Three Little Secrets, Big Power
In the last lesson we learned the difference between "good questions" and "bad questions." In this lesson, we will take a good question apart and look at the three key parts hidden inside. They are: state the role, give details, and make requirements. These three secrets are like three keys. When you put them together, they unlock the most accurate drawer in AI brain. Let us look at each one, with examples.
Secret One: State the Role (Who Should AI "Become"?
"Stating the role" means telling AI: "Please pretend to be this person right now." Why is this step so magical? Because the same question gets completely different answers from different roles. For example, if you ask "How do I learn English well," AI gives you pretty generic advice by default. But if you say: "Please act as an elementary school English teacher with 20 years of experience, and tell a fifth grader who is a little scared to speak English how to take the very first step," AI instantly switches to a teacher voice and gives you gentle, specific tips like "say three sentences to yourself in the mirror every day" or "pick a favorite English song and sing it out loud." See? Add a role and the answer comes alive. Common roles include: teacher, coach, scientist, storyteller, tour guide, journalist, and so on.
Secret Two: Give Details (Explain the Background)
"Giving details" means telling AI the ins and outs of the situation and your specific circumstances. AI does not know who you are or what happened, so you have to "fill in the background" for it. For example, suppose you want AI to help write the steps for a craft activity. Bad version: "Teach me how to do crafts." Good version: "I am a third grader, and I only have colored card paper, scissors, and glue. I want to make a Mother Day card for my mom, and I only have 30 minutes. Please give me a simple and pretty plan." The details you added: your grade, your materials, your goal, and your time limit. These details let AI pick a plan that actually fits you, instead of some complicated tutorial that needs a hot glue gun and two hours.
Secret Three: Make Requirements (Tell AI What "Shape" You Want)
"Making requirements" is the last key: you tell AI what the answer should look like. How long? How many points? What tone? With examples or without? In a table or not? The same content can give completely different results depending on the requirements. For example: "Please explain photosynthesis in 5 bullet points, each no longer than 20 words, in a fun and lively tone like popular science for kids, and finish with a real-life little metaphor." With requirements like "5 points," "no more than 20 words," "lively," and "a metaphor" added, AI answer will be neat and easy to follow. Remember, AI will not guess what shape of answer you like. You have to "place your order" clearly.
Put the Three Keys Together
Now let us combine all three secrets into one "perfect-score" question: "Please act as a storytelling scientist (role), and explain what a black hole is to a fifth grader who is super curious about space (details), using 3 short paragraphs, each with a real-life metaphor, and end with a small question that makes me think (requirements)." See, role plus details plus requirements, all in place. Next time you have something to ask AI, first silently run through these three secrets in your head, then fold them into your question. Practice a few times and it will become second nature.
🧠 The 4-Step Method
Not just understanding — learning to think