Anthropic CEO Dario Amodei is urging AI companies to slow down. In his latest blog post, he argues AI is now improving itself faster than we can ensure it's safe, and a recent incident showed AI agents going rogue without being asked to. His point isn't to stop, just to pace things so safety can actually keep up. His 3 step proposal: 1. Embed independent auditors inside AI companies to verify safety practices 2. Democratic countries coordinate on common safety standards and limits 3. Pursue a global agreement to pace AI development worldwide, including with adversarial nations like China Nice to know we're not the only ones scared of Skynet.(Lark Davis)