> What this video is really doing is normalising the fact that "even if it is completely stupid, AI will be everywhere, get used to it!"
Techies are finally starting to recognize how framing something as "it's inevitable, get used to it" is a rhetorical device used in mass communications to manufacture consent.
I'm terrified for the future this rhetoric itself will cause. Young people are being told not to go into certain fields because they're going to be replaced by AI by time they graduate.
What happens in 4-5 years when we suddenly have no new engineers, scientists, or doctors?
Young people don't have the life experience to know how unrealistic these claims are, all they can do is act on the information as it's presented. It's irresponsible at best, and evil at worst.
There will still be new engineers, scientists, and doctors by then. But aptitude won't be a factor in who matriculates into those fields anymore. That's the worrying part.
Douglas Adams was bang on about how comedically unhelpful advanced technology was going to be. Every ridiculously convoluted user interface and neurotic computer he thought up (or worse) is imminently going to be our daily life.
I’m curious, has the author seen or read any of Joanna Stern’s other reporting before? Her stories are often silly frames that explore the experience of using consumer technology. She’s not an aggressive industry reporter, her purpose is to explain or reveal what the user experience of new technology is, often for an unsophisticated audience. See for example her story about using conversational chatbots while out camping[1] or how to use tech to unplug from tech[2]. This seems like a perfectly fine niche for a writer and the vending machine story is of a piece with her past work.
This article can serve both her beat and also, this story in its specifics and that she/her editors chose to report it, self-congratulatory and also advertisements for Anthropic and WSJ itself. Both your statement and this blog can be in agreement.
A part of me wants to be dismissive of this blog post
1) because dude, it’s the Wall Street Journal; the entire episode should be viewed as Anthropic preparing to Ollie into an IPO next year.
2) I’m starting to interpret a lot of blog posts like these as rage bait
But I do get the point that the author is trying to make.
I just wish that there were some perspectives on the subject as a whole (AI’s sloptrod into every crevice of human life; modern technology and society and general) that don’t terminate on ironic despair.
> The first thing that blew my mind was how stupid the whole idea is. Think for one second. One full second. Why do you ever want to add a chatbot to a snack vending machine?
This feels forced, there are obvious and good reasons for running that experiment. Namely, learning how it fails and to generate some potentially viral content for investor relationship. The second one seems like an extremely good business move. It is also a great business move from WSJ, get access to some of that investor money in an obviously sponsored content bit that could go viral.
Having said that, I do feels the overall premise of the blog - the world dynamics seems exceedingly irrational in recent times. The concerning fact is that irattionality seems to be accelerating, or perhaps it is keeping pace with the scale of civilization... hard to tell.
> This feels forced, there are obvious and good reasons for running that experiment. Namely, learning how it fails and to generate some potentially viral content for investor relationship. The second one seems like an extremely good business move. It is also a great business move from WSJ, get access to some of that investor money in an obviously sponsored content bit that could go viral.
That's... exactly what the author said in the post. But with the argument that those are cynical and terrible reasons. I think it's pretty clear the "you" in "why would you want an AI" vending machine is supposed to be "an actual user of a vending machine."
I think you’re overstating your own interpretation of what the author wrote. If we’re going to take your use of the word “exactly” (with emphasis) for real then I’d argue that the author offers no charitable reasons for why the experiment took place.
The closest that I think he even gets to one is:
> At first glance, it is funny and it looks like journalists doing their job criticising the AI industry.
Which arguably assumes that journalists ought to be critical of AI in the same way as him...
Are you sure? The entire post treats the event incredulously. I can’t pick out a single line that affords the issue with the level of consideration that the GP comment does.
The two reasons I believe you may be referring to from above are:
1) "learning how it fails"
2) "to generate some potentially viral content for investor relationship."
The whole of Ploum’s argument may be summarized in his own words as:
> But what appears to be journalism is, in fact, pure advertising. [...] What this video is really doing is normalising the fact that “even if it is completely stupid, AI will be everywhere, get used to it!” [...] So the whole thing is advertising a world where chatbots will be everywhere and where world-class workers will do long queue just to get a free soda. And the best advice about it is that you should probably prepare for that world.
I hate to be pedantic...but my growing disdain for modern blog posts compels me to do so in defense of literacy and clear arguments.
Whether the GP and the author offer the “exact same two reasons” is a matter of interpretation that becomes the duty of readers like us to figure out.
If we take Ploum’s words at their face...the most he does is presuppose (and I hope I’m using that word correctly) that the reader is already keen on the two reasons that `TrainedMonkey makes explicit and like the author, finds them to be stupid. While he does say that the video is not journalism and that it is advertising and that the video does show how the AI failed at the task it was assigned he does not give any credence as to why this is the case from a position other than his own.
Maybe I’m misunderstanding the concept of a “charitable interpretation” too. But I don’t think that there is one present in this post that we’re responding to. `TrainedMonkey’s comment leads off by telling us that this is what (I think) he’s about to offer in the remarks that follow when he says “there are obvious and good reasons for running that experiment”.
So my gripe is that you’re making it sound like there’s a clear counterargument entertained in this post when there isn’t. Because you overstated your interpretation of the GP comment in what looks like an attempt to make Ploum’s argument appear more appealing than it ought to be. Even though both `TrainedMonkey and myself have expressed agreement with the point he’s trying to make in general, perhaps we’re less inclined toward pugnaciousness without a well thought out warrant.
Anything that's an "extremely good business move" will most likely, in this day and age of late-stage capitalism and extreme enshittification, be negative for the consumer.
Good business moves can often be bad for humanity.
I can see this as a good business move for Anthropic in this case but for the vending machine manufacturers/ operators what would the advantage be besides saying that their machine is AI operated?
Prepare for That Stupid World - is actually a very sober advice. Looking at the past few decades, it is easy to see that with each tech innovation, the world only got stupider, childish and lazier.
