Movement as Accretion

People who have tried to keep a movement habit usually know how it fails. The system works for two weeks. Keeping the system going then takes more effort than the exercise does. Eventually the habit stops, with no single moment of quitting. This failure is worth looking at because other daily activities need no system at all. For me, working with AI turned out to be one of them: I want to do it every day without any tracker. A skipped day feels wasted, and that feeling comes without any streak to break or ring to close. Movement has never worked that way. The two activities are motivated in different ways. The difference is worth working out in detail, because movement can probably be set up so that it is motivated the same way the AI work is.

The feeling of a wasted day is the thing to examine. At first it seems like guilt, but looked at closely it works differently from guilt. On a skipped day nothing breaks, nothing is lost, and no score resets. The regret is about one specific thing: the project did not progress. The day could have added something, such as a feature, a page, or a decision, and it did not, so the project has one day less work in it than it could have had. The regret is about an addition that did not happen. No rule was broken, and nothing punishes the skipped day. No tracker or reward system creates the regret. The work creates it, because every session leaves a visible result on a project that continues the next day.

This kind of motivation has three parts. First, the project is ongoing from one day to the next, and no single day is required. Second, every session leaves an artifact: something exists afterward that did not exist before, and it stays. Third, there is no quota. A very long day and a twenty-minute day both count, because both advanced the project.

Streaks and rings motivate in the opposite way. Both measure maintenance: the user keeps a number going, and the number can drop back to zero. I never liked Duolingo’s streaks. I dislike the Apple Watch rings for the same reason: they are too hard to close every day. The bar being high is only part of the problem. With a maintenance metric, the one notable event is a day of absence: the metric has something new to report only on the day the streak breaks. The only feeling it produces is fear of losing the number. That fear comes from the metric, and the work itself does not produce it. The work also never adds up to anything. After three hundred days of closed rings, the only result is the number three hundred, and that number mostly measures how much will be lost when the streak ends.

So the plan is to set up movement as a project that grows, motivated the same way as the AI work. Every session adds something that lasts. A missed day adds nothing, and it does not take away anything added before. There is no daily minimum to meet. The growth is visible enough that seeing it makes a person want to add more. The question is then no longer how to stay consistent, which has always been the hard part. The question becomes what the movement is producing.

There are three possible answers, and each should be tried for real.

Movement that produces material. A walk can produce something that lasts, and then the walk is also a way of collecting. The streets walked are marked on a map, and the things noticed on those streets go into the travel notes. When living abroad, the daily walk counts as fieldwork as soon as anything is written down. This setup also creates a useful incentive: an unexplored street adds more to the map than a familiar one, so the project encourages new routes, and new routes are also what keep walking fun. A session’s artifact is concrete: a street that has now been walked, or a scene that has now been written down. A street marked on the map stays marked.

Movement that produces data to work on together. Part of the reason the AI days are appealing is having company: working together with the AI is fun. That collaboration happens at the desk, and the walking happens alone. If the sessions produce something to bring back to the desk, such as routes, times, or whatever turns out to be interesting, then the movement gives the collaboration material to work on, and the work at the desk shapes the next sessions: patterns are found in the data, the next experiments are planned, and the numbers turn into a small shared project. Which data it is matters less than having this cycle at all. The habit to build is bringing something back from each session.

Movement as a tech tree. Physical capacity is drawn as a diagram that grows, with nodes such as the first unbroken hour, the first hill route, and the first weighted carry, and with new branches that open from earlier branches. Unlike a ring, which shows only how far short of today’s target the day is, a tree shows only what has been achieved. A node stays on the tree after a week with no training. The tree stays complete until the next node is added.

Working at the computer with the AI feels productive and good. A workout does not feel good afterwards. It makes me sweaty, so there is more laundry and more showering. Training feels more like a hassle than anything else. A session costs less when it is calm cardio. Calm cardio keeps the heart rate low, does not cause sweating, and lowers stress. Without the sweat, the session adds nothing to the laundry or the showering. Walking, walking around in a pool, and paddle boarding are examples. Hard, sweaty cardio is not needed to get lean. The rules on why to train, and on keeping training low in effort and sweat, are in Fitness Mindsets.

The walk that fills a map takes a concrete form as a walk every day, which is calm cardio. Each walk adds its streets to the map. Walking every day costs no mental energy. It is good for the joints and the mood. It does not raise hunger the way running does.

The tech tree takes a concrete form as a training log. The log records the sets, reps and weight of every exercise. Each session tries to beat the one before. After a month off, a dated best still means something, unlike a closed ring or a chain of days in a row. A sport alongside the gym, such as a combat sport, fits the tree too. The sport and the gym each teach movements the other one uses. Climbing, skating, dance and martial arts also fit, because the skill in each one keeps growing. The group of people around each of those sports provides the structure. Without the other people, starts get slower.

The test for each of these options is whether the sessions become something wanted. Completing the sessions out of obligation does not pass the test. After three weeks, the question to answer honestly is whether there is now a wish to do the sessions, and whether a skipped day brings the same specific regret, in this case about a part of the map that did not get filled in. If it does not, the artifact did not motivate anything, and the fix is to try a different artifact. More discipline is not the fix. This matters because treating the problem as a discipline problem counts every failure as a failure of willpower. A second risk is harder to notice: any of these options can turn back into a number that can be failed. If checking the map starts to feel like an obligation to the map, the option has turned back into a maintenance metric like a streak, even though it is still set up as a project.

All three options rest on the same assumption: when each session adds to something that grows, doing the session is its own reward, and a missed day needs no penalty. Whichever option is tried first, the problem is already being treated as a design problem instead of a discipline problem. As a discipline problem, the habit failed in a predictable way. The AI work never needed this kind of design. A habit that is motivated by adding to something that grows does not need to be protected, so the trackers, the streaks, and the rings are no longer needed.