The limit was real

So Minsky and Papert were right about that part.

The Perceptron can only handle problems where inputs each point toward the answer independently. The moment a problem requires understanding relationships — one input changing the meaning of another — it fails. Every time. With no way out.

That's not something you can fix by training longer or giving it more data. It comes from what multiply-and-add is at its core — and that's not going to change.

One neuron. One ceiling. Real.