Well, to an extent we know it has; they've been cutting down on chefs and have finally started exploiting Google News, it of the alleged 100 million a year revenue potential.
But now, they seem to be slowing spidering. A few days ago, I became aware of the existence of CommonQT (a Common Lisp binding for QT, a multiplatform GUI lib which recently switched from GPL to LGPL, making it practically usable for commercial apps; more on this later, CommonQT is really quite nice); Googling for it, however, turned up nothing, and I had to actually ask for the URL. Asking for URLs is very much something we did in the bad days before Google's supremacy. As it turned out, it lived on common-lisp.net; changes there normally turn up in the Google index in hours; this time it took about four days.
Google can presumably save some money doing this, as they'd save on bandwidth, and possibly idle spidering clusters and save on power too. They can presumably even get away with it; Yahoo and Microsoft search results remain dire as always. It's not a good sign, though. Google Groups has been giving dubious results for the last year, and is beginning to become useless, but fortunately, of course, barely anyone uses it anyway.