There's a pattern I kept seeing with first-year calculus and physics students:

They study for hours. They watch the lecture again. They read the textbook examples. They even ask ChatGPT to explain it. Then the exam comes and they freeze.

The problem isn't effort. It's familiarity bias.

When you re-read a solution or watch someone else solve a problem, your brain registers it as "I know this." But recognition and recall are completely different cognitive processes. Recognizing a chain rule problem when you see one ≠ setting up and solving a chain rule problem from scratch in 3 minutes.

How to test yourself honestly

Next time you're studying, try this:

1. Close everything — notes, textbook, AI chat

2. Set a timer for 5 minutes

3. Write down, from memory, the complete steps to solve a specific problem type (e.g., "related rates problem" or "free-body diagram for inclined plane with friction")

4. Compare to your notes

If you couldn't write the steps from memory, you don't know it yet. You recognize it. That's a huge difference, and it's the difference between a B- and an A.

The fix

Stop studying about problems. Start studying through problems. Every study session should have at least one problem you solve completely cold — no notes, no hints, timer running.

This is the approach behind the AI STEM Study Kit. Every tool is designed to close the gap between "I've seen this" and "I can do this under pressure."

But even without the kit: close your notes, set a timer, and test yourself. You'll learn more in 5 minutes of struggle than 30 minutes of re-reading.