Running open models from commodity hosts doesn't appear to be particularly subsidized today. It's not a massive expense, because the biggest cost that's causing the frontier labs to hemorrhage money at the moment is training. Inference isn't terribly expensive for what you get, and seems to be quite comfortably profitable.
OpenRouter is full of inference providers for open weight models that are only months behind models like Opus, with a fraction of their price per token. And those inference providers don't do charity.
If optimizing inference was a priority, OpenAI and Anthropic could also focus on it and get the prices down. They are just focusing on their dumb AGI race instead.
This is flawed thinking - the models will exist even if every company who trained them fails, and then someone else will rehost them and offer them at inference rates. Inference is _astonishingly_ cheap.
If anything, prices would go _down_ as the rehosters wouldn't be saddled with debt from training models.
We're on said plateau. In terms of normal coding usage, models have not become wildly more useful in the last few months. The existing open models that are useful today will remain useful.
Well this "silver bullet" is paid for by VC money, for now. I don't know what happens after, but VCs aren't in the business of subsidising a service forever. We're in a very temporary point in the cycle and things will change, I just can't predict how.
When Taalas burned Llama 3.1 to the HC1 they saw at least an order of magnitude more tokens per joule. That saving is sitting right there, today, ready for anyone who is willing to say "this is the model we're freezing on for a couple of years".
"Everyone globally decide to stop improving and instead compete on costs" would be an example of the kind of change I'm talking about. I don't know if that will happen but it seems somewhat plausible.
It also seems plausible that it will take many years until hardware price/performance reaches a point where companies can provide models people deem acceptable at unsubsidized prices people deem acceptable. I don't know.
I’ve lived through a few. The internet came and stayed. The bubble just wiped out the speculators, but the internet is bigger than ever. We will be using LLMs for a long time. The only way that’s not true is if a superior replacement comes along.
I've lived through all of them. This is, dare I say it, cope. In some ways I fervently wish it were true - perhaps there will still be some kind of Engineering roles (right now we're enjoying a glut via Jevon's paradox) but the point of the singularity is we can't predict what's going to happen.
What's _not_ going to happen is what's desired by this article which can basically be condensed into 3 bullet points to be appended to your initial spec to Opus 5.5.
I don't want to see this person's dreams crushed, but _log linear charts keep going up and to the right_. I also don't think we're entitled to the high status and pay jobs we currently have 'by right'. We've been useful - we've been _very_ useful and we've done very well for ourselves. We've been the shovels in the gold rush. Now we have to hope we can adapt and that we still provide value commensurate to our pay.
I didn't become a software engineer for the money - I just happened to one day have my hobby become a lucrative career, so perhaps my viewpoint is a little different. All the stuff that is important to the author is still stuff they can do, they just probably won't receive the status and remuneration they think they're entitled to.
We would have. IF a typical C codebase in its early years sucked as much as some AI generated codebases do today.
But instead, humans preferred the productivity AND the readability AND maintainability that elevated us above assembly. Maybe AI-generated code will get there also.
As someone maintaining and using agentic AI developed codebases, they’re better maintained than human stuff I worked on.
For example, we rigidly require specs that can be used to drive end to end tests. Commits can’t break builds or tests. 100% test coverage required for both unit and integration/E2E tests.
And all of this alongside adding macOS/iOS/Android to our standard builds instead of just web/electron. Oh. And moving backends to Rust/Go.
One might also notice that the internet did come to dominate our modern world, in spite of the dot com crash. A crash does not equate the end of these tools.
I really cannot wrap my mind around the idea that LLMs will just not be a part of everyday life in the future — in tech and everywhere else...
This is something I keep asking myself: why am I not scared for my job? Do I really believe I'll just manage to survive? Maybe I do?...
I can't honestly say so, though.
Yet, I can say I am so excited for the future! For once, I don't want to wake up in the future — I'm loving being at the heart of it, seeing it progress and watching it allow people to be more creative and more empowered.
Some people I know also seem to believe this? Like all the AI providers are going to go bankrupt and we'll just go back to what there was before? Do they not realize that indeed the productivity gains are there?
Sure, there's a lot to be debated about how this fell on many people's laps, on how they are being asked to do more, with the same pay and sometimes putting in more hours. These are pressing issues that have emerged as a natural (or, well, maybe forced for some) consequence of a complete shakeup in the world in general. But I don't think society (i.e. most people) will go back to what was before unless we actually have some form of devastation that really physically prevents us from doing so.
If OpenAI and Anthropic and what have you go bankrupt, others will continue. Maybe the economics will look a bit different, but to think we will just go back to what it was seems, to me, completely nonsensical. But maybe I'm the one spouting nonsense. We'll see in a few years...
Part of the current environment is the hype around how far this can go. Investment in anything else is rough because of the ballooned hardware costs and the fear that AI will eat your lunch. That's not a reality necessarily. That part of it could easily change if the mood of the investment class changes.
