The expensive part of capturing a mid-work thought is not writing it down—it is the interrogation that follows: Which notebook? What title? Which tags? Every one of those questions runs on the same attention you were using for the actual work. Answer them all, and the thought is saved but the focus is gone.
The interruption is the decisions, not the note
Ten honest words take a few seconds and barely touch your concentration. Choosing a location is different: it is a context switch, and returning from a context switch to deep focus takes minutes, not seconds. A capture tool that demands filing decisions charges you the most expensive currency you have—mid-task attention.
Capture must ask you nothing
One inbox, any format, zero required fields. Title, folder, and tags are all deferrable; the only thing that cannot wait is the thought itself. Type the fragment or say it, and go back to work.
Deferred structure is not lost structure
The system does the clerical pass after the fact:
- Transcription and text extraction run without you.
- Topics are suggested, not demanded.
- Anything that looks like a commitment is surfaced for confirmation.
Clarification finds a better moment
If the note is ambiguous—‘the second option,’ ‘ask her before Friday’—a short question arrives later, when answering costs you nothing. The same question at capture time would have cost some of the focus you were protecting.
Why the decision costs more than the typing
The asymmetry is easy to underestimate. Writing ten words is motor work you can do while still holding the problem in your head. Choosing a notebook is not: it requires you to load a second structure—your filing system—evaluate the note against it, and commit. That is a different kind of thinking, and it evicts what you were working on.
Recovering from that eviction is the expensive part, and it is slow. Getting back to the state where the problem is fully loaded again takes minutes, not the seconds the interruption appeared to cost. Do it four times in an afternoon and the capture tool has consumed more of your day than the work you interrupted it for.
This is also why ‘just be more organized at capture time’ is bad advice for anyone doing concentrated work. The discipline it asks for is spent from exactly the account you are trying to protect.
What deferring actually costs
An honest account has to include the other side. Capturing without decisions produces an unstructured stream, and a stream nobody ever looks at is a landfill with good intentions attached. Something has to happen afterward or the friction was merely moved rather than removed.
The realistic answer is that ‘afterward’ should mean automatic processing plus a small human pass. Transcription, extraction, and suggestions cost you nothing. What remains is answering the occasional clarification question and confirming the occasional detected commitment—a few minutes a week, not an inbox ritual.
There is also a class of note where deferring is wrong: if the thought contains something you need within the hour—a number to call, an address to arrive at—stopping to make it usable now is correct, because the moment it was captured for arrives before any processing would. And a small number of captures will be unreadable later no matter what, because four words typed while your attention was elsewhere sometimes fail to mean anything by evening. That is the accepted cost of protecting focus, and a much smaller loss than the notes never taken at all.
Protect the focus first and the thought second. A capture flow that asks zero questions at save time costs seconds; every question it does ask costs the concentration you were trying to keep.