Anthropic's CEO says that safety hinges on understanding how AI "thinks."

(wired.com) by rndsignals | Sep 18, 2026 | 0 comments on HN