Question
An AI assistant solved your bug by handing you a one-line Python fix using `functools.reduce` and a lambda, and it works — the test passes. But you genuinely don't understand how that line does what it does, and your team requires you to be able to explain code you merge. What do you do, and why is 'it passes, so ship it' the wrong move for a junior here?
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.