The Many Clocks — Summary

There is a paper schedule on the wall outside Dana’s office, four feet wide, one row per resident and one column per two-week block, running the whole academic year. The program bought scheduling software in 2014 and the software works. Dana maintains the paper anyway, in pencil, because the software will show you a month and will not show you a year.

Her residents applied to medical school six to nine years ago and will be practicing until roughly 2065. Training from the first day to unsupervised practice runs eleven years at the shortest. In the same nineteen years she has watched the tools change perhaps eight times in ways that mattered. The current ones are eighteen months old. She is designing a curriculum this month for a cohort that will spend most of its working life in a decade nobody can describe, and she has to decide by March.

The standard image for all of this is a race, and a race requires one track, one direction, and a finish line that means the same thing to everyone on it. None of those hold. Dana is not behind. There is nothing for her to be behind. The better image is polyrhythm: several clocks, each keeping honest time at its own rate, none of them wrong, no shared downbeat.

There are at least seven. Formation time runs about twenty years per depth and has not changed in a century. Capability time runs in months. Commoditization time, twelve to eighteen months, is the interval between a capability being scarce and being free. Release time is two clocks wearing one name, because a commercial release has a date and a counterparty and an open-weights release has neither and cannot be recalled. Deployment time is not a clock but a distribution, and the variance between jurisdictions is larger than the average says anything about. Depreciation time decides whether a capital decision was correct. Policy time is the strange one: it can run backward, because what a reversed rule destroyed was a planning horizon rather than a level, and repeal does not rebuild it.

Any two clocks running at different rates define a gap, and somebody stands in it and buys on one and sells on the other. That is not a metaphor for arbitrage. It is what arbitrage is. A staffing firm’s cost side is formation time and its revenue side is capability time, and the entire margin is the width between them. Run the commoditization clock forward and the margin goes, not because anyone competed it away but because the two clocks came back into line. Which is why advice about closing spreads mostly fails: you cannot educate your way out of a rate difference.

If the ensemble has a master, the candidate is commoditization, because it sets the interval during which anything can be charged for, and therefore the ceiling on what can be spent on all the others. Institutions are making twenty-year decisions inside an eighteen-month revenue horizon.

We wonder whether anyone standing inside a beat pattern can hear it as a pattern, or whether the recurring structure is only ever audible to someone who was not keeping any of the time.

Dana keeps the pencil. Every August she draws a new grid and the old one goes into a drawer with eighteen others, and she has never thrown one away. Laid end to end they would be nineteen years of one institution’s actual time, at the only resolution that shows both a shift and a career, which is a thing that exists nowhere else and that nobody has asked her for.