Stop relying on working memory.
Build a system that remembers context for you, reduces task-switching, and makes the next action painfully obvious.
Capture → Choose → Focus → Reset
Every distraction gets captured. Only one item gets execution rights.
Capture
Write it down immediately. Do not mentally hold future work.
Choose
Select exactly one active task.
Focus
Protect the task from Slack, tabs and rabbit holes.
Reset
Checkpoint state before changing context.
Use NOW / NEXT / LATER
The brain should execute tasks — not act as a scheduler.
Never write vague tasks.
Too abstract. Too many possible starting points.
Checkpoint before you switch.
Treat your brain like a process that can be interrupted. Persist state before losing the context.
Batch communication.
Do not turn Slack into an interrupt controller for your brain.
Reduce interrupts
Keep only urgent mentions, incidents and true on-call alerts.
Batch replies
Check Slack / email at planned intervals.
Keep escalation paths
Critical production alerts can still bypass focus mode.
Use short launch windows.
The hardest part is often starting. Lower the activation energy.
If hyperfocus is useful, don't kill it just because the timer ends. Use a 60–90 minute guardrail check instead.
Prevent rabbit holes.
Borrow external accountability.
Work silently with another person — same call, different task. The social presence can make starting and staying on-task easier.
Use one tiny operating system.
# TODAY ## 🎯 Main outcome Reduce PR pipeline runtime ## 🔥 NOW Analyze Gradle profile ## ➡️ NEXT Check test execution duration ## 📥 INBOX - review PR #184 - reply Slack - check Terraform warnings ## 🧠 Current context Unit tests = 42 min Build = 8 min
If you remember nothing else:
One active task only.
Write everything down.
Make the next action concrete.
Checkpoint before switching.
Batch communication.
Design the environment so focus is the default — not a heroic act.