Most of the proof-of-work systems out there are really demanding that you waste some amount of money, time, or both, to prevent people from just generating arbitrarily high numbers of coins (as opposed to the Hitchhiker's Guide To The Galaxy use of leaves as coins.) Bitcoin number-crunching is purely wasteful potlatching. Dogecoin is such wow, so calculation!
At least this one is doing a kind of work that's potentially valuable to the world, assuming the system collects all of it in a way that can be used to contribute to mathematical knowledge. (Yeah, yeah, this is /., and I'm commenting on the article without reading it :-) There may be other kinds of calculations that are both useful and verifiable out there. Unfortunately, protein folding and most other non-mathematical real-world applications probably aren't easily verifiable except by having N people redo the same calculation, which is a problem for currencies that need to prevent double-spending. (I ran Folding@Home for a while, as well as the GIMPS Mersenne Prime Search. For SETI@Home, which for some years was a far larger supercomputer than anything on the Top-500 list, sure, you can contend that there really aren't aliens in the chunk of sky your system was testing, but that's not the kind of verification we're looking for...)