an intuitive feel for where the ABS will kick in, and exactly what kind of stopping distance I have to work with, which makes me feel like a safer driver.
Second, going off my experience of hundreds and hundreds of ride-share rides, and maybe thirty Waymo journeys, I'd call the best 10-15% of humans better drivers than Waymo. Like, they're looking further up the road to predict which lane to be in, based on, say, that bus two blocks away. They also drive faster than Waymos do, without a perceptual decrease in safety. (I realize "perceptual" is doing some work in that sentence!) That's the type of defensive and anticipatory urban driver I try to be, so I notice when it's done well. Waymo, though, is flat-out better, in every way, than the vast majority of the ride-share drivers I see. I'm at the point where I'll choose a Waymo any time it'll go where I'm headed. This story reinforces that choice for me.
> I've read studies saying that most drivers don't brake at max effort, even to avoid a collision.
Ha! It is unbelievable how difficult it is to make someone brake hard. You'd think it's the easiest thing possible in the age of ABS - just press hard as you can.
I have a lot of experience on this, I used to teach car control both to teens and adults. One of the frequent exercises was seemingly very simple: Drive at Xmph until this spot, then brake at maximum power.
The vast majority of people can't do it on the first or second try, they'll just meekly press on the brake like they're coasting to a stop. After more coaching that hard means hard, they start to get it, but it takes many many tries.
If you brake too hard, you might fly over the handlebars.
These training sessions were on cars only, but on motorcycles yes that is possible.
The reason attentive humans don't equal the Waymo here is reaction time. When a thing happens the human takes a moment to process what it means, and choose a reaction. It's not, by our standards, a long time but it's way longer than it takes the Waymo.
Going early means you slow early, which means you also take longer to reach the child, but you're braking for all of that extra time, so you're slowing down even more.
> The car only slows down after seeing someone.
How do you know that? The article says it slowed from 17 mph. That’s cautious progress speed, not cruising speed.
> But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
Waymos do this and have for years. They know where the people are around them and will take precautionary action based on that.
Here's a video from 2019 of one understanding that a car in the bike lane means the cyclists may dart out into the lane it's in and taking action based on that. https://waymo.com/blog/2019/05/safety-at-waymo-self-driving-...
That video is nearly 7 years old at this point and they've gotten much, much better since then.
If you think a fully-attentive human driver would have done better, I think you're kidding yourself.
I know you didn't make this point, but if anyone think the average LA driver would have done better than this I've got a bridge to sell you and that's really what matters more. (I say that as someone who used to live like half a mile from where this happened)
It would be nice to see the video (although maybe there are some privacy issues, it is at a school after all).
Anyway, from the article,
> According to the NHTSA, the accident occurred “within two blocks” of the elementary school “during normal school drop off hours.” The safety regulator said “there were other children, a crossing guard, and several double-parked vehicles in the vicinity.”
So I mean, it is hard to speculate. Probably Waymo was being reasonably prudent. But we should note that this description isn’t incompatible with being literally in an area where the kids are leaving their parents’ cars (the presence of “several double parked cars brings this to mind). If that’s the case, it might make sense to consider an even-safer mode for active student unloading areas. This seems like the sort of social context that humans might have and cars might be missing.
But things speculation. It would be good to see a video.
In principle, attentive drivers, who have either somehow come independently to the appropriate understanding or have been trained in how to react to hazards ahead...
https://www.gov.uk/theory-test/hazard-perception-test
... could in some circumstances know that there's a likelihood that a child will emerge suddenly and reduce their speed in anticipation where circumstances allow.
Note that: If you cut speed but other drivers can't see why they may overtake, even unsafely, because you are a nuisance to them. Slowing in anticipation that a child will run out from behind the SUV, only for a car behind you to accelerate around you and smack straight into the child at even higher speed, is not the desired outcome even though you didn't hurt anybody...
And yes, we'd need to see the video to know. It's like that Sully scenario. In a prepared test skilled pilots were indeed able to divert and land, but Sully wasn't prepared for a test. You're trained to expect engine failure in an aeroplane - it will happen sometimes so you must assume that, but for a jet liner you don't anticipate losing both engines, that doesn't happen. There's "Obviously that child is going in the road" and "Where the fuck did they come from?" and a lot in between and we're unlikely to ever know for sure.
It was already moving slowly. 17MPH is pretty conservative. Most human drivers going past my local school are doing at least 30.
There's a bus stop right behind my house. I routinely hear the driver honking and yelling at people who ignore when the stop sign is extended (which is a misdemeanor in my state). So forgive me for not assuming a human would have done better.
I live in an area where there are pedestrians stepping into the street without looking, all over the place, and you can drive / cycle without hitting them but have to slow down appropriately if you have to go near something that you can't see behind. Like you say it would be interesting to see the video.
The car was driving 17mph before braking. I don’t think I’ve ever seen a human drive at 17mph in a school zone or other area children congregate.
Meaning you’ve never seen a human drive that slowly in such an area, or you've never seen a human exceed the speed limit in a school zone?
Drive that slowly. I’m exaggerating obviously but the norm is to speed.
I drive like this too, but I think we’re a small minority. Especially here in LA.
> But most humans would have been aware of the big picture scenario much earlier. Are there muliple kids milling around on the sidewalk? Near a school? Is there a big truck/SUV parked there?
Waymos constantly track pedestrians nearby, you can see it on the status screen if you ride in one. So it would be both better able to find pedestrians and react as soon as one was on a collision course. They have a bit more visibility than humans do due to the sensor placement, so they also can see things that aren't that visible to a person inside the car, not to mention being constantly aware of all 360 degrees.
While I suppose that in theory, a sufficiently paranoid human might outdo the robot, it looks to me like it's already well above the median here.
Do they speculate about things like “we’re near a school zone, kids are unloading, there might be a kid I’ve never seen behind that SUV?” (I’m legitimately asking I’ve never been in a Waymo).
It's not particularly meaningful to ponder the subjective experience of the waymo driving computer. Instead, focus on its externally visible behavior.
Asking whether an entity has modeled and evaluated a specific situation, using that evaluation to inform its decisions, is not about subjective experience.
If you're asking whether their training data includes situations like this, and whether their trained model/other pieces of runtime that drive the car include that feature as part of their model, the answer is yes. But not in the way a normal human driver would think about it; many of the details of its decision making process are based on large statistical collections, rather than "I'm in a school zone and need to anticipate children may be obscured and run out into traffic." There are many places where the car needs to take caution without knowing specifically it's within 50 feet of a school zone.
While the deep details are not public, Waymo has shared a fair amount of description of their system, from which you can glean some ideas about the world model it creates and the actions it takes in specific situations: https://waymo.com/blog/2024/10/ai-and-ml-at-waymo https://waymo.com/blog/2025/12/demonstrably-safe-ai-for-auto... https://waymo.com/blog/2024/10/introducing-emma
I was being informal with “speculate,” sorry. They could identify that sort of situation as a high-risk area in some way.
Are they not using a ton of ML to take exactly this sort of context into account?
>The car can react faster that I can after seeing someone
and that can potentially allow internal planning algorithm to choose more risky and aggressive trajectories/behavior, etc. say to reach target destination faster and thus deliver higher satisfaction to the passengers.
Anecdote, but I live next to an elementary school and also on a route frequented by Waymos. Human drivers routinely cruise down the 25mph roads at 40+ and blow stop signs, even during school intake and release. Waymo vehicles always seem a lot more cautious.
When thinking about these things you have to factor in the prior probability that a driver is fully attentive, not just assume they are.
If you’ve ever been in a Waymo you quickly realize their field of view is pretty good. You often see the vehicle sensing small pets and children that are occluded to a passenger or driver. For this reason and my experience with humans near aforementioned school, I doubt a human would out perform the Waymo in this particular incident and it’s debatable they even have more context to inform their decisions.
All that said, despite having many hours in a Waymo, it’s not at all clear to me how they factor in sidewalk context. You get the sense that pedestrians movement vectors are accounted for near intersections, but I can’t say I’ve experienced something like a slow down when throngs of people are about.
Precisely. Environmental context is not considered in Waymo's "peer-reviewed model" (I encourage reflexive commenters to first read it: https://waymo.com/safety/collision-avoidance-benchmarking), only basic driver behavior and traffic signal timings.
Note the weaselly "immediately detected the individual as soon as they began to emerge" in the puff piece from Waymo Comms. No indication that they intend to account for environmental context going forward.
If they already do this, why isn't it factored in the model?
How is "immediately detected the individual as soon as they began to emerge" worded weaselly?
Not OP but I interpret that as they are focusing exclusively on what happened after the car saw the kid.
And I completely agree that from that instant forward, the car did everything correctly.
But if I was the accident investigator for this, I would be far more interested in what happened in the 30 seconds before the car saw the kid.
Was the kid visible earlier and then disappear behind an obstruction? Or did the kid arrive from the side and was never earlier visible? These are the more important questions.
Possibly, but Waymos have recently been much more aggressive about blowing through situations where human drivers can (and generally do) slow down. As a motorcyclist, I've had some close calls with Waymos driving on the wrong side of the road recently, and I had a Waymo cut in front of my car at a one-way stop (t intersection) recently when it had been tangled up with a Rivian trying to turn into the narrow street it was coming out of. I had to ABS brake to avoid an accident.
Most human drivers (not all) know to nose out carefully rather than to gun it in that situation.
So, while I'm very supportive of where Waymo is trying to go for transport, we should be constructively critical and not just assume that humans would have been in the same situation if driving defensively.
Certainly, I'm not against constructive criticism of Waymo. I just think it's important to consider the counterfactual. You're right too that an especially prudent human driver may have avoided the scenario altogether, and Waymo should strive to be that defensive.
> I'm not against constructive criticism of Waymo.
I feel like you have to say this out loud because many people in these discussions don't share this view. Billion dollar corporate experiments conducted in public are sacrosanct for some reason.
> I just think it's important to consider the counterfactual
More than 50% of roadway fatalities involve drugs or alcohol. If you want to spend your efforts improving safety _anywhere_ it's right here. Self driving cars do not stand a chance of improving outcomes as much as sensible policy does. Europe leads the US here by a wide margin.
> I feel like you have to say this out loud because many people in these discussions don't share this view. Billion dollar corporate experiments conducted in public are sacrosanct for some reason.
Yes, and I find it annoying that some people do seem to think Waymo should never be criticized. That said, we already have an astounding amount of data, and that data clearly shows that the experiment is successful in reducing crashes. Waymos are absolutely, without question already making streets safer than if humans were driving those cars.
> If you want to spend your efforts improving safety _anywhere_ it's right here.
We can and should do both. And as your comment seems to imply but does not explicitly state, we should also improve road design to be safer, which Europe absolutely kicks America's ass on.
> and that data clearly shows that the experiment is successful in reducing crashes
I disagree. You need way more data, like orders of magnitude more. There are trillions of miles driven in the US every year. Those miles often include driving in inclement weather which is something Waymo hasn't even scraped the surface of yet.
> without question
There are _tons_ of questions. This is not even a simple problem. I cannot understand this prerogative. It's far too eager or hopeful.
> We can and should do both
Well Google is operating Waymo and "we" control road policy. One of these things we can act on today and the other relies on huge amounts of investments paying off in scenarios that haven't even been tested successfully yet. I see an environment forming where we ignore the hard problems and pray these corporate overlords solve the problem on their own. It's madness.
> You need way more data, like orders of magnitude more. There are trillions of miles driven in the US every year.
Absurd, reductive, and non-empirical. Waymos crash and cause injury/fatality far less frequently than human drivers, full stop. You are simply out of your mind if you believe otherwise, and you should re-evaluate the data.
> Those miles often include driving in inclement weather which is something Waymo hasn't even scraped the surface of yet.
Yes. No one is claiming that Waymos are better drivers than humans in inclement weather, because they don't operate in those conditions. That does not mean Waymos are not able to outperform human drivers in the conditions in which they do operate.
> I see an environment forming where we ignore the hard problems and pray these corporate overlords solve the problem on their own. It's madness.
What's madness is your attitude that Waymos' track record does not show they are effective are reducing crashes. And again, working on policy does not prevent us from also improving technology as you seem to believe it does.
You're moving the goalposts. The claim is that Waymos are safer than human drivers in the areas and under the conditions where they currently operate.
Yeah, I'm sure Waymos would struggle in a blizzard in Duluth, but a) so would a human and b) Waymos aren't driving there. (Yet.)
> You're moving the goalposts
No. I'm not. I'm being realistic about the technology. You're artificially limiting the scope.
> so would a human
This is goalpost moving 101. The question isn't would a human driver also struggle but _would it be better_? You have zero data.
> This is goalpost moving 101. The question isn't would a human driver also struggle but _would it be better_? You have zero data.
It is not moving the goalpost to say "so would a human". Comparison to human drivers is exactly the stated goalpost (and it should be).
> You have zero data.
Outrageously uninformed take. We have mountains of data that show Waymos in aggregate are safer drivers than humans.
>data clearly shows that the experiment is successful in reducing crashes.
That's fine. But crashes are relatively rare and what matters is accountability. Will Waymo be accountable for hitting this kid the way a human would? Or will they fight in court to somehow blame the pedestrian? Those are my big concerns when it comes to self driving vehicles, and history with tech suggests that they love playing hot potato instead of being held accountable.
And yes, better walkable infrastructure is a win for all. The minor concern I have is the notion that self driving is perfect and we end up creating even more car centric infrastructure. I'm not sure who to blame on that one.
Waymo is driving the car and should be held accountable like any other driver.
I assume that's how it works already.
I hope so too. I'll be keeping a close eye on how they handle this, though. My benefit of the doubt for tech was already long drained, and is especially critical for safety critical industries.
> More than 50% of roadway fatalities involve drugs or alcohol. If you want to spend your efforts improving safety _anywhere_ it's right here. Self driving cars do not stand a chance of improving outcomes as much as sensible policy does. Europe leads the US here by a wide margin.
Could you spell out exactly what "sensible" policy changes you were thinking of? Driving under the influence of drugs and/or alcohol is already illegal in every state. Are you advocating for drastically more severe enforcement, regardless of which race the person driving is, or what it does to the national prison population? Or perhaps for "improved transit access", which is a nice idea, but will take many decades to make a real difference?
>Driving under the influence of drugs and/or alcohol is already illegal in every state.
FWIW, your first OWI in Wisconsin, with no aggravating factors, is a civil offense, not a crime, and in most states it is rare to do any time or completely lose your license for the first offense. I'm not sure exactly what OP is getting at, but DUI/OWI limits and enforcement are pretty lax in the US compared to other countries. Our standard .08 BAC limit is a lot higher than many other countries.
That's true, but note that getting much more severe on enforcement and punishment for DUI/OWI will result in an even higher prison population, more serious life consequences for poor and minorities, etc, when the US is constantly getting trashed for how bad those things are already.
To be a bit snarkier, and not directed at you, but I wish these supposedly superior Europeans would tell us what they actually want us to do. Should we enforce OWI laws more strictly, or lower the prison population? We can't do both!
I suspect you could step up enforcement in ways that don’t involve prison time simply by taking away people’s licenses, and then having a fast feedback loop to catch people driving without a license.
In addition to what the sibling said regarding the impracticality of not driving in most of the US, which I completely agree with, I'd also ask exactly what you want to do with your "fast feedback loop to catch people driving without a license". What do you do with the people who drive anyways because not driving is so impractical and get caught?
We already took their license, we can't double-take it to show we really mean it. Fining them seems a bit rough when they need to drive to get to the job to make the money to pay those fines. Or we're right back to jail time and an even higher prison population.
> I'd also ask exactly what you want to do with your "fast feedback loop to catch people driving without a license".
Unless the vehicle is stolen, seize and impound the vehicle. If the driver is the owner, auction it off and give them back the proceeds, minus costs.
I feel like I'm living in some different world where drunk driving is a-okay when I face these types of objections to actually enforcing the rules around it.
Taking away licenses is a bad way to enforce driving rules because so many people have to be able to drive or their life collapses. The problems of aggressive license revocation are similar to the problems of aggressive prison time.
I get where you're coming from, but it's pretty hard to be sympathetic given the crimes we're talking about and the impact they have on others.
Like that would sound nuts if we applied it to other things - e.g. "take away the professional license of a mid-career pilot/surgeon/schoolteacher/engineer because he was drinking on the job and his life collapses".
Various people can't drive because of e.g. visual impairments, age, poverty, etc. - I find it an ugly juxtaposition to be asserting that we must allow people with DUIs to drive because otherwise their lives would "collapse" to the same point as those other people who can't drive.
> Like that would sound nuts if we applied it to other things - e.g. "take away the professional license of a mid-career pilot/surgeon/schoolteacher/engineer because he was drinking on the job and his life collapses".
The analogy is closer to "take away their ability to get any job" and then it sounds even more harsh.
> Various people can't drive because of e.g. visual impairments, age, poverty, etc. - I find it an ugly juxtaposition to be asserting that we must allow people with DUIs to drive because otherwise their lives would "collapse" to the same point as those other people who can't drive.
If you can't see well enough to drive, then life was unfair to you, and you can often get help with transportation that isn't available to someone that violated the law. For age, if you're young then your parents are supposed to care for you, if you're too old to drive you're supposed to have figured out your retirement by now. For poverty, you kinda still need a car no matter what, that's just how the US is set up in most areas. And it's not ugly to make the comparison to extreme poverty, to say that kicking someone down to that level is a very severe punishment.
> must allow
I wasn't saying what we should do, just that turning up the aggressiveness has serious unwanted consequences.
> The analogy is closer to "take away their ability to get any job" and then it sounds even more harsh.
If you take away the license of a pilot mid-career, they may be able to pivot to something else, but have a huge sunk cost of education and seniority where they ground out poor pay/schedules and then never made it to the part of the career with better pay. For a substantial segment of them, the career impact would be comparable to taking away the ability to drive from a random person.
> For poverty, you kinda still need a car no matter what, that's just how the US is set up in most areas.
You really don't. If you don't already live somewhere with public transit, you'll probably have to move. You'll have to make some sacrifices. But it's workable, I lived without a car and relied on city busses for all my transportation for several years. (And while I wouldn't necessarily recommend it, prior to that, I lived in a small town of ~4k people without transit service. I walked everywhere, and took the inter-city bus when I needed to leave the town.)
Absolutely, I can tell you right now that many human drivers are probably safer than the Waymo, because they would have slowed down even more and/or stayed further from the parked cars outside a school; they might have even seen the kid earlier in e.g. a reflection than the Waymo could see.
It seems it was driving pretty slow (17MPH) and they do tend to put in a pretty big gap to the right side when they can.
There are kinds of human sensing that are better when humans are maximally attentive (seeing through windows/reflections). But there's also the seeing-in-all-directions, radar, superhuman reaction time, etc, on the side of the Waymo.
