What the author says is that's it's good to have configuration, because then everyone can find what they want, but configuration alone is not enough. You need good defaults, and because "good" is subjective it means you need defaults that will please a specific category of users, and you need to go all-in on it because then your software will have its own identity. It also means that those who want to use the software another way can still do it because it's configurable.
He also says that your flexibility should be enough to fit the target audience, and not much more because you should focus on delighting that audience, not on broadening it.
I actually didn't like the article either. The means he pushes are known to not be very effective, and he makes a completely one sided analysis of a cost/benefit situation. But I think you are focusing too hard on the trees and missed the forest.