What project are they pushing forward that could possibly justify sitting on all this talent?The most likely possibility is that there isn't one. Google is a company built by academics with Ph.D.s who were so successful, so fast, right out of school that they've concluded that the secret to life is being a Ph.D.-level academic. Believe me, that's a serious occupational hazard among those of us with that much education. It's the classic trap. You're constantly tempted to fall back on the fallacy that anything you spent so much of your time on must be intrinsically valuable.
So they've built a think tank. It's no better or worse than other think tanks, like the MIT Media Lab or Xerox PARC or Bell Labs or Agilent Labs or IBM's old labs. Like those institutions of yore, Google milks its enormous cash cow and uses the money to fund pie-in-the-sky projects. Some of those projects will eventually change the world... perhaps after their creators leave Google and found startups. Others of those projects will circle round and round forever. Still others don't amount to anything, but because their inventors are under very little pressure -- why fire them, when you can just ask the cash cow for another $2 million to keep them on staff? -- they persist.
I think it's a mistake to assume that most of the Googlers are working on "destructively taking market share from Paypal". Google's core assets work too well to be the product of a giant committee. I'm guessing that the technical portion of the SEO-fighting arm of Google is just a dozen people, with perhaps another two dozen assistants. Another huge group of people is sales reps and customer service and other things that don't scale well. Most of the rest of Google is a giant research organization, dedicated to searching all of known phase space for the next cash cow. If they find it before the current cow gives out, Google gets another 20 years of profits. If not, Google will eventually explode.
Sometimes that strategy works. From the outside, it really is eerily similar to the early days of Apple Computer. Apple was much crazier, in fact: In the early '80s they were rolling in cash and therefore had absolutely no incentive to impose fiscal discipline. They reportedly ran open-loop, without a budget, for several years. The company was a free-for-all collection of quasi-independent projects that sometimes contradicted each other. And that very nearly doomed them to be squashed by IBM... except that, whoops, one of those tiny one-person pie-in-the-sky research projects was the guy working on the Macintosh. Then Steve Jobs found that guy, seized the project, and it became 12 people working to ship the Macintosh. And eventually the Mac became the next cash cow that bought the company another 20 years.
But sometimes your think tank is Xerox PARC, and it becomes most famous as a place where brilliant engineers get all their training and practice before eventually quitting in disgust and founding the startups that actually make all the money. There's real danger that Google will follow that route. Caring and feeding for a think tank is a difficult job, one that Google's founders aren't necessarily cut out for. It's not like they're magically immune to the Peter Principle.
We shall see. But I wouldn't worry that the Googlers are wasting their time. History suggests that frustrated think-tank refugees can be incredibly productive startup founders. They have Things to Prove.