[jamierumbelow]

All good points, thank you - your distinction between iterate-to-build vs iterate-to-ship is an important one, and something that was in the back of my mind when I was writing. I actually had a section on Stripe that argued that Stripe "do one thing extremely well" rather than "do lots of things poorly", but it didn't have a clear instance of a major technological breakthrough, so I cut it.

The goal here was not to argue that one OUGHT to take ages to ship - I hope I made that clear in the final section – but that there are certain types of ambitious technological visions that NEED a lot of technically risky work to achieve. And so many people shy away from these visions because they feel like they need to ship immediately!