Yeah, ethereum seems more applicable to the transfer of physical representation rather than the transfer of data conclusions. =/. It would be useful, whenever it comes about would probably be the method for computer "thinking", I cant think of how a system for data analysis would work either. It seems like the ethereum system would work for getting the analyses done and if the ioeprogram decided who to send data to and did it enough times to be sure that the data was analyzed was correct by comparing differences in analysis looking for statistically significant correlations to determine if an output was bad. Would be slow as you would have to check each level at a centralized computer =/. Too bad we don't live  in the future yet and still have to build it.

The international uses for bitcoin are staggering. Control of bitcoin via deeplearning machines and algorithms in my belief is the strongest tool for the elimination of poverty traps in third world countries. Is there any work being done on distributed data analysis? In the same way storj allows for protection from faulty data, through redundancy, is there a system being developed for data analysis that utilizes redundancy for multiple users?