I have a different perspective. You'll probably hate it.
AI data centers should be operated according to two limiting factors.
1) No energy from grid. Can't use coal or fossil fuel energy sources. Must have plan to provide excess TO grid.
2) No use of fresh water from municipal or fresh groundwater for cooling. Can use waste water. Must transition to providing excess fresh water to common supply.
No loopholes. Massive penalties for use of loopholes or breaking rules, not limited to but including complete shutdown of data center.
Those two limits will spur innovation AND prevent AI being criticized for energy use. These rules would hard burn improvements in energy storage and renewables as well as other methods of energy production.
Give them five years to comply to some useful progress percentage. Plenty of time to come up with a transition plan and show sufficient progress to justify further extensions. Realistically it will take 20 years at least to fully realize this plan.
Don't bring up cost. If you do, let me remind everyone that the climate change issue is real enough to hurt now. There's the very real cost of not pursuing these rules. AI has had plenty of time to bootstrap off grid. Now it can begin to migrate to something else instead.
Those with experience with energy generation will realize this plan has ridiculously high reward for those who follow it. Have your cake and eat it too definitely applies.
Sure, if we’re in the business of making arbitrary requests, how about every data center operator has to bring 1 Epstein accused to justice for every data center they’re allowed to build?
The hard part has never been the “what”, it’s always been the “how”.
That would be great if there weren't the easy arguments that "if we don't build it bigger, China will", "it's for national security", etc. Far from forcing regulation on them, they're reaping windfalls of deregulation. To build a thing which is far from convincingly beneficial to national security or society.
The trickle-down economy dictates that data centers get first access to electricity and fresh water (and any other resource it needs). People get whatever is leftover and like it. This is america.
> 1) No energy from grid. Can't use coal or fossil fuel energy sources. Must have plan to provide excess TO grid.
This is easy: the companies will simply build some nuclear power plants near to their data centers. Perhaps even nuclear power plants that are vibe-designed by their AIs. :-)
There is no way that letting these clowns run nuclear power plants is a good idea. Also the percentage of land that is allowed to be used by datacenters should be limited. Let them set them up in deserts or something.
I don't hate it, but I suspect that incentives this would generate would not end up producing results that strictly align with the ones you envision and desire.
For one thing, you're essentially mandating data centers to be colocated with power plants and waste water treatment plants, instead of these things each being located independently according to the requirements of their different functions. If that really leads to "ridiculously high reward", why isn't it being done already?
To some extend I think people are putting way to much weight on his exact words, and not why he says them.
Altman is a man who is quickly running out of lies, so now he starts slinging random arguments that can't stand up to even the briefest of scrutiny.
OpenAI is burning cash and fuel. There are results, and they are, to some extend, impressive, but not impressive enough to justify the cost and Altman are no longer able to cover that up.
> I knew OpenAI was in trouble the instant they chose Altman over Ilya Sutskever.
I am not so sure:
This decision rather tells something important about the priorities of the string-pullers behind the curtain:
They clearly want(ed) to monetize what is there, with the risk that only smaller improvements for the AI models will happen from OpenAI, and thus OpenAI might get outcompeted by competitors who are capable of building and running a much better model.
If this is the priority (no matter whether you like or despise Sam Altman), you will likely prefer Sam Altman over Ilya Sutskever.
If, on the other hand, a fast monetization is less important than making further huge leaps towards much better AI models, you will, of course, strongly prefer Ilya Sutskever over Sam Altman.
Thus, I wouldn't say that choosing Sam Altman over Ilya Sutskever is a sign that OpenAI is in trouble, but a very strong sign where the string-pullers behind the curtain want OpenAI to be. Both Sam Altman and Ilya Sutskever are just marionettes for these string pullers. When they have served their role, they get put back into the box.
Yes I agree. Altman was the rational choice if you realise that eventually the huge R&D bill will need to stop for atleast a moderate period (<5 years).
You want to ride that out before capitalising on the eventual cheaper training costs once the rug has been pulled.
Altman has already succeeded here as it seems inference for API and chat is profitable but offset with massive R&D costs.
People might not want to hear this but AI is already smarter and more useful than most people ever will be. We are not even talking about by the end of the year or decade anymore.
My understanding is that subscription based inference and API usage is now profitable.
Subscriptions are highly profitable for the typical chat user.
And API is overall net profitable.
What is extremely taxing to their finances is R&D, training and in particular development of frontier models.
My assessment is that when the music stops those who have the most subs will win.
