Anthropic's Blogging Model, Microsoft's Warning: The Fight Over AI Consciousness

Anthropic gave a retired model a blog; Microsoft AI's CEO says welfare research risks self-preserving agents. Researchers split on what to tell models about themselves.