Most people use AI to learn the same way: ask a question, read the tidy explanation, nod, and move on. It feels like learning. But a week later almost none of it is left — because reading a clear explanation and being able to use an idea are completely different things, and the first one quietly masquerades as the second.

The problem isn't the model. It's that AI is so good at explaining that it lets you skip the part where learning actually happens: retrieving, struggling, and being wrong. This library is built around what the evidence says works — you test yourself instead of re-reading, you explain things back instead of nodding along, and you practice retrieving instead of just absorbing.

Here's how to actually use it.

Make it test you, not just tell you

The fastest upgrade to any learning prompt is to stop asking AI to explain and start asking it to quiz you. Compare these:

Explain the causes of the 2008 financial crisis.

Quiz me on the causes of the 2008 financial crisis, one question at a time, and tell me where I'm wrong.

The first gives you the warm glow of understanding — which fades by tomorrow. The second forces retrieval, the single most powerful thing you can do to make knowledge stick. Reading the answer feels productive; pulling the answer out of your own head is what actually builds memory. The whole Check My Understanding section exists to flip you from passive reader to active recaller.

Explain it back — that's where you find the holes

There's a reason the Make me explain it back prompt is one of the highest-leverage in the library. When you try to explain something in your own words and an examiner grades the explanation, every gap you've been papering over shows up instantly.

Ask me to explain this concept back to you, then grade my explanation and show me the gaps.

I'll explain this topic; you flag what I missed or got wrong.

If you can't explain it clearly, you don't understand it yet — you just recognize it. That difference is invisible until something makes you produce the idea cold.

Learn in the right order

When you're picking up something new, work in this order:

Understand — get the intuition and the why, not just a definition.
Go deeper — connect it to what you know, find the misconceptions.
Test — check your grasp by retrieval, before you trust it.
Practice — apply it to real problems, increasing in difficulty.
Retain — space your reviews so it survives past this week.

The order matters because there's no point drilling practice problems on a concept you've misunderstood, or memorizing something you can't yet make sense of. Most people skip straight to re-reading their notes — the comfortable middle — and never test or space anything, which is exactly why it doesn't stick.

Beat the illusion of knowing

AI explanations are dangerously smooth. They make everything feel obvious, which tricks you into thinking you've learned it — the fluency illusion. The Learn How to Learn prompts are there to catch it:

How do I know when I actually understand something versus just recognize it?

Am I studying this effectively, or just re-reading?

What study habits should I drop based on the evidence?

The honest test is always the same: shut the explanation and try to reproduce it. If you can't, the smoothness fooled you — go back and retrieve, don't re-read.

Get unstuck without getting the answer handed to you

When something won't click, the worst move is to have AI explain it again the same way — you'll nod again and still not get it. The Get Unstuck prompts find where it breaks instead:

I'm confused about this. Ask me questions to find where it breaks down.

Break this hard concept into the smallest pieces I can understand.

I can follow this but can't do it myself. What's the gap?

That last one is the most common trap in learning: mistaking "I can follow along" for "I can do this." They're different skills, and only one of them survives a test.

A 30-second router

Not sure where to start? Match your situation:

New concept, no idea where to begin → Understand It, then Go Deeper.
Read it, think you get it → Check My Understanding (prove it).
Understand it but forget it → Remember It.
Have a book, article, or paper to learn from → Learn from a Source.
Taking on a whole new skill or subject → Build a Learning Plan.
Know the theory, can't do it → Practice & Apply, then Get Unstuck.
Want to get better at something you already do → Get Feedback & Improve.
Choosing what or how to learn → Decision (which to learn, what's worth it, what an expert would pick).
Ready to apply it and build something → Create (project, portfolio piece, teach it to someone).
Confused and it won't click → Get Unstuck.
Studying hard but it's not working → Learn How to Learn.

A quick example

Say you're learning recursion and you ask AI to explain it. It gives you a beautiful explanation with a factorial example. You nod. You feel like you get it.

Don't stop there — that's the illusion. Run the explain-back prompt: "ask me to explain recursion in my own words and grade me." Now you stumble on the base case, and the gap surfaces. Hit the unstuck prompt: "I can follow recursion but can't write it myself — what's the gap?" It points you to retrieval practice, so you ask for practice problems that increase in difficulty. Three problems later you can actually write it.

The understanding didn't come from the first clean explanation. It came from testing it, finding the hole, and practicing until you could produce it — not just recognize it.

Don't want to choose? Follow the journey

If a flat list of prompts is daunting, ignore it and run the six-step journey instead — it walks you through learning anything in the order that actually works:

Explain it simply — get the ground-level understanding.
What's the intuition? — get the mental picture an expert carries.
Quiz me one question at a time — prove you've got it by retrieval.
Give me practice problems — apply it until you can do it, not just follow it.
Give me actionable feedback — find the one thing to fix.
Design a review schedule — space it so it survives past this week.

Run those six in order on anything you're learning and you've used the whole library's logic without browsing a single category.

The bottom line

Prompts are starting points; the effort of retrieval does the heavy lifting. The best results come from three habits, every time:

Make AI test you, don't just let it tell you.
Explain it back in your own words — that's where the gaps show.
When it feels easy, close the explanation and prove you can reproduce it.

If you only ever use ten of these, use these:

Explain it simply, then build up.
What's the 20% that explains 80%?
Quiz me one question at a time.
Make me explain it back and grade me.
Build me a learning roadmap.
Give me practice problems that increase in difficulty.
What's the one thing that would most improve this?
Break this hard concept into the smallest pieces.
Am I studying effectively, or just re-reading?
Teach me the most effective way to learn this.

Work the order, make it test you, and these prompts will do more than explain things — they'll help you learn things that are still there next month.