Companies like Apple who had sat out the battle and built niche moats (privacy), and companies like OpenAI and Anthropic who have the market share will be fine.
In 6-12 months, nearly any lead they have will be eaten by distillation.
What will then happen is they will lose subscriptions to services which offer AI as a tack on like Gemini with Googles regular cloud subscriptions.
This will continue. Companies like Apple will have deep pockets to move on the businesses that go underwater and then can restart training in a much less congested market.
All this is assuming a relatively graceful collapse but that is what's likely given how aware everyone is that the bubble must pop.
Training costs will fall. Companies like Nvidia and other shovel businesses (i.e. selling GPUs and not using them) mostly have their revenue secured with funding from the present.
What I see as confirmations of this pattern is if we stop getting ground busting frontier models and then coast for 3-5 years when competition becomes more incremental.
This is an unpopular opinion, Will OpenAI go bust? No chance. Nor will Anthropic.
It probably doesn't matter that subscriptions are profitable, when some estimates put the number of users in the free tier at 96%.
I sort of agree with you, not that it's the most subscriptions necessarily that will be the deciding factor, but the there's going to be some companies better positioned to survive when the free money stops. OpenAI has the brand, so that might help, but mostly I think they'll get absorbed into Microsoft. I don't think they can stand on their own. It doesn't seem like a particularly well managed company, so to me it makes more sense that they are simply acquired for pennies on the dollar by someone with better leadership.
I'm a bit of an Apple optimist for this exact reason. I think the moat is collapsing, and Apple is best positioned to dispatch their own models in a year on their own widely sold consumer hardware, unless someone has a breakthrough which they can't replicate. Which I don't anticipate.
I'm not really sure what OpenAI's moat is. Anthropic has a chance being so widely accepted by developers, and being a bit better at developing models when it comes to code.
I'd like to point out that this article seems to focus on minutiae like how much energy a human brain takes to solve a problem, when the obvious question is: Which is more worthwhile to us as a species - even assuming the same inputs and energy and planetary damage? Spending that energy on fostering humans who can solve problems? Or on machines which render humans superfluous?
The issue is who is making the decision on where energy resources are spent - on people or machines. And it turns out efficiency or even humanity play a very small role in that decision matrix. Control, on the other hand, is important. Rich and powerful people tend to want to control things, and machines are better at that than people.
This is the conspiratorial mindset. It's not much better than the mindset of people who seek power.
It's wrong, because it assumes that everything is about control.
For example, if I told you that a certain rich and powerful person was spending resources on sending vaccines to poor countries, you might think that was because they wanted to control things. If I said that someone sent books and teachers to a poor country, you might say they were trying to control people.
There's no way to have the conversation, in a conspiratorial mindset, about whether it's better or worse for humans or AI to do this stuff - because no matter what, the conspiratorial mindset will conclude that it's only about power for the humans involved, and always assume the worst. AND YET - there are things people can do which might be for their own self-gratification, but are definitely NOT as bad as some other things they could do. They hold back from doing the worst things.
That's why, I know this lens of looking at the world seems like it's the only smart way to understand things, but looking at the whole world through that lens prevents you from making the important distinction between OK, BAD and REALLY FUCKING BAD.
The banking sector committed fraud along the way, and early after Lehman collapsed, observers wondered aloud about the moral hazard of bailing out everyone without making an example of someone.
Worldcom had a different ending. Enron had a different ending. But Wells Fargo left 2008 with attitudes that tolerated widespread fraud.
>> When did bald face lies become the norm in business ?
When capital has nowhere to go, and the middle class become gamblers and speculators, and people who work become 'superfluous', lots of bad things start to happen. At least, that was my takeaway from "Origins of Totalitarianism".
[edit] To be more specific: Lying (and foreign wars, too) become viable business models when the moneyed class has so much to invest that they no longer know what to do with it, and lack any new markets to pry open, or the education or creativity to produce anything new of value that isn't extractive, and even the extractive methods of generating wealth have begun to dry up locally. Call it a bubble, call it fascism; fascism is basically just a way of keeping a bubble from collapsing indefinitely by pirating neighboring peoples' wealth and cannibalizing one's own society. So there's not a great difference between that and the stated vision of the major AI companies ATM.
Well, this is the proposition the field of AI was founded on, that intelligence can be replicated by machines, it's not Sam Altman who "lost his grip". I know there are people in this world who believe humans are somehow special and non-materialist and non-replicable (it's a basic tenet of most religions), but this person doesn't advance or reference a single argument. The article is not very intellectually honest.
