The analogy to MLM doesn’t work. In a pyramid scheme, early investors are guaranteed money, whereas late investors are guaranteed losses. With a vibe-coding platform, everyone has the same (extremely low) odds of building a hit app, so is at least more equitable in that sense.
I think the point the author is trying to convey, is that they are similar to MLM schemes in that they profit off of peoples hopes and dreams. Altough there are MLMs that are pure pyramid schemes, there are some where it is also not impossible to earn some money with it. Their products do not even have to be downright bad or fake. Having read Daniel Kahneman, I suppose with (economic) desperation people become especially suceptible to hear the siren's call of such marketing. What would have made the comparison to MLM more apt tough, is if there would be a thing like: you buy X subscriptions and may resell them, or if you have Y people that you convince to also become subs/tokens sellers you also get a cut.
Not trying to be overly cynical, but isn't that just marketing?
A lot of marketing just getting people to picture a better future version of their life and then making them think that your product will get them there. They're not actually buying the product for the product, they're buying it to try to get that imaginary future. I don't really see how Repilit ad telling people they can built the app of their dreams is very different than a gym ad telling people that they can get ripped or something like The Container Store showing someone with a messy house magically getting organized and cleaned.
I'm not saying that any of those examples are particularly good or moral, but I don't get how what Repilit is doing is any different than just standard marketing tactics we see every time we watch a block of ads.
In the author’s example, Replit has a very high chance of making a profit on those folks’ desperation.
It’s not exactly an MLM. But the predatory mechanism is close. Loan sharking might be a more exact analog for the financial bit, but the social-media marketing strikes closer to MLMs.
It's similar to dating app companies selling to young desperate men, although these companies have much more insidious tactics (sending you a match right when they predict you might give up on the app) than e.g. Replit who have still not fully figured out how to hook users just enough to continue to get them to pay (not sure, though, how they can give you a hit app right before you give up on them).
Source? We would do panhandling trips as Boy Scouts in the California hills. On each outing at least someone found gold. Which means every one of us would have eventually found some given a few days.
Most panhandlers didn’t break even. But I’m challenging even that threshold for Replit. I’m guessing most of their vibe-coded businesses never make a penny of revenue.
There is already the very classical grift : you don't sell genai content, you sell "get rich quick with genai" methods. This is where the recursion starts.
I think it does. MLM is multi-level-marketing and the shovel-selling exactly IS a pyramid scheme.
You have to look more deeply into the scheme. It appears that they sell the SaaS, but what they are actually selling is the selling/referral system itself which makes it exactly a pyramid scheme.
Just like the tupperware was never the money-maker, is the referral system.
Let’s make it clear right away: none of them are building a “hit app”. Maybe one in a million does, but it’s the same as winning a lottery. TikTok grifters will just siphon commission for the video and move on. It’s a clear MLM, you can see it with peptides too.
Selling to desperation is not the same as an MLM. TikTok grifters don't expect their watchers to then go out and grift on TikTok themselves while giving the original person a cut, there are no levels to this sort of grift, as would be required in a multi level marketing scheme.
Most TikTok products I know are not of a multi level format, the originators do not care if you buy their course then make your own course as they see none of that money beyond the initial sale.
I can't help but wonder if this is really so different from an ad for any tool. If you see an ad for say kitchenware, it probably contains a lot of beautiful food with the implication that if you buy the pan you too can cook great food. Of course that takes skill and the frying pan doesn't matter too much.
I guess im wondering where is the line that makes this evil but my made up frying pan ad just a harmless exageration.
Pretty simple really: do the ads for the frying pan make the direct claim that the pan will put you on the path to becoming highly paid chef? Perhap opening a Michelin-starred restaurant with the skills your new pan has unlocked?
"you can cook better food" and "you can drastically improve your economic situation" are two very different promises (or implications) that should be held to very different standards.
> [...] serving me ads for Replit — one of the many vibe-coding startups that have emerged in the past couple of years,
Replit (formerly repl.it) has been around for a decade as an online IDE / runtime, productized from work from 2009 for editing environments on udemy and co.
But like many companies they pivoted their marketing schtick to AI because that's where the money is. Cursor and Zed are editors, but because they capitalize on the AI hype and investor money they're "worth" tens of billions.
I mean Zed is cool and all because they dared to start a new editor from scratch with performance in mind but that in itself isn't tens of billions.
