Hidden layers

And just like riding a bike, nobody programs in the exact angles to lean at or the forces to apply. The network finds those things through repetition and correction, the same way your brain did. The weights adjust, layer by layer, until something works.

This kind of structure has a name: a neural network with hidden layers. The middle layers are called hidden not because they're secret, but because you can't see them from the outside. The inputs go in. The output comes out. Everything in between is invisible unless you go looking.

The architecture was the easy part. Training it was a different problem entirely.