The way it'd work out is the same way YC itself started: attract talented, smart people who have no patience for academia's politics. Let them make their own decisions. Build a brand on the discoveries they make, which attracts more people. Change public opinion about how research should be done. Eventually, the "top academic talent" realizes they've made a horrible mistake, and scrambles to join the new system that's replaced the old.
Scientific research is by its nature risky and uncertain. The current risk-averse culture of academia with its focus on grant applications, publishing, tenure, and positive results seems to discourage the risk-taking necessary in science.
On the other hand, the culture prevalent in tech accepts and encourages risk-taking, supports fast and easy recovery from failure and generally evaluates people on merit, not success or failure of their most recent project.
Think how hard the hiring problem is. How do you hire the best people? How do you assess someone's quality based on a few bits of paper and a short interview. Allocating funding in academia consists of solving problems like that, all the time.
Those are 2 reasons off the top of my head.
If you mean "vetted by a tiny handful of fallibly-human organizations such as Nature and Cell who may or may not decide to accept your paper regardless of its actual merit and based mostly on how 'hip' the research is," well then we're back to the hazing/cliquey-clusterfuck thing again.
"Grants" seem equivalent to "angel investments" in the startup world... which seem a lot saner, btw.
http://blogs.sciencemag.org/pipeline/archives/2013/09/10/gre...
The system is far from perfect, but I don't think you'll find anyone in academia saying otherwise. I think of it being like Churchill's quote about democracy being the worst form of government, except for all the others that were tried.
Bell Labs has produced a lot of great research and many researchers have had fruitful careers there and after.
Why can't YCR be the same?
YCR could be the same, one day. Nevertheless, it's a very long and perilous road from YCR the proposal, to Bell Labs 2.0.
A YCR "fellow" (or whatever it'd be called) would probably not be a competitive candidate for an academic position. It's not because they weren't in academia, it's because they will likely not have the work product used to measure performance in a academic setting. A hiring committee is not going to risk department funds on someone with a non-existent track record when there's a line out the door of other candidates.
I think OP is correct in perceiving that selecting YCR would probably close the door to a future in academia. I will also predict that YCR will have difficulty hiring top academic talent.