Anthropic calls for global pause in AI development before humans lose control

5 pointsposted 11 hours ago
by patrickdavey

6 Comments

uejfiweun

11 hours ago

How do you resolve the security dilemma? A cursory knowledge of international relations shows why a "global pause" is basically impossible, and that's to say nothing of current international conditions. These Anthropic guys are smart, and so I can't help but think that this posture is nothing more than grandstanding to build up hype and attention ahead of the IPO.

metalcrow

10 hours ago

They mention the Nuclear Forces Treaty which manages to accomplish this for a similarly difficult problem, so i believe this is technicality a solvable problem, it's just a matter of political will.

uejfiweun

6 hours ago

I don't disagree that theoretically there could be some kind of test ban treaty type thing. But the mere fact that the primary rhetoric we see from leaders is around "winning the AI competition with China", I personally can't help but interpret this as a sign that nobody in the US is even close to ready to come to the table ready to slow things down.