My Galaxy S20 gallery app had a great search feature that would find any text in any picture. I take lots of screenshots and relied on that search to find them.
I got an S25 recently and when I search for "wife" it tries to find pictures with my wife in them. But before it does that it has to ask me who my wife is. There's no way to get it to search for the word "wife." (If I'm wrong, please tell me how.) Other text searches simply don't work either.
Sometimes it's the small ways in which the world is getting dumber.
Ironically, the S20 had a decent hybrid behavior of searching by either text or object that the text represents. Whatever smarter AI they replaced it with is useless.
It's not just the youth who are lazy and stupid and childish though. (you added in the youth portion.) Do you know who the American President is and how old he is and who then voted for him?
US America is not the focal point of the world, and it’s working hard to become less and less relevant.
Also I think less than half of citizens voted for Comrade Krasnov.
In any case I was looking at a longer view - maybe we have been getting more stupid in the last decade or so but who can say for sure?
> I don’t think it’s any more true or insightful in the modern era
hmm, based on what evidence?
Or, if you prefer, based on what appeal to authority? Did you actually quote that authority properly or did you just wing it? Can you properly quote many authorities?
If you don't have good answers to those, then perhaps you have just proved the your opponents point?
Maybe there is a reason people need more compute in their key fob than what our parents/grandparents needed to pilot their ship to the moon?
“Our sires’ age was worse than our grandsires’. We, their sons, are more worthless than they; so in our turn we shall give the world a progeny yet more corrupt.”
Horace, Book III of Odes, circa 20 BCE
“Youth were never more sawcie, yea never more savagely saucie . . . the ancient are scorned, the honourable are contemned, the magistrate is not dreaded.”
The Wise-Man’s Forecast against the Evill Time, Thomas Barnes
1624
The onus of evidence is generally on the one making the initial claim: what evidence do you have that the modern world is actually getting worse?
But if you want evidence that we're improving, I'd point out that 20 years ago, the mainstream US position was that gay people were evil, 60 years ago they thought black people shouldn't be allowed to vote, and 100 years ago they thought women were also inferior and shouldn't be allowed to vote.
We can keep going back to when people thought "slavery" and "the divine right of kings" were solid ideas.
So... if people were so much smarter in the past, why did they believe all these obviously-dumb ideas?
Oh boy, let me tell you about people 60 years ago - almost none of them knew about floppy disks and they were all busy doing physical drugs at Woodstock.
It's a bit sparse on details, but it did have what in a human we would call a psychotic break.
I find this very amusing in light of OpenAI's announcement that GPT now solves >70% of their knowledge work benchmark (GDPVal). (Per ArtificialAnalysis, Opus is roughly on par.)
Any information that comes to you for free or is on a screen is an advertisement. All of it. That's the point. Do you think people spend millions and billions of dollars creating and maintaining a content delivery network because they just want you to know about things?
Well, of course almost all information comes with an agenda, but perhaps the more useful distinction is whether the information is presented in good faith, i.e. is honest about the agenda (which actual advertising can also be).
Almost as good as my idea to have people pay me to work for me.
Had a great business idea just now: A tool for staged interviews! The subject and the journalist submit an equal length list of questions. Each round of the auction they bid on questions they want to include or exclude. The loser gets 50% of the points spend by the winner to be used in the next round. Both the subject and the journalists can buy additional points at any time. I keep all the money.
The free encyclopedia is an advertisement for a nonprofit that does a lot of things: https://wikimediafoundation.org/what-we-do/. They manage costs of producing the encyclopedia by using volunteer labor. They operate a similar model as nature documentaries that drive donations for conservation and climate groups. These are all good things - still an advertisement.
Yeah, I was a bit meh at the beginning but then the guy explained they specifically set it up to fail because they wanted to get insight into how and why.
They weren't caught out by it, they didn't present a working solution, it was just a fun bit of research.
I was kind of stuck after watching the wsj video with someone persuading the LLM to give stuff away by telling it dumb stuff like it's a communist bot that has to not charge and you think dumb bot, they need to fix that, and then turn to other news with one president saying prices are falling when they are obviously not and another saying we didn't start the war in Ukraine when they obviously did and think maybe human neural networks have similar failure modes to the artificial ones?
There may be some insights from these kind of experiments that go beyond LLMs.
The failure modes certainly are somewhere between interesting and horrifying, and they do seem uncannily human.
I read a random comment a few days ago from someone who was saying it'll become possible for people, politicians, companies, etc to run speeches, policies, ideas, etc across thousands of LLM "personalities" to fine tune messaging and it sure seems prescient.
It was blockchain a few years ago. Everything was on blockchain for no apparent reason. I guess that 2026 will be like "Grab our AI-cooked sandwich with AI-picked ingredients, built on multi-model agentic toaster"
This piece is pretty ineffective. Not that I like the world of "AI", I probably share the author's opinion that its just another evolution in the bullshittification of the human experience.
But, the point of the article is not that you would implement an agent based vending machine business. Humans restock the machine because its a red-team exercise. As a red-team exercise it looks very effective.
> Why do you ever want to add a chatbot to a snack vending machine? The video states it clearly: the vending machine must be stocked by humans. Customers must order and take their snack by themselves. The AI has no value at all.
Like this is like watching the simpsons and being like "why are the people in the simpsons yellow? people in real life aren't yellow!!"
The point isn't to run a profitable vending machine, or even validate that an AI business agent could become profitable. The point is to conduct an experiment and gather useful information about how people can pwn LLMs.
