What does it mean that externalize decisions not just information?

The most valuable thing to capture is why you chose what you chose. Decisions decay faster than facts — and unlike facts, they cannot be reconstructed after the outcome is known.

Example: A team lead chooses PostgreSQL over DynamoDB for a new service. Six months later, the service struggles under write-heavy load. Everyone says 'obviously we should have used DynamoDB.' But the original decision was made because the team had zero DynamoDB experience, the deadline was eight weeks out, and the data model was deeply relational. Without a written record of those constraints, the team learns nothing — they just blame the decision-maker. With a decision log, they can ask the right question: 'Given what we knew then, was the process sound? What new information would change the answer?'

Try this: Pick one decision you made in the last week — what to work on, which tool to use, whether to attend a meeting, anything. Write a five-line decision record: (1) What you decided. (2) What alternatives you considered. (3) What information you had. (4) What you were optimizing for. (5) What would make you revisit this decision. Do it now, not later. You will be surprised how much of your reasoning has already faded.