Thinking Tools
Eight ideas run through almost everything on this site. None of them are original — they come from decision science, forecasting research, and the older tradition of keeping a commonplace book. What is original is the insistence on writing things down rather than thinking about them.
Slow thinking
Most opinions arrive fully formed in under a second and then defend themselves. Slow thinking is the deliberate habit of separating the first reaction from the second thought — writing the reaction down, then asking what evidence would change it.
Digital garden
A feed is built for consumption: posts stream past, you glance, you scroll, you forget. A garden is built for tending: ideas are planted, watered, revised in public, and allowed to grow across years. The difference is not technical, it is about how you treat thinking.
Prediction journal
A forecast that stays in your head cannot be scored. Written down with a probability and a date, it becomes evidence about your own judgment. The point is not to be right, it is to find out where you are reliably wrong.
Brier score
A single number that grades probabilistic forecasts by punishing confident mistakes harder than uncertain ones. It turns “I had a feeling” into something that can be measured, compared, and improved.
Calibration
Being right 70 percent of the time when you say you are 70 percent sure. Confidence and accuracy are different skills, and most people are well calibrated in the middle of the range and badly overconfident at the edges.
Base rates
Before asking whether this particular project will finish on time, ask how often projects like it finish on time. The outside view is boring, statistically superior to intuition, and almost always skipped.
Open questions
Some questions are worth holding without resolving. Writing them down and revisiting them beats forcing a premature answer — and beats forgetting you ever wondered.
Epistemic humility
Not the performance of uncertainty, but the operational version: stating what would change your mind, keeping the receipts when you were wrong, and updating before you are forced to.
Where to start
If you only do one thing, write down a prediction with a number and a date. Everything else on this site follows from that habit.