Distributed history search
Rather than indexing the entire internet, index everybody's browsing history to find content
Would you share your browser history to help fellow internet users? A user would then use a special distributed search that searches peers first and then a real internet search second. This would be a form of contributing to your peers for useful content. Such as searches for products and documentation.
+[project]
// Would you share your browser history to help fellow internet users? //
Conditionally - yes, and more, there are ideas floating around of sharing DOM, parsers, etc., which are ways to taking back the Internet.