17MPH is way to fast, depending on the details. I do not think the article gives the details to know if it was a reasonable speed to be going or not, enough details to know it might be to fast, like proximity to a school and children present, yes.
And the fact that Waymo is never drunk/high/tired/texting, which an astounding portion of human drivers are.
I think my problem is that it reacted after seeing the child step out from behind the SUV.
An excellent driver would have already seen that possible scenario and would have already slowed to 10 MPH or less to begin with.
(It's how I taught my daughter's to drive "defensively"—look for "red flags" and be prepared for the worst-case scenario. SUV near a school and I cannot see behind it? Red flag—slow the fuck down.)
First, it's still the automobile's fault.
At least it was already slowed down to 17 mph to start. Remember that viral video of some Australian in a pickup ragdolling a girl across the road? Most every comment is "well he was going the speed limit no fault for him!" No asshole, you hit someone. It's your fault. He got zero charges and the girl was seriously injured.
You seem to be implying that there are no circumstances in which a vehicle can hit a pedestrian and the driver not be at fault... which is absurd.
Just about absolute. Fall off bridge onto car, I guess not. Olympic sprinter dashes out from car intentionally trying to be hit? Guess not either. Clothed mostly in black on a rainy night on a freeway? Not either.
But you hit a kid in daytime? It's your fault. Period.
A kid can dash out from a car about as fast as an olympic sprinter can. And with unlucky timing their intent doesn't matter.
sorry but that doesn’t make sense.
It’s possible a driver turns a corner (not wearing sunglasses) and suddenly the sun briefly blinds them, while a kid darts into the street.
I’ve seen kids (and ADULTS!) walk on the side of the street at night in all black or very very dark clothing. It’s extra amusing when they happen to be black (are they trying to get themselves killed?) It’s not the drivers fault if they genuinely can’t see a camouflaged person. I’ve had numerous close calls like this on rural and suburban roads and I think i’m a cautious driver. Make sure you are visible at night.
Or if a kid is riding a bicycle down a hill and flies into the middle of an intersection (dumb? brakes failed? etc). very possible to accidentally mow down the child.
HOWEVER, i do agree that 95% of the time it’s the drivers fault if they hit a kid. Poor awareness and speed are the biggest factors. It is certainly not 100% of the time the drivers fault though. That’s absurd. You really misunderstand how dumb some pedestrians (and parents) are.
But….it’s all besides the point. A child that doesn’t understand the importance of cross walks and looking both ways is too young to be walking alone, period. Yes even if they’re “right”. Being right isn’t helpful if you’re dead.
> You seem to be implying that there are no circumstances in which a vehicle can hit a pedestrian and the driver not be at fault... which is absurd.
It is absurd, but that doesn't mean that the attitude can't be useful!
In teaching my teenager to drive, I drilled into him the fact that, in every accident, regardless of who is "at fault", there is almost always something that the other party could have done to mitigate it. I gave him plenty of situations as examples...
You're going down a street that has kids on the sidewalk? You better be prepared to have one of those kids come out in front of the car while rough-housing, playing, whatever.
You had right of way? Maybe you did, but did you even look at the opposing traffic to see if it was safe to proceed or did you just look at the traffic light?
I've driven, thus far in my life, roughly 600000km (maybe more) with 2x non-trivial accidents, both ruled not my fault. In hindsight, I could have avoided both of them (I was young and not so self-aware).
I'm paranoid when driving, and my stats are much much better than Waymo's (have never injured anyone - even my 2x accidents only had me injured), even though I drive in all sorts of conditions, and on all sorts of roads (many rural, some without markings).
Most people don't drive like this though (although their accident rate is still better than Waymo's).
No it's not. The same principle applies to rules of right of way on the water. Technically the 32 foot sailboat has right of way over a triple-E because the triple-E uses mechanical propulsion.
You have a responsibility to be cautious in heavy equipment no matter what the signage on the road says, and that includes keeping a speed at which you can stop safely if a person suddenly steps onto the road in situations where people are around. If you are driving past a busy bar in downtown, a drunk person might step out and you have a responsibility to assume that might happen. If you have to go slower sometimes, tough.
I don't think that's a great analogy since a sailboat's right-of-way isn't unlimited and it can certainly be found at fault for a collision with a triple-E container ship - especially given maritime law uses the comparative fault system where fault is shared between parties.
For instance, a sailboat must alter course if a collision can't be avoided by the give-way vessel alone:
Rule 17(b):
> When, from any cause, the vessel required to keep her course and speed finds herself so close that collision cannot be avoided by the action of the give-way vessel alone, she shall take such action as will best aid to avoid collision.
So if you sail your boat into a container ship and it tries to give way, but doesn't have the ability to do so quickly enough to prevent a collision, you're violating the rules if you don't also alter course as well.
Plus, if we're going to connect this to a pedestrian, if a sailboat suddenly cut in front of a container ship with zero concern for its limited maneuverability/ability to stop, the sailboat would also violate Rule 2 by neglecting precaution required by seamen and failing to consider the limitations of the vessels involved.
And if a pedestrian jumps from a bridge to land right in front of you? or how about a passenger jumps of out the car next to you? still going to stand on your absolute?
As an aside, because it would not be germane to automotive safety…
In the Coast Guard Auxiliary “Sailing and Seamanship” class that I attended, targeting would-be sailboat skippers, we were told the USS Ranger nuclear-powered aircraft carrier had the right-of-way.
You mean the Aussie one where the guy was going an appropriate speed for the area and when the cops arrived the parents and their neighbors LIED TO THE POLICE and said he was hooning down the road at excess speed and hit the kid? And that he was only saved from prison by having a dash cam that proved the lies to be lies? That one?
That logic is utter bs, if someone jumps out when you're travelling at an appropriate speed and you do your best to stop then that's all that can be done. Otherwise by your logic the only safe speed is 0.
That’s not how fault works
It's not the drivers fault when they hit a kid who darts out in front of them and they had no time to react and weren't doing anything illegal like speeding.
They could have driven with more care and attention if they're passing large vehicles that block their view of any children.
I don't see how that's feasible without introducing a lot of friction.
Near my house, almost the entire trip from the freeway to my house is via a single lane with parked cars on the side. I would have to drive 10 MPH the entire way (speed limit is 25, so 2.5x as long).
Why can't we add friction to save lives? Automobiles are the single leading cause of death for children in the USA! We're not talking about something uncommon.
Remove the free parking if that's making the road unsafe. Or drive 10 mph. Done.
But you most likely don't have that entire road be full of little kids in the sidewalk all the way. If you did, then yes probably 10mph or less would be wise.
Yes.
- Parked cars on the street. - Drive somewhat fast. - Avoid killing people.
Pick two.
It's hard to consider it "lots of friction" in a vehicle where you press a button to go faster and another button to slow down.
A single lane residential street with zero visibility seems like an obvious time to slow down. And that's what the Waymo did.
That's why the speed limit is 25 (lower when children are present in some areas) and not 35 or 40 etc. It's not reasonable to expect people to drive at 40% of the posted speed limit the entire way. We're also not talking about zero visibility (e.g. heavy fog). We're talking about blind spots behind parked cars, which in dense areas of a city is a large part of the city. If we think as a society in those situations the safe speed is 10 mph, then the speed limit should be 10mph.
It absolutely is reasonable to expect people to drive below the limit. Speed limits are just that, maximum upper bounds of how fast you legally can go, not "recommended" speeds or minimum speeds, nor are they necessarily "safe" speeds. It's just a legal upper bounds.
It is a drivers responsibility to drive for the conditions. If conditions are calling for driving 40% slower, then that's what you do and suck it up.
If too many roads have conditions that require that, take that up with your municipality to fix the situation. Or, even better, advocate for better public transit and trains, and designing cities to move people, not move cars.
Driving for the conditions mostly means weather.
If there's cars parked on the side constantly, and that's supposed to slow you down significantly, it should be baked into the speed limit.
From what I'm aware of, you're not actually expected to slow down drastically from parked cars.
I mean, you are putting your finger right on the answer: the whole car thing doesn't work or make sense, and trying to make autonomous vehicles solve the unsolvable is never going to succeed.
Agreed.
Car culture in the US is toxic, and a lot of accidents and fatalities are a result of how poorly designed our infrastructure is. We design for cars, not for people (just one more lane bro, will totally fix traffic. Nevermind that a train can move double the capacity of that entire line of traffic).
Cars are the wrong solution, particularly in urban areas. A self driving car is still a car, and comes along with all the same problems that cars cause.
Aye, and to always look for feet under and by the front wheel of vehicles like that.
Stopped buses similarly, people get off the bus, whip around the front of them and straight into the streets, so many times I’ve spotted someone’s feet under the front before they come around and into the street.
Not to take away from Waymo here, agree with thread sentiment that they seem to have acted exemplary
You can spot someone's feet under the width of a bus when they're on the opposite side of the bus and you're sitting in a vehicle at a much higher position on the opposite side that the bus is on? That's physically impossible.
In normal (traditional?) European city cars, yes, I look for feet or shadows or other signs that there is a person in the other side. In SUVs this is largely impossible but then sometimes you can see heads or backpacks.
Or you look for reflections in the cars parked around it. This is what I was taught as “defensive“ driving.
I think you're missing something though, which I've observed from reading these comments - HN commenters aren't ordinary humans, they're super-humans with cosmic powers of awareness, visibility, reactions and judgement.
Or they hate cars/waymo/etc and will come up with any chain of reasoning that puts those things in a bad light.
>reacted after seeing the child step out from behind the SUV.
Lmao most drivers I see on the roads aren't even capable of slowing down for a pedestrian crossing when the view of the second half of the crossing is blocked by traffic (ie they cannot see if someone is about to step out, especially a child).
Humans are utterly terrible drivers.
They don't even stop when it's a crosswalk with a flashing light system installed and there are no obstructions.
>> Humans are utterly terrible drivers
Duh, driver is, essentially, a type of specialized profession. It's kinda unreasonable to think that everyone could learn to do it well
Good thing we have public transport! :D
Yes and no. Tons of situations where this is simply not possible, whole traffic goes full allowed speed next to row of parked cars. If somebody unexpectedly pops up distracted, its a tragedy guaranteed regardless of driver's skills and experience.
In low traffic of course it can be different. But its unrealistic to expect anybody to drive in expectation that behind every single car passed there may be a child jumping right in front of the car. That can be easily thousands of cars, every day, whole life. Impossible.
We don't read about 99.9% of the cases where even semi decent driver can handle it safely, but rare cases make the news.
I slow down considerably near parked cars. And I try to slow down much earlier approaching intersections where there are parked cars blocking my view of cross walk entries. I need to be able to come to full stop earlier than intersection if there happens to be a pedestrian there.
I kind of drive that way. I slow down, move as far away in my lane from the parked cars as possible. It's certainly what I would expect from a machine that would claim to be as good as the best human driver.
> a machine that would claim to be as good as the best human driver.
Does Waymo claim that? If so I haven't seen it. That should of course be the goal, but "better than the average human driver" should be the bar.
If they don't clim that, why would I be interested in their tech?
Because I'm only an average driver. Definitely not the best and have done no advanced driver training.
Because we can reduce crashes and fatalities without Waymo being Max Verstappen
Because replacing average drivers with a better than average one would save lives.
Ha ha, you should suggest that to them: "Waymo: better than average driving!"
What's the joke here? If they are better than average drivers, that's a huge win which improves road safety for everyone
This is generally the problem with self-driving cars, at least in my experience (Tesla FSD).
They don't look far enough ahead to anticipate what might happen and already put themselves in a position to prepare for that possibility. I'm not sure they benefit from accumulated knowledge? (Maybe Waymo does, that's an interesting question.) I.e., I know that my son's elementary school is around the corner so as I turn I'm already anticipating the school zone (that starts a block away) rather than only detecting it once I've made the turn.
Tesla FSD is leagues behind Waymo; generalizing based on your Tesla experience doesn't make sense.
Evidence of this? I own a Tesla (HW4, latest FSD) as well as have taken many Waymo rides, and have found both to react well to unpredictable situations (i.e. a car unexpectedly turning in front of you), far more quickly than I would expect most human drivers to react.
This certainly may have been true of older Teslas with HW3 and older FSD builds (I had one, and yes you couldn't trust it).
So... the thing is it's really not. They're behind on schedule, having just launched in public. But FSD has been showing capabilities in regular use (highway navigation, unprotected left turns in traffic, non-geofenced operation areas based solely on road markings) that Waymo hasn't even tried to deploy yet.
It's much more of a competition than I suspect a lot of people realize.
Intellectually this doesn't even compute. Waymo is tele-oped in like 5 cities and FSD works in 5 countries. It's not even close.
Yes I agree, but why 10mph? Why not 5mph? or 2mph? You'll still hit them if they step out right in front of you and you don't have time to react.
Obviously the distances are different at that speed, but if the person steps out so close that you cannot react in time, you're fucked at any speed.
10mph will do serious damage still, so please for the sake of the children please slow yourself and your daughter's driving down to 0.5mph where there are pedestrians or parked cars.
But seriously I think you'd be more safe to both slow down and also to put more space between the parked cars and your car so that you are not scooting along with a 30cm of clearance - move out and leave lots of space so there is more space for sight-lines for both you and pedestrians.
multiple children in my area have died due to being hit by distracted drivers driving near schools. One incident resulted in 2 children being dragged 60 yards. Here's a snippet from an article about the death I was referencing:
> The woman told police she was “eating yogurt” before she turned onto the road and that she was late for an appointment. She said she handed her phone to her son and asked him to make a call “but could not remember if she had held it so face recognition could … open the phone,” according to the probable cause statement.
> The police investigation found that she was traveling 50 mph in a 40 mph zone when she hit the boys. She told police she didn’t realize she had hit anything until she saw the boys in her rearview mirror.
The Waymo report is being generous in comparing to a fully-attentive driver. I'm a bit annoyed at the headline choice here (from OP and the original journalist) as it is fully burying the lede.
I usually take extra care when going through a school zone, especially when I see some obstruction ('behind a tall SUV', was the waymo overtaking?), and overtaking is something I would probably never do (and should be banned in school zones by road signs).
This is a context that humans automatically have and consider. I'm sure Waymo engineers can mark spots on the map where the car needs to drive very conservatively.
> especially when I see some obstruction ('behind a tall SUV', was the waymo overtaking?)
Yep. Driving safe isn't just about paying attention to what you can see, but also paying attention to what you can't see. Being always vigilant and aware of things like "I can't see behind that truck."
Honestly I don't think sensor-first approaches are cut out to tackle this; it probably requires something more akin to AGI, to allow inferring possible risks from incomplete or absent data.
I appreciate your sensible driving, but here in the UK, roads outside schools are complete mayhem at dropping off/picking up times. Speeding, overtaking, wild manoeuvres to turn round etc.
When reading the article, my first thought was that only going at 17mph was due to it being a robotaxi whereas UK drivers tend to be strongly opposed to 20mph speed limits outside schools.
Most US states cap speed limits around schools at 15mph when children are present. There may also be blinking lights above these signs during times that will be likely.
I'm not sure how much of that Waymo's cars take into account, as the law technically takes into account line of sight things that a person could see but Waymo's sensors might not, such as children present on a sidewalk.
> Most US states cap speed limits around schools at 15mph when children are present.
Are you sure? The ones I've seen have usually been 20 or 25mph.
Looking on Image Search (https://www.google.com/search?q=school+zone+speed+limit+sign) and limiting just to the ones that are photos of real signs by the side of the road, the first 10 are: 25, 30, 25, 20, 35, 15, 20, 55, 20, 20. So only one of these was 15.
Here's a random elementary school in Santa Monica with a 15mph sign https://www.google.com/maps/place/McKinley+Elementary+School...
I don't think this is accurate, 20mph is more common
School pick up and drop off traffic is just about the worst drivers anywhere. Like visibly worse than a bunch of "probably a little drunk" people leaving a sports stadium. It's like everyone reverts to "sixteen year old on first day behind the wheel" behavior. It's baffling. And there's always one token dad picking up his kid on a motorcycle or in a box truck or something that they all clutch their pearls at.
This exact scenario happened with my dad 50 years ago when a little girl ran out to the street from between some parked cars. It's an extremely difficult scenario to avoid an accident in.
It depends. A driver may have seen a child dart behind a car and expect them to emerge on the other side.
Does Waymo have the same object permanence and trajectory prediction (combined) to that of a human?
Once the video evidence it out, it might become evident.
Generally Waymo seems to be a responsible actor so maybe that is the case and this can help demonstrate potential benefits of autonomous vehicles.
Alternatively, if even they can't get this right then it may cast doubts about the maturity of the entire ecosystem
> Does Waymo have the same object permanence and trajectory prediction (combined) to that of a human?
On this note specifically ive actually been impressed, ie when driving down Oak st in SF (fast road, tightly parked cars) I've often observed it slow if someone on a scooter on the sidewalk turns to look toward oncoming traffic (as if to start riding), or to slow passing parked box trucks (which block vision of potential pedestrians)
“Does Waymo have the same object permanence and trajectory prediction (combined) to that of a human?”
Good technical question
>It's likely that a fully-attentive human driver would have done worse.
Maybe. Depends on the position of the sun and shadows, I'm teaching my kids how to drive now and showing them that shadows can reveal human activity that is otherwise hidden by vehicles. I wonder if Waymo or other self-driving picks up on that.
If I was a human driver in that contextual situation I wouldn't even be going 14mph in the first place...
So you say, with hindsight. Regardless, what you would do is irrelevant. The question is: what would the ~average driver do? If you tell me the average driver goes <=14mph in that situation, I'm gonna look at you funny.
I don't fail to believe that, a child running from behind an suv is really scary
A human driver in a school zone during morning drop off would be scanning the sidewalks and paying attention to children that disappear behind a double parked suv or car in the first place, no?
As described by the nhtsa brief:
"within two blocks of a Santa Monica, CA elementary school during normal school drop off hours; that there were other children, a crossing guard, and several double-parked vehicles in the vicinity"
The "that there were other children, a crossing guard, and several double-parked vehicles in the vicinity" means that waymo is driving recklessly by obeying the speed limit here (assuming it was 20mph) in a way that many humans would not.
I live near a school zone in LA and most drivers do not obey school zone speed limits.
You will get honked at by aggro drivers if you slow down to the school zone speed limit of 25mph. Most cars go 40ish.
And ofc a decent chunk of those drivers are on tiktok, tinder, Instagram, etc
Some human drivers? Yes, certainly.
Your median human driver? Sadly, I think not. Most would be rushing, or distracted, or careless.
> waymo is driving recklessly by obeying the speed limit here (assuming it was 20mph) in a way that many humans would not.
I don't think we can say at all that the Waymo was driving recklessly with the data we currently have
> It's likely that a fully-attentive human driver would have done worse.
Why is it likely? Are we taking the vendor's claims in a blog post as truth?
