Yes, the world is contested, but both Negroponte and Deibert are doing things I applaud. The common thread might be called "meshworks": massively peer-to-peer networks, sometimes bypassing the wired and wireless infrastructure. (If that term had not been overused by snake-oil salesmen, I'd grab it).
In the case of the OLPC project the mesh is that massively peer-to-peer wireless networking that is built into the OLPC XO operating system and hardware, allowing wireless LANs to be spontaneously generated to extend the reach of distant WANs. (The technology was unsuccessfully, commercialised by a brilliant Russian hand-held games console called Cybiko; and see here as much to reassure myself that it didn't just look similar; I had a classic Cybiko; I saw the future; shame 500 neighbours didn't see the same).
In the case of psiphon and the Open Net Initiative, the mesh is human to human over the internet, with people outside censored regimes (national or corporate) sharing their connections with those inside regimes that censor internet traffic and sites.
In a commercial context, the Fon project has similarities, leveraging altruism and opportunism to extend private wireless networks.
And this gets me back to Danah Boyd. Boyd was asked what Web 3.0 was. She set up a straw man saying that some people thought that MUVEs would be Web3. However, she thought that this was wrong. She offered the "mobile web" as being the Web3 thing. Negroponte and Deibert both (phrasing it differently) offered the "open web" as the Web3 thing.
For my money, Web 3.0 is the mesh: mobile and open and peer-to-peer.
There are challenges. How do you get MMUVEs to work on a mesh? I don't know, but I am willing to bet it could be done. However, the real problem must be that meshworks would be anathema to the established 3G phone networks and the wired infrastructure providers. Who would fund the research?
That will have to be the next question.