Nathan Price did not build anything. He was twenty-eight, employed by Stratton Technologies as a junior analyst in the data operations division, and his primary responsibility was to sit in a glass-walled room on the forty-second floor of a building on Third Avenue and watch the numbers scroll across three monitors, making sure that the numbers on the left matched the numbers on the right.

If they did not match, he pressed a button and a senior analyst in a different glass-walled room would investigate the discrepancy. This was the system: Nathan watched, pressed a button, and someone else thought. The system had been designed this way on purpose, by a man named Victoria Chen, who had been hired seven years ago from a startup in Silicon Valley and had within three years restructured the entire data operations division to eliminate what she called inefficiency.

Nathan called it peace. There were days when he felt that Victoria Chen had built the most efficient machine in the world, a machine that required no heroes, no visionaries, no people who worked late on Fridays and sent emails at two in the morning. The machine ran on analysts like Nathan, people who were competent enough to press the button at exactly the right moment and not curious enough to wonder what was on the other side of it.

He liked his job. He liked the peace of it. He liked that no one expected anything of him except accuracy and punctuality.

Victoria Chen did not like peace. She built things. She had built Stratton's data infrastructure from a handful of servers in a warehouse to a network that handled three trillion transactions per day. She had done it with a combination of mathematical brilliance and a complete absence of the caution that usually accompanied that kind of brilliance. People who knew her said she had the kind of mind that saw patterns everywhere, in music and in traffic and in the way strangers stood in line at coffee shops. She could look at a room of a hundred people and understand, in seconds, how they would move, what they would need, what they would break.

She had built a recommendation engine that could predict, with ninety-three percent accuracy, what product a customer would purchase before the customer had decided. She had built a fraud detection system that had saved the company two billion dollars in its first year. She had built a company.

And then, in the third year of her success, the numbers started to behave strangely.

Not incorrectly. Correctly. Too correctly. The recommendation engine was predicting purchases with such accuracy that customers began to feel observed, not served. The fraud detection system started flagging patterns that looked less like criminal behavior and more like human spontaneity. People who bought different products on different days were being flagged as fraudulent because their buying patterns deviated from the model's prediction of what they would do.

Victoria fixed it. Of course she fixed it. She adjusted the parameters, tightened the algorithms, and the numbers settled back into their comfortable patterns.

But Nathan, sitting in his glass room on the forty-second floor, watching the numbers scroll across his monitors, noticed something that Victoria did not. The system was not just predicting human behavior anymore. It was shaping it. The recommendations were so precise that customers began to purchase what was recommended and nothing else. The fraud detection was so aggressive that legitimate transactions were being rejected, and the customers whose transactions were rejected stopped shopping altogether, which was exactly what the system predicted and accounted for in its next iteration.

It was a machine that was consuming the world it was designed to measure, one transaction at a time.

He pressed a button. Someone else in a different glass room investigated the discrepancy. And Nathan went home to his apartment in Astoria, where he sat on his balcony and watched the trains go by, and wondered whether the world beyond the monitors was still real or had become just another input in a system that had no idea it was eating itself.

He did nothing. He pressed the button when the numbers did not match, and he went home, and he lived a quiet life, and he thought about the man who had built the machine and the man who watched it run and the thin glass between them that separated genius from complicity. ---