“So, if we think of it”—and here Zula held out a hand as if to deflect any objections—“if we imagine, just for the sake of argument, that we have one process, what we used to call a computer program, that does one thing only, which is to simulate the workings of one single neuron in a brain. That’s all it does.”

Sophia nodded and made a wheeling gesture with her hand, encouraging her mother to keep rolling—perhaps a little faster.

“And it’s running on a chip in a farm in Ephrata. Pretty much doing its own thing. But real neurons in the brain are connected to other neurons—that’s kind of the whole point. It’s why we spent a billion dollars scanning Uncle Richard’s connectome.”

“Right,” Sophia said. “So, somewhere else—maybe on a different thread running on the same chip, maybe on a different chip in the same farm, maybe on another farm altogether on a barge off the coast of California or whatever, there are other processes, all clones, all busy simulating the workings of other neurons. Neurons that are connected to that first one. And every so often, a message needs to be passed from one of those processes to another.”

“A neuron fires or something,” said Marcus, getting into the spirit of the thing.

Sophia paused for a second, perhaps stifling an urge to swap out what Marcus had just said for something a little more technically precise.

“Or whatever,” Marcus added. “Axons. Synapses. Dendrites.”

“Signals have to pass between neurons or you don’t have a brain,” Zula said.

“We’re getting pretty far afield from how it actually works,” Sophia said, “but no matter how you simulate a brain, the point is that you’re going to have a lot of separate, independently running processes that have to communicate over a wiring diagram that basically represents the connectome. And every signal—every separate communication—has some overhead that you have to pay in terms of computation and bandwidth. If you’re doing it the way it ought to be done, crypto-wise—”

“The way all communication happens nowadays,” Zula said, quoting Sophia’s own words back to her. She winked at Marcus.

“—yeah, the overhead is large and everything is some combination of slow and expensive. To figure out how to do those computations faster was kind of an obvious fat target for companies building quantum computers. Instead of waiting for Moore’s Law to catch up with what was needed, they were able to leapfrog it, and bring us devices that were decades or centuries ahead of the curve in terms of their ability to do this stuff.”

“Distributed secure computation,” Marcus said, just to be sure he was tracking.