Introduction

Imagine two people, same afternoon, same AI tool, same task. One types a question, reads the answer, closes the tab. Satisfied. The other reads that same answer—and then asks the AI to argue the opposite side, then explains the whole thing back in their own words, then asks what they're probably still missing. A week later, the first person can barely reconstruct the debate. The second could teach it to someone else. Same tool. Completely different outcome.

So here's the question worth the next several hours of your attention: does AI make you sharper, or does it quietly hollow you out… and is that outcome fixed, or does it depend entirely on what you do?

The answer isn't what the boosters are selling, and it isn't what the doom-sayers are warning about either. It's more precise than either camp wants to admit—and that precision is exactly what this course is built on.

There's a study you'll encounter later that should unsettle anyone who thinks AI assistance is automatically a win. Two hundred and forty-six people worked through difficult logical reasoning problems with AI help. Their scores improved—by three real, meaningful points. And when asked how well they'd done, they estimated even better than that—by four additional points on top of that. The work got better. The self-knowledge got worse. Both of those things happened simultaneously, in the same minds, with the same tool. That finding is the edge this course keeps returning to.

Then there's a randomized controlled trial out of Harvard—not a blog post, not a tech company's press release—testing an AI tutor against active learning classrooms in a real undergraduate physics course. The results were striking enough that they deserve more than a headline, and you'll get the full picture, including the part the press mostly ignored.

And there's a moment from the section on memory that will probably feel uncomfortably familiar: two focused hours with an AI, following the thread, asking follow-ups, genuinely understanding everything—and then waking up the next morning to discover that most of it is simply gone. Not fuzzy. Gone. That's not a personal failing. It's a predictable consequence of how memory is actually built, and there's a way to stop it from happening.

What you're going to understand by the end of this course isn't just a set of techniques, though you'll have those. It's a precise, research-grounded map of exactly when and how AI extends your thinking versus when it does the thinking for you—and a complete system for making sure you're always on the right side of that line.

The stakes are real. The tools are already in your pocket. The only question is whether you use them, or whether they use you.