Course

How AI Really Works

Seven short lessons go one level deeper than the basics — inside the language model itself. See how your words become numbers, what the machine does with them layer by layer, why it looks back at earlier words, and why making it bigger keeps making it smarter. Same touch-and-see format, one step under the hood.

7 lessonsIntermediate~1.1h

Start the course Get the app

Course lessons

The internet, compressed

Why an AI can't store the internet — and why the squeeze is where its understanding comes from.

›

Words become tokens

Why an AI that writes essays can't count the letters in "strawberry".

›

Inside the black box

Open the transformer: how your words become numbers, get refined layer by layer, and come back out as the next word.

›

What "it" means

How each word looks back at the others and works out which ones it depends on.

›

Rolling the dice

The model's real output isn't a word — it's the odds on every word, and it samples one.

›

Bigger, predictably better

Why so much of modern AI is just the same design — scaled up.

›

Learning on the fly

How a frozen next-word predictor picks up a brand-new task from a few examples — no retraining.

›

Take the full course

Learn it properly — tracked & personalized

Follow the whole course as a guided journey — sequential lessons, progress, streaks and a personalized path, all in the app.

Get the app