Hacker News Reader

Comment by OutOfHere | Hacker News Reader

Pacing is a really bad idea because China will not slow it down for you.

The entire point of the statement is to start international coordination.

The problem is how do you enforce coordination? This isn't like nuclear testing that can be detected around the world on seismographs or from satellites in space.

AI research happens in benign looking buildings that could be doing a hundred different non-AI workloads, much of it things that industry or governments would not want 3rd parties prying into. There is too much to potentially gain to trust coordination.

See detailed supplement to Plan A which talks entirely about this:

https://ai-2040.com/supplements/verification-plan

There is a lot of research out there with answers to your questions. GPU supply chains are extremely constrained, making it a surprisingly tractable problem at least for now.

How do you tell the difference between inference and training workloads from outside? The same GPUs can be used for both and the pacing suggested here relates to training and not inference as far as I can tell.

Random sampling of compute activity.

https://ai-2040.com/supplements/verification-plan#appendix-p...

It will never happen because there is money to be made by not coordinating, i.e. by leaping ahead. It also doesn't make any sense to hold AI back.

That's the point. Do you see any concerns by these companies? They just want to slow down pace NOW because they're losing the race

All they have to do to stay in the race is to keep the price and token efficiency competitive with China.

Huh, that's really mental gymnastics you have there :-)

> losing the race

Lol, lmao even

Reply on news.ycombinator.com