The standard advice — break it into steps, prioritize, review weekly — is not wrong. It is just written for someone whose problem is not knowing what to do. If your problem is that you know exactly what to do and still cannot start, or that you built a beautiful system in March and have not opened it since April, that advice does not reach you.
What follows is a task system built around a different assumption: that you will abandon it periodically, and the important property is not how good it is when you are on top of things but how cheap it is to come back to.
The three failures, and what each one needs
Capture fails. The thought arrives while you are doing something else, and by the time you find the right project and the right label, it is gone. Fix: capture must cost nothing. No decisions at the point of capture.
Retrieval fails. You wrote it down, filed it correctly, and never saw it again. Out of sight really is out of mind, and a well-organized archive is a very tidy way of throwing things away. Fix: today's work has to be on screen without being asked for.
Return fails. You dropped the system for eleven days. Now it is full of overdue items that all look like accusations, so you start a new one. Fix: coming back has to be a two-minute operation, not a reckoning.
Almost every ADHD task-management problem is one of those three. Notice that only the first is about capture — which is where most advice stops.
Rule 1: Capture with zero decisions
One inbox. Everything goes in it, unsorted. Not "the right project", not "the right label" — the inbox. Sorting at capture time is the tax that kills capture, because the thought you are trying to save is competing with the thought you are currently having, and it always loses.
Practically:
- Use the fastest input you have. A quick-add shortcut, a widget, a voice memo, whatever is already open. Speed beats structure every single time.
- Write it badly. "email r about thing" is a fine task. Rewriting it later takes ten seconds; not capturing it costs you the whole item.
- Never file at capture. Filing is a separate activity that happens later or never, and "never" is survivable.
An inbox with sixty unsorted items in it is not a failure state. It is a captured list, which is strictly better than a forgotten one.
Rule 2: Only today's list is allowed to be visible
The master list is for storage, not for deciding. Deciding happens against a much smaller thing.
The move is to have one screen that answers "what now" without any filtering: overdue at the top, then what is due today, then habits, then the calendar. If your tool cannot produce that view without you building it, that is a real cost — building the view is a thing you have to remember to do.
Rules for the today list:
- Three real items, maximum. Not three categories, three tasks. Anything beyond that is optimism.
- Everything else is invisible. If it is not due today, you should not be looking at it.
- A date is a promise to see it, not to do it. Dates are not deadlines; they are a scheduling mechanism for your future attention. Moving something to tomorrow is a legitimate use of the system, not a moral failure.
This is where the EvyOS Today view is deliberate: it opens on overdue-first and there is no triage step in front of it. But the principle is tool-independent. Whatever you use, the first screen has to be a short list of today, not an inbox of everything.
Rule 3: Attach tasks to something bigger, but only lightly
Fully unstructured task lists lose the thread — you cannot tell why any given item exists, so nothing feels worth doing. Heavily structured systems collapse under maintenance.
The middle: each task belongs to a project, each project belongs to a goal, and that is the entire hierarchy. Two levels. No labels, no contexts, no energy levels, no custom fields. Every additional dimension is another decision at capture time, and we have already established that decisions at capture time are where capture dies.
The point of the hierarchy is not organization. It is that when you open a project after two weeks away, the plan is still there and you are picking it up rather than reconstructing it.
Rule 4: Make starting smaller than avoiding
"Write the report" is not a task. It is a project wearing a task's clothing, and it will sit at the top of your list for three weeks radiating dread.
The reliable trick is to write down the first physical action, at a level of specificity that feels almost insulting:
- Not "write report" but "open last quarter's report and copy the headings into a new document".
- Not "do taxes" but "find the folder with the receipts".
- Not "start running" but "put the shoes by the door".
If the first action still feels heavy, it is not the first action. Go smaller. The test is whether you could do it while annoyed, which is the state you will actually be in.
Rule 5: Design for the return, not the streak
You will drop the system. Everyone does — the difference is what the system looks like when you come back.
A system designed for return has three properties:
Overdue is a state, not a punishment. Overdue items should be visible and separated, so you can see them without wading through everything else. Then reschedule the whole block in one pass. Do not read them individually and feel something about each one.
Nothing is lost, only late. If the tool deletes, hides, or "archives" things you missed, you cannot trust it, and a task system you cannot trust is worse than none, because it costs you effort and returns uncertainty.
Two-minute rebuild. Coming back should be: clear the overdue block, pick three things for today, close it. If restarting requires a full review, you will not restart.
That last point is worth sitting with. Most productivity systems are optimized for the version of you who is doing well. The one that actually helps is optimized for the version of you who has not opened it since the eleventh.
Rule 6: Review, but make it small and boring
The weekly review is real, and it is also the first thing to go. Shrink it until it survives:
- Ten minutes, same time each week, attached to something you already do — Sunday coffee, Friday afternoon, whatever exists already.
- Three questions only. What is overdue and can be dropped entirely? What is due next week? What is one thing that would matter most?
- Drop things out loud. A task list you never delete from becomes a museum of intentions and gets ignored wholesale. Deleting is maintenance, not defeat.
If ten minutes is too much, do three. A bad review that happens beats a thorough one that does not.
What a working week actually looks like
Monday you capture eleven things, do four, and move three to Tuesday. Tuesday goes well. Wednesday you do not open the app at all. Thursday you clear the overdue block in one pass, pick two things, and do both. Friday you capture six more and do one.
That is not a system failing. That is what a working system looks like from the inside — the failure would be Wednesday deleting Monday's work, or Thursday requiring an hour of triage before you could start.
Judge your setup by that pattern, not by whether you had a perfect week. Nobody has a perfect week; the people who look like they do just have systems that hold on to work while they are not looking.