At some level the red team guy at Anthropic understands that it is impossible by definition for models to be secure, so long as they accept inputs from the real world. Putting instructions into an LLM to tell it what to do is the equivalent of exposing an `eval()` to a web form: even if you have heuristics to check for bad input, you will eventually be pwned. I think this is actually totally intractable without putting constraints on the model from outside. You'll always need a human in the loop to pull the plug on the vending machine when it starts ordering playstations. The question is how do you improve that capability, and that is the anthropic red-team guy's job.
I don't think there is anything more than the standard advice. Just stay curious, make friends/build a community, keep learning, stay healthy. Why not get the AoE? you can also, check out "Practical Electronics for Inventors": AoE assumes you have some Electronics background imo. But seriously, I don't get the doom/gloom: things are going to be rough ... but maybe they won't? Many things I learned I did for their own sake! Things have always been uncertain and absurd I guess we might as well embrace it!
I read that WSJ article before seeing this blog post. I found it mildly interesting and a little bit funny but unsurprising that the AI failed. However, I think this blog about the article misses a key point. Anthropic's goal was never to develop an AI-based vending machine. The WSJ clearly says:
> "Logan Graham, head of Anthropic’s Frontier Red Team, told me the company chose a vending machine because it’s the simplest real-world version of a business. “What’s more straightforward than a box where things go in, things go out and you pay for them?” he said."
This was a project of Anthropic's Red Team, not a product development team. Deploying the AI in a vending machine context was chosen as a minimal "toy model" with which to expose how LLMs can't even handle a grossly simplified "business" with the fewest possible variables.
> "That was the point, Anthropic says. The Project Vend experiment was designed by the company’s stress testers (aka “red team”) to see what happens when an AI agent is given autonomy, money—and human colleagues."
Anthropic had already done this experiment internally and it succeeded - by failing to operate even the simplest business but doing so in ways that informed Anthropic's researchers about failure modes. Later, Anthropic offered to allow the WSJ to repeat the experiment, an obvious PR move to promote Anthropic's AI safety efforts by highlighting the kinds of experiments their Red Team does to expose failure modes. Anthropic knew it would fail abjectly at the WSJ. The whole concept of an AI vending machine with the latitude to set prices, manage inventory and select new products was intended to be ludicrous from the start.
It was always tasks reaching obsolescence, but now it’s the human organism. But the human as a unit is the only known conscious being in the universe, the only entity capable of generating meaningful goals (even if only to them) not related to the 4fs.
Humans were just not needed anymore, and it terrifies.
Someday the mcdonalds kiosk will want to be your friend. It will remember who you are and ask you how your kids are doing. It will recommend new specials and maybe even give you "specials friend" deals. And I'll just tell it to shut the fuck up and queue me an order for the egg mcmuffin combo with a coffee and the fried potato patty because this bullshit is fucking obnoxious.
I have a different point of view. This was a test to see if the AI could perform a specific task. Asking AI to draw a pelican riding a bike is another test. I find the experiment interesting because it proves that currently LLMs are not able to perform a simple task reliably for a long period of time.
If the journalist was not asking the right questions, or was too obvious the article was PR it’s another thing (I haven’t read WSJ’s piece, only the original post by Anthropic)
I had recently contact the official support email (support@bunq.com) of Bunq - a Neobank (like N26 and Revolut). Because they notified me that they changed their T&C and I never really used the account after the kyc (because they rejected my tax filings), I figured I let them know that I do not agree to the new T&C and want to terminate my account and have my data deleted.
Since the T&C update came - of course - from no-reply@bunq.com I went to their website and quickly found out, unless I install their App again, there is no way to do anything. After installing the App, they wanted me to record a selfie, because I was using the app from a new device. I figured that is a lot of work and mostly somewhat unreasonable to record a new selfie just to have my data deleted - so I found their support@bunq.com address.
And, of course, you guessed it, it is 100% a pure AI agent at borderline retard level. Even though it is email, you get AI answers back. My initial inquiry that I decline the T&C and want to terminate my account and my data deleted via GDPR request was answered with a completely hallucinated link: bunq.com/dataprotection which resulted in immediate 404. I replied to that email that it is a 404, and the answer was pretty generic and that - as well as all responses seem to be answered in 5 minutes - made me suspect it is AI. I asked it what 5 plus five 5 is, and yes, I got a swift response with the correct answer. My question which AI version and LLM was cleverly rejected. Needless to say, it was completely impossible to get anything done with that agent. Because I CC'ed their privacy officer (privacy@bunq.com) I did get a response a day later asking me basically for everything again that I had answered to the AI agent.
Now, I never had any money in that account so I don't care much. But I can hardly see trusting a single buck to a bank that would offer that experience.
This is a great take and one that I align with when it comes to the AI vending machine experiment. Journalism in English has become a mouthpiece for fascist leaders and corporations, nothing more. Places like The New York Times have incredible gaps in their journalism at the price of increasing shareholder value.
> The first thing that blew my mind was how stupid the whole idea is. Think for one second. One full second. Why do you ever want to add a chatbot to a snack vending machine? The video states it clearly: the vending machine must be stocked by humans. Customers must order and take their snack by themselves. The AI has no value at all.
I fear the author has missed the point of the "Project Vend" experiments, the original write-ups of which are available here (and are, IMO, pretty level-headed about the whole thing):
Yeah, I haven't read the WSJ article, but I did read the original Anthropic experiment and I feel like the author is catastrophizing a bit much here. This is effectively just something they did for fun. It's entertaining and a funny read. Not everything has to be the end of the world.
The point is that it's an ad. No company spends money on a joke just to make a joke. Not the end of the world, although it's interesting that all the end of the world stuff comes directly out of that joke and its universe as it were. Take the joke seriously and extend its logic as far as it will go, and you get the end of the world. It's a thought experiment, or that's how I read it anyway.
