The Ten Minutes

In the AI Story

The case has the force of paradigm. It is the specific empirical instance that Flyvbjerg's entire framework was built to analyze — the concrete mechanism through which substitution produces phronetic erosion invisible to output measurement. The engineer reported improvement because improvement was what she experienced: less tedium, higher output, expanded scope. The loss arrived only months later, when she realized she was making architectural decisions with less confidence than she used to possess, and could not explain why.

Three structural obstacles prevent the dominant research paradigm from detecting this mechanism. First, the erosion operates over years, not months — longer than most AI impact studies. Second, the loss is invisible to the person experiencing it; self-report measures cannot detect what the subject does not know has occurred. Third, the loss is qualitative, not quantitative; instruments designed to measure output cannot detect changes in the depth and reliability of judgment. The Berkeley study, careful as it was, could not have detected the ten minutes phenomenon because its instruments were not designed to look for it.

The pattern generalizes beyond software engineering. In medical training, the resident who sees hundreds of routine presentations occasionally encounters the case that is not routine — the sore throat that is actually a peritonsillar abscess, the mild infection that masks early sepsis. These rare cases build diagnostic intuition. If AI triage systems route routine presentations to automated assessment, the training pipeline loses access to the rare, unpredictable moments that the routine contains. In legal training, the junior associate reviewing documents for relevance occasionally finds the document that changes the case. Each domain has its ten minutes — formative experiences embedded in routine work, inseparable from it, destroyed by the same automation that removes the tedium.

The loss compounds generationally. The engineer whose architectural intuition has eroded does not merely make worse decisions for herself. She makes worse decisions that become the substrate for future decisions by future practitioners who inherit her systems. When an entire generation enters the profession in an AI-augmented environment — never experiencing the formative struggle, never building the embodied judgment — the profession loses a capacity that cannot be recovered through training programs after the fact. The knowledge the ten minutes built is not the kind of knowledge that can be taught in a workshop.

Origin

The specific framing originates in Segal's The Orange Pill (2026) and is elaborated through Flyvbjerg's phronetic framework in the book. The general phenomenon — formative moments embedded in routine work — has been discussed in the expertise research tradition (Anders Ericsson) and the tacit knowledge tradition (Michael Polanyi).

Key Ideas

Rare but formative. The developmental significance of these moments is disproportionate to their duration; eliminating four percent of daily work can destroy eighty percent of its developmental value.

Invisible in categories. From the system's perspective, the entire routine is the same category of work; the distinction between tedium and formation operates below the granularity of task classification.

Undetectable by self-report. The loss is not experienced as a loss; the subject reports improvement because improvement is what the surface measurements show.

Generational compounding. Loss in one generation removes the developmental substrate for subsequent generations, producing cumulative erosion that no single generation can recover.

Cross-domain pattern. The mechanism operates in every domain where expertise develops through embodied experience — medicine, law, engineering, design, management.