Your idea doesn't go far enough. What good are a couple of stale indexes of Wikipedia and Stackoverflow going to be?

Instead, I am wondering why there isnt a federated open source search engine. It would be a cloud of nodes, each node spidering and indexing a small hash bucket of urls. With a million such nodes we could have a live updated, distributed search engine to replace Google. We could run millions of queries without paying someone - we'd pay back by serving as a node, just like in BitTorrent. With all the interest in privacy, I wanted to see more discussion of replacing Google with an open, non-censured and private protocol for search.