00:36
2026-07-13
lesswrong.com
ai-policy
5 "Plan A" scenarios
The US and China could cooperate to slow transformative AI by centralizing compute in non-party countries under military oversight, controlling chips at the manufacturing level with integrated monitor…