Thoughts
You have to design for 100% of users. It’s possible this is what Theil is getting at with his argument that you should try to be a monopoly.
If you solve a problem really well for 10% of your possible users, I don’t believe that works, because you can’t get any momentum or any network effects. This is a subtle point because you do have to pick a niche. But it’s better to do that by solving a problem for 100% of people with that problem, rather than solving a common problem for a subset of people (e.g. “people on a budget” or “people running Linux”). In Theil’s words, you want a big slice of a small market, not a small slice of a big market.
This will effect, for example, how I design my Gemini client in the future.