You know the feeling. You close the document, answer the message, and come back. The file is exactly where you left it. Nothing has moved. And yet you sit there for a minute and a half reading your own last paragraph, trying to remember what you were about to do next.
That minute and a half is the whole problem.
What actually gets interrupted
The popular framing is that interruptions break concentration. That is true but not very useful, because it suggests the fix is willpower: concentrate harder, want it more. The more precise version is that interruptions destroy a working set.
When you are deep in a piece of work you are holding a small, private structure in your head. The three constraints that matter. The two things you already tried. The reason you rejected the obvious approach. None of this is written down, because writing it down would take longer than holding it.
An interruption does not pause that structure. It deletes it. Coming back is not resuming, it is rebuilding.
The work was never lost. The map you had of the work was.
Why the number is larger than it feels
Ask someone how long it takes them to get back into something and they will usually say a minute or two. Watch them and it is closer to ten, because recovery is not one event. There is the initial reorientation, then a period where you are working but working badly, making decisions you will revisit later because you have not yet reloaded the reason you made the original one.
The second part is invisible from the inside. It does not feel like recovery, it feels like work. That is what makes it expensive.
Three things that help more than they should
- Leave yourself a sentence, not a note. Not a summary of the document. A single line about what you were about to do. “Was about to cut the second example, unsure if the argument survives it.” That sentence is the working set, compressed.
- Stop at a decision, not at a paragraph break. Ending mid-thought feels untidy and is much cheaper to resume than ending at a clean boundary, because the unfinished thought carries its own context with it.
- Separate capture from filing. The reason people do not write these notes is that writing them means choosing where they go. Remove the choosing and the writing happens.
The part software can take
Most of this is a memory problem, and memory problems are the one thing software is genuinely good at. Not by summarising your document, which you can already read, but by holding the small structural facts you would otherwise have to reconstruct: what changed since you last looked, what you had flagged as unresolved, what you said you would come back to.
That is the narrow job we gave Velora. Not to think for you. To hold the map while you are gone, and hand it back when you return.
It is a smaller ambition than most tools announce. It also happens to be the expensive part.