Joe -- I like all three scenarios because they seem to have been "plucked from reality". However, I worry...

Wikipedia has shown that such massive collaboration is possible. But Wikipedia also operates under some norms that may not extend gracefully to the scenarios you suggest... I'm thinking specifically of "no original research".

arXiv.org has dealt fairly well with the question and perils of original research. I regret that they've pulled back from potential growth -- if they had found a way to limit their overhead further, that model could be used much more widely today than it is.

Some of the most powerful ideas we can pass on are ways to do important, complex, and creative things without 'overhead' that feels painful and 'expensive'. The entire edifice of academia is built on a society in which some professions are dedicated to peer review, knowledge production, teaching, writing, and dissemination. As long as we live in a society that cherishes this, anything we can dream of in the universe of knowledge organization and sharing is possible -- and already implicitly has all of the support that it needs to succeed. The greatest obstacles are those we throw up ourselves.

My suspicion is that to be successful, a massively collaborating academia will have to revise traditional assumptions of leadership

Yes. Just getting work done, allowing people of all backgrounds and ages to lead wherever they have time and inclination, is a fine first-order solution to this problem. In the journal universe, Law Reviews managed this nicely (the history of how and why they settled on student-run reviews is worth a discussion of its own).

Its a good time to think big, especially if big doesn't cost too much.

Truly priceless things rarely cost money; they are outside our financial shorthand. But they are tied to our dreams and our sense of humanity. This is why the most extraordinary dreams are often attainable, and draw whole societies with them.