x
What Happens When You Let an AI Rewrite Its Own Moral Framework (With Supervision) — LessWrong