Some years ago there were initiatives for distributed computing networks like SETI or even Fold at Home for proteins I think.
Would it be possible that the community could make a network like this for training open source models with non-problematic licenses and more than only open weights? Is that a stupid idea?
[link] [comments]