After the AI Crash
potsandpansbyccg.com
potsandpansbyccg.com
To parallel to The Big Short this is the point in the movie where folks realize it’s mathematically impossible for things to not implode and so players are quietly positioning themselves for that eventuality before things are allowed to blow.
It’s been a dramatic shift these last six months but everywhere I look now folks are quietly preparing their battle armor to survive what’s about to unfold.
The tech will stay, but the AI business landscape will have a market-cleansing forest fire.
There’s a whole generation in tech now that’s never seen what happens when a bubble like this unravels. I fully expect we’ll see the likes of offices just abandoned overnight with food still in the fridge as AI company after AI company just vaporizes.
It’s likely to be bumpy for all but stay the course with diversified strategy. The .com bust and 2008 are just blips and cheap buying opportunities for most folks with diversified portfolios and index funds. We’ll likely see similar messy markets for a while but the world will eventually recover and move on.
The ones that get truly wiped out are those with a lot of paper wealth now that implodes, but have little liquid wealth. These types go from “super rich” to can’t pay their bills almost overnight. Cash is king for times like those ahead so if you don’t have a lot of cash on hand, now is the time to secure that if you can.
There were lots of big guys that bet against the bubble, but bet too early and weren't able to keep up with margin calls etc.
Do you base your knowledge on anything other than a very mediocre movie?
If an "investor" believes the market will sell off drastically, there are many, many ways they can express that bet in the open market. They can short index futures, buy put options, long volatility... so many options that depend only on the thesis being right. It does not sound as sexy as "look kid, the billionaires rigged the system!", but it's true.
You can't time the market reliably. Ultimately the base advice remains sound even if it's couched in a bit of conspiracy: diversify and allocate your assets based on your risk profile, as informed by your retirement timeline.
There's an entire profession doing that: traders. But for most people, yes, it is not worth it. No argument there. I just don't like conspiracy theories and that line of thought, it's cheap and stupid.
Well no, there's an entire profession trying to time the market. The data is not favourable as to their ability to actually do it at a rate greater than random chance.
But.
I also don't think it's reasonable to label as a conspiracy a claim that markets are rife with cheating, insider trading, etc. Just look at the current US Presidents profit margins.
Is that systematic, organized, conspiratorial, rich v poor market rigging? No. But it would be naive, I think, to believe that the rich aren't significantly advantaged in the market in ways that the average person isn't. There is, after all, a reason payment for order flow exists as a valuable thing.
* identify a bad AI stock at $100 today
* produce an ironclad 100% guaranteed proof the company's valuation will drop below $50 by the end of 2028
* buy a bunch of long-dated puts with a break-even of $75
* see the AI investment bubble visibly falling apart in let's say March 2027
and you still lose if OpenAI acquihires the team at $80 in an attempt to prop things up.
So I expect more of a sobering process for AI companies rather than a blowup, simply because they all still will have cash in the bank and some have actual products with clients that make use of them.
Am I overstating the case? My understanding is that a data center tends to be ‘purpose built’, so they may need a gutting for repurposing, but assuming a lot of ‘sobering up’ I’m envisioning several giant cloud providers with excess capacity and a scaling potential.
Be it Jevon’s paradox sparked by cheap compute, another huge tech fad, or a ML breakthrough that brings another kind of model to the forefront, we’re likely to want a lot of compute at some point. It doesn’t seem like bad long-term positioning for the tech giants or investors.
The most straightforward thing to do that would have saved you in 2008, 2000 (and even 1929 to some extent), is to limit debt exposure and have enough cash sitting around so you don't have to sell your positions.
People who held (and invested more during the low points) did just fine. The people who got hurt the most were in a position where they had to sell (their 401k, their house etc) for a loss.
When a crash happens, it's a buyers market. In some ways it's a transfer of wealth to the top.
The problem is, timing the crash is impossible. Inflation chews into your cash portfolio, that could've been used to grow wealth.
Edit
There are of course more complex financial instruments. And they're interesting, but essentially amount to gambling. Holding short positions on sectors/companies is a zero sum game. As a retailer, you're betting against firms that have a lot of money and insider knowledge. Even if the thesis is correct, it has to be correct at the right moment.
Amen, $SIBLING. Even an idiot like me could see it coming in 2007, triggered by a young coworker asking me about interest-only housing loans. But what do you do about it? I sure didn't know how to short the real estate market, and watch The Big Short to find out how hard it was for them to do it. That, and the ol' "market can stay absolutely fucking bonkers longer than..." chestnut. I would have lost my ass anyway.
