What age should a kid start building with AI?

Around eight, for most kids. But age is the least useful thing to go on.

We write for 8–14, and that range is honest — it's where this works most reliably. It's also a blunt instrument. We've seen seven-year-olds fly through a project and twelve-year-olds bounce off it. The number on the birthday cake predicts very little.

What actually predicts readiness

Three things, in order of importance.

Can they read comfortably and independently? This is the big one. Our books are written simply, but a child who is still working hard at the mechanics of reading has no attention left over for the ideas. If reading is still effortful, do this together rather than handing it over.

Can they stay with something that isn't working? Building means things break constantly. A child who can tolerate twenty minutes of "it's still not right" will do well. A child who needs immediate success will struggle, and that's not a character flaw — it's a developmental stage, and it passes.

Do they want to make something specific? A kid who says "I want to make a game where you're a cat in a haunted mall" is ready in a way a kid who says "I guess I'll try coding" isn't. Wanting a particular thing to exist is most of the fuel.

Rough guide by age

Six to seven: possible with a parent doing it alongside them, treating it as a shared project rather than something they own. Expect to read aloud and drive the keyboard.

Eight to ten: the sweet spot for a first build. Old enough to follow written steps, young enough that nothing feels like homework. Most kids in this range need a grown-up nearby for setup and then want to be left alone.

Eleven to fourteen: capable of working entirely independently, and generally more ambitious about what they want to build. The main risk here is scope — they'll try to build something enormous and stall. Helping them cut is the most useful thing you can do.

Fifteen and up: our books will feel young. Point them at real developer tools directly — they can handle it.

The mistake to avoid

Don't start when you think they should be interested. Start when they are.

A child pushed into this at the right age with no interest will do it badly and conclude they're not a builder. A child who comes to it at eleven because they actually want something to exist will move faster and enjoy it more. There's no window closing. Nobody has ever been too late to start building things.

If you're trying to seed the interest rather than respond to it, the reliable move isn't a course — it's showing them something a kid their age made, and letting them decide they want that.