"fully attentive human driver ..." is Waymo's claim, and it could be biased in their favor.
Could be! In aggregate though, Waymos have shown to be safer than human drivers, so my prior is that that holds here.
who benefits from a statement like this?
People who reflexively assume a human driver would do better
It's possible, but likely is a heavy assertion. It's also possible a human driver would have been more aware of children being present on the sidewalk and would have approached more cautiously given obstructed views.
Please please remember that any data from Waymo will inherently support their position and can not be taken at face value. They have significant investment in making this look more favorable for them. They have billions of dollars riding on the appearance of being safe.
I remember someone using similar language when Uber self driving killed someone - and when the video was released, it was laughable.
It is also crazy that this happened 6 days ago at this point and video was NOT part of the press releases. LOL
LOL
OLO
I wonder if that is a "fully attentive human drive who drove exactly the same as the Waymo up until the point the child appeared"?
Personally, I slow down and get extra cautious when I know I am near a place where lots of kids are and sight lines are poor. Even if the area is signed for 20 I might only be doing 14 to begin with, and also driving more towards the center of the road if possible with traffic.
I do the same, and try to actively anticipate and avoid situations like this. Sadly, in my experience most drivers instead fixate on getting to their destination as fast as possible.
A fully attentive human would've known he was near a school and wouldn't have been driving at 17 mph to begin with.
Doubt
I said "fully attentive human", not "walking plant".
You clearly don't spend much time around a school measuring the speed of cars. Head on down and see for yourself how often or not a human driver goes >17mph in such a situation.
I said "a fully attentive human". At the very least try to understand my comment before spouting the usual HN snarky reply.
> It's likely that a fully-attentive human driver would have done worse.
> a huge portion of human drivers
What are you basing any of these blind assertions off of? They are not at all born out by the massive amounts of data we have surrounding driving in the US. Of course Waymo is going to sell you a self-serving line but here on Hacker News you should absolutely challenge that. In particular because it's very far out of line with real world data provided by the government.
If you have contradicting data I'd be glad to see it
>It's likely that a fully-attentive human driver would have done worse.
Is based off the source I gave in my comment, the peer-reviewed model
> a huge portion of human drivers
Is based on my experience and bits of data like 30% of fatal accidents involving alcohol
Like I said, if you have better data I'm glad to see it
> based on my experience
The data completely disagrees with you.
> Like I said, if you have better data I'm glad to see it
We all have better data. It's been here the entire time:
https://www.nhtsa.gov/research-data/fatality-analysis-report...
Your data that shows 30% of fatal crashes involve alcohol, not to mention any of the other factors I named? Seems like your data supports my conclusion!
Again, I welcome you to point to data that contradicts my claims, but it seems you are unable
What data completely disagrees with them and what does it disagree with them about?
The "Persons Killed, by Highest Driver Blood Alcohol Concentration (BAC) in the Crash"[1] report shows that in 2023, 30% of fatal crashes involved at least one driver with a BAC > 0.08 (the legal limit), and 36% involved a BAC > 0.01.
Interesting that "Non-motorist" fatalities have dropped dramatically for everyone under the age of 21, but increased for everyone between 21 and 74.[2] Those are raw numbers, so it'd be even more interesting to display them as a ratio of the group's size. Are less children being killed by drivers because there are less children generally? Changes in parents' habits? Backup cameras?
1: https://www-fars.nhtsa.dot.gov/Trends/TrendsAlcohol.aspx 2: https://www-fars.nhtsa.dot.gov/Trends/TrendsNonMotorist.aspx
Waymo is intentionally leaving out the following details:
- Their "peer-reviewed model" compares Waymo vehicles against only "Level 0" vehicles. However even my decade-old vehicle is considered "Level 1" because it has an automated emergency braking system. No doubt my Subaru's camera-based EBS performs worse than Waymo's, still it's not being included in their "peer-reviewed model." That comparison is intentionally comparing Waymo performance against the oldest vehicles on the road -- not the majority of cars sold currently.
- This incident happened during school dropoff. There was a double-parked SUV that occluded the view of the student. This crash was the fault of that double-parked driver. But why was the uncrewed Waymo driving at 17 mph to begin with? Do they not have enough situational awareness to slow the f*ck down around dropoff time immediately near an elementary school?
Automotive sensor/control packages are very useful and will be even more useful over time -- but Waymo is intentionally making their current offering look comparatively better than it actually is.
Emergency braking in non-camera/non-LIDAR cars requires a significant radar signal which you're only going to get from another vehicle (and even then it's noisy and tends to produce frustrating false positives, leading to later-than-you-want stops). It very likely won't detect a child or a dog, I'm not aware of a single instance of an EBS claiming to have done so in practice (and kids and dogs get hit every day!).
It depends on the situation, and we need more data/video. But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast, and the Waymo should have been driving more conservatively.
> But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast, and the Waymo should have been driving more conservatively.
UK driving theory test has a part called Hazard Perception: not reacting on children milling around would be considered a fail.
[0] https://www.safedrivingforlife.info/free-practice-tests/haza...
Many states in the US have the Basic Speed Law, e.g. California:
> No person shall drive a vehicle upon a highway at a speed greater than is reasonable or prudent having due regard for weather, visibility, the traffic on, and the surface and width of, the highway, and in no event at a speed which endangers the safety of persons or property.
The speed limit isn't supposed to be a carte blanche to drive at that speed no matter what; the basic speed law is supposed to "win." In practice, enforcement is a lot more clear cut at the posted speed limit and officers don't want to write tickets that are hard to argue in court.
That law seems more likely to assign blame to drivers if they hit someone. So practically it's not enforced but in accidents it becomes a justification for assigning fault.
I mean yeah. If you were traveling at some speed and caused damage to persons or property, that's reasonable, but refutable, evidence that you were traveling at a speed that endangered persons or property.
And at the same time, if you were traveling at some speed and no damage was caused, it's harder to say that persons or property were endangered.
Exactly. That’s why I’ve always said the driving is a truly AGI requiring activity. It’s not just about sensors and speed limits and feedback loops. It’s about having a true understanding for everything that’s happening around you:
Having an understanding for the density and make up of an obstacle that blew in front of you, because it was just a cardboard box. Seeing how it tumbles lightly through the wind, and forming a complete model of its mass and structure in your mind instantaneously. Recognizing that that flimsy fragment though large will do no damage and doesn’t justify a swerve.
Getting in the mind of a car in front of you, by seeing subtle hints of where the driver is looking down, and recognizing that they’re not fully paying attention. Seeing them sort of inch over because you can tell they want to change lanes, but they’re not quite there yet.
Or in this case, perhaps hearing the sounds of children playing, recognizing that it’s 3:20 PM, and that school is out, other cars, double parked as you mentioned, all screaming instantly to a human driver to be extremely cautious and kids could be jumping out from anywhere.
Slightly off topic, but it's endlessly funny to me watching people set the bar for AGI so high that only a small percentage of humans count as AGI.
humans aren't even a general intelligence at these requirements.
How many human drivers do you think would pass the bar you're setting?
IMO, the bar should be that the technology is a significant improvement over the average performance of human drivers (which I don't think is that hard), not necessarily perfect.
> How many human drivers do you think would pass the bar you're setting?
How many humans drivers would pass it, and what proportion of the time? Even the best drivers do not constantly maintain peak vigilance, because they are human.
> IMO, the bar should be that the technology is a significant improvement over the average performance of human drivers (which I don't think is that hard), not necessarily perfect.
In practice, this isn't reasonable, because "hey we're slightly better than a population that includes the drunks, the inattentive, and the infirm" is not going to win public trust. And, of course, a system that is barely better than average humans might worsen safety, if it ends up replacing driving by those who would normally drive especially safe.
I think "better than the average performance of a 75th or 90th percentile human driver" might be a good way to look at things.
It's going to be a weird thing, because odds are the distribution of accidents that do happen won't look much like human ones. It will have superhuman saves (like that scooter one), but it will also crash in situations that we can't really picture humans doing.
I'm reminded of airbags; even first generation airbags made things much safer overall, but they occasionally decapitated a short person or child in a 5MPH parking lot fender bender. This was hard for the public to stomach, and if it's your kid who is internally decapitated by the airbag in a small accident, I don't think you'll really accept "it's safer on average to have an airbag!"
The parent comment said the bar should be "significant improvement" over the average performance of human drivers.
Then you said, "this isn't reasonable", and the bar shouldn't be "slightly better" or "barely better". It should be at least better than the 75th percentile driver.
It sounds like you either misread the parent comment or you're phrasing your response as disagreement despite proposing roughly the same thing as the parent comment.
All depends on what you read as "significant improvement".
A 20% lower fatal crash rate compared to the average might be a significant improvement-- from a public health standpoint, this is huge if you could reduce traffic deaths by 20%.
But if you don't get the worst drivers to replace their driving with autonomous, that "20% less than average" might actually make things worse. That's my point. The bar has to be pretty dang high to be sure that you will actually make things better.
> In practice, this isn't reasonable, because "hey we're slightly better than a population that includes the drunks, the inattentive, and the infirm" is not going to win public trust.
Sadly, you're right, but as rational people, we can acknowledge that it should. I care about reducing injuries and deaths, and the %tile of human performance needed for that is probably something like 30%ile. It's definitely well below 75%ile.
The counterpoint, though:
> > And, of course, a system that is barely better than average humans might worsen safety, if it ends up replacing driving by those who would normally drive especially safe.
It's only if you get the habitually drunk (a group that is overall impoverished), the very old, etc, to ride Waymo that you reap this benefit. And they're probably not early adopters.
Uber and Lyft were supported by police departments because they reduced drunk driving. Drunk driving isn't just impoverished alcoholics. People go to bars and parts and get drunk all the time.
You also solve for people texting (or otherwise using their phones) while driving, which is pretty common among young, tech-adopting people.
> Drunk driving isn't just impoverished alcoholics. People go to bars and parts and get drunk all the time
Yes, but the drivers who are 5th percentile drivers who cause a huge share of the most severe accidents are "special" in various ways. Most of them are probably not autonomy early adopters.
The guy who decided to drive on the wrong side of a double yellow on a windy mountain road and hit our family car in a probable suicide attempt was not going to replace that trip with Waymo.
The bar is very high because humans expect machines to be perfect. As for the expectation of other humans, "pobody's nerfect!"
> But if there are a bunch of children milling about an elementary school in a chaotic situation with lots of double parking, 17 mph is too fast
Hey, I'd agree with this-- and it's worth noting that 17^2 - 5^2 > 16^2, so even 1MPH slower would likely have resulted in no contact in this scenario.
But, I'd say the majority of the time it's OK to pass an elementary school at 20-25MPH. Anything carries a certain level of risk, of course. So we really need to know more about the situation to judge the Waymo's speed. I will say that generally Waymo seems to be on the conservative end in the scenarios I've seen.
(My back of napkin math says an attentive human driver going at 12MPH would hit the pedestrian at the same speed if what we've been told is accurate).
Swedish schools still have students who walk there. I live near one and there are very few cars that exceed 20km/h during rush hours. Anything faster is reckless even if the max over here is 30 km/h (19 mph).
The schools I'm thinking of have sidewalks with some degree of protection/offset from street, and the crossings are protected by human crossing guards during times when students are going to schools. The posted limits are "25 (MPH) When Children Are Present" and traffic generally moves at 20MPH during most of those times.
There are definitely times and situation where the right speed is 7MPH and even that feels "fast", though, too.
> Hey, I'd agree with this-- and it's worth noting that 17^2 - 5^2 > 16^2, so even 1MPH slower would likely have resulted in no contact in this scenario.
Only with instant reaction time and linear deceleration.
Neither of those are the case. It takes time for even a Waymo to recognize a dangerous situation and apply the brake and deceleration of vehicles is not actually linear.
> It takes time for even a Waymo to recognize a dangerous situation
Reaction time makes the math even better here. You travel v1 * reaction_time no matter what, before entering the deceleration regime. So if v1 gets smaller, you get to spend a greater proportion of time in the deceleration regime.
> linear deceleration.
After reaction time, stopping distance is pretty close to n^2. There's weird effects at high speed (contribution from drag) and at very low speed, but they have pretty modest contributions.
I was thinking more that how hard the brakes are applied likely varies based on uncertainty of a collision.
Without that these vehicles could only start braking when certainty crossed some arbitrary threshold.
I think the strategy is a lot more nuanced than that.
In any case, with zero reaction time, linear deceleration time to stop is proportional to velocity squared. With reaction time, the linear deceleration time is that plus the velocity times the reaction time.
so the two cases we're comparing are 17 * r + (17^2 - 5^2) vs. 16 * r + (16^2), or 17 * r + 264 vs 16 * r + 256. As long as reaction time isn't negative, a vehicle that could slow to 5MPH starting at 17MPH could slow to 0MPH starting at 16MPH.
(There are weird things that happen at <2.5MPH reducing deceleration to sublinear, but the car moves only a few inches at these speeds during a panic stop).
Whoa! You're allowed to double park outside a school over there?!
Wait, is double parking allowed anywhere?
Pretty common at airports; of course, the `parking` only lasts a few minutes at most.
It’s common but almost always illegal based on the posted signage.
Yes, please visit LA.
Edit: Not 'allowed' but people do it constantly. Regular drivers, delivery drivers, city workers, construction trucks, etc. There may be laws but very little enforcement.
No (excluding some circumstances like delivery vehicles).
People loitering in their cars waiting for a space to pick up their kid. So not actually parked.
More like standing, and quite common in a school zone.
I would not race at 17 MPH through such an area. Of course, Waymo will find a way to describe themselves as the heroes of this situation.
An honest account of this situation would place at least some blame on there being a tall SUV blocking visibility.
These giant SUVs really are the worst when it comes to child safety
I bet we'll the the SUV mania in the future as something crazy, like smoking in a plane or using lead for gasoline. Irrational large size cars that people get because everyone it's afraid of another SUV hitting them in a sedan. The tragedy of the commons.
What I find a bit confusing is that no one is putting any blame on the kid. I did the same thing as a kid, except it was a school bus instead of SUV, and that was a fucking stupid thing to do (I remember starting to run over the street, and the next thing is that I am in the hospital bed), even though I had been told to always cross the street from behind the bus, not in front of it.
That day I learned why it was so.
Of course the kid is at fault. But everyone knows that kids do stupid and reckless things, which is why drivers are generally expected to take more care around schools and similar institutions. If robotaxis are not able to do that, then the results will be easy to predict
The best reaction from Waymo would have been to start to lobby against letting those monster-trucks park on streets near schools. They are killing so many children, I'm flabbergasted they are still allowed outside of worksites.
From a "my opinion" standpoint, yes, I would love to see this.
From a tactical PR standpoint, it would be a disaster. Muh big truuuucks is like a third rail because Americans are car obsessed as a culture. They already hit a kid, best to save some energy for the next battle.
Besides if Waymo wins (in general) private car ownership will decrease which is a win regardless. And maybe Waymo can slowly decrease the size of their fleet to ease up the pressure on this insane car size arms race.
AV’s with enough sensing are generally quite good at stopping quickly. It is usually the behavior prior to the critical encounter that has room for improvement.
The question will be whether 17 mph was a reasonably cautious speed for this specific scenario. Many school zones have 15 mph limits and when there are kids about people may go even slower. At the same time, the general rule in CA for school zone is 25 mph. Clearly the car had some level of caution which is good.
It does sound like a good outcome for automation. Though I suppose an investigation into the matter would arguably have to look at whether a competent human driver would be driving at 17mph (27km/h) under those circumstances to begin with, rather than just comparing the relative reaction speeds, taking the hazardous situation for granted.
What I would like to see is a full-scale vehicle simulator where humans are tested against virtual scenarios that faithfully recreate autonomous driving accidents to see how "most people" would have acted in the minutes leading up to the event as well as the accident itself
> a full-scale vehicle simulator
The UK is such a situation, and this vehicle would have failed a driving test there.
17 mph is pretty slow unless it’s a school zone
Indeed, 15 or 25 mph (24 or 40 km/h) are the speed limits in school zones (when in effect) in CA, for reference. But depending on the general movement and density and category of pedestrians around the road it could be practically reckless to drive that fast (or slow).
If my experience driving through a school zone on my way to work is anything to go off of, I rarely see people actually respecting it. 17 mph would be a major improvement over what I'm used to seeing.
>Though I suppose an investigation into the matter would arguably have to look at whether a competent human driver would be driving at 17mph (27km/h) under those circumstances to begin with, rather than just comparing the relative reaction speeds, taking the hazardous situation for granted.
Sure but also throw in whether that driver is staring at their phone, distracting by something else, etc. I have been a skeptic of all this stuff for a while but riding in a Waymo in heavy fog changed my mind when questioning how well I or another driver would've done at that time of day and with those conditions.
How would that help in the investigation?
For me it would be interesting to know if 17 mi/h was a reasonable speed to be driving in this environment under these conditions to begin with. In my school zones that's already close to the maximum speed allowed. What was the weather, were there cars parked which would make a defensive driver slow down even more?
The autonomous vehicle should know what it can't know, like children coming out from behind obstructions. Humans have this intuitive sense. Apparently autonomous systems do not, and do not drive carefully, or slower, or give more space, in those situations. Does it know that it's in a school zone? (Hopefully.) Does it know that school is starting or getting out? (Probably not.) Should it? (Absolutely yes.)
This is the fault of the software and company implementing it.
> Humans have this intuitive sense.
Some do, some of the time. I'm always surprised by how much credence other people give to the idea that humans aren't on average very bad at things, including perception.
It's an autonomous vehicle fitted with a gazillion of sensors and data to drive itself. We can expect better from it than humans.
I'm not sure what your point is here since this was better than many humans. Was it better than all humans? No. But there also isn't a single human who's better than all humans.
What's the success rate of this intuitive sense that humans have? Intuitions are wrong frequently.
So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
The general public is stupid.
That’s why they purchase goods and services (from others) and then cry about things they don’t and probably never will understand.
And why they can be ignored and just fed some slop to feel better.
I could lie but that’s the cold truth.
Edit: I'm not sure if the repliers are being dense (highly likely), or you just skipped over context (you can click the "context" link if you're new here)
> So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
That is the general public sentiment I was referring to.
So if they were 100% self-sufficient and understood everything they'd be smart enough to interpret a child being hit at 6 mph as progress? Fun how "general public" is always a "they" vs "you".
That's impossible though. And you and I are part of the general public as well, for things we don't understand.
It isn't me vs them. It is just me being self-aware. Clearly, you had a problem with what I said so I must have struck a nerve.
Welcome to the real world bro.
Your comment sounds like subconsciously you're trying to come off as stronger than the general public, which begs the question: Why? Why do you need to prove your strength over the populace?
