Skip to content

AI Foundations

What an LLM actually does when it 'answers' you.

AImachine learningdeep learningneural networkstransformersfoundation modelstokenscontext windowsinferencetemperature

Lessons

  1. 20.1What is an LLM, really?No magic inside: a language model guesses the next word-piece, over and over. Once you see that, cost, speed and made-up answers all make sense. 30 min Next-token sampler
  2. 20.2Machine learning, deep learning and neural networksHow a program can learn a rule from examples instead of being told it, and why your API calls never teach the model anything. 30 min
  3. 20.3Transformers, intuitivelyFollow a sentence through the engine inside every modern AI model, and see the one trick, attention, that made it all work. 30 min Next-token sampler
  4. 20.4Tokens and context windowsTokens decide what the model sees, what fits, what you pay and how long you wait. Here's how to budget them like a pro. 30 min Context window builder Next-token sampler
  5. 20.5Sampling, temperature, top-p and determinismThe model gives every possible next word a chance. Sampling decides which one you actually get, and why the same question can get different answers. 30 min Next-token sampler

On the job: try it on your own project

Pick one AI feature you use or plan to build. Ask Claude Code to count the tokens in a typical request (prompt, context and answer) and estimate cost per call. Try the same prompt at two temperatures and compare. Write a short note explaining to your PM what the model can and cannot be trusted to do in that feature, and why.

Can you do this now?

0/4. Anything you can't tick yet: reopen that lesson's practice ladder.