But back to the topic at hand, indeed, the timing is very hard. Margin calls and watching the price of a borrowed stock go up and up, those mean that if you have to ask you shouldn't be doing it. Other sibling comments have already said that the answer is a diverse portfolio. As an investor for many decades, I've seen with my own eyes the wisdom of that. Sure, you'll take a loss like everyone else, but unlike everyone else your losses will be smaller. You'll miss out on the big gains that those with good timing will get. You'll also miss out on the absolutely huge losses of those that timed it wrong. Diversify, stay the course, it'll come back: that I've never seen fail after decades of investing. (But for $DEITY's sake, stay away from AI-specific companies if you can do it.)
Buy and hold real estate, maybe?
I don't think you can really short this because you can't time it. But you can move assets to things that are 1) concrete 2) have utility and 3) that the US gov't won't allow to lose real value for political reasons.
If you knew when, it would be easy. But nobody knows the when. It might be tomorrow, next week or in 12 months or maybe two years or five. A drop is guaranteed, eventually, but nobody has any idea how much it will go up before that.
If you are truly convinced on this thesis, one could slowly keep increasing short bond allocations while moving away from tech, but no guarantees. Mainly depends how old you are. If very young, don't worry about it, stay invested. If near retirement, move to safer assets because you can't afford to loose everything now.
Microsoft can as well.
So Anthropic only has a few thousand people which is nothing and OpenAI has also only a few thousand.
A lot of people working at these companies would find a new job easily and probably making a shit ton of money anyway.
The broader AI companies do not have that much venture capital, not that many peple and potentially not even that high of risks.
Deepseek didn't do their funding round because they could'nt even get the compute if they wanted to. So they might be very lean.
And even investors for OpenAI and Anthropic might accept that they stop R&D and just keep the services up and running as they are cost effective.
The bigger problem will be the industry itself: If i'm 5x more productive, and generate now more costs on the business, they will reduce headcounts to compensate for it.
seriously, how hard is to get this? finance changed in 2020/21 forever. fiat printing level size several times bigger than all money total before
I doubt any credible analyst has claimed that. What's the total AI capex that's already been spent? About $1T? It's a pretty absurd idea that those DCs need to make $2T/year for 5-7 years -> $10T-14T over their lifetime to break even.
(Yes, this is nitpicking in the sense that there are probably analysts talking about how projected and sustained capex at >1T/year will require 2T/year in revenue, so patching the article won't be a biggie. But this article is cosplaying as financial analysis and leads off with such an obviously incorrect argument. What does that say about the credibility of the rest of the article?)
> Diseconomies of Scale.
Another very basic mistake here. The author starts talking about efficiency in the context of past technologies. That's lower unit costs as scale increases.
But for AI, they seem to switch from talking about unit costs to total costs. Or at least I can't explain what they say about models getting more expensive over time in any other way, because that is not true about unit costs.
We've never seen economies of scale as large as for AI. For a given quality level, the cost has been dropping at >10x per year, not increasing.
With what seems to be their current revenue, the frontier companies are likely better off than Uber was pre-IPO. The "circular financing" we've seen probably lowers risk relative to external financing, since both parties get aligned incentives.
It's too bad Nvidia keeps getting hammered, they're selling the shovels in a gold rush that'll last for years, and I don't think that's priced in.
That figure is $2T in the next 4 years though. $2T/year would be quite silly.
> Diseconomies of Scale. Every new technology I can think of thrived, in part, due to economies of scale, where the larger the industry grew, the more efficient it got. AI is going in the opposite direction, where every new AI model consumes more resources than its predecessors. This may turn out to be the fatal flaw – the bigger the industry gets, the more its operating costs increase.
[Agreed] Newer gens of models are more power-efficient per task, not less.
This is a low quality post full of basic errors.
You could argue that a RTX5090 needs less power than a 1996 3Dfx Voodoo 1 to render GLQuake at 320x200/60fps, but it's rarely being used for that.
Not mentioning that we are kind at the point in the Wait but Why cartoon between that's cute and wft https://waitbutwhy.com/wp-content/uploads/2015/01/Intelligen...
From the article from ten years ago https://waitbutwhy.com/2015/01/artificial-intelligence-revol...
It's interesting that it shows human level at 2025 and we've just seen a wtf story on HN https://news.ycombinator.com/item?id=49048681
I'm not going to predict how this is going to turn out in either direction but this statement gives me pause. I don't think that's ever been true. Yes, the siren song is strong but initially most new tech is met with skepticism. Are we so quick to forget "the internet/computers are just a fad"-type thinking?
Yeah, we reject anything new and scary. Gotta wonder what the cavepeople made of the wheel.
Edit: still triggered by heliocentrism in 2026? OK then.
If you have the time, I recommend reading Pliny’s Natural History. I think it really shows how much more open-minded these people were than we give them credit for. They just didn’t have the technology we have.
