Context switching can be expensive for a reason that is often skipped over: part of the cost comes from the unfinished state you leave behind. Switching away from something half-done means carrying it, and carrying it can make the return trip slow.

Which means there are two very different kinds of interruption, and only one of them is worth defending against with focus rituals.

Two kinds of interruption

External. Someone messages you, a meeting starts, a notification arrives. These are the ones every productivity guide addresses, and the standard advice (turn things off, batch communication, block time) is broadly correct.

Internal. You are working, and a thought arrives about something else entirely. You need to renew the domain. You should check whether the migration script handles empty rows. That book someone mentioned.

Internal interruptions are hard to block because there is no setting for them. You cannot mute yourself.

Why a thought is more expensive than it looks

An arriving thought has three possible endings, and two of them are bad.

Act on it. You switch tasks. The original work goes into a half-finished state you will have to rebuild, and rebuilding is the actual cost: reloading what you were doing, where you were, what you had ruled out.

Suppress it. You decide to keep going and not think about it. Now you are doing two jobs: the work, and the ongoing effort of not attending to the thing. It will come back, repeatedly, because nothing has changed about its status.

Park it. Write down enough to make the return clear, and go back to the work. The thought now has a safe place to wait.

The third avoids an immediate switch. Parking is only cheap if writing it down is genuinely simple. If it needs several steps and a decision about which list, it becomes a small version of switching, and then suppressing looks reasonable again.

Why parking works at all is worth a paragraph of its own; the Zeigarnik effect is the standard account of it.

Keep the detour short

The capture has to stay smaller than the interruption.

When capturing is one brief movement, it can feel like part of the current task instead of a new task. Add menus, filing choices, or a second app, and it starts to feel like leaving what you were doing. That is the switch you were trying to avoid.

So the design constraint on a capture tool for this purpose is unusually specific. It is not about features. It is about whether the whole interaction fits inside the gap in your attention, without opening an app, choosing a destination, or naming anything.

The same threshold explains why paper next to the keyboard works so well for so many people. It is very hard to beat on this metric.

What to write

Less than you think, but not nothing. Enough that you will not have to reconstruct it:

  • "Domain renewal: check expiry date"
  • "Migration script: empty rows?"
  • "Book M mentioned: ask her the title"

Then go back. Do not start solving it, do not tag it, do not decide whether it is important. Sorting is the review's job, and doing it now is the switch.

Where Capture fits

Capture exists at that threshold. One always-open field on iPhone, iPad, Mac, Apple Watch and Apple Vision Pro, with no list to pick and no form to fill first. On the Mac, a global keyboard shortcut opens a field over whatever application you are in and closes again on return, which is the version that matters for desk work.

It reads dates out of your own sentence, so "renew domain before the 12th" carries its own timing rather than needing a reminder set. Later, the sweep sends each item to where it belongs: Things, Apple Reminders, Todoist, Linear and Obsidian are among its 54 destinations.

One App Store reviewer, TeaMe79, described the design goal:

“This is what I’ve been searching for 2+ years for. Prioritizes simple idea capture with minimal friction and distractions. Does not try to be your “all in one productivity life system”.”

Measure the right thing

If you want to know whether this is your problem, spend one afternoon making a tally mark every time a thought about something else arrives while you are working.

Read the items back at the end of the afternoon and mark which ones were truly urgent. The gap between how loudly a thought announced itself and how much it mattered is the argument for parking rather than acting.