Chain of thought monitorability: A new and fragile opportunity for AI safety - Hacker News

6 days ago 220
Comments
Open Full Post