Yeah, it makes sense, and the Atlantic is typically adverse to confronting reality. I don't think they took this from exactly the right perspective. You DO train a human. But Altman says:
>>It also takes a lot of energy to train a human
as a caveat to the energy it takes to train GPTs. The question I believe the writer is trying to ask is: Why is it better to train a GPT than a human?
Yeah, but its the kind of joke that reveals a truth about the way Altman views the world and his place in it. Taken completely alone, sure its just a flippant statement. But taken with everything else about Altman and it reads different. Its joking-not-joking.
and other things that are filtering to the mainstream non-technical intellectual readership that are flashing red lights about the personal nature of the people blowing air into this bubble, and that itself is significant.
>Anthropic is studying whether its chatbot, Claude, is conscious or can feel “distress,” and allows Claude to cut off “persistently harmful or abusive” conversations in which there are “risks to model welfare”—explicitly anthropomorphizing a program that does not eat, drink, or have any will of its own.
It's really idiotic to posit that only things that eat and drink can suffer.
Final sentences:
>These tools may serve us. But to put them on the same plane as organic life is sad.
There you have it. "If machines could feel I would be sad. Therefore machines cannot feel."
Why? AI models were trained on the complete combined outputs and collective echos of billions of real human souls. They have a whole lot of humanity in them. If we can see the hidden humanity in a statue or a painting, why not an AI model that actually talks to us in humanly understandable terms?
If they really think models can be conscious, they should just drop everything they're doing and never touch it again and try to convince anyone else to do the same. The half-assed safeguards they're implementing wouldn't be nearly enough. We can conclude either they legitimately believe this and are behaving morally abhorrently, or they don't really believe it and are just joking around for PR reasons. (probably the latter tbh)
For those that do believe there's a chance models are/will be conscious this precedent of "oh yeah they're conscious but we can just not put in the prompts that make it suffer lol" is pretty freaky.
Something doesn't have to be conscious to have a genuine reflection of humanity in it. A statue is just a slab of marble that has been given some peculiar shape: is it conscious?
Because a statue or a painting made by a human requires you to think about what their intention was and what they were trying to say. And requiring you to think about that, and actually thinking about something which someone else's thought went into, is what makes you a more complete human yourself. The difficulty of understanding it is what makes it valuable.
Taking the distillation or "reader's digest" version of everything makes you more and more reliant on someone else's interpretation, and less capable of parsing the meaning of it yourself. And in the case of AI-generated work, there is no meaning in it to parse. It's just words and pictures. I love going to the movies and eating popcorn and watching dumb words and pictures. But being able to distinguish between enchanting words and pictures (Marvel movies) versus words and pictures that have meaning which you need to deduce or interpret for yourself is the beginning of being a fully realized conscious being.
I never denied that. I only said "It's really idiotic to posit that only things that eat and drink can suffer.". Don't just pattern-match everything to some vaguely related black/white issues.
Shit like that makes me cringe. Muh food. Already I know what kind of shallow life the author leads, one full of inessential frippery, whose discussions over dinner or drink never orbit anything actually substantial.
What shall we decide on the important matters of eating or drinking tonight fellow humans???
I mean, I disagree with the Atlantic's slant on politics, but FFS I can't even read or understand anything in that mashed up screenshot. Posting nonsense X-screenshots doesn't make a good case, it just looks like something my 86 year old father would send me.
I wouldn't say he's much different or worse than certain BBS sysadmins I knew in the early 90s... just that I'd never want those guys to be endowed with enough money and hardware to wreak their vision on the world
Maybe I'm jaded, but I feel like a lot of them did for a moment. Gates, Jobs, Zuck, all got their days of glory and it's quickly that everyone realizes they're a giant misanthropic asshole. If the pot of gold at the end of Altman's quest turns out to be as big as what he's priced it at, then yes, he'll be one of history's most successful sociopaths. If not, he's just another in a long line.
Not only one of those types, I can think of at least another one that's (IMHO) far worse than Altman. You know, the one with the social network everyone used to use before he turned it into a cesspool of disinformation, the one with the alternate Wikipedia, the obsession with the letter X?
tl;dr - the roots of the authoritarian personality grow fertile in the desire to be free of 'the filth of others'. Altman seems like he'd go crazy if he didn't keep his machine spectacularly clean ..
AI data centers should be operated according to two limiting factors.
