Sign up to keep track of your learning progress.
Log inAI Learning Guide
From someone who can use AI to someone who understands it.
Using AI and understanding it are two different things. This guide starts from the basic ideas and takes you all the way to why a model answers the way it does. Along the way you'll write prompts yourself, and at the end you can check yourself with the knowledge exam.
- Who it's for
- Middle school and up · anyone who wants to really understand AI
- Estimated time
- 3 sessions a week · about 6 weeks
- Steps
- 6 steps
Learn in this order
- 1 AI from zero Required Start with what AI is, and who built it and why. Getting one idea straight here — that AI is a machine people built, not a living thing — makes everything after it easier. Start
- 2 AI as probability Optional How a model picks the next word comes down to probability. Turn the knobs yourself and it clicks: why the same question can produce different answers. You can skip this, but doing it makes the rest much easier. Start
- 3 Talking with AI Required Test what you've learned in a real conversation. The same question phrased differently gets very different answers. Ask your Lyen learning buddy vaguely on purpose, then precisely, and feel the difference. Start
- 4 Prompt practice Required Open the lesson, compare a vague prompt with one that supplies a goal, context, and output format, then test both in Lyen. Explain which answer meets your goal and verify its claims. Start
- 5 Advanced track Required From here you look inside. Understand training data, tokens, and what a model actually is, then handle a very small one yourself in Python. If you skipped the earlier steps, this is where you'll get stuck. Start
- 6 AI knowledge exam Optional Knowing something and believing you know it aren't the same. The exam shows exactly where the gaps are. The Lyen chatbot is locked during the exam, so answer on your own. Start
Tips that make it easier
- Keep the order. Skip the basics and the advanced track will stop you cold.
- Don't just read — always try it yourself. Prompting is a skill you learn with your hands.
- Its outputs can be wrong. Investigate the input, data, model, and software rather than automatically blaming the learner.