Three modes that force active learning instead of the passive "ask AI, paste answer" loop that quietly makes you a worse thinker.
Most people use AI like a vending machine. Type the question, get the answer, paste it in, move on. It feels productive. It is the fastest known way to make yourself a worse thinker.
If you outsource the thinking, you outsource the learning. The moment the AI is gone, so is the skill. This is the opposite playbook: three modes designed to force active learning. You stay in the driver's seat. Claude becomes a tutor that refuses to do your homework.
Use when you think you know something and want Claude to find the holes — including the ones you don't know exist. Output is a map of your understanding, not a fix.
Claude asks you to explain the topic in your own words, then maps your understanding: what's solid, what's shaky, and what's completely missing. It gives you a ranked gap report — what to study first.
Use when you're working through a problem and want Claude to check your thinking — without giving away the answer. You stay in the driver's seat.
Claude reviews your reasoning step by step. It won't just fix it — it asks questions that force you to find the flaw yourself. Like a tutor who points at the wrong line and waits.
Use when you actually need to learn something new (or fix the gaps Mode 1 found). Claude teaches, then quizzes you hard, then grades you honestly, then reinforces until you've got it.
Claude teaches a concept, immediately quizzes you on it, gives you honest feedback, then escalates difficulty until you can apply it without help.
Get the full copy-paste prompts for all three modes and the Project setup instructions at the original guide. View on Notion →
The honest bit: The reason most people don't actually learn from AI is that AI is too helpful. It hands you the answer before you've earned it. These three modes force the opposite: you do the work, Claude makes sure the work is honest. That's the difference between using AI as a crutch and using AI as a coach.