I finally wrote something about my time at sail/muddy, the last startup I was at, where we were trying to build a multiplayer browser, and a few lessons that stayed with me.

I mostly just hope it’s interesting to people thinking about new ambitious interfaces right now. with AI.

This line was golden: Sometimes we were iterating on the vision when we should have been iterating closer to user signal. The difference is subtle but it matters a lot. One converges toward something people want. The other converges toward a more elegant version of something people don't.

I’m building a way for people to build on top of the browser (https://www.hyper-frame.art). One take away for me from your article was GTM is a bigger moat than technicals (which are brutal in the fork route you went)