As far as what AI actually provides, at the end of the day its cheaper software faster (or that's the promise). You still need to build products and maybe you can do that with less folks now. You're not writing assembly today. The machine does it. But you're still working on software. Maybe with AI you're programming less and working on UX more. Is that earth shattering change?
We've yet to hit a point where you any of the big software companies have been overthrown by an AI startup... Whats the revolution really? Its captured the mindshare of the industry but we're at a point of being practically disgusted by AI products already.
Things won't literally go back to 2019. But there could certainly be a return to normalcy in store.
Funny, I went to read the article, and I found it a lot less defensive than replies such as yours. The author even makes some good points.
The main issue for front end development is that companies (particularly large ones) have been releasing shitty broken UI/UX for so long that this stuff became the norm.
Either way the article does not even explain what "blows over" mean. A lot of people seem to be fearfully reacting to the title alone.
I have to say, the other day I tested out ChatGPT sites. Around a year ago I built a website for a rehearsal space - a booking system app - and spent a whole day with ChatGPT building it. It was janky, ugly, and getting there was a process. FWIW, traditional systems dev is what I learned, but I know fuck all about front-end, and I'm not too well versed on web dev in general. I did have to get under the hood and do some coding after I ran out of usage with ChatGPT back then.
But with sites? I made a big picture spec sheet, and asked ChatGPT to expand on that. Then I fed that spec sheet into sites. 30 minutes later it spit out a perfectly working site of what I had tried to make a year ago. Not only that, it was much richer, and had more features. I asked for around 10 changes and new features, and all but one worked flawlessly. One time it managed to break functionality, but it was swiftly fixed. I didn't write or edit a single line of code. NOT A SINGLE ONE.
I've previously been on the buying side, and this was a website which could have easily ended up costing $10k-$15k with all the functionality. If not more. And I ended up getting that with my measly $20 subscription.
Web dev as a career is not dead, or dying, but I predict that A LOT of the people that would make those $0 - $10k sites are going to have a tough time ahead.
EDIT: This was also on par with my general experience with AI. For the longest time you can use frontier models, be unhappy with the results and tell yourself this sucks - and then out of nowhere some updates comes along, and suddenly everything just works perfectly well.
A year ago, I asked ChatGPT to design me a 3D model for a wallet, the kind you can stick to the back of a phone with a magnet. It screwed it up over and over, never got to anything usable.
Last week I tried again, just explaining in words what to do, and it made what I had in my mind.
For websites, it's over. Any kind of brochure/simple restaurant booking site, that was easy years ago and trivial now. SaaS is going next, of course there's a lot of types of SaaS but your average "web implementation of commonly needed service" is now very easily copied with AI.
Anything like a task list or information management system with no particular constraints in terms of performance or scale, you can just make an AI customize for you.
> Web dev as a career is not dead, or dying, but I predict that A LOT of the people that would make those $0 - $10k sites are going to have a tough time ahead.
Not even the first time this happened. When I was in high school you could make great money building websites for small businesses that nowadays just have an Instagram account and no website. The number of CMSes I've built in my career is too damn high
I was able to break into software by kinda exaggerating some freelance work I did when I was a teenager. Mostly doing websites for local business. Surely not an unfamiliar story here.
While everyone has been decrying site-builders for years, I recall a conversation I had around 16 with a small business owner.
I was trying to make some money and convince this guy to let me make him a site. What he said has stuck with me. It was something along the lines of:
"Why do I need a website. Most of my customers find me, contact and and engage with me on Facebook. I need someone who can manage a social media page"
> Around a year ago [...] It was janky, ugly, and getting there was a process.
Last December/January seems to have been the tipping point from too flawed to be really useful into something pretty good for dev work, so this seems about right.
Hardware is going to get cheaper (as it always does over the long term) and even if the commercial models enshittify¹ to something too contaminated by ads to be bearable the open models won't. We're not going back.
For now it looks like the ability to lead these tools to good solutions is valuable. I guess we'll see if it carries on so. Not much we can do if it isn't except move with the times.
¹ I'm actually fairly optimistic about that at the moment because so far we are the customers.
> Accessibility has always been a foundation of web development
> CSS should be hard
I don’t see how the author can square these two. If accessibility is important, the CSS which enables it should be easy. And what about accessibility from a web dev perspective? If doing the right thing requires arcane knowledge do you think most people will buy a book on CSS to master the thing that AI is actively de-skilling?
In my opinion, the author needs to take their own advice. Talk to actual people who need web dev work, figure out what they want to pay for that can’t easily be done by machines, and get good at that.
Most web traffic is robotic. The internet is hostile to humans now. Apps gatekeep and monitor our access. LLMs build it. Nothing about the internet is evolving in the direction of a structure created by humans for humans.
i don't think we're going back to pre-ai era anytime soon.
My own experience, probably similar to a lot of people's own experience is that i haven't written code by hand (at least 10 lines) in 2026, i can work with any web/backend framework/language because LLMs can.
I'm now an architect / team leader, i manage a team of devs (LLM) and i mostly spend my day providing guidance/suggestions and making architectural choices.
