My most recent essay on AI argues that we need an AI version of the 1970s Biological Weapons Convention. Techno-optimists and p(Doom)ers agree on one thing: that unfettered Chinese AI is the biggest risk. This seems the only solution.