You know that moment when someone explains something difficult really well? You’re nodding along. The examples click. You get to the end and think, “Oh, that actually makes sense.” Then you close the book.

Can you explain what you just learned to someone else? Could you use it on a problem you’ve never seen before? Would you even notice if someone gave you a convincing but wrong explanation? Did it only make sense while it was right in front of you?

I’ve felt that gap more times than I can count. The explanation feels crystal clear in the moment, and a day later I’m staring at a blank page trying to reconstruct it. It’s so easy to confuse the feeling of understanding, with actual understanding.

With the tools we have today, that confusion gets even easier. If you can get a perfect explanation, a working example, or a full answer in two seconds, do you learn more because help is always there? Or do you start to mistake the ability to get an answer for the ability to produce one yourself?

That’s what made me want to go back to the most basic question: What does it actually mean to learn something?

For me it was always, to really learn is to use what I have learnt in practice. I’ve realized I don’t really know something until I’ve had to do something with it.

There’s a big difference between recognizing information and being able to act on it. You can memorize the quadratic formula and still freeze when the problem doesn’t look like the ones you practiced. You can memorize dates for a history test and still not understand why those events happened or how they connect.

Remembering matters. It’s just not the whole thing. The deeper learning happens when you try to use what you know. You try to solve the problem, explain the idea out loud, or make an argument. Not probably at first you’ll get it wrong. That wrong turn is often where the most valuable learning actually happens, because you have to stop and ask: “what did I miss?” “What do I need to rethink?”

That’s why practice matters. Not because repetition magically makes you remember, but because practice forces you to interact with knowledge. You have to make choices when hitting situations you haven’t seen before. There you discover the edge of what you actually understand.

My favorite test for this is simple: what happens when the original context disappears? Can I still explain it without the textbook open? Can I recognize it when it shows up in a different form? Can I take it and use it somewhere new? If I can, I’ve learned it. If I can’t, I’ve just seen it.

Knowledge still matters. A lot. This isn’t an argument against memorization, it’s an argument for it.

You can’t think deeply about math without knowing some math, and you can’t analyze history without knowing what happened. Facts, vocabulary, formulas, concepts — they’re the building blocks. Without them, there’s nothing to build with.

For most of human history, having those blocks in your head was also incredibly practical. If you didn’t know something, finding it could take hours.

Now? That part has changed. Information is instant. And with AI this goes one step further. It doesn’t just give you the information, it explains it, gives you examples, walks you through it, and adapts to how you learn.

It’s becoming a kind of external memory. That doesn’t make knowledge irrelevant. It changes our relationship to it.

What happens when answers become easy If you can always get an answer, perhaps the most important skill isn’t finding the answer anymore. It’s knowing what to do with it once you have it.

You can ask AI to solve a difficult problem. But if you don’t understand the concept underneath, how will you know if the answer makes sense, catch a mistake? How will you know if that approach will still work when the next problem looks different?

Having access to an explanation is not the same as understanding it. We can outsource finding information, but we can’t outsource the judgment needed to evaluate it.

That creates an interesting tension for all of us who are learning, and teaching. AI might make memorizing certain things less necessary, while making deep understanding more necessary than ever. When answers are cheap, the ability to question them, connect them to what you already know, and apply them in an unfamiliar situation becomes the real skill.

And here’s the tricky part: those skills need practice too. You don’t learn to think critically about an answer without trying to solve problems on your own first. You don’t learn to spot a weak argument without trying to build a strong one yourself.

So what does it actually mean to learn? I don’t think learning is a single moment when information moves from a textbook into your memory.

It’s a process. We pick up knowledge. We try to use it. Then we mess up. We connect it to other ideas. Try again. And slowly, we become able to take what we learned and carry it somewhere we’ve never been before.

I don’t think we should stop asking students — or ourselves — to remember things. And I don’t think we should keep technology away from learning. I just think we should be careful not to confuse access to knowledge with learning itself.

AI can remember for us. It can explain for us, even produce answers for us. But that doesn’t answer the most important question. It raises it:

What remains for the learner to do?

We have to think and try. We must make decisions and fail and I. The end figure things out ourselves. This is the most human part of the whole thing.