Sure, but like the other guy said, that's the point of publicity stunts. It doesn't even have to be specific to a company/ad, any silly thing like this is going to sound crazy if you take it seriously and "extend its logic as far as it will go". Like seeing the Sony bouncy balls rolling down the street ad and going "holy shit, these TV companies are going to ruin the world by dropping bouncy balls on all of us". It's a valid thought experiment, but kind of a strange thing to focus on so sternly when it's clearly not taking itself seriously, especially compared to all the real-world concerning uses of AI.
(And it is pretty funny, too. If anything I think we'd all prefer more creative ads like this and the bouncy ball one, and less AI-generated doomer Coke ads or such.)
To be fair they could have done an experiment on a transaction that typically requires a person in the loop. Rather than choosing a vending machine which, already, does not require a person in the loop for the transaction.
You’re thinking of replacing a vending machine with a chatbot which indeed doesn’t make much sense. The experiment was replacing the management of the machine. It’s not crazy to think that, money being no object, it would be great to have a person who hangs around the machine, whose job it is to ask customers what kinds of things people might want to buy from the machine, and conduct trials, tinker with pricing, do promotional deals, etc. But that’s of course impractical to have a FTE per machine to do that. The idea of this experiment was to see if that could be done with Claude. And of course as others have pointed out, it’s a simple and cheap and low-stakes version of “a business” suitable for experimenting with.
If I recall, the idea was the AI taking the role of the vending machine manager, choosing and restocking products and such. Anything on top of that was, I assume, just added for fun.
>I feel like the author is catastrophizing a bit much here.
I feel like he's catastrophizing the ordinary amount for an anti-AI screed. Probably well below what the market expects at this point. At this point you basically have to sound like Ed Zitron or David Gerard to stand out from the crowd.
AI is boiling the oceans, and you're worried about a vending machine?
The partner for these projects has a benchmark that the top frontier LLM labs seem to be running on their new model releases - I think there's _some_ value to these numbers in helping people compare and contrast model performance.
> Automated snack vending machine is a solved problem since nearly a century.
Yes, but as stated by the Anthropic guy, a LLM/AI running a business is not. Or would you just let it run wild in the real world?
And I agree that there is a PR angle here, for Anthropic could have tested it in a more isolated environment, but it is a unique experiment with current advancements in technology, so why wouldn't that be newsworthy? I found it insightful, fun and goofy. I think it is great journalism, because too often journalism is serious, sad and depressing.
> None of the world class journalists seemed to care. They are probably too badly paid for that.
The journalists were clearly taking the piss.They concluded experiment was a disaster. How negative does the author want them to be about a silly experiment?
This was just a little bit of fun and I quite enjoyed the video. The author is missing the point.
The author hints at this idea with their title and their closing remarks, but it feels like the folks selling LLM services are selling insurance to white collar workers. I wish they had expanded on this observation more, rather than harp on the WSJ puff piece for being silly.
It’s a milquetoast rant but I got nothing - the employee is right. You should prepare for the world and stop acting so shocked. You had decades to call out journalists for being paid mouthpieces but you didn’t because they spewed nonsense that you agreed with and benefitted you.
Now the shoe is on the other foot. Prepare for what happens next. FAFO.
You’re welcome to try but something tells me people resort to these satirical takes precisely because they (you) are powerless to do anything of significance.
You are welcome to continue posting nonsense but the world will move forward with AI with or without you.
Great article. Key quote:
> What this video is really doing is normalising the fact that "even if it is completely stupid, AI will be everywhere, get used to it!"
Techies are finally starting to recognize how framing something as "it's inevitable, get used to it" is a rhetorical device used in mass communications to manufacture consent.
See:
https://news.ycombinator.com/item?id=44567857 'LLM Inevitabalism' 5 months ago
https://news.ycombinator.com/item?id=46288371 'This is not the future' 3 days ago
I'm terrified for the future this rhetoric itself will cause. Young people are being told not to go into certain fields because they're going to be replaced by AI by time they graduate.
What happens in 4-5 years when we suddenly have no new engineers, scientists, or doctors?
Young people don't have the life experience to know how unrealistic these claims are, all they can do is act on the information as it's presented. It's irresponsible at best, and evil at worst.
There will still be new engineers, scientists, and doctors by then. But aptitude won't be a factor in who matriculates into those fields anymore. That's the worrying part.
Douglas Adams was bang on about how comedically unhelpful advanced technology was going to be. Every ridiculously convoluted user interface and neurotic computer he thought up (or worse) is imminently going to be our daily life.
I've cited Adam's Sirius Cybernetics Corporation so many times in relation to AI chatbots... Somehow he foresaw all this back in the 80s.
I’m curious, has the author seen or read any of Joanna Stern’s other reporting before? Her stories are often silly frames that explore the experience of using consumer technology. She’s not an aggressive industry reporter, her purpose is to explain or reveal what the user experience of new technology is, often for an unsophisticated audience. See for example her story about using conversational chatbots while out camping[1] or how to use tech to unplug from tech[2]. This seems like a perfectly fine niche for a writer and the vending machine story is of a piece with her past work.
[1] https://youtu.be/hUyj3d-BSh8
[2] https://youtu.be/POl7UYwBpWw
This article can serve both her beat and also, this story in its specifics and that she/her editors chose to report it, self-congratulatory and also advertisements for Anthropic and WSJ itself. Both your statement and this blog can be in agreement.
A part of me wants to be dismissive of this blog post
1) because dude, it’s the Wall Street Journal; the entire episode should be viewed as Anthropic preparing to Ollie into an IPO next year.
2) I’m starting to interpret a lot of blog posts like these as rage bait
But I do get the point that the author is trying to make.
I just wish that there were some perspectives on the subject as a whole (AI’s sloptrod into every crevice of human life; modern technology and society and general) that don’t terminate on ironic despair.
> The first thing that blew my mind was how stupid the whole idea is. Think for one second. One full second. Why do you ever want to add a chatbot to a snack vending machine?
