openskills.info

Resilience & Focus

Focus and Deep Work

Protecting long enough stretches of attention to do the work that actually requires thinking.

Some work is genuinely improved by being done in one uninterrupted stretch — anything where you have to hold a lot in your head at once, which is most design, most writing, most debugging of hard problems. That kind of work has a warm-up cost, and every interruption charges it again. Research on attention residue shows the cost isn't just the three minutes of the interruption; part of your attention stays with the previous task, so the recovery is slower than it feels. The practical implication isn't heroic isolation. It's that a few protected blocks per week produce more of this kind of work than a permanently available forty hours, and that protecting them is a scheduling decision rather than a matter of willpower.

Scenarios

The one hard task on your list has been carried forward for over a week while everything else gets done. What would you do — try harder to get to it, or block two hours tomorrow and turn off notifications for them?

You're on a call that only needs half your attention. What would you do — use the time for a code review, or decide whether you need to be on the call at all?

You need an answer a colleague has and could either search for it or ask them. What would you do — ask, since it's faster for you, or spend ten minutes searching first and batch anything left for one conversation?

Takeaway

  • Schedule protected blocks rather than hoping for them. Focus is a calendar problem before it's a discipline problem.
  • Turn notifications off for the block. Available-with-good-intentions is the same as interrupted.
  • Don't pair demanding work with anything. You will do both at about sixty percent and attribute the result to the tasks.
  • Start with the ninety minutes you have instead of the four hours you don't. The perfect conditions are a way of not beginning.
  • Batch your questions for other people. "Quick question" is quick for exactly one of the two participants.