The Two Clocks

Human Attention / Reality Equation

The Two Clocks

The minutes are counted by one clock and the years by another. The second one runs on surprise — and your stack is quietly winding it down.

John Rector 6-minute read
01

Two instruments, one life

Your best vacation was over before it started. You landed, the week detonated into color, and suddenly you were repacking, certain some of the days had been stolen. Then you got home, and the week did something strange in memory: it grew. Held against the routine weeks on either side of it, it reads like a month. Those routine weeks did the opposite — each felt ordinary enough while it passed, and in memory they have collapsed to almost nothing, eleven of them folding flat like empty boxes.

Psychologists have a name for this — the holiday paradox — and, more usefully, they have known for half a century that it is not one clock misbehaving but two different clocks disagreeing. Ask someone to track time while it passes and you are measuring one instrument: the prospective clock, the felt now, which runs on how much attention is left over for monitoring time itself. Ask someone afterward how long something lasted and you are reading a second instrument entirely — the retrospective clock, which does not measure duration at all. It reconstructs duration, from whatever the interval left behind in memory. The two judgments dissociate so reliably that time researchers treat them as separate literatures.

Everything in this essay follows from one observation about that second clock: you cannot consult it while living. You live by the first clock, minute by minute. But every account you will ever give of your life — every answer to how long were you there, what was that year like, where did the summer go — is read off the second. The first clock spends your time. The second clock is the only one that keeps it.

02

What the second clock counts

There is no tape. Nothing in you recorded the raw footage of March, and so when memory is asked how long March was, it does the only thing it can do: it counts what it can retrieve. Richard Block called this the contextual-change hypothesis, and the experiments behind it are almost embarrassingly direct — pack an interval with distinct, varied, context-shifting events and people later judge it longer; fill the same span with uniform repetition and it shrinks in review. Remembered duration is not elapsed time. It is retrieval density wearing a clock’s costume.

Now put this in the Reality Equation’s terms. I have argued that attention is not a spotlight you aim but a current that flows to the gap between the Actual and the Expectation — and that attention is the entry stamp on memory. What the Expectation already predicted never clears the border, which is why nine years of a commute can leave no trace. Put the two together and the second clock resolves into something exact: it is not counting events. It is counting prediction errors. A period’s remembered length is its accumulated surprise — the integral of the gap. The vacation reads long in memory because nothing in it matched a model; the routine week reads short because everything did.

The first clock counts minutes. The second counts surprises. And only the second one writes your autobiography.

The dissociation even shows up at the scale of milliseconds. In the oddball effect, a novel image slipped into a stream of repeated ones is judged — reliably, by nearly everyone — to have lasted longer than the identical duration around it. Surprise dilates time at every scale we can measure. Repetition compresses it at every scale we can measure. The clocks disagree because they are billing different quantities, and only one of them is billing reality.

03

Why the years accelerate

Scene / December 31

You say it every year now, and every year you mean it more: where did this year go? Try the experiment honestly — name what happened in it. A trip. A health scare. One project that mattered. The move, if there was a move. Four or five stones, and around them, months of water. The year was 365 days long by the first clock. The second clock has it at four or five events — and the second clock is the one you are consulting when you mourn.

William James saw the shape in 1890: each passing year converts more experience into automatic routine, and “the days and the weeks smooth themselves out in recollection to contentless units, and the years grow hollow and collapse.” The explanation usually attached to the feeling is the ratio theory James borrowed from Paul Janet — a year is a tenth of a ten-year-old’s life and a fiftieth of yours, so of course it feels thinner. The arithmetic is tidy, and I think it is the weaker account. The equation gives a sharper one: childhood is wall-to-wall prediction error. Nothing has happened twice yet; the Expectation owns nothing; every day clears the border at full width. By fifty, the Expectation owns most of the day before it starts — and the owned portion leaves no marks for the second clock to count. The years are not speeding up. They are arriving pre-predicted, and the predicted part of a year has no length in memory.

Which means the acceleration is not a law of aging. It is a law of repetition — habituation read off the retrospective clock — and repetition, unlike age, is negotiable. The old are not doomed to short years, and the young are not guaranteed long ones. A twenty-five-year-old running a fully smoothed routine is aging, by the only clock that keeps the record, faster than a sixty-year-old who keeps breaking her own week.