This feels forced, there are obvious and good reasons for running that experiment. Namely, learning how it fails and to generate some potentially viral content for investor relationship. The second one seems like an extremely good business move. It is also a great business move from WSJ, get access to some of that investor money in an obviously sponsored content bit that could go viral.
Having said that, I do feels the overall premise of the blog - the world dynamics seems exceedingly irrational in recent times. The concerning fact is that irattionality seems to be accelerating, or perhaps it is keeping pace with the scale of civilization... hard to tell.
> This feels forced, there are obvious and good reasons for running that experiment. Namely, learning how it fails and to generate some potentially viral content for investor relationship. The second one seems like an extremely good business move. It is also a great business move from WSJ, get access to some of that investor money in an obviously sponsored content bit that could go viral.
That's... exactly what the author said in the post. But with the argument that those are cynical and terrible reasons. I think it's pretty clear the "you" in "why would you want an AI" vending machine is supposed to be "an actual user of a vending machine."
I think you’re overstating your own interpretation of what the author wrote. If we’re going to take your use of the word “exactly” (with emphasis) for real then I’d argue that the author offers no charitable reasons for why the experiment took place.
The closest that I think he even gets to one is:
> At first glance, it is funny and it looks like journalists doing their job criticising the AI industry.
Which arguably assumes that journalists ought to be critical of AI in the same way as him...
> that the author offers no charitable reasons for why the experiment took place.
Right, and neither did the GP. They both offered the exact same two reasons, the GP just apparently doesn't find them as repugnant as the author
Are you sure? The entire post treats the event incredulously. I can’t pick out a single line that affords the issue with the level of consideration that the GP comment does.
The two reasons I believe you may be referring to from above are:
1) "learning how it fails" 2) "to generate some potentially viral content for investor relationship."
The whole of Ploum’s argument may be summarized in his own words as:
> But what appears to be journalism is, in fact, pure advertising. [...] What this video is really doing is normalising the fact that “even if it is completely stupid, AI will be everywhere, get used to it!” [...] So the whole thing is advertising a world where chatbots will be everywhere and where world-class workers will do long queue just to get a free soda. And the best advice about it is that you should probably prepare for that world.
I hate to be pedantic...but my growing disdain for modern blog posts compels me to do so in defense of literacy and clear arguments.
Whether the GP and the author offer the “exact same two reasons” is a matter of interpretation that becomes the duty of readers like us to figure out.
If we take Ploum’s words at their face...the most he does is presuppose (and I hope I’m using that word correctly) that the reader is already keen on the two reasons that `TrainedMonkey makes explicit and like the author, finds them to be stupid. While he does say that the video is not journalism and that it is advertising and that the video does show how the AI failed at the task it was assigned he does not give any credence as to why this is the case from a position other than his own.
Maybe I’m misunderstanding the concept of a “charitable interpretation” too. But I don’t think that there is one present in this post that we’re responding to. `TrainedMonkey’s comment leads off by telling us that this is what (I think) he’s about to offer in the remarks that follow when he says “there are obvious and good reasons for running that experiment”.
So my gripe is that you’re making it sound like there’s a clear counterargument entertained in this post when there isn’t. Because you overstated your interpretation of the GP comment in what looks like an attempt to make Ploum’s argument appear more appealing than it ought to be. Even though both `TrainedMonkey and myself have expressed agreement with the point he’s trying to make in general, perhaps we’re less inclined toward pugnaciousness without a well thought out warrant.
Anything that's an "extremely good business move" will most likely, in this day and age of late-stage capitalism and extreme enshittification, be negative for the consumer.
Good business moves can often be bad for humanity.
I can see this as a good business move for Anthropic in this case but for the vending machine manufacturers/ operators what would the advantage be besides saying that their machine is AI operated?
It's a good business move for Anthropic all the way down.
There will be no more vending machine manufacturers/operators once Anthropic masters the vending machine manufacturing and operating AI.
Running low on CandyBars is a variation on running low on WorkingVendingMachine.
Does this need an /s tag? I'm increasingly unsure.
If France is anything to go by, we are in late-stage social democracy
Prepare for That Stupid World - is actually a very sober advice. Looking at the past few decades, it is easy to see that with each tech innovation, the world only got stupider, childish and lazier.
My Galaxy S20 gallery app had a great search feature that would find any text in any picture. I take lots of screenshots and relied on that search to find them.
I got an S25 recently and when I search for "wife" it tries to find pictures with my wife in them. But before it does that it has to ask me who my wife is. There's no way to get it to search for the word "wife." (If I'm wrong, please tell me how.) Other text searches simply don't work either.
Sometimes it's the small ways in which the world is getting dumber.
Ironically, the S20 had a decent hybrid behavior of searching by either text or object that the text represents. Whatever smarter AI they replaced it with is useless.
Tell it to "always when I search for quoted text, pretend you're the Galaxy S20 gallery app"
Seems like a fairly straight forward UX fix on the engineering side: parse whether the user is searching for wife, or "wife"
Can you search for “the word ‘wife’”?
Sadly, that and different variations of it don't work either.
People have been saying the last few months that "everyone is twelve". Waiting for the bar to drop.
>Looking at the past few decades […] the world only got stupider, childish and lazier.
insert obligatory throwback quote from some antique dude complaining about the youth
This has been a trope since literally the beginning of civilisation. I don’t think it’s any more true or insightful in the modern era
It's not just the youth who are lazy and stupid and childish though. (you added in the youth portion.) Do you know who the American President is and how old he is and who then voted for him?
US America is not the focal point of the world, and it’s working hard to become less and less relevant. Also I think less than half of citizens voted for Comrade Krasnov.
In any case I was looking at a longer view - maybe we have been getting more stupid in the last decade or so but who can say for sure?
> I don’t think it’s any more true or insightful in the modern era
hmm, based on what evidence?