I've never built things that fast, it makes no sense to hand write code anymore as it's so slow. When i run out of tokens, i just wait (doing some reviews, code reads, etc ...), but it doesn't make sense to spend a day writing some code that gonna be generated in a few minutes when tokens are back ...
Obviously this is my personal experience, there's other industries / use cases / etc ... where hand writing code makes sense. But it's not my case anymore.
That's like being the amish of web workers. Good for you if you can, you may very well fit a useful niche. But most of the developers probably don't even want this.
llms / ai have already changed the landscape of tech, it's reaching into almost all industries.
a good comparison is like using a sword vs a gun. ofc in certain scenarios the sword is going to be the right tool / better than a gun. but 99.99% of the time you're gonna want to use the gun and refusing to do so bc it's different or bc you mastered the sword and have some pride isn't going to end well for you.
i find coding agents etc to be really useful and am excited about it. the thing i'm waiting to blow over is for a good work / life / human centric balance to be accepted as the norm. i'm tired of coding agents / ai being the reason behind "if you aren't shipping a critical feature like handling payments in 3 days, you're trash"
> i find coding agents etc to be really useful and am excited about it. the thing i'm waiting to blow over is for a good work / life / human centric balance to be accepted as the norm.
Job is a job. There are few reasons a new technology would change employment for better, and also quite some reasons it could change being employed for worse.
I don't know, one thing going for him is he's actively learning by doing. A lot of people appears to have mentally checked out and just press the "Solve this ticket" button and never learn anything new now. Programming skills might still be valuable down the line, he's betting on that I suppose.
I think that LLMs will be just another tool like linters or build tools. Enterprise and tech startups will still need custom solutions and people accountable for keeping them afloat. However the guy who got by making small business websites from templates will probably have a hard time. Between squarespace, shopify, and LLMs there just isn't a reason to pay more than a few bucks for a website for your business.
First, if you crack open a history book (or have an agent do it for you), you'll notice that the world goes backwards quite often. It's not pretty, but the critical mass of people who feel they have nothing to lose is pretty modest.
Second, I'm really curious what's your "adapt" path, because I suspect you're in denial too. Accept that, before long, an LLM will not need you to write good prompts, tell it to "keep going", or say "no, that's wrong". What's your value then? Does your employer keep you on the payroll as charity? For old times' sake? If your value is just supervising AI, then your value is declining rapidly.
People in unskilled professions don't have choice when automation shows up and takes their jobs. Software engineering may go down in history as one of the few highly skilled professions that didn't just keel over and die, but enthusiastically embraced its own demise.
> First, if you crack open a history book (or have an agent do it for you), you'll notice that the world goes backwards quite often.
The only thing I can think of is the reduction of truffles as an ingredient, which apparently happened around WWI. We eat a lot fewer of them than before.
> Second, I'm really curious what's your "adapt" path, because I suspect you're in denial too. Accept that, before long, an LLM will not need you to write good prompts, tell it to "keep going", or say "no, that's wrong". What's your value then? Does your employer keep you on the payroll as charity? For old times' sake? If your value is just supervising AI, then your value is declining rapidly.
Your employer cannot talk to all the LLM sessions at the same time. He also can't write a prompt that will do it for him.
> People in unskilled professions don't have choice when automation shows up and takes their jobs. Software engineering may go down in history as one of the few highly skilled professions that didn't just keel over and die, but enthusiastically embraced its own demise.
It's changed several times over the decades. How many people are working on a mainframe nowadays? How many people are writing assembly? Look at how many languages have come and gone too. LLM is a big change, but that doesn't mean software people will be gone.
You can get an Intel Arc Pro with 48G of GDDR6 for around $2000 new. That'll run up to about a 40G model at 8-bit or 80G models at 4-bit, and at probably like 100 TPS.
When the RAM bubble pops you'll be seeing cards with 4X this amount of RAM for that price.
This makes me so nostalgic for blockchain evangelists when people started to sour on the tech. Enthusiasm slowly morphed into desperate proselytizing and finally outright scamming, with the NFT clown show as a fitting headstone.
No new governments on the DAO, no ubiquitous programs running on the chain, no magical new answer to “fiat”… just a lot of empty promises and waste.
I don't personally know anyone who bought crypto for anything other than investment/gambling purposes. I know a LOT of people who have Claude or ChatGPT subscriptions because they find them useful.
Is your thesis that AI is useless? Because that seems like fantasy to me.
There is maybe a case to make around the economics of AI companies, but even there I'm very skeptical because the open models are quite capable on relatively inexpensive hardware.
My thesis is that his place is full of hype waves as people try to align themselves with the expectations and demands of the wealthy people they want money from. Who you personally know isn’t relevant to the question of how HN has reacted to this series of fads.
ML isn’t useless, that’s been obvious for a while, but it’s grossly overhyped and people here who think their vibe-coding isn’t ultimately worthless are delusional.
1) Become world class experts in one thing
2) Become full on invested in LLMs learn on how to match general intelligence on business requirements (Evals, isolation, judging,etc..) I think this will probably be still 5 years
3) Get ready to do something in real life.
