I had the same thoughts as much as 2 years ago on how this will play out. Unlike with railroads and fiber optic cable however, the core infrastructural asset of GPUs tends to become rapidly obsolete after about 5 years.
Commoditization of this scale of compute is definitely going to be a boon for many fields of research. Unfortunately fundamental public research is exactly what is being cut right now in the US.
Long term, I think the real winners are going to be in robotics. Still an unsolved field, but Waymo proves that even a nearly 20 year slog to the finish line is viable. And robotics infrastructure may be more robust to obsolescence than the underlying compute. I find it odd so many companies are making humanoid robots though... Over engineering that reeks of bubble economics and possible fraud.
I think the allure of humanoid robots is that they are drop in replacements for agents in a world desgined for humans.
If you want your robot to be a helper around the general populations houses for example, you would aim to make a general purpose bot capable of stairs, ladders, lying down, reaching high, stepping over things, holding awkward weights and loads while doing all of the above. Pinch, twist, push, pull, in all degrees of motion a human has etc.
Yes the thing that might be missed in the point about how the big buildout of GPU compute is going to be the backbone of the future etc is that, unlike railroads and dark fibre, the GPU compute gets obsolete really really quickly. So it's not the same.
I had a friend who got a Sun cluster for basically free when the 2000 dot com bubble burst. And when we were doing recreational math contests a couple of years later it was slower than our laptops.
So it is very likely that a load of today's GPU compute is very competitive next year or the year after?
The AI bubble bursting will kill investment in the next gen hardware in the west.
But china will come to market with its first gen that it is currently building to replace its dependency on the west and will leapfrog the west etc. China isn't really completely dependent on competing in our AI bubble, its using AI for its own things and will plough on even when the west bubble bursts. Seems obvious?
Still, there has been so much talk about the AI bubble bursting last week and this is the the best writeup.
The half-life of iron is pretty low too, the advantage of the rail system is what it allowed us to do when it was cheap enough.
All the investment in AI should help bring infrastructure up to a higher level, power distribution and cooling for example are at a much higher level than would have otherwise been.
Who knows what use that might have if it suddenly becomes incredibly cheap.
> however, the core infrastructural asset of GPUs tends to become rapidly obsolete after about 5 years.
Is it all about the actual GPUs though, is that the only "infrastructure" being built? A list from the top of my head of things that I'd say do last:
1. Data center buildings (take a while to build, contents completely aside).
2. Organisations and processes for running operations and procurement in said data centers - doesn't take decades to build for sure, but it's something worthwhile to already have.
3. Advances in the actual chips, i.e. more powerful processing units.
4. Advances in chip fabrication.
5. Chip fabrication facilities and organisations (similar to #1 and #2).
So sure, GPUs are highly temporary. But a lot of the things being developed and built around them much less so.
I do think one possible bubble burst scenario is that we'll have cheap compute available for decades but not a lot of great ideas of what to do with it. That is not unlike the 2000s I suppose.
Just look at every Nvidia GTC presentation for the past 3 years and Chinese investments into companies like Unitree. The plethora of robotics startups that all build humanoid machines (rather than wheeled arms)
I suspect that if AI doesn't use the compute then something else will happily fill the gap - like building more lanes on a highway. It will be interesting to see what fills the GPU compute glut though.
Regards robot form factor; I'd rather R2D2 than C3PO. I don't want anything approaching the Uncanny Valley; I want a machine that does handy things!
> The speculation democratized investing in a way never seen before. Clerks, shopkeepers, and domestic servants, people who had never owned stocks before, mortgaged their homes and borrowed money to buy railway shares.
I like the term "democratize investing" here. "We're granting the masses the privilege of dumping their lifesavings into this overhyped project, so we can make a clean exit".
I recently gave an unprompted tirade about bubbles creating pockets of fake economic activity with no “real” economic output or value creation tied to them, and how sometimes, the circulatory systems in those bubbles feed into themselves.
E.g. someone borrowing against their higher property value(s) to put a down payment on another property.
