This broke down, because most people monetizing an idea don't want others to host it outside their control. And now, even the peering agreement is breaking down, as Verizon tries to shakedown both sides of a network connection, and not just their subscribers.
We had a friendly, open Internet, and now we have ads and closed-ecosystem apps. In Africa, Facebook Zero is trying to create the illusion that there's no other Internet, just Facebook.
If things like Diaspora had the resources Facebook did, we could totally have distributed social networking, paid for by users.
Facebook Zero is no longer doing what you allege, and never actually did. It was just a plan that was scrapped.
Diaspora never had a design or plan to work. It was a couple college kids who thought it would be neat despite never having written a distributed protocol before. We've had lots of smart people attempt this, and haven't gotten anywhere close to a usable "distributed" social network prototypes.
If your idea can't bootstrap itself from zero users the way modern startups work, maybe it isn't such a great idea after all.
If there is no Internet access then there is no Internet. Critics of Facebook's initiatives should need to provide compelling alternatives if they want to be taken seriously.
If there were a well designed open source router with a decent app store with equivalents of all the web apps you use online, I'm sure you could get a lot of people to start sharing their baby pictures using software installed on that.
I'd switch to something like that in a heartbeat if it existed.
Once somebody makes an easy to use, beautiful router with a vibrant one-click-install app ecosystem (backup, calendar, photos, dropbox, email), I'm sure people will want to migrate to it. It's not just about privacy it's also about being more in control of your data, being 'closer' to your data (uploading your baby photos wouldn't take as long) and not having your experience dictated by profit seeking (your app won't mysteriously switch itself to top stories every other day).
We're just not there yet. Every kind of 'smart router' and associated OSS software is kind of a pile of crap, to be honest.
I think you're asserting a demand without any indications of its existence.