Why don't we just give AI the answers? Brendan Long lesswrong.com 2026-08-04 2026-08-05 748 words 4 min on page tags ai-safety, reward-hacking, llms, ai-agents original edit note archive.org Logged without notes.