// track · 9 lessons
Start simple
Complexity is a loan. Only take it when the simple version has actually failed you.
- 01Start simpleThe complexity you add is a loan. Only take it when the simple version has actually failed you.
- 02Complexity is a loanComplexity is borrowed, and you pay the interest forever, so add it only when it's paid for itself.
- 03When do you actually need an agent?Most tasks that feel like they need an agent need a good model call. Start with the call and graduate only when the structure demands it.
- 04YAGNI, or building for the load you actually haveBuild for the load you have, not the one you imagine, because speculative complexity rots and the real requirement rarely matches the guess.
- 0580% shipped beats 100% unshippedA hundred percent of unshipped is zero. Shipping is where the learning lives, so get the useful version out and let reality set the last twenty.
- 06Ship or open-source, because hoarding is worth zeroAn unused project is worth zero whether you protect it or not, so ship it or open-source it, because execution proof in public beats a private idea every time.
- 07The over-engineering reframe, or thoroughness versus avoidanceThoroughness on the invariants is strength; on the implementation it's avoidance. The tell is whether you can name the threat it prevents.
- 08Income before toolsBuilding tools is the most comfortable way to avoid building income, so run the thirty-day-money test and let the thing that pays go first.
- 09Cutting scope without cutting soulCutting scope isn't cutting quality, it's concentrating it on the one thing that can't fail, so find the soul, protect it, and let almost everything else go.