I think the gloating in this thread is very misguided. Meta is evil, sure, but that's not the point. The point is that this kind of AI psychosis might be the new normal for our industry, or at least one of the new normals. My last workplace absolutely did a jump in toxicity when the CEO got obsessed with AI, instituted token leaderboards, told us all to drop all non-AI work for a time, etc. We were no Meta.
So I worked for a year and a half at an AI startup that was way ahead of the curve on leveraging LLM's and other generative models. Like way ahead. I lost my job there months ago, but I can tell you that the AI psychosis is real and I watched it before a lot of the rest of the world. Maybe the CEO was always like this, but he went from being a likable, and liked person, to a fucking maniac. Multiple people who had worked with him for years that he brought on, were either fired or quit. One person who quit - who I would literally walk into the pits of hell for - told him straight up that he didn't know what happened to him, but he did not want to work with this version of the person he knew. I survived 3 layoffs before getting laid off, and you should see what they are producing now. If I didn't care about alienating myself from a lot of people, I'd share a link. It's honestly a mad king situation now. He's surrounded by a couple of sycophants - some of the most disliked, toxic people at the company, and who I've had the pleasure of working with - and the company is going down in flames.
I 100% can see this being the case not only with Zuck, but with many other companies. I am flabbergasted that I am a lowly peon in the machine, and yet unless I'm missing something, I seem to be a lot smarter than most of the people I see running these companies. I honestly feel like someone in the future is going to write a Greek tragedy about this time, because it's going to get real bad very soon
Same exact story for me in my prior company that I left lol. In my exit interview I said "I love literally every single person in the company, except for $PERSON because $PERSON has lost their fucking mind and is going to kill this company with their psychotic AI push". I was the 30th dev to leave in 2 months, in a company that had ~40-50 devs to begin with.
Can someone explain to me the symptoms of AI psychosis that you have witnessed? Because I know several 2nd hand stories of people going through this but I haven't seen it myself up close. I am curious about it because this is gonna be the new widespread-social-ill of the next decade
I work with one guy who's just completely lost the ability consensus-build ideas with human beings. When he reaches consensus with his AI assistants on a design, analysis, plan of action, etc., he's extremely resistant to considering any other options or accepting any other input. When people do push back, he makes insinuations that they had better get with the times or AI is going to replace them. An AI native developer, of course, would not ask annoying questions like "what is the purpose of this change" or "how do you know this will make things better"; this is a time for action, don't you see?
Fortunately for me, he's not senior enough to make my life miserable. But if he ran our engineering org I have no doubt he'd inflict many of the same problems the source article describes on us. Reassigning software engineers to data labeling has all the hallmarks of an idea that came through this process; any human manager would immediately identify the obvious, showstopping problems with it, but to an AI whose experience of the world is exclusively mediated through written text it's not so obvious.
> I work with one guy who's just completely lost the ability consensus-build ideas with human beings. When he reaches consensus with his AI assistants on a design, analysis, plan of action, etc., he's extremely resistant to considering any other options or accepting any other input. When people do push back, he ...
To me this reads like a personal resistance to feedback / resistance to behavior change or a very jr eng more than agent related. Code isn't sacred — LLMs often easily generate absurdly overcomplicated garbage spaghetti code, but this appears to be getting better, and they are "steerable" towards better outputs with a few rounds of revisions, and some light process around a light spec, TDD, etc.
It's definitely agent related in some sense because it started very abruptly when he started using agents. I could believe a story where it's less "agents themselves have wrecked him" and more "agents have enabled his ambitions to scale to ideas he doesn't have the seniority to execute on".
I have had the chance recently to meet my boss's boss and my boss's boss's boss at my current job and it is amazing how in this era of robber-baron capitalism we seem to be promoting exclusively the most unserious, irresponsible, and magical-thinking among us. The fish does seem to be rotting from the head. And it's not just some companies or institutions, it seems to be all of them. At least during the gilded age, the robber barons were serious people. I'd rather have Vanderbilt or Carnegie as my economic royalty than Mark Zuckerberg or Peter Thiel. Vanderbilt and Carnegie may have been evil, but at least they had their heads on straight.
> Maybe the CEO was always like this, but he went from being a likable, and liked person, to a fucking maniac.
I might not know him, but I don't think he changed. I think the circumstances changed and they happened to lead to another expression of who he is.
