I Quit Every System I Ever Loved

I have paid for eleven task managers in six years. Notion twice, because I assumed the first attempt had been the problem. Every one of them worked beautifully for about three weeks, then I found a better one, and I called that research.

The pattern stayed invisible because each switch had a real reason attached. The tags were wrong. The mobile app was slow. Recurring tasks were handled stupidly, and they genuinely were.

What I never looked at was the timing. Every reason arrived in week three. Week three was always the same week: the one where the novelty burns off and the system starts showing me the backlog I had been hauling between applications like luggage.

The migration weekend is the tell. Nothing on earth feels more like work than moving four hundred tasks between two apps, and nothing produces less.

Call it system tourism. The diagnostic: count the tools you have moved into in the last two years, then count the projects that started and finished inside one of them. If the first number is larger, the tool was never the variable.

What broke it

I lifted for four years and got almost nothing for it. Same behavior, different room. Every time a lift stalled I found a better program, and every better program opened with a deload week, so I spent four years permanently in week one.

The rule I took came from a friend who has run the same five lifts since 2019 and is stronger than everyone I know who reads about training. He put it plainly: the adaptation lives in the weeks rather than in the design, and a mediocre program run for a year beats an excellent program run for three weeks every single time.

The stall was never a signal to change the program. The stall was the program finally asking me for something, and I had been answering by leaving.

What I run now

One system. Twelve weeks. No changes, including changes that are obviously improvements, because obviously improvements are what week three produces.

Every complaint goes into a note called grievances, with the date. Grievances get read at week twelve and nowhere else. That file does most of the work, because the urge in week three is not really to fix the tool. The urge is the specific relief of a fresh empty system, and the note gives it somewhere to go that costs nine seconds instead of a weekend. Twelve weeks of a mediocre system beats three weeks of a perfect one, and the mediocre one is the one you actually fucking finish.

Results so far

I have been in the same setup for thirty one weeks. The grievance file had nineteen entries by week twelve. Two of them were real: a filter I did not know how to build, which took nine minutes once I read the documentation, and a genuine limitation I worked around with a second list. The other seventeen were week three talking, and reading them in one sitting was the most persuasive argument against myself I have ever assembled.

Two projects have finished inside it. That is a low number, and it is two more than the previous six years produced.

Eleven task managers, and I can still describe the tag model of every one of them. What I cannot describe is anything I finished in one, because the answer is nothing, and it was nothing for a reason I kept installing over. Open whichever one you have now. It is fine. Stay in it until it asks you for something, and then stay in it. [1]

[1] Some tools really are bad. Twelve weeks is how you earn the standing to say so, and the grievance file is the evidence you say it with. Quitting in week three is not a judgment about the tool, because by week three you have not gathered one.