The only ones to "stick it out" are the ones who are the most likely to profit. This tends to be apps people mostly want.
Speaking as somebody currently living on the proceeds of a software company I sold, this is a naive view.
It's not enough to have an app people want. You have to (a) sell it for enough money to make a profit and (b) keep support costs down enough so your sales profit doesn't disappear.
Right off the bat, when you sell software, it's not a matter of "a lot of people wanting" your product; it's how many want it at the price you set. Let's say you have a product that nobody would be willing to spend much money for, but you could sell it for about the price of a cup of coffee. Let's suppose the product is cheap to make and after you sell it your customers never call you. You can make money with that.
Suppose you come up with a ringtone. It takes you a week to get it into whereever you are selling it, then 5000 customers download it at $1.99, of which you clear $1.00 after the store gets its cut. $5000 for a week of work isn't going to make you rich, but it's a respectable payday. You can live off of that kind of project.
Is this something that people "want"? Well, sure, so long as its priced cheap. The key is that of those 5000 customers, you'll hear from maybe one or two, and you can just pay them $2.00 to go away.
Now suppose you (like I did) develop some kind of mobile data collection app that drives important enterprise decisions. That's pretty damned valuable. You can easily convince a company to pay you $500 *per seat*. The problem is that even if you could wish the software into existence, the customers need more than $500 per seat of support. In fact that's why an open source model works very well for critical systems -- you give the software away and charge for the real expensive parts. In any case, my calculations showed that we broke even on a $10,000 sale, after all was said and done, so we might as *well* have given the software away. We typically sold consulting services at anywhere from $20,000 to $50,000 a pop, which was where we made our money. Believe me, when you've got a team of six engineers, a $20,000 project doesn't look so big.
The point is that the "build a better mousetrap" theory is simply wrong.
Your ringtones and iFarts are bottom feeders in the world of app development. They are profitable for their developers precisely because users don't care very much about them. Price a product like that low enough and you can make money.
The kind of apps that developers garner respect and admiration for developing are a different kettle of fish. It's *hard* to make a profit selling apps that people really care about, because customers demand a relationship with you. That's expensive.
The last thing you need is a third party inserting itself into that expensive and delicate process -- especially an opaque, unpredictable one. You work with your customers and discover they really need some extra functionality. You build it, then have to wait to find out whether you can sell it? That's nuts. You need that like you need a hole in the head.
And this is even worse: you make a portfolio of apps, and then you can't sell them to a different developer? That's a critical exit strategy for many small developers. They have the vision and brains to create an app, but don't have the size to support it. So they develop and market it, and sell it to somebody who is already supporting apps for the main customer base. That's what I did when I sold *my* business. When I had more customers that I could know personally, it wasn't fun anymore so I told one company that if they didn't buy the software I'd sell it their competitor.
Basically, what Apple is telling is that the iPhone is *still* not a platform. It's a music playing phone that can also run toys like iFart.