From a certain perspective, there's nothing wrong with the mobile world today. Phones are getting more powerful, and can do lots of neat things. Sure, maybe developers have to jump through some hoops to get applications in the hands of users, but that shouldn't matter. A developer who knows deep down that he has a great idea will go the extra mile to get that application out there. Developers who aren't so sure won't bother, but that's okay since their apps are probably poor anyway. So these are all good things.
Obviously there are quite a few things wrong with that, and I don't really think anyone (or at least, anyone who's rational) would really say that; it's a straw-man argument I set up. However, I think there is a kernel of truth in there: a lot of people do think that some apps are more important than others. I'm referring to the fallacy of the "killer app".
The quest for the killer app can blind you to the real world, if you let it. True killer apps are exceptionally rare. For instance, what was the killer app for the web? Was it email? Was it mapping? Search maybe? How about social networking? With the possible (though arguable) exception of search, none of those are true killer apps. They were either "ports" of popular apps to the web, or else popular incremental additions. In fact, I assert that there's no clear killer app for the web at all, yet few people would today argue that the web is not a viable application platform.
I really believe that the history of the web backs me up when I say that this quest for "special" apps is misguided. Trying to predict in advance what the best and most disruptive applications are is a chump's game. And if you're trying to predict (let alone encourage) disruption, the people who you least want to put in charge of that are by definition the people who stand to be disrupted.
That's why I think the current state of mobile isn't working. There is absolutely no reason to believe that the next big thing is more likely to come from a well-heeled, committed, industry insider than from a hobbyist with a spouse and two kids. Yet the current mobile world, with all its restrictions and certifications and fees, favors the former.
That's why I really love what we are trying to accomplish with Android. When I read posts by people working on the Android Developer Challenge, I get much more excited by the guy who says he works on his app after he puts his kids to bed, than I am by the guy who says he's just porting something for the Challenge. (And yes, I have seen both posts -- that's not just hyperbole.) That guy with the kids is in my opinion far more dedicated and likely to come up with something clever and practical than the latter. Yet he's also the guy who is least likely to want or be able to jump through the hoops necessary to get a signed application available for users to download -- $99 annual fee or not.
In other words, the best apps are the best apps, not the biggest or most "important". I look forward to an open mobile world where apps of any size and scope -- not just the ones with big money, biz-dev teams, and CxOs smarming it up on the golf course -- can land on my phone.
 Some people I know would argue that the dark secret killer app of the web is porn. I guess I can't really disagree, although I'd argue that porn is the first-follower of new media types, rather than the driver.