You ARE the general public. _I_ am the general public.
are you being dense on purpose, or you just don't understand how context works? hint, check out the "context" link
look at what I was replying to. if you still don't get it, then yeah I'm just proving my point and you can keep crying about it.
> So the TechCrunch headline should be "Waymo hits child better than a human driver would"? Not sure if the details reflect how the general public actually interprets this story (see the actual TC headline for exhibit A).
That is the general public sentiment I was referring to.
The fact that you go around asking dumb questions in bad faith to people is enough for me, last time I engage with you.
Have a good life!
They they are being very transparent about it.
As every company should, when they have a success. Are they also as transparent about their failures?
How is hitting a child not a failure? And actually, how can you call this a success? Do you think this was a GTA side mission?
Immediately hitting the brakes when a child suddenly appears in front of you, instead of waiting 500ms like a human, and thereby hitting the child at a speed of 6 instead of 14 is a success.
What else to you expect them to do, only run on grade–separated areas where children can't access? Blare sirens so children get scared away from roads? Shouldn't human–driven cars do the same thing then?
I don't know the implementation details, but success would be not hitting pedestrians. You have some interesting ideas on how to achieve that but there might be other ways, I don't know.
>I don't know the implementation details, but success would be not hitting pedestrians.
So by that logic, if we cured cancer but the treatment came with terrible side effects it wouldn't be considered a "success"? Does everything have to perfect to be a success?
If you clearly define your goals in advance, then you can make success whatever you want. What are Waymo's goals?
Something tells me it wasn't 0 accidents, given that it's impossible.
I guess we'll never know.
The raw corporate goals? Safe enough to be allowed on roads.
The less cynical set of goals would be safer than the mean human, then safer than the median human, then safer every year indefinitely.
17 mph is way too fast near a school if it's around the time children are getting out (or in).
The limit is 20 MPH in Washington state, in California the default is 25 MPH, but is going to 20 MPH soon and can be further lowered to 15 MPH with special considerations.
The real killer here is the crazy American on street parking, which limits visibility of both pedestrians and oncoming vehicles. Every school should be a no street parking zone. But parents are going to whine they can't load and unload their kids close to the school.
On street parking is so ingrained into the American lifestyle that any change to the status quo is impossible. Cars have more rights on public property than people. Every suburban neighborhood has conflicts over people's imagined "ownership" of the street parking in front of their house. People rarely use their garages to store their car since they can just leave it on the street. There are often laws that prevent people from other neighborhoods from using the public street to park. New roads are paved as wide as possible to allow both street parking and a double-parked car to not impede traffic. And we've started building homes without any kind of parking that force people to use the street.
> On street parking is so ingrained into the American lifestyle that any change to the status quo is impossible
Plenty of American cities regulate or even eliminated, in various measures, on-street parking.
Europe is much better at this than we are. Even when you have on street parking, they make sure there are clearances around cross walks and places where there are lots of pedestrians. Most US cities don't even care, even a supposedly pedestrian friendly one like Seattle.
Lol tell that to my city. They removed the cross walks and declared the zone a "shared zone" where pedestrians have right of way.
Result? No more safe places to cross, drivers are not stopping for pedestrians when no cross walk. They added parking zones right up to the old cross walks that pedestrians still use (since it were the safest places) where vans are regularly parked and obscure the entirety one side of the road. Even outside of these shared zones, there are lots lots lots of places where parking space is obscuring the crosswalk, where huge vans park right on it, even though legally you have to be 5+ meters away. Never seen a cop give a ticket for that in my life.
One day someone will get killed right there, I'm sure of it, and it'll be mainly the city's fault.
Impossible is probably the wrong word. But where I live, a superficially "progressive" area, many of these traffic calming, road diet, etc. measures are met with regular opposition.
If it had no parking, then the parents would be parked somewhere else and loading and unloading their kids there, and then that would need to be a no-parking zone too.
I guess you could keep doing that until kids just walk to and from school?
Our local school has them unload a block away unless they are handicapped. A kid isn't going to die walking a block. But its pointless because they still allow residential on street parking around the school, and my son has to use a crosswalk where cars routinely park so close to, I had to tell him that the traffic (pretty heavy) on the road wouldn't see him easily, and he should always ease his way into a crosswalk and not assume he would be easily seen.
In the UK we have a great big yellow zig-zag road marking that extends 2/3rds the width of an average car across the road. It means "this is a school, take your car and fuck off". You find it around school gates, to a distance of a few car lengths either side of the gate, and sometimes all along the road beside a school.
It doesn't stop all on street parking beside the school, but it cuts it down a noticeable amount.
This isn't universal. The schools in our Montana town have pickup lanes and short term parking areas for pickup. Stopping on the road isn't allowed.
For a school near me, the road is no parking during pick up/drop off times. It even changes to one way traffic. The no parking windows is similar to alternate street sweeping days. There are signs posted that indicate the times.
Same for my tiny town. Stopping on the road is 100% not allowed, and parking isn't allowed there either. The school has its own parking area to park and pick up/drop off kids, and cars in there creep at 2 or 3 MPH.
"and thereby hitting the child ... is a success."
> What else to you expect them to do, only run on grade–separated areas where children can't access?
no, i expect them to slow down when children may be present
how slow?
how about 10-15 mph if directly adjacent to a school, especially during the bands before and after school stars or ends. route away from schools whenever feasible.
That's how fast it was going.
A vehicle can't go 10 to 15 miles per hour at the same time. If it was going 15, then it should have been going 10. Or driving further away from occluded spaces. And again, routing away from schools.
The simple fact is that it hit a child and even though it wasn't a serious issue due to their safety policies, there's still room for improvement in these technologies.
And since it's a robot, and not a human, you can actually make changes and have them stick. For example, routing away from schools during certain hours.
This isn't Apollo 13 with a successful failure. A driverless car hit a human that just happened to be a kid. Doesn't matter if a human would have as well, the super safe driverless car hit a kid. Nothing else matters. Driverless car failed.
If failure is defined such that failure is the only possible outcome, I don't think it's a useful part of an evaluation.
Why didn't sully just not hit the birds?
Skill issue, presumably.
They've gone to the courts to fight to keep some of their safety data secret
https://www.theverge.com/2022/1/28/22906513/waymo-lawsuit-ca...
Well, as a comparison, we know that Tesla has failed to report to NHTSA any collisions that didn't deploy the airbag.
Tesla report ids from SGO-2021-01_Incident_Reports_ADAS.csv with no or unknown airbag deployment status: 13781-13330, 13781-13319, 13781-13299, 13781-13208, 13781-8843, 13781-13149, 13781-13103, 13781-13070, 13781-13052... and more
Is this a success? There was still an incident. I'd argue this was them being transparent about a failure
Being transparent about such incidents is also what stops them from potentially becoming a business/industry-killing failures. They're doing the right thing here, but they also surely realize how much worse it would be if they tried to deny or downplay it.
> they also surely realize how much worse it would be if they tried to deny or downplay it.
Indeed. Waymo is a much more thoughtful and responsible company than Cruise, Uber, or Tesla.
"Cruise admits to criminal cover-up of pedestrian dragging in SF, will pay $500K penalty" https://www.sfgate.com/tech/article/cruise-fine-criminal-cov...
They handled an unpredictable emergency situation better than any human driver.
Was it unpredictable? They drove past a blind corner (parked SUV) in a school zone. I'm constantly slowing down in these situations as I expect someone might run out at any second. Waymo seemed to default to the view that if it can't see anyone then nobody is there.
as far as we know
even as far as we know they aren't
The Waymo blog post refused to say the word "child", instead using the phrase "young pedestrian" once.
The Waymo blog post switches to "the pedestrian" and "the individual" for the rest of the post.
The Waymo blog post also consistently uses the word "contact" instead of hit, struck, or collision.
The Waymo blog post makes no mention of the injuries the child sustained.
The Waymo blog post makes no mention of the school being in close proximity.
The Waymo blog post makes no mention of other children or the crossing guard.
The Waymo blog post makes no mention of the car going over the school zone speed limit (17 in 15).
The speed limit of a school zone in California is 25, not 15, which would explain why they didn't mention it.
> I honestly cannot imagine a better outcome or handling of the situation.
It's the "best outcome" if you're trying to go as fast as possible without breaking any laws or ending up liable for any damage.
German perspective, but if I told people I've been going 30km/h next to a school with poor visibility as children are dropped off around me, I would be met with contempt for that kind of behavior. I'd also at least face some partial civil liability if I hit anyone.
There's certainly better handling of the situation possible, it's just that US traffic laws and attitudes around driving do not encourage it.
I suspect many human drivers would've driven slower, law or no law.
"from behind a tall SUV, "
I look for shadows underneath stationary vehicles. I might also notice pedestrians "vanishing". I have a rather larger "context" than any robot effort.
However, I am just one example of human. My experience of never managing to run someone over is just an anecdote ... so far. The population of humans as a whole manages to run each other over rather regularly.
A pretty cheap instant human sensor might be Bluetooth/BLE noting phones/devices in near range. Pop a sensor in each wing mirror and on the top and bottom. The thing would need some processing power but probably nothing that the built in Android dash screen couldn't handle.
There are lots more sensors that car manufacturers are trying to avoid for cost reasons, that would make a car way better at understanding the context of the world around it.
I gather that Tesla insist on optical (cameras) only and won't do LIDAR. My EV has four cameras and I find it quite hard to see what is going on when it is pissing down with rain, in the same way I do if I don't clean my specs.
> From the Waymo blog...
I'll just remind anyone reading: they're under no obligation to tell the unvarnished truth on their blog.
Even if the NHTSA eventually points out significant failures, getting this report out now has painted a picture of Waymo only having an accident a human would have handled worse.
It would be wise to wait and see if the NHTSA agree. Would a driver have driven at 17mph in this sort of traffic or would they have viewed it as a situation where hidden infant pedestrians are likely to step out?
> reducing speed from approximately 17 mph
Isn't the speed limit normally 15 mph or less in a school zone? Was the robotaxi speeding?
Can’t trust a private company.
Where is the video recording ?
It's hardly surprising that the version of events from the PR department makes Waymo sound completely blameless.
I wonder if another waymo ahead could have seen that child earlier and told the main waymo. This would be pretty neat and have a large safety impact.
I'm picturing a 10 second clip showing a child with a green box drawn around them, and position of gas and brake, updating with superhuman reactions. That would be the best possible marketing that any of these self driving companies could hope for, and Waymo probably now has such a video sitting somewhere.
I dont think Waymo is interested in using a video of their car striking a child as marketing.
It depends on the video. What they should do is arrange for the video to get leaked and let the Internet courts argue about it, and then based on the Internet verdict, come out and claim it's real and they fired somebody for leaking it, or it's AI generated.
Love him or hate him, releasing the video is something I can see Elon doing because assuming a human driver would have done worse, it speaks for itself. Release a web video game where the child sometimes jumps out in front of the car, and see how fast humans respond like the "land Starship" game. Assuming humans would do worse, that is. If the child was clearly visible through the car or some how else avoidable by humans, then I'd be hiding the video too.
Elon has nothing to do with Waymo.
I honestly think that Waymo's reaction was spot on. I drop off and pick up my kid from school every day. The parking lots can be a bit of a messy wild west. My biggest concern is the size of cars especially those huge SUV or pickup trucks that have big covers on the back. You can't see anything incoming unless you stick your head out.
I suspect the robotaxi may have done better than a human.
Human reaction times are terrible, and lots of kids get seriously injured, or killed, when they run out from between cars.
When I was a boy, I ran into the street from between two parked cars. I did not notice the car coming, but he noticed me popping out from nowhere, and screeched to a stop.
I was very very lucky.
I saw a girl dart out between to parked cars on a strode. She was less lucky. The car did slam on their breaks. I have no idea what speed it was ultimately going when they hit the girl. It wasn't enough to send her flying but it was enough to knock her over hard. The dad, was sitting in his front yard and had her up and in his car and I'm guessing rushed to the hospital.
Those kind of neighborhoods where the outer houses face the fast large roads I think are less common now but lots of them left over from the 50+ years ago.
I once rounded a blind curve on a non-residential street only to find a man on a bicycle pulling a trailer with his baby in it, stopped in the middle of the road. I stopped and yelled at him, which surprised him.
That incident still gives me the willies.
You yelled at someone because you were in the wrong? Panic reaction?
Whatever our human laws and morality says about right and wrong and fault, the laws of physics usually judges the car a winner when it hits somebody.
Placing yourself somewhere where pedestrians are not expected (non-residental road) mostly hidden from oncoming traffic for an extended period is putting yourself in undue risk.
You don't always have a choice about where you are momentarily and anybody turning a blind corner has an obligation to immediately reduce their speed (prior to turning the corner!) to where they can safely come to a stop without endangering others. That's drivers education 101. Right after 'don't text while driving', 'don't drink while driving' and 'slow down when there are pedestrians, bicycles and other fragile road users around'.
It’s great handling of the situation. They should release a video as well.
Indeed. Rather than having the company telling me that they did great I'd rather make up my own mind and watch the video.
We should take their reporting with grain of salt and wait for official results
This is great.
what about all the traffic violations though?
https://news.ycombinator.com/item?id=46814583
> remained stopped, moved to the side of the road
Stopped or moved? Is it allowed in CA to move car at all after a serious accident happens?
If the person got up and walked away I'm not sure what damage you'd be doing by reasonably removing your car from blocking others while waiting for police.
Well done waymo!
Take that particular Waymo car off the road. Seems absurd, but they still hit someone.
The car is not the problem. The problem is the intersection of human and machine operating independently of each other with conflicting intention.
I am personally a fan of entirely automated but slow traffic. 10mph limit with zero traffic is fast enough for any metro area.
>> I honestly cannot imagine a better outcome or handling of the situation.
One better than: We investigated our own system and found ourselves to be at no fault?
EDIT: replies say I'm misremembering, disregard.
That was Cruise, and that was fixed by Cruise ceasing operations.
I don’t think that was Waymo right? Cruise is already wound down as far as I know.
Waymo driver? The vehicles are autonomous. I otherwise applaud Waymo's response, and I hope they are as cooperative as they say they will be. However, referring to the autonomous vehicle as having a driver is a dangerous way to phrase it. It's not passive voice, per se, but it has the same effect of obscuring responsibility. Waymo should say we, Waymo LLC, subsidiary of Alphabet, Inc., braked hard...
Importantly, Waymo takes full ownership for something they write positively: Our technology immediately detected the individual.... But Waymo weasels out of taking responsibility for something they write about negatively.
> Waymo driver? The vehicles are autonomous
the "Waymo Driver" is how they refer to the self-driving platform (hardware and software). They've been pretty consistent with that branding, so it's not surprising that they used it here.
> Importantly, Waymo takes full ownership for something they write positively [...] But Waymo weasels out of taking responsibility for something they write about negatively
Pretty standard for corporate Public Relations writing, unfortunately.
> I honestly cannot imagine a better outcome or handling of the situation.
> From the Waymo blog
Yeah, like, no shit Sherlock. We'd better wait for some videos before making our opinions.
In fact I would call that “superhuman” behavior across the board.
The vast vast vast majority of human drivers would not have been able to accomplish that braking procedure that quickly, and then would not have been able to manage the follow up so quickly.
I have watched other parent drivers in the car pick up line at public schools for the last 16 years and people are absolutely trash at navigating that whole process and parents drive so poorly it’s absurd. At least half parents I see on their phones while literally feet away from hitting some kid.
Human drivers are smart enough to slow down when around a school where kids are being dropped of. This piece of software wasn't. Clearly not superhuman.
How do you know how quickly the software braked? A blog post by a company selling a product is not credible material. We need independent sources.
> The vast vast vast majority of human drivers ... would not have been able to manage the follow up so quickly
You are saying the "vast vast vast majority of human drivers" wouldn't pull over after hitting a child?
I remember similar blind faith in and unlimited advocacy for anything Tesla and Musk said, and look how that has turned out. These are serious issues for the people in our communities, not a sporting event with sides.
> I honestly cannot imagine a better outcome or handling of the situation.
If it can yell at the kid and send a grumpy email to the parents and school, the automation is complete.
Most humans in that situation won't have reaction speed to do shit about it and it could result in a severe injury or death.
Yeah. I'm a stickler for accountability falling on drivers, but this really can be an impossible scenario to avoid. I've hit someone on my bike in the exact same circumstance - I was in the bike lane between the parked cars and moving traffic, and someone stepped out between parked vehicles without looking. I had nowhere to swerve, so squeezed my brakes, but could not come to a complete stop. Fortunately, I was going slow enough that no one was injured or even knocked over, but I'm convinced that was the best I could have done in that scenario.
The road design there was the real problem, combined with the size and shape of modern vehicles that impede visibility.
Building on my own experience I think you have to own that if you crash with someone you made a mistake. I do agree that car and road design for bicycles(?) makes it almost impossible to move around if you do not risk things like that.
Humans are not going to win on reaction time but prevention is arguably much more important.
How would standard automatic breaking (standard in some brands) have performed here?
I easily can: when in a school zone never every go so fast that you can't stop before hitting a kid, especially when visibility is limited.
This is the classic Suddenly Revealed Pedestrian test case, which afaik, most NCAP (like EuroNCAP, Japan NCAP) have as part of their standard testing protocols.
Having performed this exact test on 3 dozen vehicles (L2/L3/L4) for several AV companies in the Bay Area [1], I would say that Waymo's response, per their blog post [2] has been textbook compliance. (I'm not defending their performance... just their response to the collision). This test / protocol is hard for any driver (including human driven vehicles), let alone ADAS/L3/L4 vehicles, for various reasons, including: pedestrian occlusion, late ped detection, late braking, slick roads, not enough braking, etc. etc.
Having said all that, full collision avoidance would have been best outcome, which, in this case, it wasn't. Wherever the legal fault may lie -- and there will be big debate here -- Waymo will still have to accept some responsibility, given how aggressively they are rolling out their commercial services.
This only puts more onus on their team to demonstrate a far higher standard of driving than human drivers. Sorry, that's just the way societal acceptance is. We expect more from our robots than from our fellow humans.
[1] Yes, I'm an AV safety expert
[2] https://waymo.com/blog/2026/01/a-commitment-to-transparency-...
(edit: verbiage)
Waymo’s performance, once the pedestrian was revealed, sounds pretty good. But is 17mph a safe speed at an active school dropoff area? I admit that I don’t think I ever personally pay attention to the speedometer in such a place, but 17mph seems excessive even for an ordinary parking lot.
I wonder whether Waymo’s model notices that small children are present or likely to be present and that it should leave extra margin for error.
(My general impression observing Waymo vehicles is that they’ve gone from being obnoxiously cautious to often rather aggressive.)
> But is 17mph a safe speed at an active school dropoff area?
