Individual assessments were quieter.
No crowds. No shared tension. Just a room, a screen, and the uncomfortable awareness that there would be no one else to absorb attention.
He was seated across from a single terminal. No proctor. No visible observers. The glass wall reflected his posture back at him—neutral, relaxed, forgettable.
The screen activated.
[Assessment Type: Cognitive Adaptation.]
[Objective: Response Calibration.]
[Note: No optimal solution exists.]
That last line was deliberate.
He adjusted his breathing by a fraction. Not to calm himself—he was already calm—but to maintain consistency. The system loved baselines. Any deviation, even improvement, attracted scrutiny.
The first prompt appeared: a scenario involving incomplete data and conflicting incentives. Not difficult. Just layered.
He answered incorrectly.
Not obviously so. Just enough to suggest oversight.
[Response logged.]
[Confidence level: moderate.]
Good.
The second prompt arrived immediately. Different framing. Same underlying structure. This time, he corrected the error—but introduced a new inefficiency. A trade-off that looked human. Imprecise. Reasonable.
The system paused longer.
[Pattern adjustment detected.]
He waited.
The third prompt escalated. Time pressure. Reduced information. A subtle contradiction embedded in the data stream—something most people would either miss or overcorrect.
He neither rushed nor froze.
He chose the safest wrong answer.
The kind that preserved internal logic while sacrificing outcome.
The system reacted.
[Cognitive consistency: high.]
[Performance ceiling: within expected range.]
He almost smiled.
Almost.
When the assessment ended, there was no score. Only a brief acknowledgment.
[Session complete.]
As he stood, the glass wall darkened momentarily, then cleared. Somewhere above, a faculty member would flag the session for review—or dismiss it entirely.
He predicted the latter.
Outside, the corridor was empty. Classes were still in session. The academy's rhythm ensured solitude when it mattered.
His tablet vibrated again.
[Optional Consultation Available.]
[Advisor: Assigned.]
Optional consultations were never optional.
He accepted.
The advisor's office was smaller than expected. No grand design. No intimidation tactics. Just shelves, a desk, and a woman who looked up from her tablet with mild curiosity.
"Have a seat," she said. "You're… unremarkable."
She said it like an observation, not an insult.
"Yes," he replied.
"That's rare," she continued. "Most students try very hard not to be."
She studied him for a moment longer than necessary. Not hostile. Not warm. Assessing.
"Your assessment was consistent," she said. "But inconsistent with your group performance."
He tilted his head slightly. "Is that a problem?"
"Not yet."
Silence settled between them.
She tapped her screen once. "Tell me—do you prefer working alone?"
The question wasn't about preference.
It was about intent.
"I prefer clarity," he said.
She smiled faintly. "Careful answer."
The system updated quietly in the background.
[Advisor interaction logged.]
[Intent classification: inconclusive.]
The advisor leaned back. "You'll be watched a little more closely."
"I understand."
She waved him off. "You're dismissed."
As he left the office, he cataloged the interaction. Minimal risk. Slight increase in observation probability. Acceptable.
Outside, the academy continued as if nothing had changed.
But something had.
Variables had been reduced.
Which meant future errors—if any—would be easier to trace.
He adjusted again. Smaller margins. Cleaner noise.
The system believed it had narrowed the field.
It had.
Just not in the way it thought.