These interesting, but admittedly vague and flaky estimates come from computer Scientist, Peter Gutman. Although you can pick at the numbers quite easily, the guy makes a very interesting point. While projects like Seti@Home can harness a lot of computing power, a virus or worm that doesn't need to ask permission from a user could conceivably be vastly more powerful. Imagine the potential if virus writers found more interesting things to do with those cycles than send spam.
Will the first person to find extraterrestrial signals be an amateur hacker, rather than Seti? Could complex protein folding solutions be found by bored crackers? And would the benevolent act of finding a cure for a genetic illness outweigh the malevolent act of creating the worm that rounded up the processing cycles needed to do it? [Uber Review]