Now you're asking interesting questions... Technically, in CA, the speed limit in school zones are 25 mph (which local authorities can change to 15 mph, as needed). In this case, that would be something the investigation would check, of course. But regardless of that, 17 mph per se is not a very fast speed (my gut check: turning around intersections at > 10-11 mph feels fast, but going straight at 15-20 mph doesnt feel fast; YMMV). But more generally, in the presence of child VRUs (vulnerable road users), it is prudent to drive slowly just because of the randomness factor (children being the most unaware of critters). Did the Waymo see the kids around in the area? If so, how many and where? and how/where were they running/moving to? All of that is investigation data...
My 2c is that Waymo already took all of that into account and concluded that 17 mph was indeed a good speed to move at...
...which leads to your observation below:
> (My general impression observing Waymo vehicles is that they’ve gone from being obnoxiously cautious to often rather aggressive.)
Yes, I have indeed made that same observation. The Waymos of 2 years ago were very cautious; now they seem much more assertive, even a bit aggressive (though that would be tough to define). That is a driving policy decision (cautious vs assertive vs aggressive).
One could argue if indeed 17 mph was the "right" decision. My gut feel is Waymo will argue that (but likely they might make the driving policy more cautious esp in presence of VRUs, and child VRUs particularly)
> Technically, in CA, the speed limit in school zones are 25 mph
Legally a speed limit is a 'limit' on speed, not a suggested or safe speed. So it's never valid to argue legally that you were driving under the limit, the standard is that you slow down or give more room for places like a school drop-off while kids are being dropped off or picked up.
Yep, if I plow into stationary vehicles on the highway while going the "limit" that's not a very solid defense is it?
In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
> In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
That's a difficult question to answer, and the devil really is in the details, as you may have guessed. What I can say that Waymo is, by far, the most prolific publisher of research on AV safety on public roads. (yes, those are my qualifiers...)
Here's their main stash [1] but notably, three papers talk about comparison of Waymo's rider-only (i.e. no safety driver) performance vis-a-vis human driver, at 7.1 million miles [2], 25 million miles [3], 56 million miles [4]. Waymo has also been a big contributor to various AV safety standards as one would expect (FWIW, I was also a contributor to 3 of the standards... the process is sausage-making at its finest, tbh).
I haven't read thru all their papers, but some notable ones talk about the difficulty of comparing AV vs human drivers [5], and various research on characterising uncertainty / risk of collision, comparing AVs to non-impaired, eyes-on human driver [6]
As one may expect, at least one of the challenges is that human-driven collisions are almost always very _lagging indicators_ of safety (i.e. collision happened: lost property, lost limbs, lost lives, etc.)
So, net-net, Waymo still has a VERY LONG WAY to go (obviously) to demonstrate better than human driving behavior, but they are showing that their AVs are better-than-humans on certain high-risk (potential) collisions.
As somebody remarked, the last 1% takes 90% of time/effort. That's where we are...
---
[1] https://waymo.com/safety/research
[2] https://waymo.com/research/comparison-of-waymo-rider-only-cr...
[3] https://waymo.com/research/do-autonomous-vehicles-outperform...
[4] https://waymo.com/research/comparison-of-waymo-rider-only-cr...
[5] https://waymo.com/research/comparative-safety-performance-of...
[6] https://waymo.com/blog/2022/09/benchmarking-av-safety/
[edit: reference]
still, how many ppl do they kill per mile compared to humans?
> Waymo will still have to accept some responsibility
Why? This is only true if they weren't supposed to be on the road in the first place. Which is not true.
Think of it like dog ownership: if my dog hurts someone, that's on me. Property that causes harm is the owner's responsibility.
If I program a machine and it goes out into the world and hurts someone who did not voluntarily release my liability, that's on me.
There are many cases when the owner wouldn’t be liable as well, like if the victim was performing an illegal act like attacking the owner or dog, or trespassing. If a child isn’t following the law or being supervised by a parent, some consequences are inevitable and the driver isn’t instantly liable. For example, if a student jumps in front of a car in an attempted suicide, it would be very hard for a driver to avoid that in certain situations.
In a technical sense, maybe, but it's all going to be about optics. They have a responsibility to handle the situation well even if it's not their fault, and the public will hold them accountable for what they deem the involvement was, which may not be the actual scenario.
> In a technical sense, maybe, but it's all going to be about optics.
Indeed, it is, and that is exactly why Waymo will have to accept some responsibility. I can bet that internally Waymo's PR and Legal teams are working overtime to coordinate the details with NHTSA. We, the general public, may or may not know the details at all, if ever. However, Waymo's technical teams (Safety, etc) will also be working overtime to figure out what they could have done better.
As I mentioned, this is a standard test, and Waymo likely has 1000s of variations of this test in their simulation platforms; they will sweep across all possible parameters to make this test tighter, including the MER (minimum expected response from the AV) and perhaps raise the bar on MER (e.g. brake at max deceleration in some cases, trading off comfort metrics in those cases; etc.) and calculate the effects on local traffic (e.g. "did we endanger the rear vehicles by braking too hard? If so, by how much??" etc). All these are expected actions which the general public will never know (except, perhaps via some technical papers).
Regardless, the PR effects of this collision do not look good, especially as Waymo is expanding their service to other cities (Miami just announced; London by EOY2026). This PR coverage has potential to do more damage to the company than the actual physical damage to the poor traumatized kid and his family. THAT is the responsibility only the company will pay for.
To be sure, my intuition tells me this is not the last such collision. Expect to see some more, by other companies, as they commercialize their own services. It's a matter of statistics.
Bringing a vehicle onto the public roads is a privilege not a right. Any harm to pedestrians that results is your responsibility, not anyone else's.
The performance of a human is inherently limited by biology, and the road rules are written with this in mind. Machines don't have this inherent limitation, so the rules for machines should be much stronger.
I think there is an argument for incentivising the technology to be pushed to its absolute limits by making the machine 100% liable. It's not to say the accident rate has to be zero in practice, but it has to be so low that any remaining accidents can be economically covered by insurance.
At least in the interim, wouldn’t doing what you propose cause more deaths if robot drivers are less harmful than humans, but the rules require stronger than that? (I can see the point in making rules stronger as better options become available, but by that logic, shouldn't we already be moving towards requiring robots and outlawing human drivers if it's safer?)
In your experience, where do we find a credible source of info? Do we need to wait for the government's investigation to finish?
> I would say that Waymo's response, per their blog post [2] has been textbook compliance.
Remember Tesla's blog posts? Of course Waymo knows textbook compliance just like you do, and of course that's what they would claim.
> In your experience, where do we find a credible source of info? Do we need to wait for the government's investigation to finish?
Most likely, yes, the NHTSA investigation will be credible source of info for this case. HOWEVER, Waymo will likely fight it tooth-and-nail from letting it be public. They will likely cite "proprietary algorithms / design", etc. to protect it from being released publicly. So, net-net, I dunno... Will have to wait and see :shrug.gif:
But meanwhile, personally I would read reports from experts like Phil Koopman [1] and Missy Cummings [2] to see their take.
> Remember Tesla's blog posts?
You, Sir, cite two companies that are diametrically opposite on the safety spectrum, as far as good behavior is concerned. Admittedly, one would have less confidence in Waymo's own public postings about this (and I'd be mighty surprised if they actually made public their investigation data, which would be a welcome and an pioneering move).
On the other hand, the other company you mentioned, the less said the better.
[1] http://www.koopman.us/
[2] https://www.gmu.edu/profiles/cummings
There is already widespread discussion on LinkedIn about this thread... but usefully, here [1] is the NHTSA's Office of Defects Investigation report. Nothing much new there, tbh.
As I did suspect, legal scholars are already calling for "voluntary disclosure" from Waymo re: its annotated videos of the collision [2]. FWIW, my skepticism about Waymo actually releasing it remains...
[1] https://static.nhtsa.gov/odi/inv/2026/INOA-PE26001-10005.pdf
[2] https://www.linkedin.com/posts/matthew-wansley-62b5b9126_a-w...
> You, Sir, cite two companies that are diametrically opposite on the safety spectrum
Cringe. Stop it. Simping for google has stopped being cool nearly 2 decades ago.
Still relies on an actual driver.
“The event occurred when the pedestrian suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path. Our technology immediately detected the individual as soon as they began to emerge from behind the stopped vehicle. The Waymo Driver braked hard, reducing speed from approximately 17 mph to under six mph before contact was made,” a statement from Waymo explains.
"Waymo Driver" is their term for their self driving software.
Though given the situation a human driver would not have been going 17 mph in a school zone during drop-off near double parked vehicles
1. I often see signs in such areas that flash when people exceed the limit. I’d urge you to pull over and see how often humans drive above the limit. 2. I’d urge you to also pull over and watch for how many drivers are not consistently looking at the road, such as using their phones, looking down at climate/entertainment/vehicle controls, looking at a passenger, etc
Meanwhile the news does not report the other ~7,000 children per year injured as pedestrians in traffic crashes in the US.
I think the overall picture is a pretty fantastic outcome -- even a single event is a newsworthy moment _because it's so rare_ .
> The NHTSA’s Office of Defects Investigation is investigating “whether the Waymo AV exercised appropriate caution given, among other things, its proximity to the elementary school during drop off hours, and the presence of young pedestrians and other potential vulnerable road users.”
Meanwhile in my area of the world parents are busy, stressed, and on their phones, and pressing the accelerator hard because they're time pressured and feel like that will make up for the 5 minutes late they are on a 15 minute drive... The truth is this technology is, as far as i can tell, superior to humans in a high number of situations if only for a lack of emotionality (and inability to text and drive / drink and drive)... but for some reason the world wants to keep nit picking it.
A story, my grandpa drove for longer than he should have. Yes him losing his license would have been the optimal case. But, pragmatically that didn't happen... him being in and using a Waymo (or Cruise, RIP) car would have been a marginal improvement on the situation.
Err, that is not the desirable statistic you seem to think it is. American drivers average ~3 trillion miles per year [1]. That means ~7000 child pedestrian injurys per year [2] would be ~1 per 430 million miles. Waymo has done on the order of 100-200 million miles autonomously. So this would be ~2-4x more injurys than the human average.
However, the child pedestrian injury rate is only a official estimate (it is possible it may be undercounting relative to highly scrutinized Waymo vehicle-miles) and is a whole US average (it might not be a comparable operational domain), but absent more precise and better information, we should default to the calculation of 2-4x the rate.
[1] https://afdc.energy.gov/data/10315
[2] https://crashstats.nhtsa.dot.gov/Api/Public/Publication/8137...
I suspect that highway miles heavily skew this statistic. There's naturally far fewer pedestrians on highways (lower numerator), people travel longer distances on highways (higher denominator), and Waymo vehicles didn't drive on highways until recently. If you look only at non-highway miles, you'll get a much more accurate comparison.
Then you or Waymo can meet the burden of proof and present that more precise and better information. There is little reason to assume against safety at this point in time except as a intellectual exercise for how more accurate information could be found.
Until then, it is only prudent to defer snap judgements, but increase caution, insist on rigor and transparency, and demand more accurate information.
Does common sense not factor in here at all? Advocating for such rigor is fine, but a refusal to state an opinion just reeks of bias
> we should default to the calculation of 2-4x the rate.
No we should not. We should accept that we don't have any statistically meaningful number at all, since we only have a single incident.
Let's assume we roll a standard die once and it shows a six. Statistically, we only expect a six in one sixth of the cases. But we already got one on a single roll! Concluding Waymo vehicles hit 2 to 4 times as many children as human drivers is like concluding the die in the example is six times as likely to show a six as a fair die.
More data would certainly be better, but it's not as bad as you suggest -- the large number of miles driven till first incident does tell us something statistically meaningful about the incident rate per mile driven. If we view the data as a large sample of miles driven, each with some observed number of incidents, then what we have is "merely" an extremely skewed distribution. I can confidently say that, if you pick any sane family of distributions to model this, then after fitting just this "single" data point, the model will report that P(MTTF < one hundredth of the observed number of miles driven so far) is negligible. This would hold even if there were zero incidents so far.
We get a statistically meaningful result about an upper bound of the incident rate. We get no statistically meaningful lower bound.
Uh, the miles driven is like rolling the die, not hitting kids.
Sure, but we shouldn't stretch the analogy too far. Die rolls are discrete events, while miles driven are continuous. We expect the number of sixes we get to follow a binomial distribution, while we expect the number of accidents to follow a Poisson distribution. Either way, trying to guess the mean value of the distribution after a single incident of the event will never give you a statistically meaningful lower bound, only an upper bound.
The Poisson distribution is well approximated by the binomial distribution when n is high and p is low, which is exactly the case here. Despite the high variance in the sample mean, we can still make high-confidence statements about what range of incident rates are likely -- basically, dramatically higher rates are extremely unlikely. (Not sure, but I think it will turn out that confidence in statements about the true incident rate being lower than observed will be much lower.)
Would this Waymo incident be counted as an injury? Sounds like the victim was relatively unharmed? Presumably there are human-driver incidents like this where a car hits a child at low speeds, with effectively no injuries, but is never recorded as such?
If that's the case, then that's great info. Thank you for adding :)
People's standards for when they're willing to cede control over their lives both as the passenger and the pedestrian in the situation to a machine are higher than a human.
And for not totally irrational reasons like machine follows programming and does not fear death, or with 100% certainty machine has bugs which will eventually end up killing someone for a really stupid reason—and nobody wants that to be them. Then there's just the general https://xkcd.com/2030/ problem of people rightfully not trusting technology because we are really bad at it, and our systems are set up in such a way that once you reach critical mass of money consequences become other people's problem.
Washington banned automatic subway train operation for 15 years after one incident that wasn't the computer's fault, and they still make a human sit in the cab. That's the bar. In that light it's hard not to see these cars as playing fast and loose with people's safety by comparison.
>People's standards for when they're willing to cede control over their lives both as the passenger and the pedestrian in the situation to a machine are higher than a human.
Are they? It is now clear that Tesla FSD is much worse than a human driver and yet there has been basically no attempt by anyone in government to stop them.
> basically no attempt by anyone in government to stop them.
No one in the _US_ government. Note that European governments and China haven't approved it in the first place.
Do you have data to back this claim up, specifically with HW4 (most recent hardware) and FSD software releases?
FSD is already better than at least one class of drivers. If FSD is engaged and the driver passes out, FSD will pull over to the side of the road and stop. And before we leap to conclusions that it only helps in the case of drunk drivers who shouldn't be driving in the first place (which, they shouldn't be), random strokes and seizures happen to people all the time.
We should all think twice before taking a company PR statement completely at face value and praising them for slowing down faster than their own internal "model" says a human driver would. Companies are heavily interested in protecting their bottom line and in a situation like this probably had 5-10 people carefully craft every single word of the statement for maximum damage control.
Surprised at how many comments here seem eager to praise Waymo based off their PR statement. Sure it sounds great if you read that the Waymo slowed down faster than a human. But would a human truly have hit the child here? Two blocks from a school with tons of kids, crossing guards, double parked cars, etc? The same Waymo that is under investigation for passing school busses illegally? It may have been entirely avoidable for the average human in this situation, but the robotaxi had a blind spot that it couldn't reason around and drove negligently.
Maybe the robotaxi did prevent some harm by braking with superhuman speed. But I am personally unconvinced it was a completely unavoidable freak accident type of situation without seeing more evidence than a blog post by a company with a heavily vested interest in the situation. I have anecdotally seen Waymo in my area drive poorly in various situations, and I'm sure I'm not the only one.
There's the classic "humans are bad drivers" but I don't think that is an excuse to not look critically into robotaxi accidents. A human driver who hit a child next to a school would have a personal responsibility and might face real jail time or at the least be put on trial and investigated. Who at Waymo will face similar consequences or risk for the same outcome?
> Surprised at how many comments here seem eager to praise Waymo based off their PR statement.
Really? My impression is that, for the most part, HN consistently sides with the companies. I say this in the most neutral way possible.
It's going to sound batshit insane what I say - the problem is, if we don't praise company PR, the other side will use this as an excuse to push even harder regulations, not allow them in newer cities, slow down the adoption rate, while factually ignoring that this is just a safer method of transport. I wish I was not a bootlicker, but I really want robotaxis to be available everywhere in the world at some point, and such issues should not slow them down IF it's better, and especially, not worse than humans on average.
You're right, what you're saying is batshit insane.
I understand it sounds stupid, but there was huge push back for introducing Uber to the cities I lived in. And obviously this is even bigger change. However, if a private company is willing to foot the bill, go above and beyond to prove its usefulness and safety, I will be repping for it.
It’s not honestly, it’s the unspoken political battle being constantly fought over all kinds of things.
One of the few seeing through Waymo PR bullshit.
Do you know anyone who works at Waymo? The cynicism is silly. Just because some people at some companies behave horribly, it doesn't mean all or even most do.
Look at Waymo's history in the space, meet some of the people working there, then make a decision.
You don't have to think anyone is behaving horribly to acknowledge that a company's PR department will tend to put out the version of the story that makes them look best.
The a human would do it better people are hilarious. Given how many times I have been hit by human drives on my bike and watched others get creamed by a cars. One time in Boulder at a flashing cross walk a person ran right through it and the biker they creamed got stuck in the roof rack.
For real, I am convinced these are people who never walk or bike, at least around cities like Santa Monica. I am an everyday urban walker and I have to constantly be on alert not to be hit, even when I'm behaving predictably and with the right of way.
Yeah I have to wonder if any of the "humans would do it better" people actually have children and have dropped them off in a school zone. Drivers are on their phones rolling through school zones at 25-30 during pickup/dropoff hours all the fucking time.
I was just dropping my kids off at their elementary school in Santa Monica, but not at Grant Elementary where this happened.
While it's third-hand, word on the local parent chat is that the parent dropped their kid off on the opposite side of the street from Grant. Even though there was a crossing guard, the kid ran behind a car an ran right out in to the street.
If those rumors are correct, I'll say the kid's/family's fault. That said, I think autonomous vehicles should probably go extra-slowly near schools, especially during pickup and dropoff.
When my kids were school age, I taught them that the purpose of crosswalk lines is to determine who pays for your funeral.
They got the point.
This is a very good way of putting it.
We live very close to Grant. We go through this intersection to walk our kids to their schools & know the crossing guards pretty well.
This matches exactly what they said.
That kid is lucky it was a Waymo & not a human driven car.
Do you think Waymos should be banned from driving through Santa Monica?
No. They are by far the safest drivers in Santa Monica. Ideally we get to a point where human drivers are banned.
I do not like the phase "it's the kid's fault" for a kid being hit by a robot-car.
It is never a 6 year old's fault if they get struck by a robot.
Exactly. It’s his parents fault.
At some point children are capable of pursuing Darwin Awards. Parents may enable this, but ultimately if one’s child does something stupid contrary to one’s guidance and restrictions, they may end up with a Darwin for it. Two hundred years ago the child mortality rate was half, as in you lost one child per two, and most of those were not the fault of the child or parents. Society for quite some years has been pushing that down, to the point that a near-death involving a neglectful parent and a witless child is apparently (?) newsworthy — but the number of deaths will never reach zero, whether humans or robots or empty plains and blue skies. There will always be a Veruca Salt throwing themselves into the furnace no matter how many safety processes we impose onto roads, cars, drivers, and/or robots.
