I used to be in academia and I'm running a startup now. I'm absolutely certain that generating computer science research papers, even brilliant ones, is a heck of a lot easier than the absolutely insane focus required to ship breakthrough products.
From what I can tell, Yahoo now has a phenomenal environment for people with this level of focus. As for the castles-in-the-sky research people leaving Yahoo, it's probably better off for both the research people and Yahoo. It's got a new business to build -- funding research papers that go nowhere does nothing to advance that goal.