I wish I could agree, but AI (as LLMs) is absolutely not blowing over. Even if frontier labs crash and burn I'm already running a model that's maybe only a month behind them, locally, in my house. Last year this was impossible within a reasonable budget. Next year it's going to be better and cheaper to do. My phone can already run models as capable as last year's. Horse has left the barn.
React is out and no one should learn it? Really? Something like half the Javascript job listings mention React ... but someone trying to get a job should not learn it?
Now to be fair I haven't job hunted in about half a year, so the market may have changed a little now that I have my (React-based) job ... but it hasn't changed so much that the most popular web framework has somehow become dead tech.
The point isn't that React is dead tech. React is practically every AIs front-end library of choice. I think the idea is AI is too good at it and you won't get a good ROI learning it. Learn fundamentals instead and when folks get sick of React because AI loves it, you'll be ready.
Frontend development will be the first that will be completely replaced by agentic development where nobody needs to see any code at all. Its already possible and many companies are having this way of working quite succesfully. The quality is already better than 99% of what frontend engineers can achieve. Speed is just insane. I can create any ui. Whether it's web or platform specific. There are hardly any limitations. Frameworks are still useful. But I expect them to become less important.
What remains important as a dev is system knowledge. Understanding of how systems are tight together. The principles of distributed systems, security, prompt engineering, agentic automation. And that kind of knowledge.
Code syntax is from the old era. Maybe there is a niche for it. But not for the majority.
Right. My clients now get iOS, macOS, Win32 (or whatever you call it), and Android apps… for free. Same price they paid before, and now they just get it all.
Bro, I think you are too deep into tech twitter (X). React is still king , almost every job opening for web needs it. Even the ones at OpenAI and Anthropic.
Dude, it's not 2021 anymore. You can't just yell a big, bold, name like racist and expect it to stick or cause any impact. Everyone got burned out in 2021 of the million -ists, and the 2024 election showed among other things that the name calling has zero effect.
The world belongs to those who can actually explain why any particular opinion is a bad thing, or conflicts with values that the reader may otherwise hold, not those who can accuse with a cheap label.
I was skeptical so I actually read those posts and concluded that yes, they are indeed unequivocally racist (in the traditionally accepted sense of the term).
I don't care whether or not DHH is a racist. He can't stop his project from being forked or exert authority over anyone.
I care about people who think that just yelling "he's a racist" somehow scores them brownie points or does something about it. It doesn't. You're not brave, you're not noble, you're not accomplishing anything.
I am close to a free speech absolutist (i.e., I don't believe that DHH should suffer legal consequences for those posts, unless he were directly inciting violence). I also believe that those who openly promulgate racist beliefs should suffer social opprobrium. That is what makes government censorship unnecessary.
author appears to think that at some point in the future we'll all agree to go back to how we were writing code in 2019. unclear why
Ask anyone who's lived through 1-2 boom/bust decade-ish hype cycles and they'll tell you why.
There well may be an AI crash, and everyone who invested into datacenters may lose a lot of money.
But in ten years we'll be using LLMs even more than today. I'm sure of it.
This is a lot more like the dot-com boom than like crypto or IoT. The internet really was a trillion-dollar industry, and AI will be too.
When you’re paying the real, unsubsidized price? $1 for $1 of compute?
Maybe.
I’m already paying the real unsubsidized price with my own GPU.
Maybe it can’t solve a Millennium Prize, but it’s pretty good at writing CSS.
Running open models from commodity hosts doesn't appear to be particularly subsidized today. It's not a massive expense, because the biggest cost that's causing the frontier labs to hemorrhage money at the moment is training. Inference isn't terribly expensive for what you get, and seems to be quite comfortably profitable.
OpenRouter is full of inference providers for open weight models that are only months behind models like Opus, with a fraction of their price per token. And those inference providers don't do charity.
If optimizing inference was a priority, OpenAI and Anthropic could also focus on it and get the prices down. They are just focusing on their dumb AGI race instead.
Qwen 3.8 runs fine on my laptop with no subsidy.
This is flawed thinking - the models will exist even if every company who trained them fails, and then someone else will rehost them and offer them at inference rates. Inference is _astonishingly_ cheap.
If anything, prices would go _down_ as the rehosters wouldn't be saddled with debt from training models.
How long does a model stay relevant? You’re assuming that a huge and lasting plateau will emerge that can still adapt to changing conditions?
We're on said plateau. In terms of normal coding usage, models have not become wildly more useful in the last few months. The existing open models that are useful today will remain useful.
Even if models stopped improving now, I would use them. They provide so much convenience, I'll never go back. There's no need to keep changing.
In fact I'm more or less at the point where I don't even want the model to get better, I care more that it outputs faster.
For coding in existing languages? Probably indefinitely. And huge swathes of human civilization run on decades-old codebases.
After the trough of disillusionment comes the plateau of productivity.
All those previous silver bullets are still with us. This one's not going away either.
Well this "silver bullet" is paid for by VC money, for now. I don't know what happens after, but VCs aren't in the business of subsidising a service forever. We're in a very temporary point in the cycle and things will change, I just can't predict how.