If you want to see an end to this nonsensical behavior by parents, pressure your local city into having strict traffic enforcement and ticketing during school hours at every local school, so that the parent networks can’t share news with each other of which school is being ‘harassed’ today. Give license points to vehicles that drop a child across the street, issue parking tickets to double parkers, and boot vehicles whose drivers refuse to move when asked. Demand they do this for the children, to protect them from the robots, if you like.
But.
It’ll protect them much more from the humans than from the robots, and after a few thousand rockets are issued to parents behaving badly, you’ll find that the true threat to children’s safety on school roads is children’s parents — just as the schools have known for decades. And that’s not a war you’ll win arguing against robots. (It’s a war you’ll win arguing against child-killing urban roadway design, though!)
*tickets
No-fault accidents happen. Accidents can have causes that are not legal nor moral blame.
The US commercial aviation industry did not get to its excellent safety record by simply shrugging and accepting a “no-fault accident”.
There are always systemic factors that can be improved, for example working on street design to separate dangerous cars from children, or transportation policy by shifting transportation to buses, bikes, and walking where the consequences of mistakes are significantly reduced.
Cars are the #2 killer of children in the US, and it’s largely because of attitudes like this that ignore the extreme harm that is caused by preventable “accidents”
"No fault" does not mean "no cause" and air crash investigations always focus on causes, not fault. When you understand causes, you can think about how to prevent them happening again.
A human driver travelling at the same speed would have hit that child at exactly 17 mph, before their brain even registered that child was there. If that driver would also have been driving a large SUV that child would have been pushed on the ground and ran over, so probably a fatality. And also functionally nobody would have given a shit apart from some lame finger pointing at (probably) the kid’s parents.
And it is not the child’s or their parents’ fault either:
Once you accept elementary school aged children exist, you have to accept they will sometimes run out like this. Children just don’t have the same impulse control as adults. And honestly even for adults stepping out a bit from behind an obstacle in the path of a car is an easy mistake to make. Don’t forget that for children an SUV is well above head height so it isn’t even possible for them to totally avoid stepping out a bit before looking. (And I don’t think stepping out vs. running out changes the outcome a lot)
This is why low speed limits around schools exist.
So I would say the Waymo did pretty well here, it travelled at a speed where it was still able to avoid not only a fatality but also major injury.
> A human driver travelling at the same speed would have hit that child at exactly 17 mph, before their brain even registered that child was there.
Not sure where this is coming from, and it's directly contradicted by the article:
> Waymo said in its blog post that its “peer-reviewed model” shows a “fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph.” The company did not release a specific analysis of this crash.
No, Waymo’s quote supports the grandparent comment - it was about a “fully attentive human driver” - unless you are arguing that human drivers are consistently “fully attentive”?
Fair enough, so then how fast would a semi-attentive driver stop?
The comment I originally replied to makes the claim a human's brain wouldn't have even responded fast enough to register the child was there. That's going WAY further than how Waymo is claiming a human would have responded.
I don't see how that's a more reasonable assumption that a human driver actually being "fully attentive", and I'm not sure Waymo's definition of that term is the same as what you're using.
> And it is not the child’s or their parents’ fault either: Once you accept elementary school aged children exist, you have to accept they will sometimes run out like this. Children just don’t have the same impulse control as adults.
I get what you are trying to say and I definitely agree in spirit, but I tell my kid (now 9) "it doesn't matter if it isn't your fault, you'll still get hurt or be dead." I spent a lot of time teaching him how to cross the street safely before I let him do it on his own, not to trust cars to do the right thing, not to trust them to see you, not to trust some idiot to not park right next to cross walk in a huge van that cars have no chance of seeing over.
If only we had a Dutch culture of pedistrian and road safety here.
Vehicle design also plays a role: passenger cars have to meet pedestrian collision standards. Trucks don't. The silly butch grilles on SUVs and pickups are deadly. This is more of an argument for not seeing transportation as a fashion or lifestyle statement. Those truck designs are about vanity and gender affirming care. It's easier to make rational choices when it's a business that's worried about liability making those choices.
Cheers to cities pedestrianizing school streets even in busy capitals (e.g. Paris). Cars have no place near school entrances. Fix your urbanism and public transportation.
Yes, kids in developed countries have the autonomy to go to school by themselves from a very young age, provided the correct mindset and a safe environment. That's a combination of:
* high-trust society: commuting alone or in a small group is the norm, soccer moms a rare exception,
* safe, separated lanes for biking/walking when that's an option.
you're exactly right. the fixation on human vs AV error rates completely misses the point. even if we achieve 'perfect' AVs, mixing heavy machinery with children guarantees conflict. physics dictate cars can't stop instantly. the only solution is removing cars, not better drivers.
most commenters here are ignoring the structural incentives. the long term threat of waymo isn't safety, its the enclosure of public infrastructure. these companies are building a permission structure to lobby personal vehicles and public transit off the road.
transportation demand is inelastic. if we allow a transition where mobility is captured by private platforms, the consumer loses all leverage. the endgame is the american healthcare model: capture the market, kill alternatives, and extract max rent because the user has no choice. we need dense urban cores and mass transit, not a dependency on rent seeking oligopolies
Yup, nailed it.
The school speed limit there is 15 mph, and that wasn't enough to prevent an accident.
https://www.yahoo.com/news/articles/child-struck-waymo-near-...
https://maps.app.goo.gl/7PcB2zskuKyYB56W8?g_st=ac
The interesting thing is a 12 mph speed limit would be honored by an autonomous vehicle but probably ignored by humans.
If the speed limit was 15 mph, and the Waymo vehicle was traveling at 17 mph before braking, why do you believe the Waymo vehicle would honor a 12 mph speed limit? It didn't honor the 15 mph limit.
> If the speed limit was 15 mph, and the Waymo vehicle was traveling at 17 mph before braking, why do you believe the Waymo vehicle would honor a 12 mph speed limit?
+/- 2 mph is acceptable speedometer and other error. (15 mph doesn’t mean never exceed under any legal inteprerstion I know.)
It’s reasonable to say Waymo would reduce speed in a 12 versus 15 in a way most humans would not.
Ignored by some, not all humans. I absolutely drive extra slowly and cautiously when driving past an elementary school during drop off and pick up precisely because kids do dumb stuff like this. Others do too, though not everyone of course, incredibly.
The great thing about doing things like driving the speed limit in school zones is you get to witness other drivers drive even worse, like passing you in a no passing zone in front of the school, because they can't bear to drive slow for three blocks.
We are responsible for the consequences of our actions. The speed limit is almost irrelevant; drive slowly enough so you don't hit anyone - especially in a school zone.
> We are responsible for the consequences of our actions.
We're not though. Drivers are allowed to kill as many people as they like as long as they're apologetic and weren't drinking; at most they pay a small fine.
We're responsible for the consequences of our actions regardless of what anyone else says, including the law.
Also, where I live that's manslaughter, a serious felony that can put you in jail.
So drive at 0mph?
So the waymo was speeding! All the dumbasses on here defending waymo when it was going 17 > 15.
Oh also, that video says "kid ran out from a double parked suv". Can you imagine being dumb enough to drive over the speed limit around a double parked SUV in a school zone?
Depends on where the Waymo was.
The 15 mph speed limit starts on the block the school is on. The article says the Waymo was within two blocks of the school, so it's possible they were in a 25 mph zone.
https://maps.app.goo.gl/Vhce7puwwYyDYEuo6
> Can you imagine being dumb enough to drive over the speed limit around a double parked SUV in a school zone?
Can you imagine being dumb enough to think that exceeding a one size fits all number on a sign by <10% is the main failing here?
As if 2mph would have fundamentally changed this. Pfft.
A double parked car, in an area with chock full street parking (hence the double park) and "something" that's a magnet for pedestrians, and probably a bunch of pedestrians should be a "severe caution" situation for any driver who "gets it". You shouldn't need a sign to tell you that this is a particular zone and that warrants a particular magic number.
The proper reaction to a given set of indicators that indicate hazards depends on the situation. If this were easy to put in a formula Waymo would have and we wouldn't be discussing this accident because it wouldn't have happened.
> As if 2mph would have fundamentally changed this. Pfft.
According to https://news.ycombinator.com/item?id=46812226 1mph slower might have entirely avoided contact in this particular case.
That was my point. The Waymo should have been going much slower than 15 around the double-parked car. Potential speeding makes it worse.
The fact that it’s hard to turn this into a formula is exactly why robot drivers are bad.
Are you comparing robot drivers to the existing alternative? Next time you see one of those blinking speed displays, I’d urge you to pull over and see how fast many human drivers go, and watch for what percent of them aren’t consistently even looking at the road ahead.
The default, with good visibility in ideal conditions, should be to not exceed the speed limit.
In a school zone, when in a situation of low visibility, the car should likely be going significantly below the speed limit.
So, it's not a case of 17mph vs 15mph, but more like 17mph vs 10mph or 5mph.
>The default, with good visibility in ideal conditions, should be to not exceed the speed limit.
Please pass this message on to 99.999% of human drivers who think speed limit is the minimum speed.
So let me get this straight, the car should have been going less than the speed limit, but the fact that it was going a hair over the speed limit is the problem?
The car clearly failed to identify that this was a situation it needed to be going slower. The fact that it was going 17 instead of 15 is basically irrelevant here except as fodder for moral posturing. If the car is incapable of identifying those situations no amount of "muh magic number on sign" is going to fix it. You'll just have the same exact accident again in a 20 school zone.
It is a contributing factor.
If the car is going slower than the speed limit in this scenario, it is difficult to tell over the internet if that speed was appropriate. If the car is going over the speed limit, it is obviously inappropriate.
My read is the Waymo was operating normally in a situation a human would operate with heightened awareness and caution. This is not the win they think it is.
Some humans would operate with heightened awareness and caution. Some other humans might be drunk or looking at their phone.
For reference, here's a link to Waymo's blog post: https://waymo.com/blog/2026/01/a-commitment-to-transparency-...
Personally in LA I had a Waymo try to take a right as I was driving straight down the street. It almost T-boned me and then honked at me. I don’t know if there has been a change to the algorithm lately to make them more aggressive but it was pretty jarring to see it mess up that badly
It honked at you? But local laws dictate that it angrily flashes its high beams at you.
In recent weeks I've found myself driving in downtown SF congestion more than usual, and observed Waymos doing totally absurd things on multiple occasions.
The main saving grace is they all occurred at low enough speeds that the consequences were little more than frustrating/delaying for everyone present - pedestrians and drivers alike, as nobody knew what to expect next.
They are very far from perfect drivers. And what's especially problematic is the nature of their mistakes seem totally bizarre vs. the kinds of mistakes human drivers make.
The unpredictability was jarring to me as a passenger in a Waymo.
I'm curious as to what kind of control stack Waymo uses for their vehicles. Obviously their perception stack has to be based off of trained models, but I'm curious if their controllers have any formal guarantees under certain conditions, and if the child walking out was within that formal set of parameters (e.g. velocity, distance to obstacle) or if it violated that, making their control stack switch to some other "panic" controller.
This will continue to be the debate—whether human performance would have exceeded that of the autonomous system.
From a purely stats pov, in situations where the confusion matrix is very asymmetric in terms of what we care about (false negatives are extra bad), you generally want multiple uncorrelated mechanisms, and simply require that only one flips before deciding to stop. All would have to fail simultaneously to not brake, which becomes vanishingly unlikely (p^n) with multiple mechanisms assuming uncorrelated errors. This is why I love the concept of Lidar and optical together.
The true self-driving trolley problem. How many rear-end collisions and riders' annoyance caused by phantom braking a manufacturer (or a society) is going to tolerate to save one child per N million miles?
Uncorrelated approach improves sensitivity at the cost of specificity. Early sensor fusion might improve both (maybe at the cost of somewhat lesser sensitivity).
With above-average human reflexes, the kid would have been hit at 14mph instead of 6mph.
About 5x more kinetic energy.
Yeah, if a human made the same mistakes as the Waymo driving too fast near the school, then they would have hurt the kid much worse than the Waymo did.
So if we're going to have cars drive irresponsibly fast near schools, it's better that they be piloted by robots.
But there may be a better solution...
But would a human be driving at 17 in a school zone during drop off hours? Id argue a human may be slower exactly because of this scenario
> would a human be driving at 17 in a school zone during drop off hours?
In my experience in California, always and yes.
Maybe we should not only replace the unsafe humans with robots, but also have the robots drive in a safe manner near schools rather than replicating the unsafe human behavior?
One argument for the robots is that they can be programmed to drive safer, while humans cant.
But that depends on reliability, especially in unforseen (and untrained-upon) circumstances. We'll have to see how they do, but they have been doing better than expected
Depends on the school zone. The tech school near me is in a 50 zone and they don't even turn on the "20 when flashing" signs because if you're gonna walk there, you're gonna come in via residential side streets in the back and the school itself is way back off the road. The other school near me is downtown and you wouldn't be able to go 17 even if you wanted to.
Kinetic energy is a bad metric. Acceleration is what splats people.
Jumping out of a plane wearing a parachute vs jumping off a building without one.
But acceleration is hard to calculate without knowing time or distance (assuming it's even linear) and you don't get that exponent over velocity yielding you a big number that's great for heartstring grabbing and appealing to emotion hence why nobody ever uses it.
Absent more precise information, this is a statistical negative mark for Waymo putting their child pedestrian injury rate at ~2-4x higher than the US human average.
US human drivers average ~3.3 trillion miles per year [1]. US human drivers cause ~7,000 child pedestrian injurys per year [2]. That amounts to a average of 1 child pedestrian injury per ~470 million miles. Waymo has done ~100-200 million fully autonomous miles [3][4]. That means they average 1 child pedestrian injury per ~100-200 million miles. That is a injury rate ~2-4x higher than the human average.
However, the child pedestrian injury rate is only a official estimate (possible undercounting relative to highly scrutinized Waymo miles) and is a whole US average (operational domain might not be comparable, though this could easily swing either way), but absent more precise and better information, we should default to the calculated 2-4x higher injury rate; it is up to Waymo to robustly demonstrate otherwise.
Furthermore, Waymo has published reasonably robust claims arguing they achieve ~90% crash reduction [5] in total. The most likely new hypotheses in light of this crash are:
A. Their systems are not actually robustly 10x better than human drivers. Waymos claims are incorrect or non-comparable.
B. There are child-specific risk factors that humans account for that Waymo does not that cause a 20-40x differential risk around children relative to normal Waymo driving.
C. This is a fluke child pedestrian injury. Time will tell. Given their relatively robustly claimed 90% crash reduction, it is likely prudent to allow further operation in general, though possibly not in certain contexts.
[1] https://afdc.energy.gov/data/10315
[2] https://crashstats.nhtsa.dot.gov/Api/Public/Publication/8137...
[3] https://www.therobotreport.com/waymo-reaches-100m-fully-auto...
[4] https://waymo.com/blog/2025/12/demonstrably-safe-ai-for-auto...
[5] https://waymo.com/safety/impact/
I don't think this comparison is meaningful given the sample size of 1 and the differences between the between your datasets. The standard error margins from the small sample size alone are so large that you could not reasonably claim humans are safer (95% CI for Waymo is about 1 per 20 million miles to 1 per 8 billion miles). Then there are the dataset differences:
1. The NHTSA data is based on police-reported crash data, which reports far fewer injuries than the CDC reports based on ED visits. The child in this case appeared mostly unharmed and situations like this would likely not be counted in the NHTSA data.
2. Waymo taxis operate primarily in densely populated urban environments while human driver milage includes highways and rural roads where you're much less likely to collide with pedestrians per mile driven.
Waymo's 90% crash reduction claim is at least an apples-to-apples comparison.
I don't think I'd want to take much from such a statistical result yet. A sample size of 1 accident just isn't enough information to get a real rate from, not that I want to see more collisions with children. Though this is also muddied by the fact that Waymo will most likely adjust their software to make this less likely, and we won't know exactly how or how many miles each version has. I'd also like to see the data for human incidents over just the temperate suburban areas like Waymo operates in.
> child pedestrian injury rate at ~2-4x higher than the US human average.
If this incident had happened with a human driven vehicle would it even have been reported?
I don't know exactly what a 6mph collision looks like but I think it's likely the child had nothing more than some bruises and if a human has done it they would have just said sorry, made sure they were ok, and left
Do we even know that the child was injured? All I've seen anyone officially claim is that the Waymo made contact, the kid fell over, then stood up and walked to the side of the road. Assuming the Waymo was still braking hard, 6mph means it was about 1/4s and about 30cm from reaching a full stop, so it could be a very minor incident we're talking about here.
I'm not aware of any statistics for how often children come into contact with human-driven cars.
The Waymo driver tech is impressive. That said an experienced driver might have recognized the pattern where a stopped big vehicle occludes a part of the road leading to such situation, and might have stopped or slowed down almost to a halt before passing. The Waymo driver reacts faster but is not able to predict such scenarios by filling the gaps, simulating the world to inform decisions. Chapeau to Waymo anyways
There have been many instances of Waymo preventing a collision by predicting pedestrians emerging from occlusion. This isn’t new information at all for them. Some accidents are simply physically impossible to prevent. I don’t know for sure if this one was one of those, but I’m fairly confident it couldn’t have been from prediction failure.
See past examples:
https://youtube.com/watch?v=hubWIuuz-e4 — first save is a child emerging from a parked car. Notice how Waymo slows down preemptively before the child starts moving.
https://www.reddit.com/r/waymo/s/ivQPuExwNW — detects foot movement from under the bus.
https://www.reddit.com/r/waymo/s/LURJ8isQJ6 — stops for dogs and children running onto the street at night.
> detects foot movement ..
That’s probably how they do it, which is again very clever stuff, chapeau. But they do it like that b/c they can’t really predict the world around them fast enough. It might be possible in the future with AI World Models though
What do you mean “fast enough”? You can’t predict something that doesn’t exist i.e. not visible to the sensors. A Waymo wouldn’t move at all if it assumed people would always jump out of nowhere.
Even if you detect “fast enough”, there are physical limits for braking and coming to a stop.
This one should have been prevented because the Waymo should have been driving at max 10mph
I think this definitely an improvement to consider, but when comparing I think that big number, i.e. statistics are the only thing that matters. Some human could detect the pattern and come to full halt another human driver could be speeding while texting
> The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post.
The issue is that I don’t trust a private company word. You can’t even trust the president of the USA government nowadays… release the video footage or get lost.
Basically Waymo just prevented a kids potential death.
Bad any other car been there, probably including Tesla, the poor kid would have been hit with 4-10x more force.
