Jul 27, 2026
Some stubborn attachments on getting things done:
1. If you want a thing to happen it needs one owner and it’s gotta be their main job.
2. People can either lead or follow (and very rarely switch). Both are very valuable but in different ways.
3. If you ask a leader to follow or a follower to lead you’re setting them up to fail.
4. You can only do as many things at once as you have leaders. It’s very tempting to try and do more, but it never works out.
5. Every area should have a leader and a team of followers built around them.
6. For special projects you can have a killer team with multiple leaders (so long as they can set their egos aside).
7. First deliverable is a two page design doc and a demo. As quick and dirty and throwaway as possible. Like get it done in a day or two fast. For really complicated stuff maybe a week.
8. Intelligence and conscientiousness are reasonably common, but you need to select for people who have both. This is rare and extremely valuable.
9. It’s much easier to lead a world-class org with super ambitious goals than an average org with average goals. You’d expect it to be harder, but no.
10. As an aside, you should just let engineers use AI on interviews. I know it seems scary but I tested it and quality of prompts and engineering skill is r=1. You won’t lose any signal.
11. Fully remote doesn’t work, sorry.
12. Minimize cross-geo collaboration. Kinda like you’d minimize cross-region network calls.