I don't think these guys know how complex and difficult this is going to be. They might have an SDK out this year, but chances are its pretty rubbish. Its not as if others haven't tried. And having written my PhD thesis on these things I know its a complete and utter pain.
Chances for success: less than 1%.
Of course doing something hard means there is a high chance of failure. At least these guys are trying to do something that could have massive upside both to the users and the company, and not just taking the lower risk approach and building yet another social widget or game.
