A Machine Reinvented the Wave, and Nobody Taught It To

We tend to think of AI as built — engineered, line by line, like a bridge. It isn't. It's grown. You don't write the rules a large model follows; you set up a process and let it search billions of times for whatever works, and what comes out the other side is shaped by that search the way a tree is shaped by light. Mostly we can't even read what it found. But every so often researchers crack one open and see exactly what it reached for — and the answer is unsettling in the best way.

Here's the cleanest example. A few years ago, researchers trained a small network to do one boring task: modular arithmetic — clock math, the kind where after 12 you wrap back around to 1. They expected it to memorize. Instead, when they looked inside, they found the network had quietly reinvented something no one told it about. It was representing numbers as positions on a circle, and doing the arithmetic by rotating them. It had rediscovered, from scratch, the heart of the Fourier transform — the two-hundred-year-old mathematics of turning anything into a sum of waves.

Nobody put the wave in. The network found it, because for anything that repeats — and clock math repeats — representing it as a rotation on a circle is the cheapest, cleanest way to compute. The wave wasn't installed. It was discovered, as the efficient answer, by a blind process that had no idea it was doing mathematics at all.

Now — the temptation here is enormous, and worth resisting. It's tempting to say the machine "tuned into" some deep cosmic frequency, that it brushed up against the vibrational nature of reality. That's the cheap move, and it's wrong. The honest version is stranger and more grounded: the wave shows up inside the machine for the exact same reason it shows up in a plucked string and a dividing cell. It's not a hidden essence humming under everything. It's that when a problem has periodic structure, the wave is the most efficient way to handle it — so anything searching hard enough for efficiency, whether it's evolution, physics, or a neural network grinding through gradient descent, lands on the same form. Independently. Without contact.

That's the part that should give you chills. Not that a machine found magic — but that the same frugality runs through music, biology, and a pile of matrix multiplications on a server, and it's strong enough that three completely unrelated searches keep arriving at the identical shape.

We are now growing minds the way nature grows everything else. And like nature, they reach for the wave — not because it's sacred, but because it's cheap and true.

That recurrence — the same form surfacing in places that never touched — is the spine of a much larger argument. Frequentical — the full philosophy — follows this single pattern across music, biology, AI, and the physics of matter, keeping the math honest the whole way. And The Book of Life and Music, a novel, tells the same truth as a story. Both are available as PDFs at patrickwroden.com.