Here "Vibe-coded software is simply not good." is not also correct 100% why?
If all you saw means that's all there is then that itself is incorrect, why? People are building better solutions and you just havent seen them yet.
I know so because I am building one, you didn't know about it, if you are interested in seeing a sample of what it produces, the architecture etc, though not perfect but it should cause you to rethink because some of us are building the solutions to the problem you saw, you just didn't know it yet.
It's who has funding support around them that you hear of and see often.
That is simply not true. It can be better or it can be worse - depends on who directed it.
I understand where the point comes from, but someone who has coded and architected a lot of applications for many years, does get the good side. But a user who see code as an alien language - they are ultimately going to get the bad side of it.
There was a theory floated around by an youtuber (and a tech geek), on how to vibe code better - and how to let agents run the show. I tried, more than once - it failed badly. Not failed at the output or the UI - failed at writing good and well architected code.
> What happens when things go wrong
- this is the most important question - can the human step in?
For me the answer is a unequivocal yes. I may not be able to fix it in 10 minutes, but I know I will fix it in 10 hours or 100 hours - whatever it takes.
But when a user who "can't read code" comes in - and asks me to fix their problem, it is going to cost them a lot more than their total subsidized vibe coding tool cost. They're going to be like - the app cost me 100-200$ to vibe-build, but the dev is going to charge me 5-10x for a 2 line fix.
For some the decision will be like - better buy a new phone than repairing the old one, for others - they can't replace things easily.
What used to take 1.5 years to build 10 years ago, and 6-9 months to build 5 years ago, takes 1.5 months or faster to build today (if it is done with the same rigor).
> The GDPR example
How is it different from having a human dev team hired? The CEOs or founders are responsible - they can't go and say "that dev did the wrong thing, fine them" - will you work for such a person?
> the belief that AI can — and will — displace white-collar jobs is a lie
It already is displacing, unfortunately. It has been taking apart both jobs and businesses - one role at a time - within 6 months of AI coming out. Some are experiencing it now, some have experienced it earlier, some will experience it later.
For example - a good tech guy in finance domain and having good domain knowledge - gets fired. After a while, he will end up competing for jobs in the finance domain - because he needs to survive. The domino effect will be seen. And hope it does not become a race to the bottom.
And new roles are likely to come up and stabilize - but the bar will be high and you will need AI all the time. Otherwise you will be seen like ploughing the farm by hand instead of using a tractor.
> I don't like LLMs because:
>
> [ ] They steal from artists
> [ ] They're destroying the planet
> etc
Instead of these articles, just tick the boxes. Only write the article if there aren't ten repeating the same points over and over again. Are there any HN readers who aren't intimately familiar with the problems associated with LLMs?
It's just HN's Two Minutes Hate (extended). It's wise to filter out stuff that's too hot for too long on this site to focus on interesting stuff, as at this point nuanced views aren't to be found.
Of course it's all within legal limits (or at least pre-legal, as the AI people call it[0]), but it smells like MLM. They'll stretch it as far as they can until there is push-back.
Hey, don't knock Herbalife, I made good money from that!
When my neighbour decided to pack it all in, he paid me 50 quid to pull his car into my workshop and take all the branding vinyl-cut off. Less than an hour for me, he'd been at it for days.
> in part because of the huckster-like triangulations of scumbags like Sam Altman and Dario Amodei
I would happily read an AI-critical blogpost if it weren't clearly motivated by a strange, specific hatred of the prominent AI figureheads.
At this point I automatically dismiss writing like this, the motivated reasoning is palpable. Their distaste for the character and general vibes of the AI industry trap them in blatant denials of reality, like claiming that AI is a completely worthless technology or surely the bubble will pop any minute now.
I am all for well-researched criticisms of these companies and their claims, but please start at the facts and use them to derive your conclusions, rather than the other way around.
I think in the short term much of the concern for AI replacing jobs was overblown, largely due to two factors
1. Benchmark performance of LLM's and AI models did not fully represent skill in real-world domains
2. Most jobs span far more requirements than their specific job descriptions, many of which lie, even in simple jobs, in the realm of highly adaptive, context rich multi-modal information processing that most humans do still better than AI
However, there is nothing fundamental that prevents LLM's from scaling and improving, aided by better scaffolding, to the point of replacing many white-collar jobs, especially ones which have limited, specific requirements and output parameters. This is an enormous chunk of the white collar work force, and displacement is already happening in limited sections, and will surely continue as AI capabilities diffuse.
