OK, not utterly this guy, but his ideas as he talks about economics of open source:
Wonder when will one be able to download a piece of software which distributedly runs the next Google on those client machines who are actually using it for searching at a certain moment. Without any central server. Considering the ever growing information monopolies of search engines, I guess that would be a nice scenario for the future.
Update: This talk made me so enthusiastic, I've done a little research on this. I even thought to myself: why not put some effort into it and produce this software using open source components. Well, it turns out that this idea is not unique (of course) and this software already exists, thanks to Le-Shin Wu and Ruj Akavipat at the Computer Science Department of Indiana University.