You just invented a hypothetical situation in your head then drew conclusions from it. In my version, the other car misses the kid entirely.
Yeah, but Tesla has a proven bad safety record. Waymo doesn't and the GP comment is alluding to that
Evidence (preferably with recent Teslas/HW4)?
> any other car been there, probably including Tesla
Cheap shots. If this was Tesla there would be live media coverage across every news outlet around the world and congressmen racing to start investigation.
Look at any thread where Tesla is mentioned and how many waymo simps are mansplaning lidar.
The only question I have is whether the speed it was going was situationally appropriate and whether we’d expect a human to be considered reckless under the same circumstances. 17mph sounds pretty slow but it really depends on context.
Oddly I cannot decide if this is cause for damnation or celebration
Waymo hits a kid? Ban the tech immediately, obviously it needs more work.
Waymo hits a kid? Well if it was a human driver the kid might well have been dead rather than bruised.
> Waymo hits a kid? Ban the tech immediately, obviously it needs more work.
> Waymo hits a kid? Well if it was a human driver the kid might well have been dead rather than bruised.
These can be true at the same time. Waymo is held to a significantly higher standard than human drivers.
> Waymo is held to a significantly higher standard than human drivers.
They have to be, as a machine can not be held accountable for a decision.
Slowing the adoption of much-safer-than-humans robotaxis, for whatever reason, has a price measured in lives. If you think that the principle you've just stated is worth all those additional dead people, okay; but you should at least be aware of the price.
Failure to acknowledge the existence of tradeoffs tends to lead to people making really lousy trades, in the same way that running around with your eyes closed tends to result in running into walls and tripping over unseen furniture.
But we have no way of knowing whether robotaxis are safer. See, for example, the arguments raised here: https://www.bloomberg.com/news/features/2026-01-06/are-auton...
We can't blindly trust Waymo's PR releases or apples-to-oranges comparisons. That's why the bar is higher.
You may not have any way of knowing but the rest of society has developed all sorts of systems of knowing. "Scientific method", "Bayesian reasoning", etc. or start with the Greek philosophy classics.
Waymo is not a machine, it is a corporation, and corporations can, in fact be held accountable for decisions (and, perhaps more to the point, for defects in goods they manufacture, sell, distribute, and/or use to provide services.)
The promise of self-driving cars being safer than human drivers is also kind of the whole selling point of the technology.
What? No? The main selling point is eliminating costs for a human driver (by enabling people to safely do other things from their car, like answering emails or doomscrolling, or via robotaxis).
Sure, but the companies building them are just shoving billions of dollars into their ears so they don't have to answer "who's responsible when it kills someone?"
The question of responsibility, while philosophically interesting wrt. increasingly autonomous machines, is not going to be an issue in practice. We'll end up dealing with it like we always do with multi-party responsibility in complex systems: regulators setting safety standards and outlining types and structures of liabiliy, contracts shifting the liability around, and lots and lots of insurance.
In fact, if you substitute "company providing self-driving solution (integrated software + hardware)" for "company renting out commercial drivers" (or machine operators), then self-driving cars already fit well into existing legal framework. The way I see it, the only change self-driving cars introduce here is that there is no individual operator we could blame for the accident, no specific human we could heavily fine or jail, and then feel good about ourselves because we've issued retributive justice and everything is whole now. Everything else has already long been worked out.
> They have to be, as a machine can not be held accountable for a decision
This logic applies equally to all cars, which are machines. Waymo has its decision makers one more step removed than human drivers. But it’s not a good axiom to base any theory of liability on.
In traditional vehicles, liability is structurally centered on the human driver, and product liability exists but is difficult to assert absent clear or systemic defects.
Autonomous driving systems disrupt this by directly assuming the driving function, forcing liability upstream (where it's significantly more difficult to navigate).
So now it's driver vs injured party. Self-driving makes it trillion dollar company vs injured party. Night and day difference.
That sucks, and I love to hate on "self driving" cars. But it wasn't speeding to start with (assuming speed limit in the school zone was 20 or 25), braked as much as possible, and the company took over all the things a human driver would have been expected to do in the same situation. Could have been a lot worse, probably wouldn't have been any better with a human driver (just going to ignore as no-signal Waymo's models that say an attentive human driver would have been worse). It's "fine". In this situation, cars period are the problem, not "self driving" cars.
I don't like even the very idea o self-driving cars, but based on the description of the accident, I think the machine passed this with flying colors.
It's hard to imagine how any driver could have reacted better in this situation.
The argument that questions "would a human be driving 17mph in a school zone" feels absurd to the point of being potentially disingenuous. I've walked and driven through many school zones before, and human drivers routinely drive above 17mph (in some cases, over the typical 20mph or 25mph legal limit). It feels like in deconstructing some of these incidences, critics imagine a hypothetical scenario in which they are driving a car and its their only job to avoid a specific accident that they know will happen in advance, rather than facing the reality of what human drivers are actually like on the road.
Who is legally responsible in case a Waymo hits a pedestrian? If I hit somebody, it's me in front of a judge. In the case of Waymo?
When I was a kid (age 12, or so), I got hit by a truck while crossing the road on my bike.
In that particular instance, I was cited myself -- after the fact, at the hospital -- and eventually went before a judge. In that hearing, it was established that I was guilty of failing to yield at an intersection.
(That was a rather long time ago and I don't remember the nature of the punishment that resulted. It may have been as little as a stern talking-to by the judge.)
A person who hits a child, or anyone, in America, with no resulting injury, stands a roughly 0% chance of facing a judge in consequence. Part of Waymo's research is to show that even injury accidents are rarely reported to the police.
Are you thinking of civil liability or criminal liability?
Waymo is liable in a civil sense and pays whatever monetary amount is negotiated or awarded.
For a criminal case, some kind of willful negligence would have to be shown. That can pierce corporate veils. But as a result Waymo is being extremely careful to follow the law and establish processes which shield their employees from negligence claims.
Waymo is going to make sure they are never criminally liable for anything, and even if they were, a criminal case against a corporation just ends up being a modest fine.
This is sad but unfortunately probably happens more frequently with human drivers and people walking out into traffic and you never hear about it.
Straw man argument.
It's interesting how polarized this comments section is. Lots of people claiming a human driver would definitely have been driving slower. Lots of people claiming statistics show that human drivers do worse in this scenario aggregate. Of course neither side presenting convincing evidence.
I tried to find some data: In a 20 mph school zone, ~70-80% of drivers drove 26 mph or faster, regardless of whether there were flashing lights, an always-20 zone, or a 20-when-schoolchildren-present zone. [1]
Although note that ~70-80% of drivers drove 6 or more mph over the speed limit in a school zone, while it seems like the claims in some of these comments are that a human driver would drive at less than 17 mph out of an abundance of caution.
[1] https://wtsc.wa.gov/wp-content/uploads/dlm_uploads/2023/12/1...
And a truly disappointing number of people just accepting company PR as a complete account.
Waymo failed to stop and hit a child. Normal person would drive carefully around blind spots. I wonder what would comments be if Tesla hit a child.
> Normal person would drive carefully around blind spots.
I can't tell if you are using sarcasm here or are serious. I guess it depends on your definition of normal person (obviously not average, but an idealized driver maybe?).
From the Waymo blog...
> The Waymo Driver braked hard...
By Waymo Driver, they don't mean a human, do they?
No, this is the term they use to refer to their package of sensors, compute, and software.
> Waymo said its robotaxi struck the child at six miles per hour, after braking “hard” from around 17 miles per hour. The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post. Waymo said its vehicle “immediately detected the individual as soon as they began to emerge from behind the stopped vehicle.”
As this is based on detection of the child, what happens on Halloween when kids are all over the place and do not necessarily look like kids?
These systems don't discriminate on whether the object is a child. If an object enters the path of the vehicle, the lidar should spot it immediately and the car should brake.
It is more complicated than that. Deepends on size of object and many other factors.
The object could be a paper bag flying in the wind, or leaves falling from the tree.
You're right: a quick search shows that pedestrian fatalities are 43% higher on Halloween.
That's probably more a function of more people being in the road than people not understanding what object they're about to hit.
Sorry, I was being oblique. Humans kill other humans with cars every day. They kill even more on Halloween. Let's start addressing that problem before worrying whether Waymos might someday decide it's OK to drive through ghosts.
Autonomous vehicles won't be perfect. They'll surely make different mistakes from the ones humans currently make. People will die who wouldn't have died at the hands of human drivers. But the overall number of mistakes will be smaller.
Suppose you could wave your magic wand and have a The Purge-style situation where AVs had a perfect safety record 364 days of the year, but for some reason had a tricky bug that caused them to run over tiny Spidermen and princesses on Halloween. The number of fatalities in the US would drop from 40,000 annually to 40. Would you wave that wand?
"The number of fatalities in the US would drop from 40,000 annually to 40. Would you wave that wand?"
This strawmam is bordering offtopic fiction. Only the 40k yearly deaths is based on reality. https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in...
Lidar would pick up a moving object in 3D so unlikely to just keep going.
"Oh that obstructing object doesn't look like a child? Gun it, YOLO." Lmao.
I suspect the cars are trying to avoid running into anything, as that's generally considered bad.
If what Waymo wrote is true, this sounds more like kids fault or guardian’s.
related: https://www.cbsnews.com/news/ntsb-investigation-waymo-robota...
Will post it here:
> In October 2025, a Waymo autonomous robotaxi struck and killed KitKat, a well-known bodega cat at Randa's Market in San Francisco's Mission District, sparking debates over self-driving car safety
It's a child now. All I wanna ask - what should happen, so they stop killing pets and people?
The real but contentious answer is to change our street and urban design. You can only do so much to make a giant metal machine safe for children and small animals to be struck by. Reducing the frequency of cars and pedestrians occupying the same space will go further than trying to engineer the equivalent of a pool that is impossible to drown in.
Do you think that a company that operates autonomous vehicles will support legislation that makes it easier and safer to move around on foot without getting hit by a car? Or will they lobby for car-centric urban design, like many many companies before them?
Absolutely. Because the next step is to ban human driven cars from those areas, and in that case, who makes boat loads of money?
I know submissions are not meant to contain modifications to article titles, but would it be so bad to have added "at 6mph" and/or "minor injuries" to the title?
I don't disagree with you but unfortunately I needed to keep from editorializing and I was restricted by a strict title length limit.
Seems the robotaxi saved a life.
hmm idk how i feel about taking one in the freeway anymore.
I'm a big fan of Waymo and have enjoyed my Waymo rides. And I don't think Waymno did anything "bad" here.
> The young pedestrian “suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle’s path,” the company said in its blog post. Waymo said its vehicle “immediately detected the individual as soon as they began to emerge from behind the stopped vehicle.”
BUT! As a human driver, I avoid driving near the schools when school's letting out. There's a high school on my way home and kids saunter and jaywalk across the street, and they're all 'too cool' to press the button that turns on the blinking crosswalk. So I go a block out of my way to bypass the whole school area when I'm heading home that way.
Waymos should use the same rationale. If you can avoid going past a school zone when kids are likely to be there, do it!
> Waymos should use the same rationale. If you can avoid going past a school zone when kids are likely to be there, do it!
I can see that, prioritize obstacle predictability over transit time. A school zone at certain times of day is very unpredictable with respect to obstacles but a more car congested area would be easier to navigate but slower. Same goes for residential areas during Halloween.
Waymo will 100% go down a route human drivers avoid because it will have "less traffic".
Laws broken: 0
Nothing to see here.
Amazing response to this situation.
Great job, Waymo, for maybe hitting a little kid less than your study assumes a human would have! Is that study legit? Who cares, we trust you!
If this had been Tesla, HN would have crashed from all the dunking.
OK Its like this!, If I hit a child in a school district, I loose my licence for many years, and if I dont or cant show remourse, it could be longer, I pay fines, etc Therefor waymo, must have it's algorythm terminated, ie: totaly destroyed, all the hardware smashed, and they never get to try again with any derivitive of this technology, as there is no reasonable, understandable path towards repentence and rehabilitation, it is litteraly a monster running over children. or was it carrying an ICE team, then nevermind.
>To put this in perspective, our peer-reviewed model shows that a fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph. This significant reduction in impact speed and severity is a demonstration of the material safety benefit of the Waymo Driver.
Our car hits better is a win, I guess?
Glad the child is okay.
When is enough, enough? Software devs working on autonomous driving: look in your soul and update your resume.
The statistically relevant question is: How many human drivers have hit children near elementary schools, since Waymo's last accident?
If Waymo has fewer accidents where a pedestrian is hit than humans do, Waymo is safer. Period.
A lot of people are conjecturing how safe a human is in certain complicated scenarios (pedestrian emerging from behind a bus, driver holds cup of coffee, the sun is in their eyes, blah blah blah). These scenarios are distractions from the actual facts.
Is Waymo statistically safer? (spoiler: yes)
This is wrong, although something quite like it is right.
Imagine that there are only 10 Waymo journeys per year, and every year one of them hits a child near an elementary school, while there are 1000000 non-Waymo journeys per year, and every year two of them hit children near elementary schools. In this scenario Waymo has half as many accidents but is clearly much more dangerous.
Here in the real world, obviously the figures aren't anywhere near so extreme, but it's still the case that the great majority of cars on the road are not Waymos, so after counting how many human drivers have had similar accidents you need to scale that figure in proportion to the ratio of human to Waymo car-miles.
(Also, you need to consider the severity of the accidents. That comparison probably favours Waymo; at any rate, they're arguing that it does in this case, that a human driver in the same situation would have hit the child at a much higher and hence more damaging speed.)
Please read this article: https://www.bloomberg.com/news/features/2026-01-06/are-auton...
Spoiler: we definitely don't know yet whether Waymo is statistically safer
So confident yet so wrong.
Who is liable when FSD is used? In Waymo's case, they own and operate the vehicle so obviously they are fully liable.
But in a human driver with FSD on, are they liable if FSD fails? My understanding is yes, they are. Tesla doesn't want that liability. And to me this helps explain why FSD adoption is difficult. I don't want to hand control over to a probabilistic system that might fail but I would be at fault. In other words, I trust my own driving more than the FSD (I could be right or wrong, but I think most people will feel the same way).
I believe Mercedes is the only consumer car manufacturer that is advertising an SAE Level 3 system. My understanding is that L3 is where the manufacturer says you can take your attention off the road while the system is active, so they're assuming liability.
https://www.mbusa.com/en/owners/manuals/drive-pilot
A human driver would most likely have killed this child. That's what should be on the ledger.
That's pretty hyperbolic. At less than 20 mph, car vs pedestrial is unlikely to result in death. IIHS says [1] in an article about other things:
> As far as fatalities were concerned, pedestrians struck at 20 mph had only a 1% chance of dying from their injuries
Certainly, being struck at 6 mph rather than 17 mph is likely to result in a much better outcome for the pedestrian. And that should not be minimized; although it is valuable to consider the situation (when we have sufficient information) and validate Waymo's suggestion that the average human driver would also have struck the pedestrian and at greater speed. That may or may not be accurate, given the context of a busy school dropoff situation... many human drivers are extra cautious in that context and may not have reached that speed; depending on the end to end route, some human drivers would have avoided the street with the school all together based on the time, etc. It's certainly seems like a good result for the premise, child unexpectedly appears from between large parked vehicles, but maybe there should have been an expectation.
[1] https://www.iihs.org/news/detail/vehicle-height-compounds-da...
There's a 50/50 chance that a distracted driver wouldn't have slowed at all and run the child over.
> To estimate injury risk at different impact speeds, IIHS researchers examined 202 crashes involving pedestrians ages 16 or older
A child is probably more likely to die in a collision of the same speed as an adult.
How many human drivers do under 20mph, like ever?
Plenty. Have you ever driven on a freeway at rush hour? Have you driven in a pickup/dropoff line at a school or an airport? You may or may not want to go 100, but when there's a vehicle in front of you going 20mph or less, you're kind of stuck.
Yes, I have seen how people struggle to drive at low speed and can only do start/stop, accelerating to a way too high speed only to slam on brakes, repeating endlessly.
But really, did you seriously read my post as meaning people literally can't go slower than 20 so just plough into whatever is in the way? I'm obviously talking about an open road situation. Hardly any human drivers go under 20 by choice.
What does an open road situation have to do with the incident situation?
For me, the policy question I want answered is if this was a human driver we would have a clear person to sue for liability and damages. For a computer, who is ultimately responsible in a situation where suing for compensation happens? Is it the company? An officer in the company? This creates a situation where a company can afford to bury litigants in costs to even sue, whereas a private driver would lean on their insurance.
So you're worried that instead of facing off against an insurance agency, the plantiff would be facing off against a private company? Doesn't seem like a huge difference to me
Is there actually any difference? I'd have though that the self-driving car would need to be insured to be allowed on the road, so in both cases you're going up against the insurance company rather than the actual owner.
Personally I'm a lot more interested in kids not dying than in making income for injury lawyers. But that's just me.
Your comment implies that they are less interested in kids not dying. Nowhere do they say that.
I'm not interested in the policy question.
Then don't reply??
That still doesn't excuse trying to make them look bad.
It was a reply to my comment.
Waymo hits you -> you seek relief from Waymo's insurance company. Waymo's insurance premium go up. Waymo can weather a LOT of that. Business is still good. Thus, poor financial feedback loop. No real skin in the game.
John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
NOW ... add criminal fault due to driving decision or state of vehicle ... John goes to jail. Waymo? Still making money in the large. I'd like to see more skin in their game.
> John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
John probably (at least where I live) does not have insurance, maybe I could sue him, but he has no assets to speak of (especially if he is living out of his car), so I'm just going to pay a bunch of legal fees for nothing. He doesn't car, because he has no skin in the game. The state doesn't care, they aren't going to throw him in jail or even take away his license (if he has one), they aren't going to even impound his car.
Honestly, I'd much rather be hit by a Waymo than John.
> John probably (at least where I live) does not have insurance, maybe I could sue him, but he has no assets to speak of
https://en.wikipedia.org/wiki/Judgment_proof
I see. Thank you for sharing. Insurance here is mandatory here for all motorists.
If you are hit by an underinsured driver, the government steps in and additional underinsured motorist protection (e.g. hit by an out of province/country motorist) is available to all and not expensive.
Jail time for an at-fault driver here is very uncommon but can be applied if serious injury or death results from a driver's conduct. This is quite conceivable with humans or AI, IMO. Who will face jail time as a human driver would in the same scenario?
Hit and run, leaving the scene, is also a criminal offence with potential jail time that a human motorist faces. You would hope this is unlikely with AI, but if it happens a small percentage of the time, who at Waymo faces jail as a human driver would?
I'm talking about edge cases here, not the usual fender bender. But this thread was about policy/regs and that needs to consider crazy edge cases before there are tens of millions of AI drivers on the road.
