🧠 (Designed for reviewing AI outputs that include claims, data, or citations) When I get output from a colleague, an AI or a report I've taken to using this prompt to help me. It acts like an intellectual audit for AI-generated research. Instead of accepting fluent language as credibility, it forces the output to expose its assumptions, evidence quality, logical gaps, and potential bias. It also helps me distinguish between what is supported, what is inferred, and what is si

In your most used LLM ask the following prompt Please describe our relationship in a single metaphor an honest evaluation of how we work together and find a suitable analogy and a short rationale. The output can be very entertaining (it's a great icebreaker for an online group) This is what mine said - how about you? Metaphor: We’re like a workshop with a long wooden bench  — you bring in half-built ideas, strange components, and occasionally something that looks finished but

(And learning as you go) I recently opened up my favourite LLM and asked it this question What are some playful light hearted prompts users can give their LLM that will entertain and teach them about how to prompt And here's what it gave me: 🎭 1. “Same task, different style” (teaches tone control) Prompt: “Explain how a refrigerator works in 3 ways: like a pirate like a Shakespearean actor like a bored teenager” 👉 Teaches: tone instructions, formatting, multi-output control