← All lessons

The big picture — software & AI in plain terms

Two clear pictures, in order: first what software actually is, then what AI is — kept separate, because they are not the same thing. Imagine it explained to a curious 15-year-old who has never coded. Nothing to memorize.

Software is genuinely hard to picture, for one reason: you can't see it. Let's start exactly there.

1. Software is invisible — so what is it, really?

You can hold a phone or a game console in your hand. That's hardware — the physical machine. But the app or the game running on it isn't something you can touch. So what is it?

Software is a big set of instructions — a long, exact list of "when this happens, do that" — written down and stored as a file. On its own it just sits there, like a recipe in a drawer or the rulebook of a board game: words on a page, doing nothing. The magic only happens when the machine runs those instructions — then things start appearing and responding on your screen.

So hold onto this: software = a pile of instructions, stored somewhere, that a machine follows. Invisible on its own; alive when it runs.

2. A video game makes it click

A game is software — and it lets you actually see every piece of software at once:

A banking app, a chat app, this very project — they're all the same shape as that game. Different job, same parts.

3. The two halves: front end and back end

Engineers split every app into two halves, and you only ever touch one of them:

And the back end keeps two different kinds of stored stuff — worth keeping straight:

Both have to live somewhere. Exactly where — on your own device, or on a powerful computer out on the internet — is what the later lesson Foundations unpacks. For now: front end = where you talk to it; back end = where it works and remembers.

4. How you talk to software — and how it talks back

You've always had many ways to interact with software, long before AI existed:

A calculator, a music player, a video game — all software, all interactive, and none of them AI. That matters, because it tells us exactly what AI adds.

5. Now — what is AI?

Everything so far ran on rules a human wrote out by hand: fixed and exact, "if X, then Y." That's ordinary software — like a vending machine: press B4, you always get B4. Dependable, but it can only do what someone spelled out in advance.

AI is a new kind of back-end worker. Instead of following only hand-written rules, it has learned patterns from a massive pile of examples — so it can handle the fuzzy, human requests nobody can write exact rules for: "understand what I just typed," "draw a cat wearing a hat," "answer this in plain English." Less vending machine, more clever, very well-read assistant — you ask in everyday words, it gives its best response. (And, like a person, it can be confidently wrong — so you stay the judge.)

The key point: AI doesn't replace software — it slots into it. It's a new kind of part you add to the back end for the fuzzy, human jobs, bolted onto the very same front-end / back-end / data shape you just learned.

(There's also a second role — the one you'll live in here: AI as the helper that builds the software with you, where you describe and it assembles the parts. More on that across the next lessons.)

6. The whole picture, in one breath

Software is a set of invisible instructions that work on your data, split into a front end (where you see and interact) and a back end (out of sight, where the instructions run and both the program and your data are stored). AI is a new kind of worker you can add to the back end for fuzzy, human tasks — and also the helper that builds the whole thing with you.

Everything after this just zooms into one piece at a time.

Next: The AI side — a closer look at those AI "workers": models, agents, and why we put many of them in one place.