The factory was big and the fluorescent lights were bright and Malcolm Grey sat in row fourteen, column six, and labeled pictures.

Cat. Car. Dog. Person. Traffic light. Fire hydrant. Person again. Backpack. Laptop. Person holding a cup.

Ten hours a day. Two hundred pictures an hour. Six days a week. The supervisor walked around with a clipboard and a stopwatch and did not smile.

Malcolm did not mind the work. It was easy. It was repetitive. It was the kind of work that let your mind wander while your hands kept moving, like a man mowing a lawn who thinks about nothing and everything.

He thought about the machine. His company called it Precog—a prediction system that analyzed human behavior using data labeling and pattern recognition. Malcolm called it what it was: a fancy way of looking at pictures of cats and telling the boss what he wanted to hear.

"Seventy-three percent accuracy on this batch," the supervisor said, stopping at row fourteen. "You're falling behind."

"Okay," Malcolm said. He labeled three more pictures faster. Cat. Person. Cup. The supervisor nodded and walked away.

At lunch, Malcolm ate a sandwich from the convenience store next to the factory. It cost $3.49 and consisted of bread that was too soft and turkey that was too thin and lettuce that had been cut into rectangles. He ate it on a bench in the parking lot, watching the other workers smoke and text and stare at nothing.

A guy named Ricky sat down next to him. Ricky had been Malcolm's coworker for two years and they had never had a conversation that didn't involve work.

"Hey," Ricky said.

"Hey."

"Watch this." Ricky pulled out his phone and opened an app. It showed Malcolm's face, Malcolm's name, and below it, a number: 67 percent. "They updated my scores today. Mine said 82. Must be a bug."

Malcolm looked at the screen. "What does 82 mean?"

"Employment stability. Higher is better. 82 means I'm probably gonna keep my job. Your—" Ricky squinted. "Yours says 34. You know what that means?"

Malcolm took the phone and looked. Thirty-four percent. A number that described his future, generated by a system that had looked at his address and his education and his work history and decided that he was probably going to lose his job within the next twelve months.

"I know what it means," Malcolm said.

He ate the rest of his sandwich in silence. The lettuce was cut into rectangles.

After work, Malcolm drove home in his 2003 Toyota. The radio only picked up three stations: a country station that played the same four songs, a talk radio station that yelled at people, and a Spanish station Malcolm could not understand. He kept it on the talk station because the yelling was easier to tolerate than the silence.

His apartment was in a complex off I-75 that had been built in the nineties and had been declining ever since. The walls were thin. The neighbor upstairs played football on his TV at 2 AM. The pool was closed because the filter broke and nobody replaced it.

Malcolm made dinner: pasta with jarred sauce. He ate it at the kitchen table, watching a crack in the wall that looked like the state of Missouri if you squinted.

After dinner, he turned on the TV. Sports. He did not watch sports. He turned it on because he liked the noise. It made the apartment feel occupied, even when he was the only one there.

A news segment came on: "Mirror Technologies' Precog system achieves 97.3 percent accuracy in predicting consumer behavior, analysts say. The system, which analyzes purchasing patterns, social media activity, and demographic data, is now used by over two hundred corporations nationwide..."

Malcolm turned the volume down but not off. He sat at his kitchen table and looked at the crack in the wall that looked like Missouri.

He thought about the thirty-four percent. He thought about what Ricky had said: probably gonna lose your job. He thought about the word probably.

Probability. That was what the machine dealt in. Not certainties. Probabilities. A system that looked at a million people like him—same address range, same education, same job—and found that thirty-four of them lost their jobs within twelve months.

But Malcolm was not thirty-four of anything. He was one person. He had thoughts. He had plans. He was thinking right now about looking for another job, or learning a new skill, or—

Or was that thought itself a product of his environment? Was the idea of changing his life just another data point, another pattern the machine had already identified and factored into the thirty-four percent?

