It's different and I like the thesis behind it: "decision engine". A good way to understand search is to see how people make the decisions to go about what they're searching for. Keyword they start out with, its variants, the final answer/result, and what form of search they should be using. travel? local? image? etc. It's all going to come down to execution though (Cliche, but relevant here more than ever). If it sucks, then 80 mil in advertising wont work well. BUT if it is something that people have to say "yo, you NEED to try this", then things could get interesting.