[lifeisstillgood]

On the token limits etc - one assumes that OpenAI et al are able to “hire expert in field, and let them spend the equivalent of a million dollars of tokens” because they are not actually selling their complete compute 24 hrs a day, so the cost internally is a negligible (ish) electricity bill.

Which is very suggestive - if after everything they are not fully loaded then the next gazillion data centres being built look unlikely to be needed.

[lwansbrough]

For OpenAI, research is marketing. I’m sure they’ve got plenty of budget for that.

[paxys]

No such thing as free, even internally at a company. All such use of resources is accounted for, assigned a dollar value and billed to some department. Someone ran the numbers and figured that whatever they spent on these GPU cycles was worth it.

[traes]

Presumably it's a rounding error compared to their full output, and they're making sure they have enough compute set aside for research by limiting public models. The more datacenters they build the less they have to limit them.

[Davidzheng]

RL training can use all of them - idk what needed means.

[simianwords]

I love how people come up with creative ideas to prove the bubble. This one is even more ridiculous - that OpenAI had spare compute to advance mathematics proves that data centres will not be needed. WHAT.

If anything it proves more data centres are needed. That's literally the only reasonable conclusion from this news.

[lifeisstillgood]

Sorry I thought that a bubble was widely accepted.

Are you arguing there is not an AI bubble, and that all the DC buildout is fine, going to be profitable etc?

I am not looking for a online slanging match - just looking for a different point of view

[simianwords]

It’s not obvious at all. If it were obvious to you, it would’ve been to OpenAI. It’s in their interest to accurately predict demand. The assumption that OpenAI/Sam is both really powerful but simultaneously ignorant to know what others know as obvious is well.. just strange. Especially strange when OpenAI has more information on models, breakthrough and usage patterns and we don’t.

I’m not participating in the slinging match but it’s very very weird that you think it’s some established thing that these companies won’t make profit. A lot of hubris must go in this kind of thought. Like.. do you all think everyone’s playing musical chairs?

[svieira]

They very often have been in the past. Why do you think this time is different?

[simianwords]

“Often” is load bearing. I don’t think markets are more likely than not to be musical chair shaped. To make this conversation more concrete, give me a falsifiable prediction on there existing a bubble. And then I’ll tell you if I believe in it or not.