Reclaiming the
machinery of attention

ChronoZero starts from an observation about social media. Its platforms are the most effective behavior-shaping systems ever deployed at global scale. No school, calendar, productivity tool or public institution has shaped daily behavior with anything near their precision.

Entertainment explains little of that. Platforms got there by building a loop: observe a person, estimate their internal state, choose an intervention, measure the response, and adjust the next one. Every swipe and every return counts as evidence. The system keeps learning what a person is susceptible to at a given moment, and then it acts on that prediction.

None of the parts are harmful on their own. Pattern recognition, behavioral psychology and well-timed prompts are neutral techniques, and the harm comes from what they are pointed at. A platform uses what it learns about you to keep you on the platform longer and in front of more advertising, because that is where its revenue comes from. Whether the loop serves you or traps you depends on whose interests set its reward function.

The inversion

ChronoZero points the same kinds of technology the other way. The goal is to help a person protect their attention, recover at the right time, and keep making progress on work and learning that still matter after today. We keep the loop platforms perfected and give its reward function to the person it runs on.

That changes what success looks like. A platform does well when you come back compulsively; ChronoZero does well when you close it. Often the right outcome for us is the one a platform wants to avoid: one useful recommendation, a clearer picture of the next few hours, a decision you commit to, and then the app gets out of the way. It is a control panel for the rest of your day and was never meant to be where you spend it.

Autonomy arrived before the structure for it

The second half of the problem has nothing to do with platforms. In the past few years a large share of office work moved into rooms where nobody sees when you start. Remote and hybrid jobs, flexible hours and teams judged on output instead of presence have given many people near-total control over the shape of their own day. The shift doesn't look temporary, and the work most likely to stay valuable is the work nobody can supervise by the hour.

Autonomy is a real gain, and it comes with a cost. The office used to supply structure whether you wanted it or not. A commute marked the start of the day, colleagues set an implicit pace, and the building made it obvious when work had ended. Take that away and each of those decisions becomes yours, all day, on top of the work itself. Procrastination, work that drifts into the evening, and days that end without a clear sense of where they went are what unstructured freedom predictably produces. They are not character defects.

Students get the sharpest version of this. School hands you a timetable. University, or any self-directed study, hands you a deadline several months away and all the time in between. The freedom arrives at once, usually before anyone has learned to plan around their own energy, and a badly spent week stays invisible until the exam.

This is the situation ChronoZero is built for. The scarce resource here is structure. It used to come from outside, and now a person has to rebuild it every day while unable to see their own state very clearly; extra hours or more willpower don't replace it. A system that estimates capacity and proposes a shape for the day can give that structure back without handing it to an employer, a streak counter, or a rigid template that ignores how you are today.

The harder road: sparse signals

Platforms work with abundant data. Their interfaces produce a near-continuous stream of behavior, and the reward shows up within seconds: you kept watching or you left. Optimizing under those conditions is comparatively easy.

ChronoZero turns that setup down on purpose. A system meant to send you away from the screen can't also ask you to tap, rate and monitor yourself all day. So the signal is sparse. The model sees occasional behavior blocks and the fixed points in your calendar. It gets the Vitality slider, an optional 45-second readiness check, and whether you followed a recommendation. You can write in the Journal, and nothing writes back. The reward is slow as well. A break that looks unproductive at 15:00 may be the reason the evening holds up, and a focused session can pay off weeks later.

Sparse observations with delayed reward make the learning problem much harder, and accepting that is an ethical position expressed as an engineering constraint. Platforms reduce uncertainty by collecting more of your behavior. ChronoZero reduces it with structure. A simulated physiology carries a plausible picture of you through the gaps, the estimate keeps its uncertainty in view, and the planner looks at consequences hours ahead. We try to infer more from less, because asking for more would turn us into the thing we're arguing against.

Stewardship instead of extraction

Underneath all of this are two ways of treating human focus. Extraction treats it as something to capture, fragment and sell, so a platform profits when you come back often and struggle to leave. Stewardship treats focus as a limited resource that belongs to you. A system built on that idea is useful when you can direct your attention on purpose, protect it from interruptions, recover it when it runs low, and spend it on things that last.

The difference shows up in specific decisions. A platform optimizes the next interaction, while ChronoZero has to optimize the course of the whole day. A platform that notices fatigue serves easier, more stimulating content. When ChronoZero estimates fatigue, it lowers what the plan asks of you, schedules recovery, or tries to head off a distraction spiral it can see coming. Platforms benefit when you return compulsively. We benefit when you leave with something to do.

What we refuse

The inversion only holds if the product proves it in its details, so ChronoZero keeps a short list of refusals that work as design rules. There are no streaks and no badges. Nothing in the product uses guilt, and recovery is never described as falling behind. There are no dark patterns, no manufactured urgency, and no endless feed dressed up as self-improvement. Distraction is logged without penalty. Getting pulled in by systems built with billion-dollar reinforcement loops is not a personal failing, and we won't pretend it is. Statistics are there for orientation and never as a score to chase. Every recommendation shows its reasoning and can be overridden, corrected or ignored.

The same humility applies to the model. ChronoZero doesn't assume it knows better than you, and your behavior and your own sense of how you feel are part of the ground truth. A slip, a rejected recommendation and an unplanned break all tell the model something. The system should learn from the person more than it trains them.

The wager

ChronoZero is an attempt at a new kind of personal system: a closed loop built around the fact that energy, focus and fatigue change through the day, with the reward function owned by the individual. The predictive machinery that keeps people inside platforms today could instead help them get back to their work and the rest of their lives.

The next generation of personal technology should know more than what captures our attention. It should understand what our attention can do, notice when it needs protecting or rest, and help us point it at the futures we choose for ourselves.

Already approved? Sign in