It seems however deeply entrenched in many people's identity to deny this fact, because to accept it requires accepting that many of the essential claims of AI CEO's are somewhat true to a degree, and that LLM's are a genuinely useful technology.
We're what, 2 years in? Or 6 months since the coding models became reasonably proficient. No one knows what AI will look like 6 months from now let alone 5 years so the present jobs data can't tell you anything other than that nothing catastrophic has happened yet.
No, but we know what it’s looked like over the past 6 months. We also have two years of Altman preening about imminent, massive job losses. Per his own timelines, those haven’t manifested.
Predicting indeterminate catastrophe isn’t a prediction, it’s a scam.
And its a quite insane thing to push, even if true - AI spend has outgrown wages (even in software) to such a degree, that basically if AI allowed companies to fire every human employee, they wouldn't be saving that much.
"Basically" is doing a lot of work there. What are the profit margins at FooCorp, and how much time and money is spent on messy human drama? If you could replace every human employee with AI, then you could fire the entire HR department and replace it with one AI "HR" that just spins up new AI employees like we do EC2 instances today. Since there are no more messy human resources to manage, anyway. Shutting down the entire department should boost profits by quite a bit!
Agreed but the converse is also true. Predicting that everything will be sunshine and rainbows merely because the sky hasn't fallen yet doesn't pass the sniff test.
> we know what it’s looked like over the past 6 months.
Oddly I seem to interpret that in the opposite manner that you do. The output became noticeably more cohesive sometime around the new year although it certainly still has plenty of shortcomings.
> Predicting that everything will be sunshine and rainbows merely because the sky hasn't fallen yet doesn't pass the sniff test
I’m not predicting it will be fine. I’m saying we have enough evidence to cast the opposite prediction, made on specific timescales in the past, as BS.
> I see no argument whatsoever in the piece as to why this is the case, just an emphatic declaration.
We’ve heard Claude is good enough now that you no longer need SE and founders will just do their coding for a year already. It hasn’t happened. And guess who develops CC and other Anthropic/OpenAI products?
People who talk about the masses are at the same time completely distanced from them but also completely tuned in to their mass psyche, or so they claim.
Most reasoning outside of philosophy departments is motivated and it is fine as long as it is either reflected on or the audience knows where they are coming from. This is just having skin in the game. You know, needing a job to survive, which is much more immediate than the motivation of society disruptors who either get to become fantastically rich or else move on to some other executive role.
I would also have wished for some substantiation for how this and that were a lie.
As it stands though it’s an argument against multi-level marketing. And it really doesn’t hold up that your everyman will be able to make money off of vibe coded apps. Maybe build their own very personal software gadget? Yeah, but there’s no money in that.
The Reasoned Case Against AI Disruptors does not need to be covered in such a piece.
> I would happily read an AI-critical blogpost if it weren't clearly motivated by a strange, specific hatred of the prominent AI figureheads.
The thing is... many prominent figures, both individuals and companies, in the AI industry just lend themselves to being hateable.
Sam Altman personally completely fucked up the RAM market with his double dealing. Every single one of us felt the consequences of that and we will feel it for years to come. And it is why I will call for his arrest, speedy trial and imprisonment every time I have the misfortune of reading his name. I wish to see this person suffer from the bottom of my heart.
Elon Musk, well, there are so, so many valid reasons to hate him. Regarding AI itself, the mechahitler incident and the "undress her in a bikini" CSAM generator are the worst issues. Regarding him as a person, the "pedo diver" incident, the shady stuff surrounding virtually all of his children's mothers, the complete clusterfuck around his daughter, DOGE, the right-arm salute his fanbase keeps denying being a nazi salute, him stoking racial riots in the UK twice, his constant overpromising in all of his ventures (some of which would normally fall under "investor defrauding" claims if there were a functioning legal and regulatory world), the trashcontainer on wheels...
Google has no (notable) individual persons to raise the pitchforks against, but as a company, they severely degraded the quality of their "ordinary" search and have gone to steal clicks and thus money from creators by distilling their work into the AI results box at the top of every search.
Microsoft keeps shoveling AI down everyone's throats no matter if we want it.
Anthropic has literally ripped books apart to scan them for Claude. Google, back when they created the dataset for Google Books, at least didn't destroy the books. Destroying books at that scale is a sacrilege.