Leverage is the amplifier. And I don’t see many self-circulating capital flows. I expect contractions to be reasonable for this bubble, or more realistically industry stagflation.
I think this is directionally correct. But we tend to remember the technologies which became mainstream after the bubble burst and forget those that fizzled or found a much less world-changing niche.
Blockchain, NFTs and 3D printing are still around and have vacuumed up billions and billions without the average person being able to tell an impact on their lives.
Mostly agree with the author. But I also believe that this is very emergent tech and we’re still figuring out the best way to actually make use of it
The clearest example is in AI generated visual content. If you dig through what people are doing, its clear that only a very small % of users are actually getting truly high-quality, ready-for-production content, while the rest are just prompting in pure slop. There is a skill level to this that hasn’t really permeated the mainstream
Once that happens, we might see some of that 95% waste figure change to, maybe, 50% waste
> Creative destruction is brutal math. The capital? Gone. Completely vaporized. But infrastructure isn’t stock certificates. Those fiber optic cables didn’t vanish when Pets.com did. The data centers kept humming after Webvan went dark. All that ‘wasted’ investment had already transformed into something physical. The pipes, servers, and networks that would become the foundation for Google, Facebook, Amazon Web Services, and the digital transformation that actually did change everything. The bubble’s victims unknowingly funded the future. They just paid a decade too early.
The flow of money to spur innovation is exactly like "Cambrian Explosion". We should do this more often, with biotech and future fields to come.
> Creative destruction is brutal math. The capital? Gone. Completely vaporized. But infrastructure isn’t stock certificates. Those fiber optic cables didn’t vanish when Pets.com did.
I read quotes like this and reminded that it is common that people forget money is just a competitive resource we use to outbid each other for _real_ things. Money moves around, it isn't lost or "Completely vaporized", someone receives it at the other side of the transaction. It is still in circulation, it can still be used to outbid people for real things, just by different people.
Also, pets.com still exists, it just forwards to petsmart.com.
> I read quotes like this and reminded that it is common that people forget money is just a competitive resource we use to outbid each other for _real_ things.
But money is an abstraction of wealth, and wealth absolutely can be destroyed, in multiple ways:
1. It can be physically destroyed - if I break a window, that's wealth that is destroyed. That window now needs to be replaced, which costs materials and labor, which could've gone to building something new instead.
2. It can be spent on things that end up not used. If five years from now, those millions of GPUs are no longer in use, we created them for nothing instead of creating more of something people would use.
3. Wealth can be spent on the less important things, rather than the more important things. This is not exactly wealth being destroyed, just built more slowly, because instead of building lots of new wealth (via innovation, say) we're creating less valuable things.
I don't think any of the above are relevant to AI, btw.
There is hubris, but calling it a bubble simply does not check out, for one reason alone: If AI did absolutely nothing from here on out but give maybe a somewhat better version of current claude code (and it confuses me to no end that some people still refuse to see what is going on there, which admittedly are increasingly few of those who try, which makes sense because stuff gets better) that leads to, say, a ~2x dev speed up it, given the size of the market and how much software is missing still, AI as a whole would still be undervalued.
Of course, assuming that this would be the only thing where economic gains come from is already such a laughably bearish vision. It's just that that's all you need for the bubble-thesis to fall flat.
> Here’s what’s never happened before: everyone knows the script.
I'm don't think this is unique, most bubbles historically as far back as the South Sea bubble have had a lot of people aware of the irrationality, but investing in an attempt to profit from it.
I'd even go so far as to say, this is exactly what makes bubbles so volatile as opposed to normal "market corrections". If the dotcom boom had been all people who really believed they were sensibly evaluating the internet's financial potential, I don't think we'd have seen them jump ship quite so quickly.
I won't predict the future, but another point about historic bubbles: they almost all go on much further than people think they will before collapse.
Is AI a bubble? Probably.
Does that make it meaningless? Not at all.
Carlota Perez’s Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages (2003) is strikingly prescient--and worth revisiting
[thread]
I think we should be weary of the fact that Sam Altman having an interest in cutting VC funding for new AI products and competitors in order to condensate most AI funding into OpenAI and kill any new competitors, doesn't make him exactly a neutral player in all of this.
