You're Absolutely Right Linch lesswrong.com 2026-08-10 2026-08-12 2,290 words 11 min on page tags ai-safety, chain-of-thought, interpretability, reward-hacking, ai-governance original edit note archive.org Logged without notes.