Every single AI company is guilty of using questionably sourced materials - either outright stolen or human input based on exploitation.
And on top of that, it's not just RAM that has gotten expensive. The entire rest of the economy - both individuals and companies of all sizes - are priced out of personal and even cloud compute, as the blown-up AI giants scoop up everything they can and the scraps and aged hardware that's available gets fought over by everyone else.
> blatant denials of reality, like claiming that AI is a completely worthless technology or surely the bubble will pop any minute now.
It is undeniable that the entire AI sphere is a bubble, artificially propped up by circular investments and wash trades, and that is now poised to raid pension funds.
> Vibe-coded software is simply not good. Let’s suppose that someone deploys an app and there’s a critical security vulnerability that allows a threat actor to, say, exfiltrate all their customer information. How would they know? And if they became aware of it (presumably because said threat actor exploited said vulnerability), how would they fix it?
Is all non-vibe-coded software "good"? Are so-called real companies (with professional software development and IT staff) impervious to these threats?
> Also, would the person who developed the app know that, under legislation like GDPR, they can be financially liable for data breaches? Because they would be! And the whole point of the financial penalty system (at least, with respect to GDPR) is to be dissuasive — to act as a deterrent to other people who would be cavalier with other people’s data.
> I can very easily imagine a national data protection authority — like the UK’s ICO — giving someone a massive, massive fine in order to dissuade other people from deploying their own AI-generated, unvetted slop code.
Thousands of companies have been hit with GDPR fines, including some of the biggest companies in the world. Why the apparent assumption that vibe-coders are any more cavalier about people's data than companies that in many cases exist solely to profit from it?
I think you can make a legitimate argument that companies selling vibe coding dreams to laypeople are selling something generally unrealistic but the tone of this person's article seems like gatekeeper bait. It feels like he just doesn't like the idea that non-engineers might try to use tools (oversold or not) that allow them to do things he thinks non-engineers shouldn't do.
It's very much a "keep out", "stay in your lane" vibe.
It doesn't fucking matter to the success of a business.
I spent much of my early career unfucking large codebases that had been thrown together by sysadmins or teenagers or HTML guys who knew a bit of Perl on which an enterprising person had built very, very successful businesses. The software got fixed by pros long after profitability when it started to matter.
There are very few businesses where the quality of the software makes any real difference. What matters is execution, marketing, commercialization, but programmers see every business problem as a technical problem requiring technical excellence, because they're gigantic hammers.
I think it does matter. Bad software can make peoples lives worse, especially when they depend on it. Really bad software can kill people. Maybe it doesn't matter to the bottom line of a business, but that's a particularly low bar to meet.
Someone in a genuine health crisis seeking help but can't receive it could have deadly consequences. It doesn't matter to the business, but it does matter to people. Life is messy and complex, and if our software doesn't work correctly it does add to the suffering of others. Maybe it pushes some people past the point of no return. There are consequences to what we do, good or bad.
This is what I fear most about the rise of vibe coding. Businesses profit, people get hurt, and the incentives are all wrong.
The analogy to MLM doesn’t work. In a pyramid scheme, early investors are guaranteed money, whereas late investors are guaranteed losses. With a vibe-coding platform, everyone has the same (extremely low) odds of building a hit app, so is at least more equitable in that sense.
I think the point the author is trying to convey, is that they are similar to MLM schemes in that they profit off of peoples hopes and dreams. Altough there are MLMs that are pure pyramid schemes, there are some where it is also not impossible to earn some money with it. Their products do not even have to be downright bad or fake. Having read Daniel Kahneman, I suppose with (economic) desperation people become especially suceptible to hear the siren's call of such marketing. What would have made the comparison to MLM more apt tough, is if there would be a thing like: you buy X subscriptions and may resell them, or if you have Y people that you convince to also become subs/tokens sellers you also get a cut.
> they profit off of peoples hopes and dreams
Not trying to be overly cynical, but isn't that just marketing?
A lot of marketing just getting people to picture a better future version of their life and then making them think that your product will get them there. They're not actually buying the product for the product, they're buying it to try to get that imaginary future. I don't really see how Repilit ad telling people they can built the app of their dreams is very different than a gym ad telling people that they can get ripped or something like The Container Store showing someone with a messy house magically getting organized and cleaned.
