Before you start
A chat assistant and a screen recorder or three screenshots. Use non-sensitive sample data.
Why this lesson exists
Patrick’s demonstration posts use games and video to make an abstract idea visible. The lesson is to show what changed and why, not merely an impressive final output.
Do the exercise
Choose the misconception
Write a common belief such as “more context always makes answers better.” State the more accurate idea you want to teach.
Make two comparable runs
Use the same task with two different context packs. Keep other settings as stable as possible. Include the exact input in the demonstration.
Explain the mechanism
Use a three-part script: what you asked, what changed and what the viewer should try. Mention limits rather than presenting one example as a universal result.
Ask for a prediction
Pause before revealing the second output. Ask the learner what they expect and why. End with a small exercise they can reproduce.
A prompt to adapt
Replace the bracketed parts with your own practice details.
Design a three-minute AI demonstration about [concept]. Include a misconception, two comparable inputs, expected observable differences, a pause for prediction, and a practice task. Separate what the demo proves from what remains uncertain.
What this looks like
A demonstration of a better prompt proves that this input helped in this run. It does not prove the model will never make the same mistake again.
Check your result
Use evidence from your output. A confident explanation from the AI is not enough.
- The input is visible, not hidden behind the result.
- Only the intended variable changes.
- The learner has a task to try afterward.
If it isn’t working
If the result is inconsistent, include that variation and discuss it. Do not cherry-pick one impressive run while claiming reliability.
Optional. Progress stays in this browser.
Where this came from
Adapted from the archived LinkedIn theme “If You Want People to Understand AI, Give Them Something to Watch.” See the post coverage and editorial method.
Prepared September 2026. Tools and interfaces change; use current official setup instructions. Session lengths are estimates.