5 ms·

i think it's better to approach this from the other side - there are already many distributed computing projects like folding@home, etc that are trying to solve

i think it's better to approach this from the other side - there are already many distributed computing projects like folding@home, etc that are trying to solve "problems that are cpu and memory hard" - and will realistically have no ending (unless one day there is no disease, or we attain world peace, find aliens, etc). How about we assign value and proof of work to something tangible? Like if your computer folds a correct protein that solves a problem re cancer or aids, you get 1 "humanity coin" ?

aaron-lebo 13y ago
That is ideal but not especially easy. Hashing as they are doing it requires several factors: what you need to compute is not handed out by a central authority, it is hard to compute, and is easily provable once computed. folding@home and others do not usually fit 1 and 3. However, there have been some attempts, see Curecoin and Primecoin to a lesser extent.
bmm6o 13y ago
While protein folding calculations are useful, they can't be used as a foundation of PoW systems because there is no easy way to verify a solution.
- barkingcat 13y ago
  Yes there is. It's the most easy way to verify - see if the protein matches with the problem receptor. It's pretty much black or white if something binds to a receptor. I think the perspective just needs to change. Maybe a little handheld "protein receptor match" kit that you plug your numbers into, and it will "make" your particle to try to match in a small lab kit that is portable.
  - possibilistic 13y ago
    First you have to verify that the conformation of the protein is correct. What use is an improperly or incompletely folded protein with regard to ligand binding?
  - pavel_lishin 13y ago
    So, a machine that can print prions?
TheLoneWolfling 13y ago
For something to work as a proof of work, it needs to have a couple of properties that most @home projects don't have: There needs to be a quick way to verify a correct solution. There needs to be a way to pre-estimate the difficulty of the problem. There should be a way to adjust said difficulty. Ideally, there shouldn't be a time/memory tradeoff.
KVFinn 13y ago
>Like if your computer folds a correct protein that solves a problem re cancer or aids, you get 1 "humanity coin" ? There have been a couple of attempts. GridCoin comes to mind. It works like a regular coin but you get a huge speedup in mining from Boinc credits. Probably easy to abuse and depends on a centralized server though so I'm not sure it could actually work.

aaron-lebo 13y ago

That is ideal but not especially easy. Hashing as they are doing it requires several factors: what you need to compute is not handed out by a central authority, it is hard to compute, and is easily provable once computed. folding@home and others do not usually fit 1 and 3. However, there have been some attempts, see Curecoin and Primecoin to a lesser extent.

bmm6o 13y ago

While protein folding calculations are useful, they can't be used as a foundation of PoW systems because there is no easy way to verify a solution.

barkingcat 13y ago
Yes there is. It's the most easy way to verify - see if the protein matches with the problem receptor. It's pretty much black or white if something binds to a receptor. I think the perspective just needs to change. Maybe a little handheld "protein receptor match" kit that you plug your numbers into, and it will "make" your particle to try to match in a small lab kit that is portable.
- possibilistic 13y ago
  First you have to verify that the conformation of the protein is correct. What use is an improperly or incompletely folded protein with regard to ligand binding?
- pavel_lishin 13y ago
  So, a machine that can print prions?

TheLoneWolfling 13y ago

For something to work as a proof of work, it needs to have a couple of properties that most @home projects don't have: There needs to be a quick way to verify a correct solution. There needs to be a way to pre-estimate the difficulty of the problem. There should be a way to adjust said difficulty. Ideally, there shouldn't be a time/memory tradeoff.

KVFinn 13y ago

>Like if your computer folds a correct protein that solves a problem re cancer or aids, you get 1 "humanity coin" ? There have been a couple of attempts. GridCoin comes to mind. It works like a regular coin but you get a huge speedup in mining from Boinc credits. Probably easy to abuse and depends on a centralized server though so I'm not sure it could actually work.