Your next step: keep learning and keep responsibility
You've reached the last lesson of Stage 0. Congratulations! In the SDE and SDD lesson, you met the name for the way our example was built and checked: specify, build, verify, own.
This lesson gathers the whole stage together. We'll:
- turn what you've seen into three steps you can use with any new topic;
- meet the four skills the rest of the book teaches;
- see how deep each topic goes, and why;
- find exactly where to go next.
Why it matters: Stage 1 asks for real effort, every day. You'll learn far better if you know why each topic is there, and how it connects to what you've already seen.
How to approach every topic: three steps
Everything in this book follows three steps:
- Identify what AI can do and what it can't, for the task in front of you.
- Understand, learn and master what AI can't do. Understand, and get exposure to, what it can.
- Learn to use AI for what it can do.
Don't mix these up with SDE's four steps. SDE is how you build one piece of work. These three steps are how you decide what to learn, and how deeply.
The verification lesson showed how to begin step 1: ask how reliably AI does a task, what checking costs, and how bad a mistake would be. The answers change as AI improves, so ask them again whenever the tools change.
Step 2 doesn't mean ignoring what AI can do. If you never understand how something works, you can't direct the AI or judge what it gives you. You learn it deeply enough that the AI becomes your tool instead of your competition.
Which four skills the rest of the book builds
The rest of the book builds four skills:
| Skill | What it means for you |
|---|---|
| Software engineering foundations | How software is made, structured, checked and changed: everything Course 1 introduced |
| Building agentic AI | Making AI that plans steps and uses tools, like the agents you met in Course 2 |
| Using agentic AI exceptionally well | Directing and checking AI at a high level, with the method from Course 2 |
| Advanced web development | Building the web pages and apps people use to reach AI: the chat window, the dashboard, the page where you sign in |
The third skill holds the others inside it. You can't direct AI well at building software, agents or web apps without understanding those things yourself.
How deep each topic goes
Not every topic needs the same depth. The book measures depth on a scale from 0.1 to 3.0. The number describes how deeply a topic needs to be understood. It is not a grade. Here are a few landmarks:
| Depth | What it asks of you |
|---|---|
| 0.5 | Know that something exists, and what it's for |
| 1.0 | Understand its purpose, and judge a result you can see. Stage 0 works at about this level |
| 1.3 | Read common code line by line, predict its result and spot obvious bugs. This is the Python target of Stage 1's first course |
| 2.0 | Guide AI through the work, and check its choices |
| 3.0 | Own the topic fully, including its difficult cases |
Stage 0 introduced these ideas; it didn't train you in them. Reading a small program and checking its results is a real start. Handling an app that uses real money needs much more. And at every depth, checking what you accept is your job: you don't wait for 3.0 to start.
If you want the whole argument behind these choices in one place, the Thesis makes it. Parts of it will read very differently now that you've seen each idea at work.
Start Stage 1 in learn mode
Stage 1's first course goes back to Python, aiming for depth 1.3. Its first lesson introduces a Python function: a named piece of code that does one job. You'll follow its names, predict a result, and then check it.
Remember the two modes from the coding agents lesson. While you learn, let AI explain things and ask you questions, but do the thinking yourself. Stage 1 asks for steady daily study, about four hours a day if you can manage it. Short, regular sessions beat one long session now and then.
What you can now explain about Stage 0
Before you move on, answer these without looking back:
- What are the three steps, and how would you begin the first one for a new task?
- Which of the four skills holds the others, and why?
- At what depth does Stage 0 work, and what does that mean for what you can do now?
- In the committee example, what did you add that the AI couldn't add alone?
If an answer feels thin, go back to that lesson first. When you're ready, continue to Stage 1. I'll see you there!