Labs

Hands-on, run-them-yourself notebooks. The thing nobody else writing about AI can hand you.

Reading about AI is fine. Running it yourself is better. These are live notebooks built to teach: short wall time, heavily commented for narration, runs on Google Colab’s free tier, and requires no Python or AI knowledge to use. Click a card, then a badge, and you’re inside the demo.

Every lab is self-contained and reusable — the same notebooks power the talks.

Mine the Room

Call out ten numbers and watch a model search them for the pattern. A few minutes of pixels that make ‘training’ feel mechanical instead of magical.

nanoGPT

A ~10-million-parameter transformer, trained from scratch on free cloud hardware. Watch the loss fall from random noise to recognizable prose.

How to use

Open a lab, then click its Open in Colab badge.
Read the welcome cell at the top — it explains notebooks and Colab if you’ve never touched either.
Click Runtime → Run all and watch the output.

You don’t need to read or understand the code. Every block is commented for the curious, but the point is to watch the model do its thing.

Feels like magic, but it’s math.