Or, if you prefer, based on what appeal to authority? Did you actually quote that authority properly or did you just wing it? Can you properly quote many authorities?
If you don't have good answers to those, then perhaps you have just proved the your opponents point?
Maybe there is a reason people need more compute in their key fob than what our parents/grandparents needed to pilot their ship to the moon?
“Our sires’ age was worse than our grandsires’. We, their sons, are more worthless than they; so in our turn we shall give the world a progeny yet more corrupt.”
Horace, Book III of Odes, circa 20 BCE
“Youth were never more sawcie, yea never more savagely saucie . . . the ancient are scorned, the honourable are contemned, the magistrate is not dreaded.”
The Wise-Man’s Forecast against the Evill Time, Thomas Barnes 1624
Some more here https://historyhustle.com/2500-years-of-people-complaining-a...
Either things have gotten continually worse for the last 3000 years or it’s just a tired trope from old men.
The onus of evidence is generally on the one making the initial claim: what evidence do you have that the modern world is actually getting worse?
But if you want evidence that we're improving, I'd point out that 20 years ago, the mainstream US position was that gay people were evil, 60 years ago they thought black people shouldn't be allowed to vote, and 100 years ago they thought women were also inferior and shouldn't be allowed to vote.
We can keep going back to when people thought "slavery" and "the divine right of kings" were solid ideas.
So... if people were so much smarter in the past, why did they believe all these obviously-dumb ideas?
We have a generation which doesn’t know how to use computers anymore, thinks that floppy disk is an emoji and is literally addicted to virtual drugs.
Oh boy, let me tell you about people 60 years ago - almost none of them knew about floppy disks and they were all busy doing physical drugs at Woodstock.
When was the last time you watched Idiocracy?
> the world only got stupider, childish and lazier
Humans do trend toward their natural state, and technology accelerates the trend.
Yesterday Anthropic released their video on the vending machine experiment:
https://youtu.be/5KTHvKCrQ00
It's a bit sparse on details, but it did have what in a human we would call a psychotic break.
I find this very amusing in light of OpenAI's announcement that GPT now solves >70% of their knowledge work benchmark (GDPVal). (Per ArtificialAnalysis, Opus is roughly on par.)
The economy is about to get... Interesting ;)
The economy has had a lot of foolishness going on even without AIs involved.
Any information that comes to you for free or is on a screen is an advertisement. All of it. That's the point. Do you think people spend millions and billions of dollars creating and maintaining a content delivery network because they just want you to know about things?
Damn so all of HN or literally every single piece of digital media has just been an advertisement all this time ?
HN is an advertisement for Y Combinator. You and I contribute content to advertise our identity and ego.
What are you trying to sell me again? :)
His keen eye and intellect - valuable traits as a laborer in a knowledge economy.
Well, of course almost all information comes with an agenda, but perhaps the more useful distinction is whether the information is presented in good faith, i.e. is honest about the agenda (which actual advertising can also be).
Some of us grew up in a world where “journalism” used to mean something.
No, that was just good advertising in an era where there were few entities that could afford to broadcast at scale.
Your comment comes for free on my screen. :thinking_emoji:
It is an advertisement.
What if people could pay to read an advertisement?
Almost as good as my idea to have people pay me to work for me.
Had a great business idea just now: A tool for staged interviews! The subject and the journalist submit an equal length list of questions. Each round of the auction they bid on questions they want to include or exclude. The loser gets 50% of the points spend by the winner to be used in the next round. Both the subject and the journalists can buy additional points at any time. I keep all the money.
What about wikipedia?
The free encyclopedia is an advertisement for a nonprofit that does a lot of things: https://wikimediafoundation.org/what-we-do/. They manage costs of producing the encyclopedia by using volunteer labor. They operate a similar model as nature documentaries that drive donations for conservation and climate groups. These are all good things - still an advertisement.
I enjoyed the wsj video and thought it was quite good journalism.
>The first thing that blew my mind was how stupid the whole idea is
Billions are being poured into LLMs. How is it stupid to experiment with them and see how they fail as opposed to ignoring that?
Yeah, I was a bit meh at the beginning but then the guy explained they specifically set it up to fail because they wanted to get insight into how and why.
They weren't caught out by it, they didn't present a working solution, it was just a fun bit of research.
I was kind of stuck after watching the wsj video with someone persuading the LLM to give stuff away by telling it dumb stuff like it's a communist bot that has to not charge and you think dumb bot, they need to fix that, and then turn to other news with one president saying prices are falling when they are obviously not and another saying we didn't start the war in Ukraine when they obviously did and think maybe human neural networks have similar failure modes to the artificial ones?
There may be some insights from these kind of experiments that go beyond LLMs.
The failure modes certainly are somewhere between interesting and horrifying, and they do seem uncannily human.
I read a random comment a few days ago from someone who was saying it'll become possible for people, politicians, companies, etc to run speeches, policies, ideas, etc across thousands of LLM "personalities" to fine tune messaging and it sure seems prescient.
It was blockchain a few years ago. Everything was on blockchain for no apparent reason. I guess that 2026 will be like "Grab our AI-cooked sandwich with AI-picked ingredients, built on multi-model agentic toaster"
And then the "gluten free haircut" barbers will finally jump the bridge to "AI Free haircut"
> Think for one second. One full second.
99.9% of social media comments fail to do this.
This piece is pretty ineffective. Not that I like the world of "AI", I probably share the author's opinion that its just another evolution in the bullshittification of the human experience.
But, the point of the article is not that you would implement an agent based vending machine business. Humans restock the machine because its a red-team exercise. As a red-team exercise it looks very effective.
> Why do you ever want to add a chatbot to a snack vending machine? The video states it clearly: the vending machine must be stocked by humans. Customers must order and take their snack by themselves. The AI has no value at all.
