Trust calibration across expertise domains
You're a backend engineer who has spent six years in payments and almost none in front-end CSS. In one afternoon you ask an AI agent to (a) implement a new idempotency-key strategy for your refund endpoint in Go, and (b) build a responsive pricing-table layout in CSS/React. How much should you trust the AI's output on each, and how should your verification effort differ given that you're an expert in one domain and a novice in the other?
verification_plan(expertise_years: int, outcome_observable: bool, high_severity: bool) → list[str][6,false,true]out["read_line_by_line","add_regression_test"][0,true,false]out["observe_across_cases"][0,false,true]out["get_expert_review","add_regression_test"]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 & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.