Before you touch a tool, get this one idea straight, because everything else in the course rests on it: you are not going to write software. You are going to describe what you want, clearly, and direct AI tools to build it. The job you are learning is not coding. It is directing, and then judging what came back.

To direct well, you need a picture of what you are directing. Every app you will ever build, however simple or complex, is made of the same four parts.

The frontend is what the user sees and touches. The screens, the buttons, the forms, the colours. When you open any app and look at it, you are looking at the frontend. Its only jobs are to show things and to collect what the user types or taps. It decides nothing on its own.

The backend is the brain. When a user clicks a button, the frontend hands the request to the backend, and the backend decides what happens next. It runs the rules. It does the calculations. It is the part that says "this password is correct, let them in" or "this order is over the limit, reject it." The user never sees the backend, but nothing meaningful happens without it.

The data is the memory. It is where your app keeps everything it needs to remember between visits: users, their inputs, their history. When you log into an app a week later and your stuff is still there, that is the data layer doing its job. When an app "forgets" what you typed the moment you refresh, that is a data-layer story, every time.

The auth (short for authentication) is the gatekeeper. It answers two questions: who is this person, and what are they allowed to see. Sign-up, login, "you can edit your own profile but not someone else's," all of that is auth. It is the layer people underestimate most, and the one that quietly separates a real product from a demo.

Here is the whole thing in motion. A user types their email and password and clicks Log in. The frontend collects those and hands them over. The backend takes them and checks them against the data, where the real account records live. Auth confirms the person is who they say and decides what they are allowed to do. The backend sends the answer back, and the frontend shows the logged-in screen. Four layers, one click. That loop, in some form, is happening behind every action in every app.

"Full-stack" simply means all four layers working together. When someone can open your URL, sign up, do something, and come back tomorrow to find it still there, you have shipped a full-stack app. That is your Week 1 goal, and you will hit it without writing code, because the tools build these layers for you when you describe them well.

Which is the last piece to internalise now: the quality of what you get depends almost entirely on the quality of what you ask for. Treat the AI like a brilliant, fast junior who takes everything literally and has no idea what is in your head. Give it a vague instruction and it will invent something plausible and wrong. Give it a specific one and it will build the right thing. Most of what feels like "the tool is broken" is really "my brief was unclear," and learning to tell the difference is most of what this course teaches. When a build misbehaves, naming the layer that is failing (is this a frontend, backend, data, or auth problem?) is how you know what to ask the AI to fix.

That is the entire mental model. Keep it in your head and the rest of the course is you getting fluent at directing each layer.