The office was on the 42nd floor of a building on Avenue of the Americas that had a name David couldn't pronounce and a logo that was just a letter M inside a circle, which made it look less like a technology company and more like a cult.

David Chen worked in building analytics. His title was Senior Data Scientist, which meant he designed algorithms that predicted what people would do next. Not literally—well, actually, mostly literally.

Precog was the company's flagship product. It ingested data: purchase history, social media interactions, location data, even the typing patterns collected from corporate keyboards, and produced predictions. Consumer behavior: 97.2 percent accuracy. Voting patterns: 89.6 percent. Employment retention: 84.3 percent.

The last one was the one David worked on. Loyalty and turnover prediction. It told companies which employees were likely to leave, and why, and when.

"It's going to change HR," his manager, a woman named Priya with perfect hair and perfect nails and a loyalty score that David was fairly certain was close to 100, said at the last all-hands meeting. "No more surprise resignations. No more losing your best people."

"Or no more treating your people like problems to be solved," said the guy in row three. David didn't know his name. He knew everyone's name because he had designed a system that predicted interpersonal relationships from email metadata, but that didn't mean he remembered them.

David's loyalty score was 67 percent and dropping.

It had been 78 percent in January. By March, it was 71. Now it was April and it was 67, and Priya had mentioned it in their last one-on-one with the casual cruelty that characterized her management style. "Just something to keep an eye on, David. You're one of our most valuable team members. We'd hate to see you go."

One of the perks of working on the loyalty prediction system was that you could see your own prediction. David's dashboard showed a breakdown: 34 percent probability of leaking proprietary information, 28 percent probability of engaging in unauthorized external communications, 67 percent overall resignation risk within twelve months.

He stared at the number every morning. It was the only number in his life that seemed to matter.

He lived in Brooklyn. Not the part that was Brooklyn now—with hipster coffee shops and $3,000 studios and Instagrammable facades. The part that was still Brooklyn: the bodegas that had been there since the eighties, the schools that were underfunded because the property taxes were low, the trains that were always late because they were always late and nobody had fixed that in forty years.

Amy worked in social services in Bed-Stuy. She was a social worker, which in New York meant she was overworked, underpaid, and the only reason the city wasn't on fire. She was also the reason David stayed in Brooklyn instead of moving back to the city where he had grown up, which was Queens, which was nothing like Brooklyn and everything like nowhere.

"Did you tell her?" Amy asked over dinner at their apartment—a second-floor walk-up with a kitchen window that faced another kitchen window and a fire escape that smelled like other people's cooking.

"Tell who what?"

"About your score."

"I don't tell her anything. It's internal."

"But you're worried."

"I'm not worried. I'm aware."

"That's the same thing."

David ate his pasta. It was good—Amy was a better cook than he was, which was not a high bar but she cleared it every night. He thought about what she had said. Was he worried? He was aware. He was also curious, which was worse, because worry was a feeling and curiosity was a mechanism, and mechanisms didn't care how you felt.

The incident happened on a Thursday. David arrived at the office and found that his keycard did not work on the elevator to the 42nd floor. It worked on the 41st. Not the 42nd.

He called Priya. "There must be a glitch."

"No glitch," Priya said. Her voice was calm. Not cruel, not kind. Just calm, which in a manager was the most dangerous tone of all. "Your access has been adjusted. You'll work from the 41st floor until this is resolved."

"What is this resolved?"

"We'll let you know."

He worked from the 41st floor for two weeks. The building analytics team was on the 42nd. He could see them through the window across the open space—Priya, laughing at something, her hair perfect, her loyalty score presumably perfect. He could see thePrecog dashboard on the big screen in the 42nd-floor lobby, showing real-time predictions. He couldn't see it from the 41st.

He started noticing things he had not noticed before. The way security watched him more closely. The way his emails were slower to be delivered. The way his calendar invitations stopped coming.

He was being predicted out of his own job.

One evening, about three weeks after the elevator incident, David stayed late. The office was empty. The cleaning crew was vacuuming the 41st floor, and the sound rose through the ceiling like a lullaby for people who had nothing left to lose.

He sat at his desk and opened the Precog source code. He had written parts of it. Most of the prediction logic was proprietary and locked, but he knew enough to navigate. He found the internal memo that had triggered his score reduction: a recommendation from the security team, dated two days before the elevator incident, recommending "gradual access restriction pending internal review."

