Principles
A living document. These are the operating principles that have held up across domains and over time — extracted from experience, stress-tested through practice, revised when they stop being true.
Optimize for learning rate, not current knowledge
What you know depreciates. How fast you can learn compounds. Given the choice between depth in a known domain and the ability to navigate an unknown one, choose navigation.
Work at the level of principles, not procedures
Procedures are fragile — they break when context shifts. Principles transfer. Understand the constraints, incentives, and feedback loops underneath, and you can reconstruct the procedure for any new context.
Seek structural analogies across domains
The same patterns — feedback loops, diminishing returns, phase transitions, principal-agent problems — show up everywhere. Recognizing them across fields is where the deepest insight lives.
Test through small experiments, not big bets
Conviction is cheap. Evidence is expensive. Run small, reversible experiments before committing to large, irreversible ones. Let reality do the editing.
Hold models loosely
Every mental model is a simplification. Useful, but incomplete. The moment a model stops predicting well, update it or discard it. Identity should never be tied to a framework.
Push authority to where the information lives
Centralized decision-making creates bottlenecks and information loss. The people closest to the problem usually have the best signal. Give them the context and the authority to act on it.
Simplify relentlessly
Complexity is the default. Simplicity requires effort. Every system, explanation, and process should be as simple as it can be — but no simpler. If you can't explain it clearly, you don't understand it well enough.