I recently saw Microsoft's Mike Gannotti speak about building Sharepoint portals for large organizations, and he said "if a user has to search you have failed them." The argument was that if a proper hierarchical navigation was in place a user should never have to search.
To me that's heresy, and I wasn't shy about saying so. The retort was that I saw search as a panacea, or worse as an excuse to be lazy about information architecture. Of course neither is true. I believe great IA begets great search and that both are important and worth getting right.
So I'm cautiously optimistic. I'm hoping Bing can work on both fronts, by applying taxonomy to unstructured information when it makes sense and staying out of the way the rest of the time. With the wealth of talent at Microsoft I think they are capable of building a great search engine, but only if they can really see search as more than an ad platform.
The thing about Google is that it is a very intelligent but very minimal interface to the whole Internet itself. This thing, the whole Internet, has been expanding faster than any "portal" or other kind of "walled garden" could. I hope this continues, for the variety involved but also for the freedom - knowing that no entity (even Google) will be able to sweep the results away.
When Google is just providing access to something others can give access to as well, it keeps them somewhat honest even if they are 90% dominant. When Google or Microsoft or Wolfram or anyone both hosts content and provides the interface, it makes censorship much easier. Even Google is moving more towards doing combining hosting and search through owning Youtube but Google still looks like least likely to create a "wall garden".
I'm also glad they went with Bing and not Sift, as we bought BingSEO.com recently. ;)