Question
You're using an AI agent to backfill a new denormalized `last_active_at` column across a 50M-row users table in a live SQL database, then keep it updated going forward. Decompose this into steps with checkpoints, and specify where a human must verify before letting the agent run the next step against production data. The system is read-heavy and can't take downtime.
Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it — tests, edge cases, reading critically — and how you’d re-prompt or decompose to get it right.
Vibe coding: describe the solution in plain language (or narrate it) and the coach grades your approach. Generating runnable code from your description is coming next.