He picked up his beer. It was warm. He drank it anyway.

On Friday, Malcolm called in sick. He did not feel sick. He just didn't want to go to the factory and label pictures and watch a machine predict things about people who couldn't predict themselves.

He drove through the abandoned industrial zone on the edge of town. Factories that had been empty since the nineties. Windows broken. Walls covered in graffiti. A Chevrolet sign lying in a field like the skeleton of some enormous animal.

He drove slowly, watching the ruins pass by his window. He thought about how everything here had been predicted too. The factories were built because someone predicted demand would be high. They closed because someone predicted demand would be low. The town was dying because someone predicted it was more profitable to move the jobs to another country.

Everything predicted. Everything decided before it happened.

He pulled over at a overlook on the edge of the Detroit River. The water was brown and slow and smelled like things that had been dumped in it and time that had washed over them and changed them and changed nothing.

Malcolm sat in his car and looked at the river and thought about the thirty-four percent.

Then he thought about something else. He thought about the fact that he had called in sick today. The machine had not predicted that. Or had it? Had the machine predicted that Malcolm would call in sick on a Friday because he felt uneasy about his employment score and needed to avoid his coworkers for one more day?

He pulled out his phone and opened the Precog app. Thirty-four percent. Same number as yesterday. The machine was always right, but it never changed its mind. It just kept saying the same thing, day after day, like a man who only knew one joke and told it at every party.

Malcolm closed the app. He started the car. He drove home. He made dinner. He turned on the TV. He went to sleep.

On Monday, he went back to work. He sat in row fourteen, column six. He labeled pictures.

Cat. Car. Dog. Person. Traffic light. Person again.

He labeled them all day. At 3 PM, the supervisor stopped by. "Your scores dropped today," the supervisor said. "Thirty-two percent."

Malcolm nodded. He labeled three more pictures. Cat. Person. Cup.

At 5 PM, he drove home. At 6 PM, he ate pasta. At 7 PM, he watched TV. At 10 PM, he went to sleep.

On Tuesday morning, he woke up at 6 AM, made coffee, and drove to work. He sat in row fourteen, column six, and labeled pictures.

He thought about calling in sick again. He did not. He thought about quitting. He did not. He thought about breaking into the server room and deleting the system. He did not.

He labeled pictures. Cat. Car. Dog. Person.

The machine knew what he would do. It had always known. And today, for the first time, Malcolm knew that the machine knew.

He did not know if that made him free or trapped. He did not know if it mattered.

He labeled a picture of a cup.

---

## OTMES V2 Objective Tensor Encoding

**Code**: `OTMES-v2-931869-94-M3-10E-6R3673-01A` **Title**: What the Machine Knew **Variant**: V-5

### Tensor Parameters - **Overall Literary Potential (E_total)**: 9.4 - **Dominant Mode**: MDOM (intensity: 70%) - **Dominant Angle**: 270.0deg - **Tensor Rank**: 9 - **Dominance Ratio**: 0.7 - **Irreversibility (I)**: 0.6

### Mode Vector M (10-dimensional) [[6.0, 0.5, 2.0, 7.0, 1.5, 3.0, 2.0, 5.0, 1.0, 3.5]]

| Mode | Dimension | Value | |------|-----------|-------| | M0 | Tragedy | 6.0 | | M1 | Comedy | 0.5 | | M2 | Satire | 2.0 | | M3 | Poetry | 7.0 | | M4 | Power/Strategy | 1.5 | | M5 | Suspense | 3.0 | | M6 | Horror | 2.0 | | M7 | Sci-Fi | 5.0 | | M8 | Romance | 1.0 | | M9 | Epic | 3.5 |

### Action Source Vector N [[0.35, 0.65]] (Active / Passive)

### Value Carrier Vector K [[0.7, 0.3]] (Individual / Trans-individual)

### Style Classification - **Western Style**: E - Dirty Realism - **Genre**: Working-Class Fiction