Invoking “heliocentrism” is the HN version of “Ive already drawn you as the virgin and me as the Chad.”
But it's humorous that Sergey Brin has a company dedicated to bringing back zeppelins and that's a lot more fun than Zuck's doomsday bunker or Bezos's doomsday yacht.
They had a ton of fundamental problems despite seeming like a good idea at first. It was only after we started using them did we realize they weren’t actually as good of an idea as they seemed.
When was that?
It was getting huge in the late 90s and we all know now it was indeed the future.
But it still took a huge dot.com crash because the valuations got ahead of the curve.
But agree with you on all the other examples.
This doesn't mean LLMs will be as useless as the metaverse, but a lot of the biggest proponents of the technology within companies are those most clueless about it which feels like a red flag.
> This doesn't mean LLMs will be as useless as the metaverse, but a lot of the biggest proponents of the technology within companies are those most clueless about it which feels like a red flag.
wether this actually reliably works out, or it just turns out to be a useful plugin/assistant is what will pop the bubble quicker rather than later imho
Maybe those older than me have been through this kind of thing before in 2008-2009 and in the early 2000s but the state of the IT job market in the last year or two has been really concerning to me. Anecdotally I’ve also heard it’s very tough for new graduates these days.
If humans are machines executing on some medium eventually it will be possible to emulate them on some other medium. Believing otherwise requires some essentially religious convictions on the existence of the soul or similar.
My doubt has nothing to do with religion.
I believe we'll never be able to harness the amount of technology required to replicate human intelligence.
The only reason to emulate human intelligence is to capture it, exploit it, and reduce the negotiating power of those who have it naturally. Anyone who wants this is not a friend of human beings
This could not be further from the truth.
Creating a new educated, competent human takes ~25 years, and significant costs and resources.
AGI can be just copy-pasted at essentially no cost. Even a robot that is mass-produced for tens of thousands of dollars is still much less expensive and takes much less time to make than a competent human.
> Creating new humans is easy
If you did a survey I suspect agreement with this statement would be rather bimodal.
I am not calling you out personally here, but this comment bears many hallmarks of the "accelerationist disease": thinking in abstract models rather than in concrete terms; optimising for the fangs of capital rather than human beings, imaginary infinite bets that lead participants to nonsensical conclusions
Not exactly true. If AI goes from 0 to 100 quickly (human level AGI in a few years), we will all lose our jobs together and will have to confront this fact. There is no ignoring such massive societal elephant in the room.
If instead AI improves slowly, lots of people will lose their jobs, but majority will not. Those still with jobs may not be persuaded by the sizable jobless minority to support any societal changes, since they are still OK. This means the jobless will be screwed.
Secondly, we already know what happens when sectors face mass unemployment from the 1980s, when steel and coal workers were laid off en-masse: nobody came to help
Accelerationists want to see their jobs go up in smoke because they think that if enough white collar workers are impoverished, everyone who was formerly poorer than them, will chip in to rescue us
Just like they didn't in 1985...
In a context where the state already has less tax revenue due to mass unemployment! It's completely bananas thinking that only survives on internet brainrot forums
Human-level AGI is by definition able to take away all human jobs.
> Secondly, we already know what happens when sectors face mass unemployment from the 1980s, when steel and coal workers were laid off en-masse: nobody came to help
There is a massive difference between 20-30% of people in a country being unemployed (historical sector-wide collapses), and 80-100% of people in a country being unemployed (human-level AGI unemployment). The second situation is completely historically unprecedented and will *not* lead to the same results.
> In a context where the state already has less tax revenue due to mass unemployment!
If human-level AGI is an actual reality, taxation (and money itself) is no longer relevant. Only energy and resources may still matter.
This would mean that there are no AI companies at all, because there is no one to buy their products. So, they have nowhere to get money for whatever they want to sell.
But my point really was that it is economic nonsense.
That depends on what those companies do in response. Giving away robot-produced stuff for free could quicky earn them praise, instead of hate. If human-level robots enable massive increases in efficiency and production, at least one of the companies will decide to mitigate any risks in this way.
And they will be hated, because they stripped people of future and bought in violent hellscape for literally everyone.
You don't need to persuade all of them.
> To achieve that, they would need to steal farms from current owners.
Steal, or buy? And perhaps not. With robotic labor massively increasing economic efficiency, other forms of production could become viable, such as vertical farming.
So, you have starwing hopeless population not buying your products. Which means you cant produce. Which means, you eventually become part of that starwing population (or they kill you in some outburst).
In a monetary economy, if you need X, you need the money to buy it. In a post-AGI economy, if you need X, you ask your human-level robots to make it (including making more robots). No money involved at all.
They can also perform work with the aim to earn money to buy more land.