Koomey's law is a pretty safe bet.
When Taalas burned Llama 3.1 to the HC1 they saw at least an order of magnitude more tokens per joule. That saving is sitting right there, today, ready for anyone who is willing to say "this is the model we're freezing on for a couple of years".
"Everyone globally decide to stop improving and instead compete on costs" would be an example of the kind of change I'm talking about. I don't know if that will happen but it seems somewhat plausible.
It also seems plausible that it will take many years until hardware price/performance reaches a point where companies can provide models people deem acceptable at unsubsidized prices people deem acceptable. I don't know.
Take a look at this: https://en.wikipedia.org/wiki/Koomey%27s_law
2.6 years is about 16x in a decade.
How's that Metaverse looking? Bought any NFTs in it?
This is nothing like that stuff. It's incredibly useful, while that stuff was useless.
This is a lot more like the PC or the Internet than cryptocurrency.
The metaverse never had a remote inkling of PMF
You think LLMs are on a par with NFTs?
Huh.
I’ve lived through a few. The internet came and stayed. The bubble just wiped out the speculators, but the internet is bigger than ever. We will be using LLMs for a long time. The only way that’s not true is if a superior replacement comes along.
I've lived through all of them. This is, dare I say it, cope. In some ways I fervently wish it were true - perhaps there will still be some kind of Engineering roles (right now we're enjoying a glut via Jevon's paradox) but the point of the singularity is we can't predict what's going to happen.
What's _not_ going to happen is what's desired by this article which can basically be condensed into 3 bullet points to be appended to your initial spec to Opus 5.5.
I don't want to see this person's dreams crushed, but _log linear charts keep going up and to the right_. I also don't think we're entitled to the high status and pay jobs we currently have 'by right'. We've been useful - we've been _very_ useful and we've done very well for ourselves. We've been the shovels in the gold rush. Now we have to hope we can adapt and that we still provide value commensurate to our pay.
I didn't become a software engineer for the money - I just happened to one day have my hobby become a lucrative career, so perhaps my viewpoint is a little different. All the stuff that is important to the author is still stuff they can do, they just probably won't receive the status and remuneration they think they're entitled to.
None of us have gone back to writing 100% assembler.
We would have. IF a typical C codebase in its early years sucked as much as some AI generated codebases do today.
But instead, humans preferred the productivity AND the readability AND maintainability that elevated us above assembly. Maybe AI-generated code will get there also.
As someone maintaining and using agentic AI developed codebases, they’re better maintained than human stuff I worked on.
For example, we rigidly require specs that can be used to drive end to end tests. Commits can’t break builds or tests. 100% test coverage required for both unit and integration/E2E tests.
And all of this alongside adding macOS/iOS/Android to our standard builds instead of just web/electron. Oh. And moving backends to Rust/Go.
the author's web-dev job only ever existed because we didn't all stop using the web after the dotcom crash
Did things ever go back? Like did the PC industry after 2001 go back to the way it was before the Internet?
Assuming you lived through those cycles, can you tell us why?
One might also notice that the internet did come to dominate our modern world, in spite of the dot com crash. A crash does not equate the end of these tools.
I really cannot wrap my mind around the idea that LLMs will just not be a part of everyday life in the future — in tech and everywhere else...
"People will go back to reading print newspapers, just you wait"
Why? Same reason as with most people about most things: because their income depends on it.
This is something I keep asking myself: why am I not scared for my job? Do I really believe I'll just manage to survive? Maybe I do?...
I can't honestly say so, though.
Yet, I can say I am so excited for the future! For once, I don't want to wake up in the future — I'm loving being at the heart of it, seeing it progress and watching it allow people to be more creative and more empowered.
This is something that genuinely baffles me.
Some people I know also seem to believe this? Like all the AI providers are going to go bankrupt and we'll just go back to what there was before? Do they not realize that indeed the productivity gains are there?
Sure, there's a lot to be debated about how this fell on many people's laps, on how they are being asked to do more, with the same pay and sometimes putting in more hours. These are pressing issues that have emerged as a natural (or, well, maybe forced for some) consequence of a complete shakeup in the world in general. But I don't think society (i.e. most people) will go back to what was before unless we actually have some form of devastation that really physically prevents us from doing so.
If OpenAI and Anthropic and what have you go bankrupt, others will continue. Maybe the economics will look a bit different, but to think we will just go back to what it was seems, to me, completely nonsensical. But maybe I'm the one spouting nonsense. We'll see in a few years...
Part of the current environment is the hype around how far this can go. Investment in anything else is rough because of the ballooned hardware costs and the fear that AI will eat your lunch. That's not a reality necessarily. That part of it could easily change if the mood of the investment class changes.
As far as what AI actually provides, at the end of the day its cheaper software faster (or that's the promise). You still need to build products and maybe you can do that with less folks now. You're not writing assembly today. The machine does it. But you're still working on software. Maybe with AI you're programming less and working on UX more. Is that earth shattering change?