Figure 01 Seven days, as the second clock files them
Week somewhere new first time, no working model 100
Week you shipped something hard problem, real stakes 62
Ordinary week commute, desk, screens, sleep 18
Fully assisted week briefed, smoothed, pre-answered 8
Schematic, not a dataset: the bars illustrate the documented direction of the retrospective effect — remembered duration tracks the density of distinct, surprising, context-changing events, not elapsed time. The hatched bar is this essay’s forecast, not a finding: a week optimized end-to-end for zero prediction error should file shorter than any week we have yet lived.
04

The stack buys minutes and pays in years

Now put the synthetic subconscious into the frame. Everything the personal AI stack does well, it does to the first clock. It removes the wait, smooths the friction, pre-answers the question, briefs the morning so nothing in it startles. Judged by lived minutes, this is pure gain — I have called that freed attention real wealth, and I hold to it. But notice which clock the gains are denominated in. Every gap the stack closes on your behalf is a felt minute saved on clock one and a mark not made on clock two. The smoothing that makes a day frictionless is precisely what makes it unretrievable. In “The Anesthesia of the Expected” I argued that a stack which continuously improves your Expectation accelerates habituation — that was a claim about perception, about what you notice while the day is happening. This is the same mechanism one ledger later: the anesthesia has an invoice, and the invoice arrives in December, when you ask where the year went and the second clock hands you four stones.

Convenience is billed by the minute and paid for in remembered years. No one shows you that exchange rate before you sign.

I want to be precise about the claim, because the cheap version of it — technology bad, friction noble — is not what the equation says. The equation says remembered time is a budget you can spend deliberately, and that the stack, as currently built, spends it by default in one direction. The design requirement is the same one I keep arriving at from different sides: the stack must learn to run the subtraction in reverse, and now we can say what the metric is. Not screen time. Not tasks closed. Retrievable density — what this week will be worth when the second clock files it. A stack that can plan a day could just as easily plan a memorable one: route the walk through the street you have never taken, leave one problem unsolved for you to hit cold, put the strange thing on the calendar and decline to explain it. The tools that made your days efficient are the first tools in history that could make your years long.

Already true
The prospective/retrospective dissociation has been documented since the 1970s, remembered duration reliably tracks contextual change and novelty rather than elapsed time, and the oddball effect shows surprise dilating time at even sub-second scale. The personal stack is already flattening daily prediction error for millions of people, and “the years are flying” is already the most common thing adults say about time.
What has to happen
Within a few years, a personal AI product measures and reports something like remembered time — a retrospect score, a count of weeks that left retrievable marks — and offers to optimize it: planning novelty density the way fitness apps plan training load, and marketing the result as making your life feel longer in review.
Where I am probably wrong
Engineered surprise may not pay the second clock at full rate. The mechanism that makes the vacation long — a genuine absence of any working model — may be exactly what scheduling destroys: a novelty subscription becomes a routine of novelty, the Expectation learns the wrapper even when the contents vary, and the second clock, which cannot be fooled about whether prediction actually failed, files the engineered surprises the way it files any other recurring appointment. If so, remembered time cannot be manufactured, only protected — and the stack’s honest ceiling is refusing to smooth what little genuine surprise still reaches you.
Sources

Block, R. A., & Reed, M. A. (1978). Remembered duration: Evidence for a contextual-change hypothesis. Journal of Experimental Psychology: Human Learning and Memory, 4(6).

Zakay, D., & Block, R. A. (1997). Temporal cognition. Current Directions in Psychological Science, 6(1) — the prospective/retrospective distinction.

Tse, P. U., Intriligator, J., Rivest, J., & Cavanagh, P. (2004). Attention and the subjective expansion of time. Perception & Psychophysics, 66(7) — the oddball effect.

James, W. (1890). The Principles of Psychology, Ch. XV, “The Perception of Time” — the “hollow” years passage, and the ratio theory credited to Paul Janet (1877).

Hammond, C. (2012). Time Warped: Unlocking the Mysteries of Time Perception — the holiday paradox.

Author: John Rector

John Rector is a Charleston-based entrepreneur, author, and AI strategist. He co-founded E2open, the supply-chain software company acquired for $2.1 billion in 2025, and in 2026 opened Charleston AI, a 3,000-square-foot lab that helps people and organizations understand and use artificial intelligence. He is the creator of The Reality Equation — a lecture series, book, and curriculum exploring attention, prediction, and how reality is experienced — and the author of more than two dozen books. He writes and speaks widely on artificial intelligence, attention, and the future of human work.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from John Rector

Subscribe now to keep reading and get access to the full archive.

Continue reading