On who's land? With who's energy?
> They can also perform work with the aim to earn money to buy more land.
Weren't we just saying money was gonna be obsolete?
Land is an inherently limited resource, so some kind of money might still exist for that, even after the singularity. The point is: whatever process exists to obtain land after the singularity, the robots can do it.
Yes, that's my point exactly.
> whatever process exists to obtain land after the singularity, the robots can do it
Who's robots? Who will get all this land? And its spoils, and the robots that multiply them?
Because your utopia sounds like a recipe for making our oligarchs even more powerful at the expense of the rest of us. It sounds like a slide into neofeudalism. The already existing don't even need labor to multiply their riches anymore, the rest of us will be robbed of being able to trade our labor for even basic subsistence.
Worst case scenario: you will live exactly as you live today, with one change: instead of you having to do your job, your robot will do it for you, and you gain more free time.
You've already conceded that there is inherently limited resources like land. These are going to be the limiting factor in any economic system, the key question is always how you distribute these.
Robots will not magically produce more land. What they will do is that the ones with the capital to buy many robots will be able to use those robots to then buy up all the land and then extract the resources from there to build even more robots etc.
No resources for you and me. No robots for us either. Why are you talking about prices going down? We're making money obsolete, remember? Why would they even sell us robots when they already own all the land and all the resources to make as many robots as they want? Sure, the pie will get bigger. But you ain't getting a slice.
What leverage would the vast majority of people have? If the oligarchs own all the robots and land and resources, how do we even ensure basic subsistence? Our labor is worthless. Our money is worthless. What stops them from... simply not sharing? The goodness of their hearts?
The speed at which the availability of the robots to the general population increases will be much faster than the speed at which the initial robot owners would be able to buy land, and exclude others from using it. There are trillionaires today, extremely rich people have existed for a long time, yet they still own only a small fraction of all available land.
There could be a class of people who do not own any robots and thus are dependent on the goodwill of the robot owners to get anything they need to live, but this will be a small fraction of the population, equivalent to the homeless today. Definitely not the majority.
What are you basing this prediction on?
If you're a robot maker, already part of the oligarchy class, and you knew that robots would rapidly accelerate resource acquisition, why sell the robots at all? Money is about to become useless, remember? Why let the plebes have robots? They'll just use up resources you could hoard.
I feel like you predict the end of market economics, but at the same time make a bunch of assumptions based on them. Why would availability increase, price decrease etc etc if we're ending that whole resource allocation system? It'll be a giant landgrab by the ultra rich to grow and cement their empires, why wouldn't it be?
The idea that economics goes out the windows in a post-AGI world is highly speculative and not a situation that has ever happened in the real world.
With cheap robotic labor, many resource mines that are currently not economically viable will become viable, so basic resources will become much more plentiful.
The new economics will probably be based on needed energy, direct tally of needed material resources, or the number of robot-hours something takes to do.
Will they create also water that will be getting more expensive due data center consumption?
This is absurd.
Not from nothing. How do humans create metal and food? Mining and farming. You ask your human-level robots to do the same. Human-level robots can by definition do everything humans can do, including gathering basic resources and growing food.
> Will they create also water that will be getting more expensive due data center consumption?
Data center consumption of water is miniscule compared to other uses. But yes, you ask the robots to make you water, for example by building desalination plants.
Even if there is still monetary economy, and you come to the conclusion that buying something is the best way to get it, you simply ask your robots to think of a way to earn you money by performing work for others, or making things to sell on the market. If they are truly human-level, they can do it.
> This is absurd.
This is the reality, if human-level AI and robotics is developed.
So, the plan is that AI companies will steal the mines and land from their current owners? Like, you plan to use those robots to beat up the owners and take their stuff? If your actual plan is such fascist takeover where AI companies bully everyone else, then everyone else is entirely valid to wage war against you before you are close to that goal.
> Data center consumption of water is miniscule compared to other uses. But yes, you ask the robots to make you water, for example by building desalination plants.
The water is getting too expensive now. Why are they not doing that now already? Technology is there and work is not the most expensive part of this.
Same with energy. So, you plan to take over the power plants by force? And take over the land to build them on? These AI companies could be building power plants, could even pick places where they dont give kids asthma already ... but they are choosing not to.
> This is the reality, if human-level AI and robotics is developed.
No, it is not reality. You are ignoring reality. Robotics does not imply end of economics. The end of economics will happen only if you win war and install an odious violent autocracy. Which I understand is an intentional goal of current CEOs and their fans. I both disagree with that goal and find it, frankly, stupid.
What? Why the hell would they need to steal or beat up people? They can buy existing mines, or establish new, robotic mines.
> The water is getting too expensive now. Why are they not doing that now already? Technology is there and work is not the most expensive part of this.