We've yet to hit a point where you any of the big software companies have been overthrown by an AI startup... Whats the revolution really? Its captured the mindshare of the industry but we're at a point of being practically disgusted by AI products already.
Things won't literally go back to 2019. But there could certainly be a return to normalcy in store.
Funny, I went to read the article, and I found it a lot less defensive than replies such as yours. The author even makes some good points.
The main issue for front end development is that companies (particularly large ones) have been releasing shitty broken UI/UX for so long that this stuff became the norm.
Either way the article does not even explain what "blows over" mean. A lot of people seem to be fearfully reacting to the title alone.
i so desperately want this to be true
i've held on hope for a year, but can no longer
we're a non-unionized industry up against very powerful people who want to deskill the field
i want hope, if you have any send my way.. unfortunately things are not looking good
> we're a non-unionized industry up against very powerful people who want to deskill the field
And have tools to do so
I have to say, the other day I tested out ChatGPT sites. Around a year ago I built a website for a rehearsal space - a booking system app - and spent a whole day with ChatGPT building it. It was janky, ugly, and getting there was a process. FWIW, traditional systems dev is what I learned, but I know fuck all about front-end, and I'm not too well versed on web dev in general. I did have to get under the hood and do some coding after I ran out of usage with ChatGPT back then.
But with sites? I made a big picture spec sheet, and asked ChatGPT to expand on that. Then I fed that spec sheet into sites. 30 minutes later it spit out a perfectly working site of what I had tried to make a year ago. Not only that, it was much richer, and had more features. I asked for around 10 changes and new features, and all but one worked flawlessly. One time it managed to break functionality, but it was swiftly fixed. I didn't write or edit a single line of code. NOT A SINGLE ONE.
I've previously been on the buying side, and this was a website which could have easily ended up costing $10k-$15k with all the functionality. If not more. And I ended up getting that with my measly $20 subscription.
Web dev as a career is not dead, or dying, but I predict that A LOT of the people that would make those $0 - $10k sites are going to have a tough time ahead.
EDIT: This was also on par with my general experience with AI. For the longest time you can use frontier models, be unhappy with the results and tell yourself this sucks - and then out of nowhere some updates comes along, and suddenly everything just works perfectly well.
I've noticed the same.
A year ago, I asked ChatGPT to design me a 3D model for a wallet, the kind you can stick to the back of a phone with a magnet. It screwed it up over and over, never got to anything usable.
Last week I tried again, just explaining in words what to do, and it made what I had in my mind.
For websites, it's over. Any kind of brochure/simple restaurant booking site, that was easy years ago and trivial now. SaaS is going next, of course there's a lot of types of SaaS but your average "web implementation of commonly needed service" is now very easily copied with AI.
Anything like a task list or information management system with no particular constraints in terms of performance or scale, you can just make an AI customize for you.
> Web dev as a career is not dead, or dying, but I predict that A LOT of the people that would make those $0 - $10k sites are going to have a tough time ahead.
Not even the first time this happened. When I was in high school you could make great money building websites for small businesses that nowadays just have an Instagram account and no website. The number of CMSes I've built in my career is too damn high
I was able to break into software by kinda exaggerating some freelance work I did when I was a teenager. Mostly doing websites for local business. Surely not an unfamiliar story here.
While everyone has been decrying site-builders for years, I recall a conversation I had around 16 with a small business owner.
I was trying to make some money and convince this guy to let me make him a site. What he said has stuck with me. It was something along the lines of:
"Why do I need a website. Most of my customers find me, contact and and engage with me on Facebook. I need someone who can manage a social media page"
Of course, that definitely was not my forte.
But but but Facebook is the devil.
> Around a year ago [...] It was janky, ugly, and getting there was a process.
Last December/January seems to have been the tipping point from too flawed to be really useful into something pretty good for dev work, so this seems about right.
Hardware is going to get cheaper (as it always does over the long term) and even if the commercial models enshittify¹ to something too contaminated by ads to be bearable the open models won't. We're not going back.
For now it looks like the ability to lead these tools to good solutions is valuable. I guess we'll see if it carries on so. Not much we can do if it isn't except move with the times.
¹ I'm actually fairly optimistic about that at the moment because so far we are the customers.
[dead]
> Accessibility has always been a foundation of web development
> CSS should be hard
I don’t see how the author can square these two. If accessibility is important, the CSS which enables it should be easy. And what about accessibility from a web dev perspective? If doing the right thing requires arcane knowledge do you think most people will buy a book on CSS to master the thing that AI is actively de-skilling?
In my opinion, the author needs to take their own advice. Talk to actual people who need web dev work, figure out what they want to pay for that can’t easily be done by machines, and get good at that.
"the web is a human creation for human needs"
Most web traffic is robotic. The internet is hostile to humans now. Apps gatekeep and monitor our access. LLMs build it. Nothing about the internet is evolving in the direction of a structure created by humans for humans.
Maybe it's primed for disruption, who knows?
Perhaps the disruption is to something even worse.
i don't think we're going back to pre-ai era anytime soon.
