I Asked AI to Build Me a Rubik's Cube Simulator. It Worked. I Had No Idea Why.

#webdev #python #ai #basic

So there I was, staring at a fully working Rubik's Cube simulator on my screen.

Colors, grid, faces — everything exactly where it should be. It ran first try.

And I had absolutely no idea what any of the code did.

The Problem With "It Works"

I'm an IB student trying to build a CS portfolio for university applications. Everyone says the same thing: build projects, put them on GitHub, show you can code.

So I did what most students do. I described what I wanted to an AI, it gave me the code, I ran it, it worked. Done. Next project.

Except — when I looked at the code, it was completely foreign to me. I knew what I asked for. I had no idea what the AI had actually built. I couldn't explain a single function if someone asked me to.

That's not a portfolio. That's a magic trick I didn't understand.

The Rebuild

I decided to start over. Same project, same AI tools — but this time with one rule: I don't move to the next line until I understand the current one.

The difference was in how I prompted. Instead of just saying "build me a Rubik's cube simulator in pygame," I wrote something like this:

"Add a comment above every function explaining what it does. Add inline comments on any line that isn't obvious. After each stage, pause and ask me if I understand before continuing."

That one change made everything different. Suddenly the AI wasn't just handing me a finished product — it was teaching me as it built.

What I Actually Learned

This isn't really about Rubik's Cubes or pygame. It's about context engineering — the quality of what AI gives you is entirely determined by the quality of how you ask.