IBM wants to sell the stuff that's Watson's based on to companies. Companies will get their own 'answer engine'. I guess they think that licenses will make them more money than ad-money for search engines. Which could well be true in this case. Watson's insane knowledge still won't give me the cat pictures I want.
QA technology is very different than search because the former makes strong assumptions about quality of the underlying data. Google's search is much better at filtering out and dealing with low quality content (e.g. blogs) and spam.
IBM should use this tech to one-up the content-scraper SEO junk sites. They should put out sites that are what the scraper sites purport to be and put them out of business.
Or conversely, maybe Google should buy them?