is perhaps the single most important thing to internalize about AI output. The more authoritative an AI response sounds, the more carefully you should verify it, not the less.