Practical AI literacy

Build real AI skills, one exercise at a time.

Short lessons and hands-on practice that teach you to prompt, verify, research and work with AI. No signup, no API keys — and your progress stays on your device.

Try it before you read a thing

Watch the model predict the next word

This is the real diagram from the first track: AI writes by scoring every plausible continuation and sampling one. That single step explains hallucinations, why answers vary, and why verification is a skill — not a trick. Sample it yourself.

Open the lesson this comes from

“The cat sat on the”

  • mat 41%
  • floor 24%
  • windowsill 15%
  • sofa 11%
  • everything else 9%

The model scores every plausible continuation — then picks one to continue the sentence. The most likely token (mat, 41%) is the usual pick, not the only possible one.

Sample it a few times — a low-probability token can still win a round.

Tracks

How it works

  1. Read the concept.

    Two minutes of plain language — with a weak and a stronger example side by side.

  2. Practice immediately.

    Every lesson ends in a real exercise; feedback explains why, never just “wrong”.

  3. Transfer it anywhere.

    Skills that work in any AI tool — principles, not one magic prompt.

Why you can trust this practice