The "fearmongering" he is trying to create, can be seen as self-serving, so his opinions should be taken with a very big grain of salt.
A golden rule that has been very decent in the last decades - When my mom starts asking me "What's up with this tech" - it's most usually is a bubble ;)
The thing is, this is a financial bubble, not a capabilities bubble.
If you're worried about money, this is horrible news. If you just want to get shit done, it's great news, only if you can avoid losing personal agency long enough to survive the crash.
We're headed towards the hockey stick in terms of what people using AI can do. I'm rapidly learning that even ChatGPT5 can get confused, and lose sight of goals, but not in the hallucination variety, just the bog standard way people end up trapped in rabbit holes. I'm learning how to talk to it and get it back on track.
AI really can be productive, but it still needs guidance to be really useful.
My thoughts: I'm expecting the LLM-driven bubble to burst. Not AI in general, but it seems LLM has been soaking up all the investment dollars and LLM will eventually hit a wall in critical fields where an expert can poke holes at the results. The data centers go poof!
Now, the contracted land rights, water rights, and electricity rights, highly discounted will be of value. Kind of like the defunct railroad rights in your community that can't easily be repurposed into a bike path as the city does not have the rights.
So, I wonder if there will be companies formed to buy up the infrastructure rights (the contracted cheap power and water) and re-sell them to the city and counties that originally generated the contracts. Its too easy to see the naïve city and county officials succumbing to the hookers and other marketing techniques when these contracts were issued. Now they can be sold back at a profit.
The thing is the infrastructure that is being built now is owned by a small number of already large companies. Is it really that likely that Microsoft, Amazon, Meta/Facebook, Alphabet/Google, Oracle, Nvidia and all the companies Elon Musk is involved with are going to go bankrupt?
34 comments
[ 3.5 ms ] story [ 50.7 ms ] threadCommoditization of this scale of compute is definitely going to be a boon for many fields of research. Unfortunately fundamental public research is exactly what is being cut right now in the US.
Long term, I think the real winners are going to be in robotics. Still an unsolved field, but Waymo proves that even a nearly 20 year slog to the finish line is viable. And robotics infrastructure may be more robust to obsolescence than the underlying compute. I find it odd so many companies are making humanoid robots though... Over engineering that reeks of bubble economics and possible fraud.
If you want your robot to be a helper around the general populations houses for example, you would aim to make a general purpose bot capable of stairs, ladders, lying down, reaching high, stepping over things, holding awkward weights and loads while doing all of the above. Pinch, twist, push, pull, in all degrees of motion a human has etc.
I had a friend who got a Sun cluster for basically free when the 2000 dot com bubble burst. And when we were doing recreational math contests a couple of years later it was slower than our laptops.
So it is very likely that a load of today's GPU compute is very competitive next year or the year after?
The AI bubble bursting will kill investment in the next gen hardware in the west.
But china will come to market with its first gen that it is currently building to replace its dependency on the west and will leapfrog the west etc. China isn't really completely dependent on competing in our AI bubble, its using AI for its own things and will plough on even when the west bubble bursts. Seems obvious?
Still, there has been so much talk about the AI bubble bursting last week and this is the the best writeup.
All the investment in AI should help bring infrastructure up to a higher level, power distribution and cooling for example are at a much higher level than would have otherwise been.
Who knows what use that might have if it suddenly becomes incredibly cheap.
(this is my silver lining thinking)
Is it all about the actual GPUs though, is that the only "infrastructure" being built? A list from the top of my head of things that I'd say do last:
1. Data center buildings (take a while to build, contents completely aside).
2. Organisations and processes for running operations and procurement in said data centers - doesn't take decades to build for sure, but it's something worthwhile to already have.
3. Advances in the actual chips, i.e. more powerful processing units.
4. Advances in chip fabrication.
5. Chip fabrication facilities and organisations (similar to #1 and #2).
So sure, GPUs are highly temporary. But a lot of the things being developed and built around them much less so.