I'm not saying that any of those examples are particularly good or moral, but I don't get how what Repilit is doing is any different than just standard marketing tactics we see every time we watch a block of ads.
> I think the point the author is trying to convey, is that they are similar to MLM schemes in that they profit off of peoples hopes and dreams
Every scam plays to people's hopes and dreams. As do most legitimate opportunities.
> everyone has the same (extremely low) odds
In the author’s example, Replit has a very high chance of making a profit on those folks’ desperation.
It’s not exactly an MLM. But the predatory mechanism is close. Loan sharking might be a more exact analog for the financial bit, but the social-media marketing strikes closer to MLMs.
It's similar to dating app companies selling to young desperate men, although these companies have much more insidious tactics (sending you a match right when they predict you might give up on the app) than e.g. Replit who have still not fully figured out how to hook users just enough to continue to get them to pay (not sure, though, how they can give you a hit app right before you give up on them).
I think the standard analogy would be selling shovels to the gold miners.
> standard analogy would be selling shovels to the gold miners
There was actually gold in the California hills. Nobody is starting a business the way the Replit ads seems to be pitching.
This one’s a simple pump and dump before the attorneys general get wiser.
I'm sure someone somewhere at some point has made some money on replit. There is always an outlier.
Keep in mind the vast majority of gold miners found no gold.
> the vast majority of gold miners found no gold
Source? We would do panhandling trips as Boy Scouts in the California hills. On each outing at least someone found gold. Which means every one of us would have eventually found some given a few days.
Most panhandlers didn’t break even. But I’m challenging even that threshold for Replit. I’m guessing most of their vibe-coded businesses never make a penny of revenue.
There is already the very classical grift : you don't sell genai content, you sell "get rich quick with genai" methods. This is where the recursion starts.
MLM — multi-level marketing
I think it does. MLM is multi-level-marketing and the shovel-selling exactly IS a pyramid scheme.
You have to look more deeply into the scheme. It appears that they sell the SaaS, but what they are actually selling is the selling/referral system itself which makes it exactly a pyramid scheme.
Just like the tupperware was never the money-maker, is the referral system.
Let’s make it clear right away: none of them are building a “hit app”. Maybe one in a million does, but it’s the same as winning a lottery. TikTok grifters will just siphon commission for the video and move on. It’s a clear MLM, you can see it with peptides too.
Selling to desperation is not the same as an MLM. TikTok grifters don't expect their watchers to then go out and grift on TikTok themselves while giving the original person a cut, there are no levels to this sort of grift, as would be required in a multi level marketing scheme.
They absolutely do, you’re just looking at the leafs of the tree.
Most TikTok products I know are not of a multi level format, the originators do not care if you buy their course then make your own course as they see none of that money beyond the initial sale.
I can't help but wonder if this is really so different from an ad for any tool. If you see an ad for say kitchenware, it probably contains a lot of beautiful food with the implication that if you buy the pan you too can cook great food. Of course that takes skill and the frying pan doesn't matter too much.
I guess im wondering where is the line that makes this evil but my made up frying pan ad just a harmless exageration.
Pretty simple really: do the ads for the frying pan make the direct claim that the pan will put you on the path to becoming highly paid chef? Perhap opening a Michelin-starred restaurant with the skills your new pan has unlocked?
"you can cook better food" and "you can drastically improve your economic situation" are two very different promises (or implications) that should be held to very different standards.
read the article. it stated clearly why vibe-coded apps do not work. a frying pan does its intended purpose.
> [...] serving me ads for Replit — one of the many vibe-coding startups that have emerged in the past couple of years,
Replit (formerly repl.it) has been around for a decade as an online IDE / runtime, productized from work from 2009 for editing environments on udemy and co.
But like many companies they pivoted their marketing schtick to AI because that's where the money is. Cursor and Zed are editors, but because they capitalize on the AI hype and investor money they're "worth" tens of billions.
I mean Zed is cool and all because they dared to start a new editor from scratch with performance in mind but that in itself isn't tens of billions.
The Replit CEO also harassed a kid on HN who made an open source version of it.
https://news.ycombinator.com/item?id=27424195
It's not just AI, these are the consequences of affiliate marketing. Just look at the crap that is YouTube nowadays.
Here "Vibe-coded software is simply not good." is not also correct 100% why?
If all you saw means that's all there is then that itself is incorrect, why? People are building better solutions and you just havent seen them yet.