My own experience, probably similar to a lot of people's own experience is that i haven't written code by hand (at least 10 lines) in 2026, i can work with any web/backend framework/language because LLMs can.
I'm now an architect / team leader, i manage a team of devs (LLM) and i mostly spend my day providing guidance/suggestions and making architectural choices.
I've never built things that fast, it makes no sense to hand write code anymore as it's so slow. When i run out of tokens, i just wait (doing some reviews, code reads, etc ...), but it doesn't make sense to spend a day writing some code that gonna be generated in a few minutes when tokens are back ...
Obviously this is my personal experience, there's other industries / use cases / etc ... where hand writing code makes sense. But it's not my case anymore.
That's like being the amish of web workers. Good for you if you can, you may very well fit a useful niche. But most of the developers probably don't even want this.
Looks like OP is never returning from the Winchester.
llms / ai have already changed the landscape of tech, it's reaching into almost all industries.
a good comparison is like using a sword vs a gun. ofc in certain scenarios the sword is going to be the right tool / better than a gun. but 99.99% of the time you're gonna want to use the gun and refusing to do so bc it's different or bc you mastered the sword and have some pride isn't going to end well for you.
i find coding agents etc to be really useful and am excited about it. the thing i'm waiting to blow over is for a good work / life / human centric balance to be accepted as the norm. i'm tired of coding agents / ai being the reason behind "if you aren't shipping a critical feature like handling payments in 3 days, you're trash"
> i find coding agents etc to be really useful and am excited about it. the thing i'm waiting to blow over is for a good work / life / human centric balance to be accepted as the norm.
Job is a job. There are few reasons a new technology would change employment for better, and also quite some reasons it could change being employed for worse.
When it blows over? What is the post current phase scenario exactly?
It won't be 2019, that I am sure. There are brain worms around, sure. But this genie is not going back in the box.
This dude is DEEP into the denial stage. Adapt or die, we're(the world) is not going backwards.
I don't know, one thing going for him is he's actively learning by doing. A lot of people appears to have mentally checked out and just press the "Solve this ticket" button and never learn anything new now. Programming skills might still be valuable down the line, he's betting on that I suppose.
I think that LLMs will be just another tool like linters or build tools. Enterprise and tech startups will still need custom solutions and people accountable for keeping them afloat. However the guy who got by making small business websites from templates will probably have a hard time. Between squarespace, shopify, and LLMs there just isn't a reason to pay more than a few bucks for a website for your business.
First, if you crack open a history book (or have an agent do it for you), you'll notice that the world goes backwards quite often. It's not pretty, but the critical mass of people who feel they have nothing to lose is pretty modest.
Second, I'm really curious what's your "adapt" path, because I suspect you're in denial too. Accept that, before long, an LLM will not need you to write good prompts, tell it to "keep going", or say "no, that's wrong". What's your value then? Does your employer keep you on the payroll as charity? For old times' sake? If your value is just supervising AI, then your value is declining rapidly.
People in unskilled professions don't have choice when automation shows up and takes their jobs. Software engineering may go down in history as one of the few highly skilled professions that didn't just keel over and die, but enthusiastically embraced its own demise.
> First, if you crack open a history book (or have an agent do it for you), you'll notice that the world goes backwards quite often.
The only thing I can think of is the reduction of truffles as an ingredient, which apparently happened around WWI. We eat a lot fewer of them than before.
> Second, I'm really curious what's your "adapt" path, because I suspect you're in denial too. Accept that, before long, an LLM will not need you to write good prompts, tell it to "keep going", or say "no, that's wrong". What's your value then? Does your employer keep you on the payroll as charity? For old times' sake? If your value is just supervising AI, then your value is declining rapidly.
Your employer cannot talk to all the LLM sessions at the same time. He also can't write a prompt that will do it for him.
> People in unskilled professions don't have choice when automation shows up and takes their jobs. Software engineering may go down in history as one of the few highly skilled professions that didn't just keel over and die, but enthusiastically embraced its own demise.
It's changed several times over the decades. How many people are working on a mainframe nowadays? How many people are writing assembly? Look at how many languages have come and gone too. LLM is a big change, but that doesn't mean software people will be gone.
Having AI “blow over” is overblown but OP does have some suggestions (except maybe CSS) that should be AI-resistant.
What sphinx of cement and aluminum…
Dude is overdosing on copium
"When all this blows over" you will still have open source models that are at least as capable as right now, probably more.
Qwen3.8 on a large laptop can create sites, CSS, and basic JS/React wiring.
No kidding. The world’s never going back to a state where $10k doesn’t buy me enough hardware to run very capable agents.
You can get an Intel Arc Pro with 48G of GDDR6 for around $2000 new. That'll run up to about a 40G model at 8-bit or 80G models at 4-bit, and at probably like 100 TPS.
When the RAM bubble pops you'll be seeing cards with 4X this amount of RAM for that price.
So… then agentic AI will be even more ubiquitous and accessible.
This makes me so nostalgic for blockchain evangelists when people started to sour on the tech. Enthusiasm slowly morphed into desperate proselytizing and finally outright scamming, with the NFT clown show as a fitting headstone.