Because human-level robots do not exist yet. And human work is very often the most expensive part of any advanced technological endeavour. It is the limiting factor. Good, qualified workers are very expensive.
> Robotics does not imply end of economics.
Monetary economics only exists because human work is needed to obtain resources and produce stuff, and human work, especially educated human work is currently in limited supply, so we need to allocate this limited resource efficiently. Remove this limitation and there is no longer the need for monetary economics.
There are no money in your world, they cant buy them. Why would the owners sell anyway? Selling the land would mean they have absolutely nothing except money they can not start new business with.
>The water is getting too expensive now. [...] Because human-level robots do not exist yet
We are not carrying water in buckets. It is all done with technology now.
> Monetary economics only exists because human work is needed to obtain resources and produce stuff, and human work
No. It exists because people who have thing too much of a thing A and a little of the a thing B. Before money, people would barter - I will give you 5 eggs and you will give me 10 apples. Money were invented to facilitate the barter.
Human-level robots can mine resources and build power plants.
Perhaps you need some initial land to build your first robotic base, but robots can also perform work to earn you money to buy more land afterwards.
Also, if there are no money so 80%-100% of people move to barter instead of money to trade stuff. Same thing, just less effective.
What more does it require?
Spoiler: now you tell me what more is needed, and I would then tell you how to use intelligent robots to obtain it.
What definition is this? At best it would be able to take jobs that currently rely on human intelligence. There's a lot of ditch diggers, prostitutes, athletes, models, acrobats, painters, etc, etc out there who have nothing to fear from the first-order availability of disembodied intelligence. (There may be impacts from second-order and higher effects, but this is hard "by definition.")
AGI-level robots can do this.
> prostitutes, models, acrobats, painters, etc
You have a point. Perhaps I should have said "overwhelming majority of human jobs", instead of "all human jobs".
If AI replaces all labor in two years, unemployment insurance and savings can comfortably carry me over the transitional period to universal income or obsolescence of money.
Should AI stagnate at the point where 70% of software engineering jobs are gone, I am going to have a problem.
Concretely, I imagine that I would need to learn other knowledge work. If AI can also perform it, I would have to go into blue collar work.
You are in denial. If you don’t like this the answer was what the Unabomber proposed the whole time.
The Unabomber (Theodore Kaczynski) killed 3 people and injured 23 others across 16 bombings between 1978 and 1995
It's absolutely incredible to see this talked about as a "reasonable" position on AI taking jobs.
edit: updated as it is not clear if Mosura is being serious or exaggerating for rhetoric
You cannot stop the advancement of technology towards AI without the threat of violence towards those that pursue technology. The Unabomber was at least honest about that.
> edit: updated as it is not clear if Mosura is being serious or exaggerating for rhetoric
You are the HN manifestation of that woman’s face when interviewing Sydney Sweeney.
As an already slightly outdated senior dev in the UK rapidly approaching 50 and not wanting to touch generative AI / agentic dev… … I'm increasingly looking forward to waiting tables, being a hospital porter, or anything else away from tech!
I'm thinking that if I get out soon, I might beat most of the rush and be safely employed before many of the rest of you are forced down the same path.
[more than a little defeatist, yes, but I've be been unhappy in dev for years so for me the AI driven displacement is just the final nail rather than the main big new problem]
Just out of curiosity, do you have a specific objections to using the technology? I totally understand the burnout however (I'm rapidly approaching 50 myself).
TBH I consider myself a senior in name only, or because I've been around a long time! I've done my best to avoid anything that looks like managing or leading people (or things!). When I was first given such a title I think people were somewhat offended that I wasn't overjoyed (“you can call me what you want”)! I want to do things, tinker with details, etc, not manage people/stuff.
> Just out of curiosity, do you have a specific objections to using the technology?
I want to do things, tinker with details, etc, not manage people/stuff.
I also don't like how the tech came to be, in part via rampant piracy, and how it continues to grow by sucking resources out of everything else, so in at least small part there are principles involved. I know some SRs who railed against generative AI when it was taking jobs of their designer friends, but now they can do the same work with less JRs it is all fine and dandy.
My biggest problem before the Great Displacement was upon us was that I'm one of those people who really don't get on with remote work. While it was a lovely convenience to use occasionally (for instance when I need to work from elsewhere while looking after ill relatives, or when people were getting on my tits and I needed to get away to concentrate) doing it the majority of the time does not work for my mental health. I actually go into the office still, but when the people I'm working most directly with don't I'm still effectively working remote (thought at least I have the work/home separation, I suppose). Teams/equiv doesn't work for me: people on there are just names and faces on a screen and don't register in my head the same as someone in the same room does unless I specifically concentrate on that, especially those I've never met and likely never will. Since the family got its first landline I've not liked talking on phones, I've always been an “in person or in writing” type, so this isn't a new sensation for me. If agentic development continues along its current path, I see a not-too-distant future where I'd just be sat with nothing but 'king Claude for company day-in-day-out, and I'll be buggered if I'm going to do anything to help that come to pass.
