“It was very hard. Six next-gen supercomputer research projects were launched not long after you went into hibernation. Three of them employed traditional architecture, one used non–Von Neumann architecture, and the other two were quantum and biomolecular computing projects. But two years later, the lead scientists of those six projects told me that the computing power we desired was impossible. The quantum computing project was the first to be terminated, because it failed to find sufficient support in current theoretic physics: Research had run into the sophon barrier. Next, the biomolecular project was discontinued. They said it was only a fantasy. The last to end was the non–Von Neumann computer. Its architecture was actually a simulation of the human brain, but they said it was a shapeless egg that would never turn into a chicken. Only the three traditional architecture projects were still ongoing, but for a long time there was never any progress.”

“So that’s it.… I ought to have been with you the whole time.”

“It would have been no use. You only would have wasted eight years. It was only recently, during a period of time when we were totally discouraged, that we came up with the crazy idea of simulating the human brain in a practically barbaric way.”

“And what was that?”

“To put the previous software simulation into hardware by using a microprocessor to simulate one neuron, letting all the microprocessors interact, and allowing for dynamic changes to the connection model.”

Hines thought about this for a few seconds, then realized what she meant. “Do you mean manufacturing a hundred billion microprocessors?”

She nodded.

“That’s … that’s practically the sum total of all the microprocessors that have been manufactured in human history!”

“I didn’t run the numbers, but it’s probably more than that.”

“Even if you really had all those chips, how long would it take to connect them all together?”

Keiko Yamasuki smiled wearily. “I knew it wasn’t workable. It was just a desperate idea. But we really thought about doing it back then, and making as many as we could.” She pointed around her. “This here is one of the thirty virtual brain assembly shops we had planned. But it’s the only one that got built.”

“I really should have been here with you,” Hines repeated with more emotion.

“Fortunately we still got the computer we wanted. Its performance is ten thousand times better than when you entered hibernation.”

“Traditional architecture?”

“Traditional architecture. A few more drops squeezed out of the lemon of Moore’s law. It astonished the computing community—but this time, my love, we’ve really come to the end.”

A peerless computer. If humanity failed, it would never be equaled, Hines thought, but did not say it out loud.