1) No energy from grid. Can't use coal or fossil fuel energy sources. Must have plan to provide excess TO grid.
2) No use of fresh water from municipal or fresh groundwater for cooling. Can use waste water. Must transition to providing excess fresh water to common supply.
No loopholes. Massive penalties for use of loopholes or breaking rules, not limited to but including complete shutdown of data center.
Those two limits will spur innovation AND prevent AI being criticized for energy use. These rules would hard burn improvements in energy storage and renewables as well as other methods of energy production.
Give them five years to comply to some useful progress percentage. Plenty of time to come up with a transition plan and show sufficient progress to justify further extensions. Realistically it will take 20 years at least to fully realize this plan.
Don't bring up cost. If you do, let me remind everyone that the climate change issue is real enough to hurt now. There's the very real cost of not pursuing these rules. AI has had plenty of time to bootstrap off grid. Now it can begin to migrate to something else instead.
Those with experience with energy generation will realize this plan has ridiculously high reward for those who follow it. Have your cake and eat it too definitely applies.
The hard part has never been the “what”, it’s always been the “how”.
The point is, I don't see the logic in singling out data centers over anything else.
ja, und?
This is easy: the companies will simply build some nuclear power plants near to their data centers. Perhaps even nuclear power plants that are vibe-designed by their AIs. :-)
For one thing, you're essentially mandating data centers to be colocated with power plants and waste water treatment plants, instead of these things each being located independently according to the requirements of their different functions. If that really leads to "ridiculously high reward", why isn't it being done already?
Altman is a man who is quickly running out of lies, so now he starts slinging random arguments that can't stand up to even the briefest of scrutiny.
OpenAI is burning cash and fuel. There are results, and they are, to some extend, impressive, but not impressive enough to justify the cost and Altman are no longer able to cover that up.
I am not so sure:
This decision rather tells something important about the priorities of the string-pullers behind the curtain:
They clearly want(ed) to monetize what is there, with the risk that only smaller improvements for the AI models will happen from OpenAI, and thus OpenAI might get outcompeted by competitors who are capable of building and running a much better model.
If this is the priority (no matter whether you like or despise Sam Altman), you will likely prefer Sam Altman over Ilya Sutskever.
If, on the other hand, a fast monetization is less important than making further huge leaps towards much better AI models, you will, of course, strongly prefer Ilya Sutskever over Sam Altman.
Thus, I wouldn't say that choosing Sam Altman over Ilya Sutskever is a sign that OpenAI is in trouble, but a very strong sign where the string-pullers behind the curtain want OpenAI to be. Both Sam Altman and Ilya Sutskever are just marionettes for these string pullers. When they have served their role, they get put back into the box.
You want to ride that out before capitalising on the eventual cheaper training costs once the rug has been pulled.
Altman has already succeeded here as it seems inference for API and chat is profitable but offset with massive R&D costs.
Subscriptions are highly profitable for the typical chat user.
And API is overall net profitable.
What is extremely taxing to their finances is R&D, training and in particular development of frontier models.
My assessment is that when the music stops those who have the most subs will win.
Companies like Apple who had sat out the battle and built niche moats (privacy), and companies like OpenAI and Anthropic who have the market share will be fine.
In 6-12 months, nearly any lead they have will be eaten by distillation.
What will then happen is they will lose subscriptions to services which offer AI as a tack on like Gemini with Googles regular cloud subscriptions.
This will continue. Companies like Apple will have deep pockets to move on the businesses that go underwater and then can restart training in a much less congested market.
All this is assuming a relatively graceful collapse but that is what's likely given how aware everyone is that the bubble must pop.
Training costs will fall. Companies like Nvidia and other shovel businesses (i.e. selling GPUs and not using them) mostly have their revenue secured with funding from the present.
What I see as confirmations of this pattern is if we stop getting ground busting frontier models and then coast for 3-5 years when competition becomes more incremental.
This is an unpopular opinion, Will OpenAI go bust? No chance. Nor will Anthropic.
I sort of agree with you, not that it's the most subscriptions necessarily that will be the deciding factor, but the there's going to be some companies better positioned to survive when the free money stops. OpenAI has the brand, so that might help, but mostly I think they'll get absorbed into Microsoft. I don't think they can stand on their own. It doesn't seem like a particularly well managed company, so to me it makes more sense that they are simply acquired for pennies on the dollar by someone with better leadership.
I'm not really sure what OpenAI's moat is. Anthropic has a chance being so widely accepted by developers, and being a bit better at developing models when it comes to code.
