Each time a LLM model answer start with "Here’s the clean, no‑nonsense explanation you want..." I imagine the engineer pre-promt to make it behave like: "...and please, answer the question with a clean, no‑nonsense explanation." ...and that's funny.