Skip to content
Hacker News · AI·· 6 h agoAI score64

Paul Christiano: A 1-in-3 chance of superintelligence within a year, and labs should pause now

Paul Christiano: 12 months away from AIs that can overthrow the government

AI summary

Paul Christiano, head of the Alignment Research Center and a recent OpenAI board member, says there is roughly a 1 in 3 chance that superintelligence capable of overthrowing the government arrives within a year if development does not slow. He argues frontier labs should pause further algorithmic progress on pre-training and RL until they can defend their weights and research secrets, warning that otherwise China could steal the technology mid-intelligence-explosion and the labs would lose the option to slow down. He also says no current alignment approach has better than even odds of controlling superintelligence, and that the hope is AIs solve the problem during the explosion itself.

Source:Hacker News · AI · dwarkesh.com