I don't disagree with anything PB says. I say to do hard things if they benefit the user proportionately. He says not to do hard things just because they're impressive. Since in practice impressing people (including oneself) is the main reason people do hard things that don't benefit users, these two principles yield the same advice.
I suppose I took PB's post more broadly than it was intended, then. If he was only talking about solving hard problems that don't benefit the user proportionally, I see that there's no contradiction [and have a clear answer to the question of when to follow which advice ;-)].
