Skip to content

How Language Models Actually Work

BeginnerLesson 2

Not yet recorded

This lesson is on its way

The notes below are ready — the video lands soon.

Prediction, tokens, and training — the three ideas that explain almost everything a chatbot does. Enough depth to be useful, not so much that it stops being interesting.

What you’ll learn

  • What a token is and why it explains a lot of odd behavior
  • Why the model has no memory between conversations by default
  • The difference between training data and what happens at the moment you ask

Keep going

All lessons

More lessons from the same tracks.

Newsletter

The email list opens soon

I'm still setting up the newsletter. Subscribe on YouTube in the meantime and I'll announce it there first.