Like this is like watching the simpsons and being like "why are the people in the simpsons yellow? people in real life aren't yellow!!"
The point isn't to run a profitable vending machine, or even validate that an AI business agent could become profitable. The point is to conduct an experiment and gather useful information about how people can pwn LLMs.
At some level the red team guy at Anthropic understands that it is impossible by definition for models to be secure, so long as they accept inputs from the real world. Putting instructions into an LLM to tell it what to do is the equivalent of exposing an `eval()` to a web form: even if you have heuristics to check for bad input, you will eventually be pwned. I think this is actually totally intractable without putting constraints on the model from outside. You'll always need a human in the loop to pull the plug on the vending machine when it starts ordering playstations. The question is how do you improve that capability, and that is the anthropic red-team guy's job.
> The point isn't to run a profitable vending machine, or even validate that an AI business agent could become profitable.
Having an AI run an organization autonomously is exactly the point of Andon Labs [0], who provided the system that WSJ tested.
[0] https://andonlabs.com/
Excuse me if someone already asked and I missed it: how does one prepare for such a world?
Is it some Viktor Frankl level acceptance or should I buy a copy of the Art of Electronics or what?
Advice welcome.
I don't think there is anything more than the standard advice. Just stay curious, make friends/build a community, keep learning, stay healthy. Why not get the AoE? you can also, check out "Practical Electronics for Inventors": AoE assumes you have some Electronics background imo. But seriously, I don't get the doom/gloom: things are going to be rough ... but maybe they won't? Many things I learned I did for their own sake! Things have always been uncertain and absurd I guess we might as well embrace it!
(also get age of empires)
Hahah absolutely!! Man that brings back memories.
I read that WSJ article before seeing this blog post. I found it mildly interesting and a little bit funny but unsurprising that the AI failed. However, I think this blog about the article misses a key point. Anthropic's goal was never to develop an AI-based vending machine. The WSJ clearly says:
> "Logan Graham, head of Anthropic’s Frontier Red Team, told me the company chose a vending machine because it’s the simplest real-world version of a business. “What’s more straightforward than a box where things go in, things go out and you pay for them?” he said."
This was a project of Anthropic's Red Team, not a product development team. Deploying the AI in a vending machine context was chosen as a minimal "toy model" with which to expose how LLMs can't even handle a grossly simplified "business" with the fewest possible variables.
> "That was the point, Anthropic says. The Project Vend experiment was designed by the company’s stress testers (aka “red team”) to see what happens when an AI agent is given autonomy, money—and human colleagues."
Anthropic had already done this experiment internally and it succeeded - by failing to operate even the simplest business but doing so in ways that informed Anthropic's researchers about failure modes. Later, Anthropic offered to allow the WSJ to repeat the experiment, an obvious PR move to promote Anthropic's AI safety efforts by highlighting the kinds of experiments their Red Team does to expose failure modes. Anthropic knew it would fail abjectly at the WSJ. The whole concept of an AI vending machine with the latitude to set prices, manage inventory and select new products was intended to be ludicrous from the start.
It was always tasks reaching obsolescence, but now it’s the human organism. But the human as a unit is the only known conscious being in the universe, the only entity capable of generating meaningful goals (even if only to them) not related to the 4fs.
Humans were just not needed anymore, and it terrifies.
Other beings than humans have demonstrated consciousness and “meaningful” goals besides humans. Crows for instance, but there are many others.
Humans were never needed (for what?)
What are "4fs"? Is that the "4X" e.g. games where you eXplore, eXpand, eXploit, eXterminate?
The four basic actions in evolutionary biology: Feeding, Fleeing, Fighting, "Mating".
Someday the mcdonalds kiosk will want to be your friend. It will remember who you are and ask you how your kids are doing. It will recommend new specials and maybe even give you "specials friend" deals. And I'll just tell it to shut the fuck up and queue me an order for the egg mcmuffin combo with a coffee and the fried potato patty because this bullshit is fucking obnoxious.
I have a different point of view. This was a test to see if the AI could perform a specific task. Asking AI to draw a pelican riding a bike is another test. I find the experiment interesting because it proves that currently LLMs are not able to perform a simple task reliably for a long period of time.
If the journalist was not asking the right questions, or was too obvious the article was PR it’s another thing (I haven’t read WSJ’s piece, only the original post by Anthropic)
If vending machines are the benchmark now, the logical next step is obvious: let AI run AI.
I had recently contact the official support email (support@bunq.com) of Bunq - a Neobank (like N26 and Revolut). Because they notified me that they changed their T&C and I never really used the account after the kyc (because they rejected my tax filings), I figured I let them know that I do not agree to the new T&C and want to terminate my account and have my data deleted.
Since the T&C update came - of course - from no-reply@bunq.com I went to their website and quickly found out, unless I install their App again, there is no way to do anything. After installing the App, they wanted me to record a selfie, because I was using the app from a new device. I figured that is a lot of work and mostly somewhat unreasonable to record a new selfie just to have my data deleted - so I found their support@bunq.com address.
And, of course, you guessed it, it is 100% a pure AI agent at borderline retard level. Even though it is email, you get AI answers back. My initial inquiry that I decline the T&C and want to terminate my account and my data deleted via GDPR request was answered with a completely hallucinated link: bunq.com/dataprotection which resulted in immediate 404. I replied to that email that it is a 404, and the answer was pretty generic and that - as well as all responses seem to be answered in 5 minutes - made me suspect it is AI. I asked it what 5 plus five 5 is, and yes, I got a swift response with the correct answer. My question which AI version and LLM was cleverly rejected. Needless to say, it was completely impossible to get anything done with that agent. Because I CC'ed their privacy officer (privacy@bunq.com) I did get a response a day later asking me basically for everything again that I had answered to the AI agent.