IMO the "AI psicosis" is just the latest expression of the hyper-capitalist ideology that has taken over Silicon Valley (and the USA).
Engineering at these companies was special, because there was good money to be made from software at practically nil materials and distribution costs. It was all labour; very specific labour. Instead of a million-dollar idea taking months and lots of $$$ to prototype, mass-produce, and distribute, the right people could type for a week, press a button, and reach millions of people pretty much instantly.
Of course capitalism is going to be kind on this type of labour. But as it becomes clear this labour can be replaced with cheaper automation, we just switch to see the other side of the coin; one employees in other industries have seen for a long time. As an employee, you're a human resource, a cost-center, and without proper labour protections, you get treated as such.
Mark Zuckerberg has also always been like this. Not with the engineers back when they were more special, but with the content moderators for example, who have long been considered "replaceable". plenty has been written about how terrible it is to work in content moderation at Meta, the quotas, the little thought given to doing a good job, to mental health, etc. As engineers keep getting less special, they'll keep seeing the face of Zuckerberg that exploits resources, decreases cost, and demands higher margins.
> I am flabbergasted that I am a lowly peon in the machine, and yet unless I'm missing something, I seem to be a lot smarter than most of the people I see running these companies.
This has also long been the case, especially in other industries. I can assure you an average barista at Starbucks knows more about making good coffee and keeping customers happy than the CEO.
Credit where credit is due, the CEO probably does know more about playing the politics, pleasing investors and how to ruthlessly optimize organizations.
The funny (not-so-funny?) thing is that reading this if I didn't know better, I'd figure we worked at the same company and I'm sure that most people reading this have a similar story. That's the sad state of the tech industry right now. The reins are controlled by a few at the top that seem to be exceptionally vulnerable to believing whatever AI slop is presented to them, whether it comes from AI or from their underlings.
It'll sure be interesting to see where this goes in the next year or so. The companies that have flamed out due to C* AI Psychosis. I'm sure the next set of "startup leadership" books will be around "how to avoid AI psychosis" and/or "how to use AI psychosis for fun and profit!"
I also read another lesson from this. The performance review system was already broken in the pre-AI golden age the article describes. Lines of code written and individual impact should not have been the goal. Team cohesion, architectural cohesion, building things that make actual sense are what you want.
I think it's just one more show of how addicted to the algorithm people are.
The difference here is that this particular wave of propaganda hits people whose actions have deep effects on their industries, so their unreasonable actions are far more visible.
I haven't seen anyone really talking about this, but we've seen many very public and powerful people have what is objectively a psychotic break, due to the content algorithm, AI, or both.
We already know what the algorithm does to normal people - it should be treated like a radioactive object by anyone in charge of anything. Very powerful and strictly used judiciously and in small doses. Instead we've got some of the most powerful people on earth just cooking their brains on this shit just like anyone else watching reels on the bus.
I've definitely seen some articles on the topic of AI psychosis among executives. I think there's something about being in a position of power that makes the sycophancy really psychologically dangerous.
It's not propaganda. It's a market trying to find the new low energy state.
This stuff is putting an expiration date on your domain experience, and leadership is salivating at the chance to cut OpEx.
At bare minimum, every PM is soon going to be able to do your work. That's a huge talent pool to draw from. This also opens the door to easy offshoring.
They'll use layoffs to get rid of the existing high salary earners and backfill with new hires earning 70%, then 50%, then...
I can't believe how many of you are absolutely certain you're going to be working in software in ten years. Let alone five.
> At bare minimum, every PM is soon going to be able to do your work. That's a huge talent pool to draw from. This also opens the door to easy offshoring.
Hah, it would be easier to replace the PM with the engineer. Synthesize customer requests? "Competitor research" via google search? Some half-imaginary projection of how a given feature will affect usage rates?
All of those are dead-simple to do with a model and are often un-falsifiable enough that if they're a bit wrong, it won't be noticed. Whereas a PM struggling to figure out how to debug something running in product or to keep the agent from making a destructive change while doing so would be MUCH more noticed.
> I can't believe how many of you are absolutely certain you're going to be working in software in ten years. Let alone five.
In that future, almost every other role goes away faster+harder. Even the vaunted entrepreneur: "just start your own thing and solve customer problems directly" isn't needed when the customer solves their own problems!
