Being a great AI partner.
Anyone can use AI. Few people partner with it. This chapter is about the difference.
What this chapter will cover
1. The partner mindset
The shift from asking AI for answers to collaborating with AI to think better. A user types and reads. A partner argues, refines, and uses AI to sharpen their own thinking.
2. Push back. Don't accept the first answer.
Pros never stop at the first response. They ask the model to:
- "Argue against your own answer."
- "Where could you be wrong?"
- "What did you miss?"
- "Make the case for the opposite position."
3. Use AI to teach, not just answer
The single most underused prompt pattern: "Teach me this assuming I'm smart but ignorant. Use one analogy. Then test me with 3 questions."
That one move turns AI from a search engine into a tutor that adapts to you in real time.
4. The productive-struggle rule
When NOT to use AI: when the struggle itself is the point. Learning to write. Learning to do math. Learning to debug code. Wrestling with hard questions.
The research is now clear: people who outsource thinking before they've built the skill get measurably worse at the skill. This is called cognitive offloading, and it's real.
The rule: build the foundation yourself, then use AI to scale it. Not the other way around.
5. Knowing when AI is wrong
Three reliable warning signs:
- Specific facts you can't easily verify (dates, quotes, citations, names).
- Confident answers in a niche domain you don't know well.
- Answers that feel too tidy. Reality is messier than AI's first draft.
6. The 30-minute experiment
A concrete exercise to do this week. Pick one:
- Become an expert in 30 minutes. Pick a topic you know nothing about. Use AI to teach it to you. Take notes. Write three paragraphs in your own words. Notice what happened to your understanding.
- Build a thing. Pick a small problem in your life. Use AI as a thinking partner to design a tool, list, study aid: whatever solves it. Don't have AI do it for you; argue with it about how.
- Stress-test it. Pick a topic you know deeply. Find where AI gets it wrong. Form an opinion about where it's reliable and where it isn't.
7. The line nobody crosses
Where's your line? A use of AI that feels wrong to you? Most people have one. Discovering it is part of becoming a thoughtful partner.