Day 1 · Start here

Welcome to the
AI Masterclass

Before we touch a single tool — what AI actually is, and how we'll learn to work with it.
Format Two-day practical intensive By the end you'll build and test a real website with an AI agent

What you'll be able to do

  • Explain what AI is — and why it can be confidently wrong
  • Operate an AI coding agent safely — and control what it can do
  • Direct it with clear instructions, context and memory
  • Use Git to protect and track your work
  • Build, test and refine a real website with Codex or Claude Code

And leave with a repeatable process you can use on your next project.

The big idea

This isn't a "learn to prompt" course. We're learning to use AI as an assistant and a development tool.

That means more than typing a good request: directing it, giving it the right context, controlling what it can do, and verifying what it produces.

What is AI? (the 30-second version)

  • Software that has learned patterns from an enormous number of examples
  • A generative AI writes text and code by predicting what plausibly comes next
  • That's powerful — and exactly why it can sound certain yet be wrong

We'll go properly under the hood in Module 4 — What is an LLM. For now: it predicts; it doesn't "know".

The model is not the whole system

The model

The brain that generates text. Clever, but on its own it only talks.

The system

Instructions, context, tools, files and permissions — what turns a clever model into something useful and safe.

A powerful model does not automatically make a reliable system. That gap is what this course is really about.

Seven principles we'll keep returning to

  1. The model is not the whole system
  2. Prompting is only one part of good AI work
  3. Markdown makes work explicit and portable
  4. Memory must be designed — it doesn't just happen
  5. More capability requires more control
  6. Work must exist outside the conversation
  7. AI output is not automatically a finished result

How we'll work together

The rhythm

Explain briefly → demonstrate → you practise → we review what happened.

One project

We carry a single project across both days, so each step builds on the last.

We get hands-on very soon — but grounded first, so you know what you're steering.

One habit from minute one

You'll run a real agent that can read and change files. So: work in a dedicated practice folder, review actions before you approve them, and never put real passwords or private data in front of it.

We'll go deep on safety in Module 10 — but the habit starts today.

Now — let's begin

Module 1 — Launch
your first AI agent

You understand the shape of it. Time to open a tool and make your first file.