I can't think of much to use Alpha for yet either, but I think it represents a significant accomplishment nevertheless, in a frontiers of knowledge representation/interaction sort of way. Doug Lenat describes it as halfway between Google and his Cyc project on the generality <-> sophistication spectrum (see link), which is a fairly impressive accomplishment (it can do a lot more than just 'sin 5'). It may not end up being much more useful than Cyc is at the moment, but it's still an advance. This stuff will get better, but we still have to go through the intermediate steps.

http://www.semanticuniverse.com/blogs-i-was-positively-impre...