Course 1 · AI Foundations · Lesson 06
Prompting: Stop Typing Wishes
Two people, same model, wildly different results — and the difference is about thirty seconds of writing. The brief, the three levels, and why your first prompt is never your last.
The one mental model
The model predicts from what's in front of it — it cannot read your mind. So every detail you don't give, it fills with the most average choice. No audience? It writes for everyone. No tone? Corporate AI voice. No length? Whatever's typical. It isn't failing you — it's averaging. A brief removes the average.
Key terms
Wish vs. brief
A wish says what you want to be true (“make it better”). A brief says what to actually do. Five lines: role, context, task, format, limits.
The averaging rule
Anything unspecified gets the most statistically average filler. Generic in, generic out.
Steering
Treating the first answer as a first draft. “Shorter.” “Warmer.” “Like this example.” Corrections land because the conversation is the context (the desk from Lesson 04).
Show, don't describe
One pasted example of the tone or format you want beats a paragraph of adjectives.
The misconception to drop
✕
“Good prompting is about magic words — and if the answer is bad, the model just isn't good enough.”
✓
There are no magic words — the skill is clarity. The official guides from the makers of ChatGPT and Claude say the same thing: be specific, give context, show examples, iterate. Same model, better brief, completely different answer.
Put it to work — the brief
R
Role — who should the model be? A fitness coach answers differently than a doctor.
C
Context — the situation, and who it's for.
T
Task — precisely what to do.
F
Format — length, structure, style of the output.
L
Limits — what to avoid, and what must stay true. Then: steer.
Ask the AI Tutor
Pause the video and ask anything from this lesson — the tutor answers from this lesson’s material.
Why do my answers come out so generic?
What goes into a good brief?
Turn my last prompt into a level-3 brief
When should I start over vs. steer?
Next lesson
07 — Memory, Projects & System Prompts