From my experience most people seem to have their wall at which point they can't possibly conceive of a better way to scale X until someone comes along and goes "Try this". My experience also says this level is pretty low for most people.
Google :)
> The world is absolutely littered with fantastic clever well written programs, that no one uses.
The world is also littered with twitter/facebook/youtube clones that no one uses.
What sets apart the wannabes from the successful ones is (among other things) the implementation.
I find much joy in doing clever things. If those things end-up being used by other people then I literary have an orgasm. But doing something simple that happens to end-up popular because of happening at the right time ... that's really a turn-off (look honey, this website of mine got really popular so now I have to maintain it ... oh joy). Fortunately that happens less and less because the low-hanging fruits are already taken.
The hard bit of course would be getting people to use it.
The world is absolutely littered with fantastic clever well written programs, that no one uses. This is why I think programming is far easier than convincing someone to use it and all that other non-programming stuff, and why I personally am far more interested in the non-technical side of the equation. In comparison, the technical side is trivial.