Originally Posted by
videotape
It doesn't matter if it's 2D or 3D. Richness of experience can be loosely correlated with the amount of data the game has to shuffle around. Millions of users times lots of data shuffling in real-time = very difficult problem to solve. You either solve it by being extremely clever in ways which defy our current understanding of computational theory, or you solve it with lots of hardware and ultimately lots of money.
Simply being optimistic is actually not enough to overcome the immense technical hurdles. What works much better is a deep understanding of the problems that need to be solved and a realistic view of the costs and benefits of attempting to execute an actual plan. Given that the project in question seems to have more hope and hype than technical acumen, I'm going to stick with my original response of "NOT GOING TO HAPPEN." Not now, and not from this company.
Someone like Google could pull it off, since they probably already have the infrastructure and the money. But they probably wouldn't do it because it's probably not worth doing as a business.