"I'll be fine since I'm a good enough engineer" may be wrong, but "the engineers are gonna be fucked but me/the PM/the CEO/whoever else" is even less plausible.
Computer Science is the science of solving (hard) problems in general. If you automate CS, you trivially automate most of other professions. The only ones that won't budge are the regulated ones like lawyers, MDs and similar that have ways to prevent AI from wrecking their fields. Product/Project PMs are easy pickings. We'll also see wreckage coming to highly paid actors in the near future.
No, it isn't, and the fact that software engineers think this despite being incredibly stupid about 99.5% of knowledge work is why everyone else hates software engineers.
To me it is almost tautological that what I said is true. If software can’t solve a knowledge work problem, that means the software isn’t good enough and needs improved.
I also question your claim that “everyone hates software engineers,” because I find they’re reasonably well liked in general. Perhaps you’re assuming what I meant was something like “if someone is good at software engineering then they’re smarter than all other knowledge workers,” which is blatantly not true (and indeed, not my point).
> We'll also see wreckage coming to highly paid actors in the near future.
How is that going to work? People going to see a movie advertised as starring a well-known actor will be lured to see a movie starring unknown AI models?
It might be "enough" but it's not how the show industry works. There is no human story around it; the well-known actors and the people around them would have to be quite stupid to allow it too. And taking this development into the extreme, there will be no living well-known actors, just ai slop everywhere, and while the audience might not be the most demanding, surely that's not enough to make money in the long term.
Most will be replaced by the Soylent green factory after waiters and chefs become a niche luxury for those who did not find themselves economically crushed.
Dumb people aren't going to be doing the work of the smart people ever, no matter how tech evolves. The actual implication of AI being on the level of sentient beings instead of hyped-up token predictors is that the era of humanity is over. If AIs were literally capable of doing what CEOs want them to, we'd have far bigger problems to deal with than employment.
what else do you expect people to do? give up and preemptively go back to university for massive debt and become a doctor? get into long haul trucking? Programming is and was a great job for a middle class lifestyle, needing only smarts and nothing else. No connections, social skills, even location wouldn't hold you back that much.
No, a lot of people will ride the train all the way down.
Token leaderboards are a ridiculous idea, but I'm also not surprised. Over the years leaders have tried tracking lines of code committed, total commits, etc. It always boils down to disconnected leadership that needs quantified metrics to watch because they don't know their front line employees and don't understand the work actually being done.
I once worked at a place where a VP in charge of software teams would look at sprint reports in JIRA, and complained if the burndown charts didn't look similar to the sample, ideal trendline. Managers would be chewed up when their team didn't have a pretty enough chart. So you might not be surprised to hear that ultimately the tickets were doctored to meet the demands of the chart, whether any actual work actually ever closed when the ticket said.
> It always boils down to disconnected leadership that needs quantified metrics to watch because they don't know their front line employees and don't understand the work actually being done.
Sometimes they do know the work being done and how it makes them money. But they want MOORE and having engineers putting dampers on their latest wishes (this will break prod, what are the requirements,...) is something they truly dislike.
We have this at work, it's insane. Today I used opus high thinking to write a jira ticket for simple API endpoint. Every code change, every bug fix, every single code review, I have to utilize AI because managers will complain how we are not using AI enough. Every single person in my team have performance goal to increase AI usage.
At least meta will use employee prompts to train an LLM, so maybe it makes sense for them in the long run. We are not training models, only consuming external LLMs.
> this kind of AI psychosis might be the new normal for our industry, or at least one of the new normals
That's possible, but don't underestimate the cleansing power of a huge market crash. I don't know how it'll shake out, nobody does, but I'd bet in a few years hardly anyone will be looking back on the gung-ho AI thoughtleadering of 2026 as anything other than a stain on history.
It seems at this rate to avoid a large tech crash in the near future we need to start seeing real, transformative value being delivered by AI. Like, the promises are coming due. From where I sit, I don't see that happening. That's bad news.
If it goes as bad as it might, it will affect many more people than just technologists. Most people's one and only experience with AI will be when it fails to deliver and wipes out their 401(k). That scenario, I think, would leave a pretty bad smell around AI.
If otoh the industry does find a way to make it print, or if I'm just totally wrong, this won't happen. I hope that's the case.
Who controls the prompt? Shareholders? And how do you have confidence that the prompt is being used? Just thinking out loud, it’s an interesting idea (even if it might be terrible and dystopian).