Insurance here is also mandatory for all motorists. Doesn't matter if the rules aren't actually enforced.
Waymo has deep pockets, so everyone is going to try and sue them, even if they don't have a legitimate grievance. Where I live, the city/state would totally milk each incident from a BigCo for all it was worth. "Hit and run" by a drunk waymo? The state is just salivating thinking about the possibility.
I don't agree with you that BigCorp doesn't have any skin in the game. They are basically playing the game in a bikini.
> Insurance here is mandatory here for all motorists.
You do know that insurance being mandatory doesn't stop people from driving without insurance, right?
> If you are hit by an underinsured driver, the government steps in and additional underinsured motorist protection (e.g. hit by an out of province/country motorist) is available to all and not expensive.
Jolly good for you.
If I don't carry underinsured coverage, and someone totals my car or injures me with theirs, I'm basically fucked.
Even if you have uninsured coverage, you might still be dead.
>John Smith hits you -> you seek relief from John's insurance company. John's insurance premium goes up. He can't afford that. Thus, effective financial feedback loop. Real skin in the game.
Ah great, so there's a lower chance of that specific John Smith hitting me again in the future!
Yes, that is the specific deterrence effect.
The general deterrence effect we observe in society is that punishment of one person has an effect on others who observe it, making them more cautious and less likely to offend.
No, "the ledger" should record actual facts, and not whatever fictional alternatives we imagine.
Fact: This child's life was saved by the car being driven by a computer program instead of a human.
No, the fact is that the child sustained minor injuries. And, fact: no human driver made the decision to drive a vehicle in that exact position and velocity. Imagining a human-driven vehicle in the same place is certainly valid, but your imagination is not fact. I imagine that the kid would be better off if no vehicle was there. But that's not a fact, that's an interpretation -- perhaps the kid would have ended up dead under an entirely different tire if they hadn't been hit by the waymo!
This is a good point. Cars should be banned near schools. That is a logical conclusion given the facts.
Considering school buses kill children every year, do we ban them near schools as well?
That is not a logical conclusion given the facts.
Instead of a human who was driving exactly the same as the Waymo up until the instant the child ran out. Important distinction.
Would have. Could Have. Should have.
Most humans would be halfway into other lane after seeing kids near the street.
Apologist see something different than me.
Perception.
Disagree, most human drivers would notice they are near an elementary school with kids coming/going, crossing guard present, and been driving very carefully near blocked sight lines.
Better reporting would have asked real people the name of the elementary school, so we could see some pictures of the area. The link to NHTSA didn't point to the investigation, but it's under https://www.nhtsa.gov/search-safety-issues
"NHTSA is aware that the incident occurred within two blocks of a Santa Monica, CA elementary school during normal school drop off hours; that there were other children, a crossing guard, and several double-parked vehicles in the vicinity; and that the child ran across the street from behind a double parked SUV towards the school and was struck by the Waymo AV. Waymo reported that the child sustained minor injuries."
We're getting into hypotheticals but i will say in general i much much prefer being around Waymos/Zooxs/etc. than humans when riding a bicycle.
We're impatient emotional creatures. Sometimes when I'm on a bike the bike lane merges onto the road for a stretch, no choice but to take up a lane. I've had people accelerate behind me and screech the tyres, stopping just short of my back wheel in a threatening manner which they then did repeatedly as i ride the short distance in the lane before the bike lane re-opens.
To say "human drivers would notice they are near an elementary school" completely disregards the fuckwits that are out there on the road today. It disregards human nature. We've all seen people do shit like i describe above. It also disregards that every time i see an automated taxi it seems to drive on the cautious side already.
Give me the unemotional, infinite patience, drives very much on the cautious side automatic taxi over humans any day.
Couldn’t be anymore callous and clinical. This press release alone makes me want not to use their service.
* “Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene,” Waymo wrote in the post.*
The 'next' comment after yours is "Alternate headline: Waymo saves child's life" and the 'prev' comment is "A human driver would most likely have killed this child. That's what should be on the ledger." - would either of those be less 'callous and clinical'?
Other accident reports I've seen (NTSB, etc) often seem to take a similar approach - is it a bad thing?
Or what kind of language wouldn't make you 'want not to use their service'?
Alternate headline: Waymo saves child's life
In this timeline, we want our headlines to somehow reflect the contents of the story.
Saved child from what? From themselves. You can't take full credit for partially solving a problem that you, yourself, created.
can we just get waymo tech in busses?
Big vehicles that demand respect and aren't expected to turn on a dime, known stops.
Dog. Bites. Man.
Q: Why did the self-driving car cross the road?
A: It thought it saw a child on the other side.
That's Tesla. Waymo seems mostly ok.
Waymo is a subsidiary of Alphabet Inc. the same parent company as Google LLC
It was formerly known as the Google Self Driving Car Project
Wow this is why I feel comfortable in a Waymo. Accidents are inevitable and some point and this handling was well-rehearsed and highly ethical. Amazing company
I’m actually pretty surprised Waymo as a general rule doesn’t completely avoid driving in school zones unless absolutely unavoidable.
Any accident is bad. But accidents involving children are especially bad.
That would be one hell of a convoluted route to avoid school zones. I wonder if it would even be possible for a large majority of routes, especially in residential areas.
It might not be possible for a lot of places — I don’t really know.
But I know when I drive, if it’s a route I’m familiar with, I’ll personally avoid school zones for this very reason: higher risk of catastrophe. But also it’s annoying to have to slow down so much.
Maybe this personal decision doesn’t really scale to all situations, but I’m surprised Waymo doesn’t attempt this. (Maybe they do and in this specific scenario it just wasn’t feasible)
Most people prefer the shortest ride. Circling around school zones would be the opposite of that. Rides are charged based on distance, so maybe this would interest Waymo, but one of the big complaints about taxi drivers was how drivers would "take them for a ride" to increase the fare.
Seems like a solvable problem: make it clear on the app/interior car screens that a school zone is being avoided — I think most riders will understand this.
You also have to drive much more slowly in a school zone than you do on other routes, so depending on the detour, it may not even be that much longer of a drive.
At worst, maybe Waymo eats the cost difference involved in choosing a more expensive route. This certainly hits the bottom line, but there’s certainly also a business and reputational cost from “child hit by Waymo in school zone” in the headlines.
Again, this all seems very solvable.
Well, I'm a human and I figure out how to avoid school zones.
And before the argument "Self driving is acceptable so long as the accident/risk is lower than with human drivers" can I please get that out of the way: No it's not. Self driving needs to be orders of magnitude safer for us to acknowledge it. If they're merely as safe or slightly safer than humans we will never accept it. Becase humans have a "skin in the game". If you drive drunk, at least you're likely to be in the accident, or have personal liability. We accept the risks with humans because those humans accept risk. Self driving abstracts the legal risk, and removes the physical risk.
I'm willing to accept robotaxis, and accidents in robotaxis, but there needs to be some solid figures showing they are way _way_ safer than human drivers.
I think those figures are already starting to accumulate. Incidents like this are rare enough that they are news worthy. Almost every minor incident involving Waymo, Tesla's FSD, and similar solutions gets a lot of press. This was a major incident with a happy end. Those are quite rare. The lethal ones even rarer.
As for more data, there is a chicken egg problem. A phased roll out of waymo over several years has revealed many potential issues but is also remarkable in the low number of incidents with fatalities. The benefit of a gradual approach is that it builds confidence over time.
Tesla has some ways to go here. Though arguably, with many hundreds of thousands of paying users, if it was really unsafe, there would be some numbers on that. Normal statistics in the US are measured in ~17 deaths per 100K drivers per year. 40K+ fatalities overall. FSD for all its faults and failings isn't killing dozens of people per years. Nor is Waymo. It's a bit of an apples and oranges comparison of course. But the bar for safety is pretty low as soon as you include human drivers.
Liability weighs higher for companies than safety. It's fine to them if people die, as long as they aren't liable. That's why the status quo is tolerated. Normalized for amounts of miles driven with and without autonomous, there's very little doubt that autonomous driving is already much safer. We can get more data at the price of more deaths by simply dragging out the testing phase.
Perfect is the enemy of good here. We can wait another few years (times ~40K deaths) or maybe allow technology to start lowering the amount of traffic deaths. Every year we wait means more deaths. Waiting here literally costs lives.
> ~17 deaths per 100K drivers per year. 40K+ fatalities overall.
I also think one needs to remember those are _abysmal_ numbers, so while the current discourse is US centric (because that's where the companies and their testing is) I don't think it can be representative for the risks of driving in general. Naturally, robotaxis will benefit from better infra outside the US (e.g. better separation of pedestrians) but it'll also have to clear a higher safety bar e.g. of fewer drunk drivers.
Also fun to calculate how this compounds over say 40 years. You get to about 1 in 150 drivers being involved in some kind of deathly accident. People are really bad at numbers and assessing risk.
It will also never get worse. This is the worst the algorithms from this point forward.
I am not sure. Self-driving is complex and involves the behavior of other, non-automated actors. This is not like a compression algorithm where things are easily testable and verifiable. If Waymos start behaving extra-oddly in school zones, it may lead to other accidents where drivers attempt to go around the "broken" Waymo and crash into it, other pedestrians, or other vehicles.
I know Tesla FSD is its own thing, but crowdsourced results show that FSD updates often increase the amount of disengagements (errors):
https://electrek.co/2025/03/23/tesla-full-self-driving-stagn...
And we haven't reached the point where people start walking straight into the paths of cars, either obliviously or defiantly. https://www.youtube.com/shorts/nVEDebSuEUs
There are already anecdotes of people aggressively jaywalking in front of a Waymo because they know it will stop, and people driving more aggressively around Waymos because it will always defer to them.
Has this been true of other Google products? They never get worse?
> I'm willing to accept robotaxis, and accidents in robotaxis, but there needs to be some solid figures showing they are way _way_ safer than human drivers.
Do you mean like this?
https://waymo.com/safety/impact/
Yes but ideally from some objective source.
Like this? https://waymo.com/blog/2024/12/new-swiss-re-study-waymo
Maybe an objective source that isn't on the waymo.com domain?
"We find that when benchmarked against zip code-calibrated human baselines, the Waymo Driver significantly improves safety towards other road users."
https://pmc.ncbi.nlm.nih.gov/articles/PMC11305169/
If waymo is to be believed, they hit the kid at 6mph and estimated that a human driver at full attention would have hit the kid at 14 mph. The waymo was traveling 17mph. The situation of "kid running out between cars" will likley never be solved either, because even with sub nanosecond reaction time, the car's mass and tire's traction physically caps how fast a change in velocity can happen.
I don't think we will ever see the video, as any contact is overall viewed negatively by the general public, but for non-hyperbolic types it would probably be pretty impressive.
That doesn't mean it can't be solved. Don't drive faster than you can see. If you're driving 6 feet from a parked car, you can go slow enough to stop assuming a worst case of a sprinter waiting to leap out at every moment.
If we adopted that level of risk, we'd have 5mph speed limits on every street with parking. As a society, we've decided that's overly cautious.
But with waymos it would be possible. Mark those streets as "extremely slow" and never go there unless you are dropping someone off. (The computer has more patience than human drivers.)
If that's too annoying then bad parking by school areas so the situation doesn't happen.
I don't know if you've been to some cities or neighborhoods but almost every street has on-street parking in many of them.
And why would you make Waymo's go slower than human drivers, when it's the human drivers with worse reaction times? I had interpreted the suggestion as applying to all drivers.
Oh I have no problem believing that this particular situation would have been handled better by a human. I just want hard figures saying that (say) this happens 100x more rarely with robotaxis than human drivers.
> The situation of "kid running out between cars" will likley never be solved
Nuanced disagree (i agree with your physics), in that an element of the issue is design. Kids running out between cars _on streets that stack building --> yard --> sidewalk --> parked cars --> driving cars.
One simple change could be adding a chain link fence / boundary between parked cars and driving cars, increasing the visibility and time.
How do you add a chain link fence between the parked and driving cars for on-street parking?
there's still an inlet and outlet (kinda like hotel pickup/drop off loops). It's not absolutely perfect, but it constrains the space of where kids can dart from every parked car to 2 places.
Also the point isn't the specifics, the point is that the current design is not optimal, it's just the incumbent.
Ok, that's not really a simple change anymore, because you need more space for that. Unless it's really just a drop off queue, but then it's not parked cars, since a parked car blocks the queue.
We would really need to see the site to have an idea of the constraints, Santa Monica has some places where additional roadway can be accomodated and some places where that's not really an option.
Second-order benefit: More Waymos = fewer parked cars
In high parking contention areas, I think there's enough latent demand for parking that you wouldn't observe fewer parked cars until reduce demand by a much greater amount.
>We accept the risks with humans because those humans accept risk.
It seems very strange to defend a system that is drastically less safe because when an accident happens, at least a human will be "liable". Does a human suffering consequences (paying a fine? losing their license? going to jail?) make an injury/death more acceptable, if it wouldn't have happened with a Waymo driver in the first place?
I think a very good reason to want to know who's liable is because Google has not exactly shown itself to enthusiastically accept responsibility for harm it causes, and there is no guarantee Waymo will continue to be safe in the future.
In fact, I could see Google working on a highly complex algorithm to figure out cost savings from reducing safety and balancing that against the cost of spending more on marketing and lobbyists. We will have zero leverage to do anything if Waymo gradually becomes more and more dangerous.
> Wherever I'm going, I'll be there to apply the formula. I'll keep the secret intact. It's simple arithmetic. It's a story problem. If a new car built by my company leaves Chicago traveling west at 60 miles per hour, and the rear differential locks up, and the car crashes and burns with everyone trapped inside, does my company initiate a recall?
> You take the population of vehicles in the field (A) and multiple it by the probable rate of failure (B), then multiply the result by the average cost of an out-of-court settlement (C). A times B times C equals X. This is what it will cost if we don't initiate a recall. If X is greater than the cost of a recall, we recall the cars and no one gets hurt. If X is less than the cost of a recall, then we don't recall.
-Chuck Palahniuk, Fight Club
Even in terms of plain results, I'd say the consequences-based system isn't working so well if it's producing 40,000 US deaths annually.
That’s the fault of poor infrastructure and laws more than anything else. AV’s must drive in the same infrastructure (and can somewhat compensate).
Yes
Orders of magnitude? Something like 100 people die on the road in the US each day. If self-driving tech could save 10 lives per day, that’s wouldn’t be good enough?
"It depends". If 50 people die and 50 people go to jail, vs. 40 people die and their families are left wondering if someone will take responsibility? Then that's not immediately standing out as an improvement just because fewer died. We can do better I think. The problem is simply one of responsibility.
If the current situation was every day 40 people die but blame is rarely assigned, would you recommend a change where an additional 10 people are going to die but someone will be held responsible for those deaths?
Yes
People don't usually go to jail. Unless the driver is drunk or there's some other level of provable criminal negligence (or someone actively trying to kill people by e.g. driving into a crowd of protesters they disagree with), it's just chalked up as an accident.
Apart from a minority of car related deaths resulting in jail time, what kind of person wants many more people to die just so they can point at someone to blame for it? At what point are such people the ones to blame for so many deaths themselves?
In such situations it’s useful to put yourself in a hypothetical situation. Rules: you can’t pick who you will be: one of the dead or alive. It will be assigned randomly.
So would you pick situation 1 or 2?
I would personally pick 1.
Do they go to jail?
That is not my experience here in the Bay Area. In fact here is a pretty typical recent example https://www.nbcbayarea.com/news/local/community-members-mour...
The driver cuts in front of one person on an e-bike so fast they can’t react and hit them. Then after being hit they step on the accelerator and go over the sidewalk on the other side of the road killing a 4 year old. No charges filed.
This driver will be back on the street right away.
Ugh. That is so despicable both of the driver and as a society that we accept this. Ubiquitous Waymo can't come soon enough.
Have you been in a self driving car? There are some quite annoying hiccups, but they are already very safe. I would say safer than the average driver. Defensive driving is the norm. I can think of many times where the car has avoided other dangerous drivers or oblivious pedestrians before I realized why it was taking action.
I generally agree the bar is high.
But, human drivers often face very little accountability. Even drunk and reckless drivers are often let off with a slap on the wrist. Even killing someone results in minimal consequences.
There is a very strong bias here. Everyone has to drive (in most of America), and people tend to see themselves in the driver. Revoking a license often means someone can’t get to work.
That’s an incentive to reduce risk, but if you empirically show that the AV is even 10x safer, why wouldn’t you chalk that up as a win?
> Self driving needs to be orders of magnitude safer for us to acknowledge it. If they're merely as safe or slightly safer than humans we will never accept it
It’s already accepted. It’s already here. And Waymo is the safest in the set—we’re accepting objectively less-safe systems, too.
> Self driving needs to be orders of magnitude safer for us to acknowledge it
All data indicates that Waymo is ~10x safer so far.
"90% Fewer serious injury or worse crashes"
https://waymo.com/safety/impact/
> The vehicle remained stopped, moved to the side of the road
How do you remain stopped but also move to the side of the road? Thats a contradiction. Just like Cruise.
My reading of that is that they mean stopped the progression of the journey rather that made no movement whatsoever.
I agree, it’s poorly worded but I think that’s what they mean.
I also assume a human took over (called the police, moved the car, etc) once it hit the kid.
They mean the vehicle didn't drive away. It moved to the side of the road and then stopped and waited.
So many tech lovers defending waymo.
If you drive a car, you have a responsibility to do it safely. The fact that I am usually better than the bottom 50% of drivers, or that I am better than a drunk driver does not mean that when I hit someone it's less bad. A car is a giant weapon. If you drive the weapon, you need to do it safely. Most people these days are incredibly inconsiderate - probably because there's little economic value in being considerate. The fact that lots of drivers suck doesn't mean that waymo gets a pass.
Waymos have definitely become more aggressive as they've been successful. They drive the speed limit down my local street. I see them and I think wtf that's too fast. It's one thing when there are no cars around. But if you've got cars or people around, the appropriate speed changes. Let's audit waymo. They certainly have an aggressiveness setting. Let's see the data on how it's changing. Let's see how safety buffers have decreased as they've changed the aggressiveness setting.
The real solution? Get rid of cars. Self-driving individually owned vehicles were always the wrong solution. Public transit and shared infra is always the right choice.
> The fact that lots of drivers suck doesn't mean that waymo gets a pass.
But that fact does mean that we should encourage alternatives that reduce fatalities, and that not doing so results in fatalities that did not need to occur.
> The real solution? Get rid of cars.
I also support initiatives to improve public transit, etc. However, I don't think "get rid of cars" is a realistic idea to the general public right now, so let's encourage all of the things that improve things - robot drivers if they kill people less often than humans, public transit, etc. - let's not put off changes that will save lives on the hope that humanity will "get rid of cars" any time soon. Or when do you think humanity will "get rid of cars"?