Loading lesson...

Vague vs Good Prompts — Maîtrise de l'étude avec l'IA

Module 4: Writing better prompts — AI is just a tool. I'll show you how to write prompts that actually get you what you want.

Why do vague prompts like 'write an email' always result in robotic, generic sounding text?
Because the AI has to guess the context, so it plays it safe by using average language from its training data. — In the absence of rules, the AI falls back to the middle of the road.
What is the consequence of leaving 'blank spaces' in your instructions to the AI?
The AI fills those blanks with its own assumptions, which rarely align with what you actually want. — It's an autocorrect engine—if you don't steer it, it drives itself.
The 'Good Approach' in the lesson succeeded because the prompt acted like a what?
A mold — it gave the AI an exact role, audience, and format to restrict its guessing. — Think of a physical object used to force liquid metal into a specific shape.
What is 'The Golden Rule of AI'?
You are programming a computer using plain English. Stop asking and start directing. — You aren't asking a human for a favor. You are operating a machine.
When fixing a bad prompt, what is the fastest way to get a human-sounding result?
Tell the AI exactly what role to play and who the audience is. — By defining the character and the target, you severely restrict the language it can pull from.