I know so because I am building one, you didn't know about it, if you are interested in seeing a sample of what it produces, the architecture etc, though not perfect but it should cause you to rethink because some of us are building the solutions to the problem you saw, you just didn't know it yet.
It's who has funding support around them that you hear of and see often.
> Vibe-coded software is simply not good.
That is simply not true. It can be better or it can be worse - depends on who directed it.
I understand where the point comes from, but someone who has coded and architected a lot of applications for many years, does get the good side. But a user who see code as an alien language - they are ultimately going to get the bad side of it.
There was a theory floated around by an youtuber (and a tech geek), on how to vibe code better - and how to let agents run the show. I tried, more than once - it failed badly. Not failed at the output or the UI - failed at writing good and well architected code.
> What happens when things go wrong
- this is the most important question - can the human step in?
For me the answer is a unequivocal yes. I may not be able to fix it in 10 minutes, but I know I will fix it in 10 hours or 100 hours - whatever it takes. But when a user who "can't read code" comes in - and asks me to fix their problem, it is going to cost them a lot more than their total subsidized vibe coding tool cost. They're going to be like - the app cost me 100-200$ to vibe-build, but the dev is going to charge me 5-10x for a 2 line fix.
For some the decision will be like - better buy a new phone than repairing the old one, for others - they can't replace things easily.
What used to take 1.5 years to build 10 years ago, and 6-9 months to build 5 years ago, takes 1.5 months or faster to build today (if it is done with the same rigor).
> The GDPR example
How is it different from having a human dev team hired? The CEOs or founders are responsible - they can't go and say "that dev did the wrong thing, fine them" - will you work for such a person?
> the belief that AI can — and will — displace white-collar jobs is a lie
It already is displacing, unfortunately. It has been taking apart both jobs and businesses - one role at a time - within 6 months of AI coming out. Some are experiencing it now, some have experienced it earlier, some will experience it later.
For example - a good tech guy in finance domain and having good domain knowledge - gets fired. After a while, he will end up competing for jobs in the finance domain - because he needs to survive. The domino effect will be seen. And hope it does not become a race to the bottom.
And new roles are likely to come up and stabilize - but the bar will be high and you will need AI all the time. Otherwise you will be seen like ploughing the farm by hand instead of using a tractor.
We need a form people can fill out:
Instead of these articles, just tick the boxes. Only write the article if there aren't ten repeating the same points over and over again. Are there any HN readers who aren't intimately familiar with the problems associated with LLMs?It's just HN's Two Minutes Hate (extended). It's wise to filter out stuff that's too hot for too long on this site to focus on interesting stuff, as at this point nuanced views aren't to be found.
For those saying it's not MLM, it's not not MLM: https://replit.com/refer
Of course it's all within legal limits (or at least pre-legal, as the AI people call it[0]), but it smells like MLM. They'll stretch it as far as they can until there is push-back.
[0] https://web.archive.org/web/20260515043739/https://www.revsw...
Hey, don't knock Herbalife, I made good money from that!
When my neighbour decided to pack it all in, he paid me 50 quid to pull his car into my workshop and take all the branding vinyl-cut off. Less than an hour for me, he'd been at it for days.
> in part because of the huckster-like triangulations of scumbags like Sam Altman and Dario Amodei
I would happily read an AI-critical blogpost if it weren't clearly motivated by a strange, specific hatred of the prominent AI figureheads.
At this point I automatically dismiss writing like this, the motivated reasoning is palpable. Their distaste for the character and general vibes of the AI industry trap them in blatant denials of reality, like claiming that AI is a completely worthless technology or surely the bubble will pop any minute now.
I am all for well-researched criticisms of these companies and their claims, but please start at the facts and use them to derive your conclusions, rather than the other way around.
The preceding clause is just as bad:
>> the belief that AI can — and will — displace white-collar jobs is a lie that’s been accepted by the masses
I see no argument whatsoever in the piece as to why this is the case, just an emphatic declaration.
Anyway. The belief that the author isn’t talking out of their ass is an insidious lie spread by dark forces. QED.
