In 2005 YC's thesis was that the smartest technical people would make the best founders. That's why when you applied for startup school, they asked you what your favorite technologies were. And while there were certainly Java people there, the expected answers were clearly Python, Lisp, or Haskell.
The reason it may feel less exclusive now is because they're now actively encouraging people to apply rather than (arguably) actively discouraging people from applying. But that's because A) they know the original thesis was wrong B) you can basically now acquire the technical skills needed to build a startup by spending a few hours watching YouTube video about Django, so for most startups there isn't really any advantage conferred by being some sort of technical genius.
While it's true there are more startups now, it's also much harder to get in. Back in 2005 - 2007, probably a third of the startups went out of business without ever even launching. And it was expected that another third would never get any users. The stuff that would get you funded in 2006 not only wouldn't get you an interview today, it wouldn't even get you the email saying you were close to getting an interview.
This is probably a good thing given that the original value proposition (accelerating web startups) makes zero sense today.
Twelve years ago there were an enormous number of people connected to the web but not much good content. The network was primed for hyper virality, so the only viable strategies were to raise huge amounts of money on a deck or an MVP, or else take on five cofounders (and later fire three of them). Any other strategy and you'd get completely destroyed.
These days this is no longer the case. The web is already saturated with content, so there is zero risk of having a competitor's site go super viral overnight. And the fact that consumer expectations are much higher and you need to design for 3+ screen sizes means it takes much longer to reach product market fit.
Being bootstrapped is no longer a huge disadvantage. Neither is not being a genius, the tools are all easy now. There are still going to be new startups that reach the size of Facebook, but they're going to get there by grinding, not overnight by using some party trick.
The new YC makes a lot more sense in this new landscape. Most other investors are still looking for startups that fit the pattern of Facebook and Twitter, and these are the ones who are going to lose all their money.