Honestly, I've never been happier — that might technically say more about how far from happy I was before, rather than the extent to which I am happy now. But the main point is that I am not less happy.
I still work on side-projects but it's strictly for leisure or occasionally some freelance stuff for long-standing relationships — just pocket money and nothing I'd lament being rid of.
I wish the SEC would start auditing claims about AI for job cuts, not sure if its within their ball park but there's people claiming this, when they have no receipts, and then stock goes up, if that's not market manipulation, I'm not sure what is.
And yet...
(Not saying AI helps with that yet.)
> then having more employees would be beneficial, because each can produce even more output
Junior programmers will be in demand again. Someone will have to fix the mountains of vibe coded technical debt.
I have a different opinion. I agree senior programmers will be in demand again, but not at junior price. Possibly even at a premium senior price. Why would we fix the vibe coded systems for less pay?
My thinking is most seniors could stop working and continue living just fine. Sure, we’d have to cut down travel vacation and other luxuries, but most if not all of us own our homes and have enough passive income to not have to worry about basic bills and food, so I don’t see us working for less.
We already have disruptions in Software engineering, image generation, video generation, etc.
Humans might not like AI, Companies do and with Claude we are no longer sitting on the right side, if a Product manager tells a Boss that the PM now can do also code, the boss will listen.
But we will see other effects in the next 5-15 years: population pyramid is crashing which will increase demand for people. We already have high demand for other jobs.
so i can see a transition for a lot of white collar workers to blue collar workers which will continue the disruption for everyone.
Then the robots come along and AI integration is done and we will have a real issue on our hand. When will this come? No clue i don't think its unrealistic to expeirernece this in my lifetime.
Does anyone remember expert systems?
Heinlein in his book The Moon Is A Harsh Mistress imagined that people in the future would find it worthwhile, even necessary, to learn an artifical language LogLan so they could talk with computers. LogLan (Logical Language) would be designed to be totally unambiguous, a necessity for speaking with a computer. He didn't imagine that we'd simply throw so much compute at the problem that we'd teach computers to understand our vague, mumbled natural languages in all their diversity.
https://en.wikipedia.org/wiki/Lojban
Sadly this does not let you speak to computers, just other language nerds with a lot of time.
Investors for sure don't care about people. Replace Accenture with OpenAI and you replace 800k people immediadly.
Geoffrey Hinton said it in his Videos: a normal person costs A LOT of money to train, educate, onboard, keep etc. they forget things etc.
AGI you train once, teach once then clone and copy and just run it.
We even might already crossed the line were it is already more cost efective to train 1-5 frontier models something it doesn't know yet than teaching this to a 1000 humans.
Unless if they take down the flat rate subs and you have to pay api prices, usage will drop a lot. Right now I freelance using gpt, if I had to pay api prices I'd probably lose 50% of the income, if not more, with the ~40% tax on top I might as well spend my time doing something else
This is a little dicier in post-agent AI, because it's easier for casual users to automate power-user consumption, but the providers have done decently in discouraging that.
I doubt opanai and anthropic will go that road
To give you an example of model in this class, the DeepSeek V4 Flash preview is a 284B A13B model, with a native quantization mixing 4 bit and 8-bit values. You can easily run it on an RTX Pro 6000 Blackwell (or 2) at a reasonable quant, especially if you offload the MoE weights to 48-64GB of system RAM. This costs US$11,800 at Microcenter right now, and it will work in any gaming box with decent cooling and a modern 1000W power supply. Over the lifetime of the card, an entire system would cost you under $4,000/year. Power is about 300W for the card (either a blower model, or a workstation model with the power cap), and another 150W or so for the rest of the server.
Or you could buy it on Open Router from dozens of different commodity vendors, starting around $0.09 per million tokens input, $0.18 per million tokens output. This is a competitive market price, so some of the providers might be losing money or reselling surplus capacity. But given the underlying hardware costs, the numbers are in the ballpark. In other words, if you're willing to settle for lower-quality tokens, you can get roughly Sonnet 4.5 for close to free, or as a modest capital expense for a successful freelancer. Halfway decent tokens are cheap, and you can generate them in-house!
Sure, and people can just build their own dropbox for like $250 too.
How many people will bother though
Or you could just take your credit card and spend $20 on credits at https://openrouter.ai/deepseek/deepseek-v4-flash. I'm not sure that I could manage to spend even a $1/day at those rates.