I think in the short term much of the concern for AI replacing jobs was overblown, largely due to two factors
1. Benchmark performance of LLM's and AI models did not fully represent skill in real-world domains
2. Most jobs span far more requirements than their specific job descriptions, many of which lie, even in simple jobs, in the realm of highly adaptive, context rich multi-modal information processing that most humans do still better than AI
However, there is nothing fundamental that prevents LLM's from scaling and improving, aided by better scaffolding, to the point of replacing many white-collar jobs, especially ones which have limited, specific requirements and output parameters. This is an enormous chunk of the white collar work force, and displacement is already happening in limited sections, and will surely continue as AI capabilities diffuse.
It seems however deeply entrenched in many people's identity to deny this fact, because to accept it requires accepting that many of the essential claims of AI CEO's are somewhat true to a degree, and that LLM's are a genuinely useful technology.
> see no argument whatsoever in the piece as to why this is the case
It hasn’t happened. We have AI rolling out and the jobs data aren’t showing this effect.
We're what, 2 years in? Or 6 months since the coding models became reasonably proficient. No one knows what AI will look like 6 months from now let alone 5 years so the present jobs data can't tell you anything other than that nothing catastrophic has happened yet.
> No one knows what AI will look like 6 months
No, but we know what it’s looked like over the past 6 months. We also have two years of Altman preening about imminent, massive job losses. Per his own timelines, those haven’t manifested.
Predicting indeterminate catastrophe isn’t a prediction, it’s a scam.
And its a quite insane thing to push, even if true - AI spend has outgrown wages (even in software) to such a degree, that basically if AI allowed companies to fire every human employee, they wouldn't be saving that much.
Are you perhaps confounding inference purchased by clients with R&D and capex invested by frontier providers?
"Basically" is doing a lot of work there. What are the profit margins at FooCorp, and how much time and money is spent on messy human drama? If you could replace every human employee with AI, then you could fire the entire HR department and replace it with one AI "HR" that just spins up new AI employees like we do EC2 instances today. Since there are no more messy human resources to manage, anyway. Shutting down the entire department should boost profits by quite a bit!
Agreed but the converse is also true. Predicting that everything will be sunshine and rainbows merely because the sky hasn't fallen yet doesn't pass the sniff test.
> we know what it’s looked like over the past 6 months.
Oddly I seem to interpret that in the opposite manner that you do. The output became noticeably more cohesive sometime around the new year although it certainly still has plenty of shortcomings.
> Predicting that everything will be sunshine and rainbows merely because the sky hasn't fallen yet doesn't pass the sniff test
I’m not predicting it will be fine. I’m saying we have enough evidence to cast the opposite prediction, made on specific timescales in the past, as BS.
most ai use and spend is still by humans. laying off humans would (ironically) reduce ai spend and use.
> No one knows what AI will look like 6 months
Which means that they were lying.
Or they were wrong. But best to assume malice given their incentives. (No, that certain saying is just stupid.)
> I see no argument whatsoever in the piece as to why this is the case, just an emphatic declaration.
We’ve heard Claude is good enough now that you no longer need SE and founders will just do their coding for a year already. It hasn’t happened. And guess who develops CC and other Anthropic/OpenAI products?
The false part is that the masses do not accept the lie, as most don't actually believe those companies.
People who talk about the masses are at the same time completely distanced from them but also completely tuned in to their mass psyche, or so they claim.
Most reasoning outside of philosophy departments is motivated and it is fine as long as it is either reflected on or the audience knows where they are coming from. This is just having skin in the game. You know, needing a job to survive, which is much more immediate than the motivation of society disruptors who either get to become fantastically rich or else move on to some other executive role.
I would also have wished for some substantiation for how this and that were a lie.
As it stands though it’s an argument against multi-level marketing. And it really doesn’t hold up that your everyman will be able to make money off of vibe coded apps. Maybe build their own very personal software gadget? Yeah, but there’s no money in that.
The Reasoned Case Against AI Disruptors does not need to be covered in such a piece.
> I would happily read an AI-critical blogpost if it weren't clearly motivated by a strange, specific hatred of the prominent AI figureheads.
The thing is... many prominent figures, both individuals and companies, in the AI industry just lend themselves to being hateable.
Sam Altman personally completely fucked up the RAM market with his double dealing. Every single one of us felt the consequences of that and we will feel it for years to come. And it is why I will call for his arrest, speedy trial and imprisonment every time I have the misfortune of reading his name. I wish to see this person suffer from the bottom of my heart.
