Re:Top Bitcoin questions I’ve been asked
by UweSchmidt on Thursday April 03, 2014 @06:39AM
(#7521957)
It was just an example to clarify the criticism of "cryptocurrencies waste energy". It would be up to the cryptocurrency advocates to show such a function.
On the plus side, the problem does not necessarily be "proportional to D", in fact, finding a better function would be incentivized. Since the computer cycles would produce a "real" value (curing cancer), the upper limit for these coins would be the amount of money society would spend on a cancer treatment.
/speculation
A good proof of work function needs two things (among other desirable properties I will elide):
1) For some difficulty factor D, you should be able to generate an instance of the problem that takes time proportional to D to solve
2) The solution should be verifiable in time much less than D (preferably constant time)
So until you can show me a foolproof way to generate protein folding problems, genetics problems, or SETI problems, etc. that have these properties (I haven't found one myself), it seems quite difficult to make a "useful" PoW.
Now, theoretically, if you found a "useful" problem that had property 1, but not property 2, you could convert it into a viable PoW using efficient zero knowledge proof techniques [0], but at the moment that isn't really viable.
[0] - https://eprint.iacr.org/2013/507.pdf