Question
You ask an AI to optimize a C++ hot loop that processes 10M particles per frame. First pass it 'optimizes' by replacing your `std::vector` with `std::list` ('faster insertion'); frame time gets worse. You say slower; it adds `#pragma omp parallel for` over a loop with a shared accumulator — now it's a data race and the output flickers. It's confidently making it worse each turn. How do you recognize the model lacks the real bottleneck and reset?
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.