A short list. Revised often.

A short list of things I try to operate by. The earlier version of this page sounded much more certain than I actually am about any of it, which was the first thing worth fixing.

Time is the input that matters

Capital can be raised, people can be hired, most mistakes can be corrected. Time can't. Most of the decisions I've gotten wrong were trades where I exchanged years for something I could have bought cheaper elsewhere — money, status, comfort, the appearance of progress. The discipline is to keep noticing when that trade is happening, and to correct it without turning the correction itself into a performance.

Impact has to be observable

A useful test for the work I want to spend my own time on: is someone meaningfully better off because it exists, and is the change large enough to measure if you actually looked. Most things fail this test, which is fine. Not everything has to be impactful. But this is the bar I want to hold the things with my name on them to.

Originality is a hedge, not an aesthetic

If you're solving a problem the same way as everyone else, your upside is capped by the field. The interesting upside lives slightly off-consensus. I'd rather be wrong in a direction nobody is looking than right in one everyone is crowded into. This isn't about disagreeing for the sake of disagreeing — most off-consensus takes are wrong, including most of mine. It's about being willing to hold the unusual answer long enough to find out.

Hold beliefs lightly, including these

The version of this page from two years ago contained a lot of certainty I no longer have. The version from two years out will look at this one the same way. That's the system working.