Pacing the Frontier: A statement from 1,132 employees of frontier AI companies

»Slow down AI« sounds noble until you ask who actually slows down: Democracies and companies that comply simply hand the advantage to rivals that don’t. History suggests voluntary technological restraint rarely wins against geopolitical competition. The race doesn’t stop: it just changes leaders.

This all presupposes that this “arms race” is going somewhere. Far as I can tell, trillions of dollars will be incinerated making a better word-guesser while claiming it will lead to some nebulous “AGI”.

How much more effective have the large models become for you over the last three years? Do you think development might come to a halt in China, for example, if US providers were indeed to stop developing or releasing them?

I think they’ve hit the point of diminishing returns and want to frame it as “we’d better stop before we all die”. That is, incremental improvement is all they’re getting from throwing exponentially increasing compute at these things.

So have they been getting better and better?