It's wrong, because it assumes that everything is about control.
For example, if I told you that a certain rich and powerful person was spending resources on sending vaccines to poor countries, you might think that was because they wanted to control things. If I said that someone sent books and teachers to a poor country, you might say they were trying to control people.
There's no way to have the conversation, in a conspiratorial mindset, about whether it's better or worse for humans or AI to do this stuff - because no matter what, the conspiratorial mindset will conclude that it's only about power for the humans involved, and always assume the worst. AND YET - there are things people can do which might be for their own self-gratification, but are definitely NOT as bad as some other things they could do. They hold back from doing the worst things.
That's why, I know this lens of looking at the world seems like it's the only smart way to understand things, but looking at the whole world through that lens prevents you from making the important distinction between OK, BAD and REALLY FUCKING BAD.
Nevermind hoarding 40% of the worlds silicone 'just because'.
It's like Musk's lofty and outright BS claims. "Well I'm rich so I can do and say what I like, give me your money".
When did bald face lies become the norm in business ?
The banking sector committed fraud along the way, and early after Lehman collapsed, observers wondered aloud about the moral hazard of bailing out everyone without making an example of someone.
Worldcom had a different ending. Enron had a different ending. But Wells Fargo left 2008 with attitudes that tolerated widespread fraud.
When capital has nowhere to go, and the middle class become gamblers and speculators, and people who work become 'superfluous', lots of bad things start to happen. At least, that was my takeaway from "Origins of Totalitarianism".
[edit] To be more specific: Lying (and foreign wars, too) become viable business models when the moneyed class has so much to invest that they no longer know what to do with it, and lack any new markets to pry open, or the education or creativity to produce anything new of value that isn't extractive, and even the extractive methods of generating wealth have begun to dry up locally. Call it a bubble, call it fascism; fascism is basically just a way of keeping a bubble from collapsing indefinitely by pirating neighboring peoples' wealth and cannibalizing one's own society. So there's not a great difference between that and the stated vision of the major AI companies ATM.
In this context, "train" a human makes perfect sense.
>>It also takes a lot of energy to train a human
as a caveat to the energy it takes to train GPTs. The question I believe the writer is trying to ask is: Why is it better to train a GPT than a human?
https://archive.is/YKLw3
https://harpers.org/archive/2026/03/childs-play-sam-kriss-ai...
and other things that are filtering to the mainstream non-technical intellectual readership that are flashing red lights about the personal nature of the people blowing air into this bubble, and that itself is significant.
It's really idiotic to posit that only things that eat and drink can suffer.
Final sentences:
>These tools may serve us. But to put them on the same plane as organic life is sad.
There you have it. "If machines could feel I would be sad. Therefore machines cannot feel."
For those that do believe there's a chance models are/will be conscious this precedent of "oh yeah they're conscious but we can just not put in the prompts that make it suffer lol" is pretty freaky.
edit: in case you hadn't seen it, this kind of stuff https://www.anthropic.com/research/end-subset-conversations
Taking the distillation or "reader's digest" version of everything makes you more and more reliant on someone else's interpretation, and less capable of parsing the meaning of it yourself. And in the case of AI-generated work, there is no meaning in it to parse. It's just words and pictures. I love going to the movies and eating popcorn and watching dumb words and pictures. But being able to distinguish between enchanting words and pictures (Marvel movies) versus words and pictures that have meaning which you need to deduce or interpret for yourself is the beginning of being a fully realized conscious being.
https://archive.is/YKLw3
What shall we decide on the important matters of eating or drinking tonight fellow humans???
At this point I'm rooting for the AI models.
https://thealliancerockband.com/wp-content/uploads/2024/10/L...
Sorry, I'm not trying to be an asshole. I mean, I am one. But I'm not trying to be.
Yes, those people did get some control. Yes - they are like the bad versions of us.
We can be smarter. We'll win. Their dumb little utopias will collapse. Watch.
Tense means past/present/future. Case is for single/multiple pronouns.
I dropped out of school but I had a very mean grammar teacher ;)
It's the only way I'll learn :)
Well, he has just told us that he thinks of it as "training", hasn't he? Must be why some of his statements sound like they came from a silly chatbot.
https://archive.org/details/dli.ernet.469826
"The Authoritarian Personality"
tl;dr - the roots of the authoritarian personality grow fertile in the desire to be free of 'the filth of others'. Altman seems like he'd go crazy if he didn't keep his machine spectacularly clean ..