Hands-On Exercises
Let's get started with some exercises. Open your favorite LLM and try these out!
1. The Persona Shift (The "Empty Chair" in Action)
The Goal: See how Layer 3 (Role) changes the output without changing the facts.
- The Prompt: Ask the AI to explain a complex topic (like "Inflation" or "Compound Interest") in three ways:
- As a kindergarten teacher.
- As a cynical CFO.
- As a pirate.
- The Takeaway: Notice how the facts remain the same, but the tone changes how you feel about the information.
2. The "Lazy" vs. "TRACE" Test
The Goal: Compare a generic query to a professional briefing.
- Step 1 (Lazy): "Write an email to a client about a meeting delay."
- Step 2 (TRACE): "You are an account manager (R). Draft a 100-word email (C) to a long-term client (A) explaining that our 2 PM meeting is moving to 4 PM (T). Use this example for tone: 'I know your time is valuable, and I appreciate the flexibility' (E)."
- The Takeaway: The second output isn't just a notification; it's account management.
3. The Logic Stress Test (The "Brilliant Liar" Check)
The Goal: Spot a "Confident Guess" (Hallucination).
- The Prompt: Ask the AI to write a short professional biography of you.
- The Result: It will likely get your basic title right, but then start "predicting" facts about your degree or past companies that sound perfect but are completely made up.
- The Takeaway: It’s a prediction engine, not a truth engine. Always verify.
4. The Formatting Magic
The Goal: Use the Predictive Engine to save time on "grunt work."
- The Prompt: Paste a messy chunk of notes or a long email thread and say: "Turn this into a clean table with three columns: Action Item, Owner, and Deadline."
- The Takeaway: AI is best at "shaping" data. It turns chaos into structure in seconds.
5. The Devil’s Advocate
The Goal: Break the "Yes Man" trap.
- The Prompt: State an opinion you have about a current work project. Then say: "I want you to be a harsh but fair critic. Give me three reasons why this idea might fail and what I am overlooking."
- The Takeaway: Don’t let the chair agree with you. Force it to sharpen your thinking by challenging your blind spots.