Aaron Swartz has outlined some semantic web projects but is worried that people will steal those ideas, implement them and take the credit. What he has is something that sounds a lot like other P2P system (not the popular ones though), a UI that isn't hierachical (I liked polyarchies) for storing notes, a Python library backed onto an RDF persistance layer, and hyperliterate programming (maybe something like Maven's, specifically XRef). I'm afraid stuff won't be implemented so for the six people that read my ideas please steal my thunder!
A lot of companies that are doing the Semantic Web like products (knowledge management, text mining, persistent layers, document management even) seem to be taking the high-right strategy. While the early developers have done tools for themselves (Java APIs, ontology management, etc) there's not much for the end user. So I do hope some idea is implemented and some of this stuff gets into everyones hands (like Apple tries to do with iBooks and eMacs now).