RSA puts up a challenge for each encryption algo they come up with, in this case, the RC5-72 algo. distributed net provide client software for participating in each challenge. I've been running the RC5-72 client on a number of machines (including an Amiga) since the challenge was announced -- about 16 years ago. Even with thousands of people participating to find the correct 72-bit key in what is essentially a trial and error test, only about 7% of the total number of possible keys have been generated, and so far nobody has found the right key.
If you do find the right key, I think the reward is $10K, which (IIRC) $8K of which goes to distributed.net and a charity or two, the remaining $2K goes directly to the participant running the client that generated the key.
Unfortunately (for the planet) it's unlikely that this $2K is enough to pay for all the electricity used to find the key. So really the only winners are RSA and dnet -- maybe just RSA, as dnet had to pay someone to code the clients. Whether RSA actually cares at all is another question.
A side benefit of the running the client on a variety of machines is that I can comparatively benchmark their performance, in case anyone ever asks me "Which machine is faster, A or B?". Which my boss does from time to time.