My larger point is that while frontier tokens are a near-monopoly and who knows what they really cost, many real-world workflows can be run using commodity tokens, or even served in-house by anyone who can afford to hire US or EU programmers. And if you're willing to settle for what would have been a state-of-the-art coding model in October 2025, commodity tokens are close to free.
No, there's is a way to stop it. As obvious counterexample: go all Butlerian Jihad: death sentence for anyone caught "making a machine in the likeness of a human mind." It's doable with a few laws and treaties. I'm not saying that's the only way it could be stopped, it's just a way it could.
I'm fucking sick of TINA, especially with arguments from vigorous assertion.
So the moment investment dries up, models stop improving.
However, it's likely we'll get good 80/20 solutions where you get most of the performance of the then-unsustainable high end models for significantly less compute.
Nearly there today. https://www.anthropic.com/institute/recursive-self-improveme...
Uber's a poor example because it did IPO in 2019 and its stocks are up around 75% since then.
Did you? If you keep on reading, the parent comment further implies that both AI companies and uber are "vibes based". But that's hard to square with the fact that 7 years since IPO the share price is up 75%. Of course, you can argue we're still in the irrational exuberance stage, but that just creates a situation where you can never be called wrong.
Uber has severely underperformed the market. Passive S&P500 index funds had a total return of more than double that in the same time period, while also having less risk than investing in a single stock.
Obviously much worse outcomes are possible than Uber's, but it's far from a massive success, especially after all that hype! This isn't to say the old common wisdom about excessive funding rounds was valid, but it's not completely invalid either.
I don't think so. They say the rules are vibes-based (or just wrong). That means you cannot use the rule to draw conclusions. That does not mean "the opposite of the rule is true" or anything like that. Nor does it mean that the AI companies resemble Uber in any significant way. They are just saying that since the rules are not real, you cannot say something like "this business will obviously fail, because it has bad fundamentals". The fundamentals are only part of the equation.
Absolutely not a reasonable assumption at all. Most people are not very good at their jobs.
> What's the steelman argument here?
If you're right about your prediction, but wrong about the timing, then you're wrong. The markets can stay irrational longer than you can stay solvent. Meaning, even if you know it's going to crash eventually, you don't know if it's tomorrow or in three years. Betting on stocks going down is extremely risky -- you have to get the timing pretty much exactly right, or you lose everything (potentially much more than you put down, depending on how you do it). Meanwhile, bull positions are basically free money until this point. Getting completely out of the market means you lose to inflation, although many investors have taken this strategy anyway.
These people think they're on the verge of creating a technology which, at a minimum, would constitute an unprecedented superweapon (and, at the extreme, would usher in a new era of civilization). Even if you don't buy-in to the take that one of these companies will reach a singularity and create a superintelligence, the cybersecurity implications alone is enough to put these products into a category outside the confines of profitability. We're already starting to see these implications become reality.
If the US NEEDS an advanced AI on an existential level, then it doesn't really matter how much it costs to make or whether or not it can produce a profit. It'll be valuable one a scale where financials like that just don't apply.
They will release MacBook Pro M10 or whatever that can comfortably run Opus5 levels of performance local model for software development/general ai that is baked in the MacOS for 7k USD.
> Increasing Corporate Skepticism: The news is full of stories of corporations that are throttling the employee use of AI since the costs to use the software are a lot higher than expected.
Real AI spend is out of control, with the news is full of stories about corporations trying to keep a lid on it, but also real AI spending is low and concentrated to a few firms.
I don't know. This doesn't feel very coherent to me, but rather like a collection of assertions that are adopted because they individually say something bearish about the industry.
Certainly some investments will have been overreaches, but I find it pretty unlikely that any of the compute build-out to date is going to be left sitting idle one or two or five years from now.
It’s not the same people doing both.
Also, cheaper RAM would be nice.
> The news is full of stories of corporations that are throttling the employee use of AI since the costs to use the software are a lot higher than expected
is not something I've seen. News is not "full of stories" of this behaviour. There are a few anecdotal stories here and there.
People say the AI bubble will burst, and I also think it will. But I want to think about how it might differ from other bubbles.
One positive factor is that during the internet startup era, there was almost no revenue. But now, big tech companies are generating profits and can absorb AI-scale losses.
The dot-com bubble burst, but the internet ended up being far more valuable than the bubble itself. I think the value was priced in early, and AI will be similar.
The core issue is always the same: there's a bubble and there are warnings, but you never know when it will collapse. And the people who act first on the warnings have to give up on the upside. The only winners are those who sell everything just before the crash. The problem is knowing when that moment is.
