Somebody's going to do much better searches than Google, which is after all, a word-based search ranked by a linking scheme.
But it's going to be hard to tell the world about it when it happens because of the enormous advantage Google has in indexing. It's so fast, it's almost an eye on the web, and that will take more hardware than will fit in anyone's garage. No one's going to finance the billions needed on the basis of a limited, meagre sample (like this very informal study).
Then there's the name recognition and holier-than-thou reputation -- tarnished by privacy issue but few people seem to care about that.
You don't really believe Microsoft's going to do it. It's got the money, but it's a big corporate bureaucracy that won't overcome the herd mentality either in business matters or in science and engineering, and therefore in R&D.