TR published an article titled: "Building a Better Search Engine". There's a lot of buzz going on around "beating Google search". Wikia is one of the names that's often mentioned. Powerset is too. (The TR article also mentions Avatar by IBM.)
The TR article is disappointing due to the fact that nothing new is mentioned. All the issues in natural language technology (which is part of my educational background) and search mentioned are old (and very difficult). And progress is slow. Up until now, we haven't seen a convincing demo by, for instance, Powerset. They say on their blog they will be demo-ing soon. I know my post sounds cynical, but the fact that I write about this topic is because I really hope search will be improved by companies at Powerset, IBM, etc. Or by older Dutch companies as Irion and Collexis.