One difference from the dot-com bubble is that after it burst, a second wave of entrepreneurs leveraged the cheap telecom infrastructure that was left behind. So after the AI bubble bursts, will OpenAI and Anthropic fall and new companies emerge? I'm skeptical about that. Something feels different this time.
My biggest concern, though, is that if I suddenly couldn't use AI anymore, I'm worried about how long it would take to recover the coding skills I've lost after nearly a year of barely hand-coding. AI has become too deeply embedded in my life
This hasn't happened yet for AI, and there's a good chance it won't happen. Ie neither OpenAI nor Anthropic go public, or if they go public, the reception is "meh"
Nevertheless, independent of wider market sentiment, imho there is still a bubble in "closed models". One could, eg, pay attention to OpenAI/Anthropic/Copilot/Google's monthly new consumer subscriptions (Oversubscribed but in a hyperhyperreal sense lol. It seems that there are too many people queuing for this new restaurant, and those that have tasted the food all think it's ok but not healthy enough ("productivity mirage"). but people on the street can't see it because the bookings are all online. It may be that Gemini pro's subscriptions may be the last one to fall off the cliff. When that happens even the "open cloud models" bubble may pop (after everyone sees Google sunsetting Gemini, or making it exclusive to Apple lol)
US, not Korea https://archive.ph/WBSCj
For example:
- Anthropic makes a profit right now and is seemingly on an exponential upward trajectory, so the debt being too much for it doesn't seem compelling to me.
- AI technology continues to get better exponentially and doesn't have any clear sign this trend is flattening. If anything, it's accelerating. So it's plausible the investor value is legitimate for these companies given the massive potential for continued profitability.
- I would say markets are typically very good predictors of the future. Many sophisticated investors know about the case for the future crash and are still buying at these valuations.
I am open to being wrong, but the assessment in the blog seems one-sided to me.
Polymarket currently puts the chance of such a downturn at 20% by December 2026. Seems like most people would put higher chances here, but I'm not convinced by the arguments. (https://polymarket.com/event/ai-bubble-burst-by)
- if every autonomous self-driving car is going to need some sort of edge AI data center nearby (for instantaneous exceptional incident handling)
- if every household robot is going to need some sort of edge AI data center nearby (for instantaneous exceptional incident handling)
- if there are going to be millions of self-driving cars and autonomous robots joining us in the next decade
...then we will need lots more semi-conductor chips, and data centers in lots more places, in the next decade.
But those silver boxes at the side of the road? Might see many more of those, and much larger, maybe red hot on top
https://www.advisorpedia.com/media/media/2026/02/23/trailing...
Somehow we've managed to turn software from a machine for printing money to a machine for setting it on fire.
Btw the $2T/year number is just wrong, thats higher than current capex.
I often see developers giddy about the idea of AI being a bubble and waiting for the crash, but I suspect this event could actually be particularly terrible for us.
Let’s argue about SQL vs NoSQL again. I miss that.
https://startupfortune.com/cursors-ai-agents-rebuilt-sqlite-...
And this: https://github.com/swapnil404/mineSQL (SQL implemented in a minecraft world).
If you haven't noticed it's because you've stayed too close to the homeland.
The US is in a position where they have to debase the dollar all the way to zero and the Japan carry trade will slowly unwind for the worst as soon as the Bank of Japan (BOJ) begins to increase rates in more than 30 years.
Someone has to pay for all the trillions of dollars of debt for the data center build out that is being hidden from the balance sheets of the big tech companies.
So its cleaning the house of parasites o clock.
The US has lost its appeal as a hegemon. Constantly threatening allies with invasion isn't helping either.
It's lost those titles when it started behaving that way.
I'm already on it, man, don't need to convince me!
The US produces stuff, and it produces ideas. Yeah, it imports a lot of stuff too. But it exports things like food, movies, and software.
Without AI, the financial sector may take a huge hit. But the US economy? It will still be here, still producing things that are valuable.
All four frontier models contracted with the Pentagon. Claude already having been involved in Operation Abduct Sitting President & Use Active Denial Systems to Make People Vomit and Shit Themselves 'because'. Institutional and corporate capture . Foreign nations now having our models trained to facilitate tailored political orientation. And the fact that these systems, for all their limitations perennially bewailed, have tremendous capacity as force multipliers for all the things civil rights activists were previously terrified about apparently just before all our problems were magically solved.
Snowden docs x³
Things that once took fusion centers, thinktanks, research, etc and thousands of manual processing hours -- now real-time, instant, with real-time 'understanding'.
I think the AI crash will be primarily comprised of collectively lamenting all the obvious things we overlooked.
And with great irony, the only competitor (or savior) to total US asymmetric top-down domination is China. Or wait, I think I got that backward... https://www.wired.com/story/super-pac-backed-by-openai-and-p...