Fable and Mythos: Model Welfare — LessWrong lesswrong.com 2026-07-24 tags ai-safety, llms, ai-governance original edit note archive.org Logged without notes.