I would write more, but I'm at a loss for words. See below.
DistributedComputing can now provide massive CPU power for use on a problem, and the major obstacles now become providing sufficient computation matter to use those CPU cycles (Seti@home has been accused of artificially inflating the computational work needed to handle their data in order to make up for a shortage of incoming radio astronomy data) and analyzing the results. In other words, you've got one heckuva computer, but you still have to have data to feed it, and the intelligence to interpret the results.
Publicity also feeds distributed computing - media attention and grassroots "buzz" greatly affect how many cycles a given effort will receive (people can't contribute cycles if they haven't heard of you).