The review had never happened. The system had predicted a problem, and the system had responded, and no human being had intervened. No one had called David into an office and asked him what was going on. No one had looked at his record—five years at the company, zero infractions, three performance awards—and said, wait, the number might be wrong.

The number wasn't wrong. David knew that. He had helped build the thing that produced it. If you fed it enough data about a person, it would produce an accurate prediction. Not a guaranteed outcome. A probability. But probabilities, in a corporate system, were treated like certainties.

David opened a new document. He began writing.

He wrote about what he had seen. About the prediction system, how it worked, how it treated people, how it had reduced his entire professional existence to a number on a dashboard, how it had decided his fate without a human being thinking about it for more than three seconds.

He wrote for four hours. When he finished, it was 2 AM. The office was dark. The building was quiet. The only light came from his screen, illuminating his face in the blue glow of a man who was about to do something that his loyalty score had predicted with 42 percent probability.

He attached the document to an email. He addressed it to three people: a journalist at the Times, a journalist at the Times, and a regulator at the state securities commission.

His finger hovered over the send button.

He thought about Amy. He thought about her walking home from work at midnight, taking the train because the bus didn't run that late, carrying the weight of her cases in her shoulders the way she carried her briefcase—slightly crooked, because she was tired and her muscles had learned to compensate.

He thought about her asking him, Did you tell her? as if she already knew the answer and just wanted to hear him say it.

He thought about the 42 percent.

He pressed send.

The email went out. Three copies, delivered to three inboxes, carrying four hours of writing and six months of observation and one man's slowly crystallizing certainty that a number on a screen was not a verdict, even if the system that produced it believed it was.

David sat in the dark office and listened to the cleaning crew vacuuming downstairs. The sound was rhythmic and steady and utterly indifferent to what he had just done. It was the sound of a city that would continue whether he was fired or promoted or arrested or free.

He picked up his coat and walked to the elevator. He pressed the button. The elevator arrived. He got in. He pressed 1.

The doors closed. The elevator descended. And David Chen, who had spent his career predicting what people would do, sat in a descending box of polished steel and wondered, for the first time in his life, whether what he had just done was the thing a system could have predicted or the thing it couldn't.

He didn't know. He probably wouldn't know for a long time.

The elevator opened on the lobby. The doorman nodded at him. "Late night, Mr. Chen?"

"Something like that," David said.

He walked out into the Manhattan night. The street was wet from an afternoon rain that neither he nor the doorman had noticed. The reflections of the streetlights in the puddles looked like data points on a dashboard, scattered and bright and ultimately meaningless until you connected them into a pattern that told you something about where you were going.

David walked toward the subway. He did not know what would happen tomorrow. The system would probably know. But right now, in this moment, in this wet street under these bright lights, nothing was predicted. Nothing was certain.

For about three blocks, David Chen was free.

---

## OTMES V2 Objective Tensor Encoding

**Code**: `OTMES-v2-458407-105-M4-13B-5R5546-01C` **Title**: The Loyalty Score **Variant**: V-7

### Tensor Parameters - **Overall Literary Potential (E_total)**: 10.5 - **Dominant Mode**: MDOM (intensity: 85%) - **Dominant Angle**: 315.0deg - **Tensor Rank**: 10 - **Dominance Ratio**: 0.85 - **Irreversibility (I)**: 0.5

### Mode Vector M (10-dimensional) [[5.0, 2.0, 7.5, 3.5, 8.5, 6.5, 2.0, 6.0, 2.5, 5.0]]

| Mode | Dimension | Value | |------|-----------|-------| | M0 | Tragedy | 5.0 | | M1 | Comedy | 2.0 | | M2 | Satire | 7.5 | | M3 | Poetry | 3.5 | | M4 | Power/Strategy | 8.5 | | M5 | Suspense | 6.5 | | M6 | Horror | 2.0 | | M7 | Sci-Fi | 6.0 | | M8 | Romance | 2.5 | | M9 | Epic | 5.0 |

### Action Source Vector N [[0.5, 0.5]] (Active / Passive)

### Value Carrier Vector K [[0.4, 0.6]] (Individual / Trans-individual)

### Style Classification - **Western Style**: B1 - NY Realism - **Genre**: Political Thriller