Adding capabilities hasn't really resulted in much. A bunch of titles appear that include video or Fancy New Widget™ but most of them are either unpolished or badly thought out. None of them play a major role in the sales charts.
What else do you expect when we are given poor documentation and no tools? What else do you expect when you can't even trust basic CSS formatting to be preserved across all of the major platforms?
What else do you expect when we have to implement fixed layout books at least twice?
Time that is wasted working blind, testing across half a dozen platforms with no tools to help you, and reimplementing the same fixed layout again and again, is time that doesn't go into product development, doesn't go into polishing the book, doesn't go into play-testing widgets, and doesn't go into turning a half-baked design into something fantastic that will blow the reader's mind.
News flash: Incompatible platform capabilities are a crap way to differentiate your product.
Polish, however, is a great way to differentiate, to rise above the rest. If your implementation is the best, that's a differentiator. Having the capability in the first place is a differentiator—“X now supports videos in e-books!” But incompatible capabilities aren't, because they won't get used.
All of the major platforms, Kindle, Kobo, Nook, and iBooks, can trust that—no matter how market share is distributed among them—publishers and e-book developers will test and try to make sure their books work on all four.
In fact, that's kind of the problem. Unless a capability is compatible across all of the major platforms that implement it, that capability will never see widespread and profitable use, because developer will either target a compatible subset of the feature, or they will do a simple design that is easy to port.
So, a plea: when you add a capability to your platform, either make it compatible with existing capabilities on other platforms, base it on an existing standard, or make sure it can degrade gracefully. Then write some documentation. Then make some nice tools.
(That is, don't do what Apple did with .ibooks, and do a wholesale, incompatible, fork of a standard format. Extend the standard format in ways that tie into the built-in fallback capabilities.)
Do this wrong and you'll end up with a bunch of badly thought out titles languishing unsold on your servers. Do this right and you might end up enabling e-book genres that grow the e-book market by appealing to non-readers or satisfy needs readers didn't know they had.