Some analysis from my Burton Group colleague Guy Creese; see the full post for more context-setting
I would not be surprised if we now see a replay of that episode here. With its ability to buy its way into a market blocked, Microsoft went back and fixed Money, to the point where in feature bakeoffs it's typically considered either slightly better or slightly worse than Quicken from year to year. If Microsoft did the same here--if it took even a chunk of the $44 billion it was going to use to buy Yahoo! and applied it to basic block and tackle improvements--it could improve from being a distant number three. However, to do so it cannot mimic Google--it must be significantly better than Google in some areas--to get people to try it. That's why Google became pre-eminent in the first place--it was so much better than the incumbent at the time (Alta Vista).
For example, because of its software and services strategy, Microsoft could mine what users do on their PCs (with appropriate permissions from the user, of course) and use that information to give context to web searches (e.g., realizing that user A is a banker, the system would return "automated teller machines" results to an "ATM" query and "asynchronous transfer mode" results to a network engineer). This is a capability that would be difficult for either Google or Yahoo! to counter and would give better search results. So while Microsoft is currently down, it's not necessarily out.
One tangential note in this context: I'm surprised there hasn't been more coverage of the Microsoft/FAST acquisition the press/punditry/blogosphere, along with the speculation about all of the feasible MSFT/GOOG/YHOO permutations. The technology and expertise Microsoft acquired in FAST will likely play a pivotal role in the next few chapters in the broader MSFT/GOOG story -- a story in which, at current course and speed, Yahoo! has just elected to relegate itself to a supporting rather than leading role, imho.