Elon Musk, well, there are so, so many valid reasons to hate him. Regarding AI itself, the mechahitler incident and the "undress her in a bikini" CSAM generator are the worst issues. Regarding him as a person, the "pedo diver" incident, the shady stuff surrounding virtually all of his children's mothers, the complete clusterfuck around his daughter, DOGE, the right-arm salute his fanbase keeps denying being a nazi salute, him stoking racial riots in the UK twice, his constant overpromising in all of his ventures (some of which would normally fall under "investor defrauding" claims if there were a functioning legal and regulatory world), the trashcontainer on wheels...
Google has no (notable) individual persons to raise the pitchforks against, but as a company, they severely degraded the quality of their "ordinary" search and have gone to steal clicks and thus money from creators by distilling their work into the AI results box at the top of every search.
Microsoft keeps shoveling AI down everyone's throats no matter if we want it.
Anthropic has literally ripped books apart to scan them for Claude. Google, back when they created the dataset for Google Books, at least didn't destroy the books. Destroying books at that scale is a sacrilege.
Every single AI company is guilty of using questionably sourced materials - either outright stolen or human input based on exploitation.
And on top of that, it's not just RAM that has gotten expensive. The entire rest of the economy - both individuals and companies of all sizes - are priced out of personal and even cloud compute, as the blown-up AI giants scoop up everything they can and the scraps and aged hardware that's available gets fought over by everyone else.
> blatant denials of reality, like claiming that AI is a completely worthless technology or surely the bubble will pop any minute now.
It is undeniable that the entire AI sphere is a bubble, artificially propped up by circular investments and wash trades, and that is now poised to raid pension funds.
> Vibe-coded software is simply not good. Let’s suppose that someone deploys an app and there’s a critical security vulnerability that allows a threat actor to, say, exfiltrate all their customer information. How would they know? And if they became aware of it (presumably because said threat actor exploited said vulnerability), how would they fix it?
Is all non-vibe-coded software "good"? Are so-called real companies (with professional software development and IT staff) impervious to these threats?
> Also, would the person who developed the app know that, under legislation like GDPR, they can be financially liable for data breaches? Because they would be! And the whole point of the financial penalty system (at least, with respect to GDPR) is to be dissuasive — to act as a deterrent to other people who would be cavalier with other people’s data.
> I can very easily imagine a national data protection authority — like the UK’s ICO — giving someone a massive, massive fine in order to dissuade other people from deploying their own AI-generated, unvetted slop code.
Thousands of companies have been hit with GDPR fines, including some of the biggest companies in the world. Why the apparent assumption that vibe-coders are any more cavalier about people's data than companies that in many cases exist solely to profit from it?
I think you can make a legitimate argument that companies selling vibe coding dreams to laypeople are selling something generally unrealistic but the tone of this person's article seems like gatekeeper bait. It feels like he just doesn't like the idea that non-engineers might try to use tools (oversold or not) that allow them to do things he thinks non-engineers shouldn't do.
It's very much a "keep out", "stay in your lane" vibe.
> Vibe-coded software is simply not good
It doesn't fucking matter to the success of a business.
I spent much of my early career unfucking large codebases that had been thrown together by sysadmins or teenagers or HTML guys who knew a bit of Perl on which an enterprising person had built very, very successful businesses. The software got fixed by pros long after profitability when it started to matter.
There are very few businesses where the quality of the software makes any real difference. What matters is execution, marketing, commercialization, but programmers see every business problem as a technical problem requiring technical excellence, because they're gigantic hammers.
This is the attitude that gives us daily data breaches.
It is not an attitude, it is a clear statement of fact. That you don't like it is just shouting at clouds.
I think it does matter. Bad software can make peoples lives worse, especially when they depend on it. Really bad software can kill people. Maybe it doesn't matter to the bottom line of a business, but that's a particularly low bar to meet.
Take this article for example
https://ericwbailey.website/published/modern-health-framewor...
Someone in a genuine health crisis seeking help but can't receive it could have deadly consequences. It doesn't matter to the business, but it does matter to people. Life is messy and complex, and if our software doesn't work correctly it does add to the suffering of others. Maybe it pushes some people past the point of no return. There are consequences to what we do, good or bad.
This is what I fear most about the rise of vibe coding. Businesses profit, people get hurt, and the incentives are all wrong.
It matters when all your data ends up on pastebin.
[flagged]
> "I’ve noticed that TikTok..."
Problem found. NEXT!