Frequently asked questions about thinking, epistemology, and cognitive tools. 9738 answers
Add buffer to every estimate and use reference class forecasting.
Pick a project or task you are currently planning. Write down your gut estimate for how long it will take. Now find three to five comparable past projects — yours or others' — and record how long each actually took from start to finish. Calculate the average actual duration and compare it to your.
Turning countermeasures into cynicism. The point is not to assume everything will be a disaster and pad every estimate with so much buffer that you never commit to anything ambitious. The point is to replace optimistic fiction with calibrated realism. Over-buffering is its own failure mode — it.
Add buffer to every estimate and use reference class forecasting.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
Open your current task list, project board, or backlog. Go through every item and estimate — honestly, without rounding up — how long each one would actually take to complete. Mark every item that falls under two minutes. Now set a timer and dispatch them all, one after another, without pausing to.
Treating the two-minute rule as a license to handle every incoming task the moment it appears, regardless of what you are currently doing. The rule is an administrative-time heuristic, not a blanket interrupt policy. If you apply it during deep work, you destroy the maker-time blocks you built in.
If a task takes less than two minutes do it immediately rather than scheduling it — because the overhead of capturing, organizing, and tracking it exceeds the cost of doing it now.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Identify three to five categories of small, recurring tasks that you currently handle as they arrive — email, Slack messages, administrative approvals, filing, errands, phone calls, social media responses. For one full week, instead of handling them on arrival, capture each one into a simple list.
The most dangerous failure mode is over-batching: grouping so many tasks into such infrequent blocks that urgent items wait too long and relationships suffer from delayed responsiveness. Batching email once per day sounds efficient until a client sends a time-sensitive request at 9:05 AM and does.
Group similar small tasks together and process them in one dedicated block, so that setup costs are paid once instead of once per task.
Every meeting needs a purpose, an agenda, a time limit, and clear outputs — and most meetings fail not because they exist but because they lack these structural elements.
Every meeting needs a purpose, an agenda, a time limit, and clear outputs — and most meetings fail not because they exist but because they lack these structural elements.