Three caveats to build something people want

I sat on a panel for the Anthropic Claude Community event in Santa Barbara. I shared with the builders in the room the messiness involved and mistakes I've made getting to PMF with Falconer.

I concur that YC's "Build something people want" is still the best axiom to live by. However, I have some important caveats to add:

  1. Build something you want and would kill to have. Otherwise the constant ups and downs will crush you instead of energizing you.
  2. Build something people will pay for, otherwise you're a hobby shop, not a business.
  3. Build something that other people hate building. Otherwise you'll hear, "We can just build this ourselves" as a constant objection.

It may seem like a scary time to build, but go look at your suite of tools and list of problems. How many of your tools are A+? How many of your problems are solved? There's plenty of stuff still to build.