Homeostasis

Every productivity tool I have used can hold a list. None of them can answer the only question that matters at eight in the morning: of everything I could do today, which thing actually moves the rest of it?

That question is not about discipline. It is a dependency problem. Tasks have prerequisites, costs, energy requirements and downstream consequences, and a list flattens all of that into a row of identical checkboxes.

Homeostasis keeps the structure. It knows what blocks what, what today's work unlocks in six months, and where the plan and the person have quietly diverged.

Task software assumes the hard part is remembering. It is not. The hard part is ordering — knowing that this week's boring administrative task is the gate on something you actually care about, and that the ambitious thing you want to start cannot begin for another four months regardless of how motivated you feel today.

Strip that structure out and every task looks equally urgent, which is the same as none of them being urgent. So you work on whatever is loudest, and the plan quietly stops matching the life.

The system makes strong decisions inside rules you set. You do not micromanage it and it does not micromanage you: you write the constitution, it does the arithmetic.

It must always be able to explain itself. If it puts something at the top of your day, it can show the dependency chain, the deadline pressure and the opportunity cost that put it there. A recommendation you cannot interrogate is just an opinion with a progress bar.

It must shrink when you slip. A system that greets a bad week with thirty unchecked boxes is not a system, it is a debt collector. Adherence drops, the daily load drops with it, and the plan re-forms around where you actually are.

Each task carries its prerequisites, expected duration, energy cost, location, financial requirement and what it unlocks. Goals carry their cost, their skill requirements and their gates. The scheduler resolves that graph against the time and capacity you actually have.

It then forecasts twice. The realistic forecast uses your demonstrated consistency. The ideal forecast uses near-perfect execution. Both are shown, because the distance between them is the single most useful number in the system — it is the price of the gap, in months, and it is either worth closing or it is not.

In active development and in daily use by exactly one person. The scheduling core and the dependency model work; the forecasting layer is being rebuilt around real adherence data rather than estimates.

It will be opened up when it survives ordinary use without becoming a second job. That is the bar, and it has not cleared it yet.