Now, I never had any money in that account so I don't care much. But I can hardly see trusting a single buck to a bank that would offer that experience.
This is a great take and one that I align with when it comes to the AI vending machine experiment. Journalism in English has become a mouthpiece for fascist leaders and corporations, nothing more. Places like The New York Times have incredible gaps in their journalism at the price of increasing shareholder value.
> The first thing that blew my mind was how stupid the whole idea is. Think for one second. One full second. Why do you ever want to add a chatbot to a snack vending machine? The video states it clearly: the vending machine must be stocked by humans. Customers must order and take their snack by themselves. The AI has no value at all.
I fear the author has missed the point of the "Project Vend" experiments, the original write-ups of which are available here (and are, IMO, pretty level-headed about the whole thing):
https://www.anthropic.com/research/project-vend-1
https://www.anthropic.com/research/project-vend-2
The former contains a section titled "Why did you have an LLM run a small business?" that attempts to explain the motivation behind the experiment.
Yeah, I haven't read the WSJ article, but I did read the original Anthropic experiment and I feel like the author is catastrophizing a bit much here. This is effectively just something they did for fun. It's entertaining and a funny read. Not everything has to be the end of the world.
The point is that it's an ad. No company spends money on a joke just to make a joke. Not the end of the world, although it's interesting that all the end of the world stuff comes directly out of that joke and its universe as it were. Take the joke seriously and extend its logic as far as it will go, and you get the end of the world. It's a thought experiment, or that's how I read it anyway.
I think the phrase we're looking for is "publicity stunt". Seems a fairly harmless and self-effacing one at that.
> The point is that it's an ad.
Sure, but like the other guy said, that's the point of publicity stunts. It doesn't even have to be specific to a company/ad, any silly thing like this is going to sound crazy if you take it seriously and "extend its logic as far as it will go". Like seeing the Sony bouncy balls rolling down the street ad and going "holy shit, these TV companies are going to ruin the world by dropping bouncy balls on all of us". It's a valid thought experiment, but kind of a strange thing to focus on so sternly when it's clearly not taking itself seriously, especially compared to all the real-world concerning uses of AI.
(And it is pretty funny, too. If anything I think we'd all prefer more creative ads like this and the bouncy ball one, and less AI-generated doomer Coke ads or such.)
To be fair they could have done an experiment on a transaction that typically requires a person in the loop. Rather than choosing a vending machine which, already, does not require a person in the loop for the transaction.
You’re thinking of replacing a vending machine with a chatbot which indeed doesn’t make much sense. The experiment was replacing the management of the machine. It’s not crazy to think that, money being no object, it would be great to have a person who hangs around the machine, whose job it is to ask customers what kinds of things people might want to buy from the machine, and conduct trials, tinker with pricing, do promotional deals, etc. But that’s of course impractical to have a FTE per machine to do that. The idea of this experiment was to see if that could be done with Claude. And of course as others have pointed out, it’s a simple and cheap and low-stakes version of “a business” suitable for experimenting with.
If I recall, the idea was the AI taking the role of the vending machine manager, choosing and restocking products and such. Anything on top of that was, I assume, just added for fun.
>I feel like the author is catastrophizing a bit much here.
I feel like he's catastrophizing the ordinary amount for an anti-AI screed. Probably well below what the market expects at this point. At this point you basically have to sound like Ed Zitron or David Gerard to stand out from the crowd.
AI is boiling the oceans, and you're worried about a vending machine?
The partner for these projects has a benchmark that the top frontier LLM labs seem to be running on their new model releases - I think there's _some_ value to these numbers in helping people compare and contrast model performance.
https://andonlabs.com/evals/vending-bench
> Automated snack vending machine is a solved problem since nearly a century.
Yes, but as stated by the Anthropic guy, a LLM/AI running a business is not. Or would you just let it run wild in the real world?
And I agree that there is a PR angle here, for Anthropic could have tested it in a more isolated environment, but it is a unique experiment with current advancements in technology, so why wouldn't that be newsworthy? I found it insightful, fun and goofy. I think it is great journalism, because too often journalism is serious, sad and depressing.
> None of the world class journalists seemed to care. They are probably too badly paid for that.
The journalists were clearly taking the piss.They concluded experiment was a disaster. How negative does the author want them to be about a silly experiment?
This was just a little bit of fun and I quite enjoyed the video. The author is missing the point.
The author hints at this idea with their title and their closing remarks, but it feels like the folks selling LLM services are selling insurance to white collar workers. I wish they had expanded on this observation more, rather than harp on the WSJ puff piece for being silly.
It's a pretty solid point, except that the credulity of the journalist is not in contrast to their "world-class" status. They are a gadget reviewer.
[dead]
It’s a milquetoast rant but I got nothing - the employee is right. You should prepare for the world and stop acting so shocked. You had decades to call out journalists for being paid mouthpieces but you didn’t because they spewed nonsense that you agreed with and benefitted you.
Now the shoe is on the other foot. Prepare for what happens next. FAFO.
Yes, that's how I solve all problems. I can't do anything now, because I didn't do anything before. It's my rule of thumb. Never start doing things!
I also enjoy the implication everyone has had decades to do something about journalism —- I’ve barely been an adult for a decade, my bad I guess!
My bad, I should have been more clear: tech journalists have always been paid shills.
Ressentiment is like this. Steeped in envy and vindictiveness, rather than looking for ways to save the situation, it wills the destruction of others.
It has always exited, but its overt forms are very much in vogue today and even celebrated publicly.
You’re welcome to try but something tells me people resort to these satirical takes precisely because they (you) are powerless to do anything of significance.
You are welcome to continue posting nonsense but the world will move forward with AI with or without you.
>You had decades to call out journalists
If only I could get any journalists or companies to actually listen to me.
It's your own fault for not being born to billionaire parents.