No new governments on the DAO, no ubiquitous programs running on the chain, no magical new answer to “fiat”… just a lot of empty promises and waste.
No one here ever seems to learn.
I don't personally know anyone who bought crypto for anything other than investment/gambling purposes. I know a LOT of people who have Claude or ChatGPT subscriptions because they find them useful.
Is your thesis that AI is useless? Because that seems like fantasy to me.
There is maybe a case to make around the economics of AI companies, but even there I'm very skeptical because the open models are quite capable on relatively inexpensive hardware.
My thesis is that his place is full of hype waves as people try to align themselves with the expectations and demands of the wealthy people they want money from. Who you personally know isn’t relevant to the question of how HN has reacted to this series of fads.
ML isn’t useless, that’s been obvious for a while, but it’s grossly overhyped and people here who think their vibe-coding isn’t ultimately worthless are delusional.
[delayed]
A sustainable phone operator career, for when all this “world wide web” thing blows over
beautiful site though
Honestly the best red pill advice is:
1) Become world class experts in one thing 2) Become full on invested in LLMs learn on how to match general intelligence on business requirements (Evals, isolation, judging,etc..) I think this will probably be still 5 years 3) Get ready to do something in real life.
> Become world class experts in one thing
Good luck, unless you're very high IQ or a wealthy teenager.
Also you have to become a world class expert in something deemed valuable.
This was...not a great read
I wish I could agree, but AI (as LLMs) is absolutely not blowing over. Even if frontier labs crash and burn I'm already running a model that's maybe only a month behind them, locally, in my house. Last year this was impossible within a reasonable budget. Next year it's going to be better and cheaper to do. My phone can already run models as capable as last year's. Horse has left the barn.
Terrible advice in this article.
React is out and no one should learn it? Really? Something like half the Javascript job listings mention React ... but someone trying to get a job should not learn it?
Now to be fair I haven't job hunted in about half a year, so the market may have changed a little now that I have my (React-based) job ... but it hasn't changed so much that the most popular web framework has somehow become dead tech.
The point isn't that React is dead tech. React is practically every AIs front-end library of choice. I think the idea is AI is too good at it and you won't get a good ROI learning it. Learn fundamentals instead and when folks get sick of React because AI loves it, you'll be ready.
No it's all still React
Even CLI apps are now written in React...
Frontend development will be the first that will be completely replaced by agentic development where nobody needs to see any code at all. Its already possible and many companies are having this way of working quite succesfully. The quality is already better than 99% of what frontend engineers can achieve. Speed is just insane. I can create any ui. Whether it's web or platform specific. There are hardly any limitations. Frameworks are still useful. But I expect them to become less important. What remains important as a dev is system knowledge. Understanding of how systems are tight together. The principles of distributed systems, security, prompt engineering, agentic automation. And that kind of knowledge. Code syntax is from the old era. Maybe there is a niche for it. But not for the majority.
Right. My clients now get iOS, macOS, Win32 (or whatever you call it), and Android apps… for free. Same price they paid before, and now they just get it all.
Bro, I think you are too deep into tech twitter (X). React is still king , almost every job opening for web needs it. Even the ones at OpenAI and Anthropic.
Oh boy
Didn't know DHH was a fucking racist, but I'm not surprised. Never heard about Guillermo Rauch before this post, and now I wish I hadn't.
Reads like copium to me.
ok, the politics sections is a bit much lol
lol
[dead]
"DHH is a racist!"
Dude, it's not 2021 anymore. You can't just yell a big, bold, name like racist and expect it to stick or cause any impact. Everyone got burned out in 2021 of the million -ists, and the 2024 election showed among other things that the name calling has zero effect.
The world belongs to those who can actually explain why any particular opinion is a bad thing, or conflicts with values that the reader may otherwise hold, not those who can accuse with a cheap label.
Don’t try to shout people down. If you think someone is wrong, go read DHH’s writings on race and come back and defend them.
I think anyone reading DHH’s blog would come to the same conclusions — particularly the posts on wolves and Roma, or the post on London
I was skeptical so I actually read those posts and concluded that yes, they are indeed unequivocally racist (in the traditionally accepted sense of the term).
I don't care whether or not DHH is a racist. He can't stop his project from being forked or exert authority over anyone.
I care about people who think that just yelling "he's a racist" somehow scores them brownie points or does something about it. It doesn't. You're not brave, you're not noble, you're not accomplishing anything.
You seem to be ignoring the possibility that the author genuinely believes this.
I am close to a free speech absolutist (i.e., I don't believe that DHH should suffer legal consequences for those posts, unless he were directly inciting violence). I also believe that those who openly promulgate racist beliefs should suffer social opprobrium. That is what makes government censorship unnecessary.
You either don't think he's a racist or you don't care he's a racist. You can't have it both ways expressing an opinion while also being above it all.
Why is all of your personal opinion on if people are doing it for "brownie points" important and not just being pompous?
Why do you care so much?
What a bunch of nonsense.