I do think one possible bubble burst scenario is that we'll have cheap compute available for decades but not a lot of great ideas of what to do with it. That is not unlike the 2000s I suppose.
I think at least with CPU:s the depreciation has slowed down a lot compared to 15 years ago.
Regards robot form factor; I'd rather R2D2 than C3PO. I don't want anything approaching the Uncanny Valley; I want a machine that does handy things!
I like the term "democratize investing" here. "We're granting the masses the privilege of dumping their lifesavings into this overhyped project, so we can make a clean exit".
What year is this from? The author might want to do a recent news search.
E.g. someone borrowing against their higher property value(s) to put a down payment on another property.
Leverage is the amplifier. And I don’t see many self-circulating capital flows. I expect contractions to be reasonable for this bubble, or more realistically industry stagflation.
Blockchain, NFTs and 3D printing are still around and have vacuumed up billions and billions without the average person being able to tell an impact on their lives.
The clearest example is in AI generated visual content. If you dig through what people are doing, its clear that only a very small % of users are actually getting truly high-quality, ready-for-production content, while the rest are just prompting in pure slop. There is a skill level to this that hasn’t really permeated the mainstream
Once that happens, we might see some of that 95% waste figure change to, maybe, 50% waste
This article is based off of the Altman bubbly comment.
The flow of money to spur innovation is exactly like "Cambrian Explosion". We should do this more often, with biotech and future fields to come.
I read quotes like this and reminded that it is common that people forget money is just a competitive resource we use to outbid each other for _real_ things. Money moves around, it isn't lost or "Completely vaporized", someone receives it at the other side of the transaction. It is still in circulation, it can still be used to outbid people for real things, just by different people.
Also, pets.com still exists, it just forwards to petsmart.com.
But money is an abstraction of wealth, and wealth absolutely can be destroyed, in multiple ways:
1. It can be physically destroyed - if I break a window, that's wealth that is destroyed. That window now needs to be replaced, which costs materials and labor, which could've gone to building something new instead.
2. It can be spent on things that end up not used. If five years from now, those millions of GPUs are no longer in use, we created them for nothing instead of creating more of something people would use.
3. Wealth can be spent on the less important things, rather than the more important things. This is not exactly wealth being destroyed, just built more slowly, because instead of building lots of new wealth (via innovation, say) we're creating less valuable things.
I don't think any of the above are relevant to AI, btw.
Of course, assuming that this would be the only thing where economic gains come from is already such a laughably bearish vision. It's just that that's all you need for the bubble-thesis to fall flat.
https://en.wikipedia.org/wiki/AI_winter
I knew about since like 2010 or before, anti-tech Luddite will act like it's never a thing, shatup.
I'm don't think this is unique, most bubbles historically as far back as the South Sea bubble have had a lot of people aware of the irrationality, but investing in an attempt to profit from it.
I'd even go so far as to say, this is exactly what makes bubbles so volatile as opposed to normal "market corrections". If the dotcom boom had been all people who really believed they were sensibly evaluating the internet's financial potential, I don't think we'd have seen them jump ship quite so quickly.
I won't predict the future, but another point about historic bubbles: they almost all go on much further than people think they will before collapse.
Perez goes back to the technology of canals:
* https://en.wikipedia.org/wiki/Technological_Revolutions_and_...
People getting excited for something (perceived as) new is part of the human character.
The "fearmongering" he is trying to create, can be seen as self-serving, so his opinions should be taken with a very big grain of salt.
If you're worried about money, this is horrible news. If you just want to get shit done, it's great news, only if you can avoid losing personal agency long enough to survive the crash.
We're headed towards the hockey stick in terms of what people using AI can do. I'm rapidly learning that even ChatGPT5 can get confused, and lose sight of goals, but not in the hallucination variety, just the bog standard way people end up trapped in rabbit holes. I'm learning how to talk to it and get it back on track.
AI really can be productive, but it still needs guidance to be really useful.
“When I see a bubble forming, I rush in to buy, adding fuel to the fire,” goes one of George Soros’s well-known quotes. “That is not irrational.”