Dr. Elara Vance had spent ten years in computational dermatology, but her latest project was different. She called it Eidos , a “normal human face simulator” built not to beautify or exaggerate, but to generate the profoundly unremarkable.

A man this time. Fortyish. Receding hairline, ears that stuck out just a little, tired but kind eyes. She stared. He looked like her seventh-grade math teacher, Mr. Hamada, who’d let her borrow his protractor when she’d lost hers.

The room was silent. Then a woman in the back, an engineer from a major social-media company, raised her hand. “Can I license this?”

“The hook,” Elara said, “is that these people exist. Or they could. And no algorithm has ever been trained to care about them.”

“No,” Elara said, closing her laptop. “But you can look at someone today without trying to improve them. That’s the simulator.”

Eidos wasn’t creating faces. It was remembering them. Every face it generated felt like a person Elara had once glimpsed on a bus, or stood behind in line, or sat next to in a waiting room. She realized, with a strange ache, that her simulator had done what no AI art tool had ever done before: it had made the invisible visible.