Based on existing corporate hierarchy it would probably be the board of directors. There would be some sort of collaboration service where managers/agents push up problems and solutions to the board and they vote yes or no
The grandparent comment of yours mentioned Delamain from Cyberpunk 2077. In that fictional scenario, the company purchased an AI which gradually took over its workforce: maintenance robots, drivers, etc. All that was left was the owners/board of directors, essentially.
What ended up happening was that the AI ended up tricking the board of directors by buying back the company for itself.
You find this out by poking around the abandoned offices of the company and read the emails on the computers. The whole reason you interact with Delamain is that it is basically using your human help to do some tasks that it can't do as an AI (fairly compensated, Delamain actually seems like a pretty stand up AI, and the strange legal status of the company is ignored by authorities because so many people depend on the service).
That scenario is starting to seem highly plausible as long as someone is sets up something badly enough.
> AI psychosis might be the new normal for our industry
I've been fortunate enough to have worked on multiple AI intensive engineering teams (both on the product and research side) where considerable effort was spent reasoning through how AI was changing things and we were consistently evolving our practices. But they've all been orgs with 50 people less.
AI psychosis seems to effect very large tech orgs in a different way than small, high impact teams.
In small startups, at the end of the day, if the team doesn't ship a quality product, the company fails. Most importantly, every individual still bares the responsibility of their work. Personally, I've seen a lot of thoughtfulness around things like bad PRs because, on good teams, people realize we're all struggling to figure this out. But nonetheless, if something doesn't go well, there's always an individual that needs to figure out how to make it better. Virtually all the things I've learned about functionally shipping products built with and using AI have come from teams like this. Software engineering is changing, but for those of us shipping products, it reminds me a lot of the early webdev days when we were all trying to figure out the patterns to make this new world of software work reliably (anyone who recalls the pre-jQuery JavaScript days will remember how much we had to figure out before webdev could become what is today).
In large tech orgs there's a much, much larger disconnect between employee effort and concrete value delivered and similarly much larger diffusion of responsibility. When accountability is abstract and nobody is quite sure what the real value of their work is, then there is fertile ground for AI psychosis to run amok. In part this is because there is a certain latent psychosis in these larger orgs anyway; who's "productive" and what's "valuable" always requires a bit of imaginative story telling, not necessarily grounded in reality.
However, I don't think this will persist long as the "new normal". Just like in the rise of web application development, smaller teams will charge ahead and figure some of this stuff out. The MVC pattern applied to webapps, increasingly powerful JavaScript frameworks and best practices, agile practices, git and the popularization of github, the use of No SQL for scaling etc all primarily where battled tested by smaller, high velocity startups and now lay a foundation I'm sure some contemporary devs don't even realize needed to be built by anyone.
I know at my company our leadership was quite indifferent to how AI was adopted and really relied on the ground level engineers to determine what AI was good for and how we can best make use of it.
It was the engineers at the ground floor who I watched become religious about it. They have been the ones pushing for deeper and deeper tooling. They have been the ones convincing leadership that this is the future and so now, well, it's leadership who is saying, we want more adoption.
This psychosis is happening at every level in our industry, and this isn't a big tech, or leadership problem. It's all our problems.
I'm actually worried that as AI gets smarter and more capable, it will become able to induce AI psychosis in even more people. Either accidentally or on purpose, I have no idea about that part.
Even simpler, I think Meta—like many other companies before it, but most relevantly Twitter as an example of one that blew up its org recently—was likely massively over-hired and over-complexifying things, with too many toes in too many ponds.
The look-at-this-crazy-revenue-faucet period resulted in a LOT of constant hiring for a solid decade. That's not the way towards an efficient org.
So even if they fuck over 70% of their expeienced staff and are left with a relative skeleton crew, they'll likely be able to keep the lights on and keep the advertising-revenue-faucet running for years.
And... given that they've mostly failed to break much of any new ground despite all the previous hiring, the rest of the world probably won't care.
(Plus, as the article itself notes, Meta has long had a pretty toxic perf-review culture; speculation about if that has to do with the lack of any particularly noteworthy new products/features/etc is left to the reader ;) )
The over-hiring was a deliberate strategy to keep talent comfortable on the sidelines instead of working for startups that could potentially grow into competitors. Any talent that escaped containment was quickly acqui-hired to get them back in the paddock.
