The more i think about it the more I think a random subset merkle tree p2p system sort of avoids the established method of copyright prosecution set up by bittorrent: with bittorrent, you establish yourself as a seeder by saying that you have a particular file (or some of its pieces). If instead a seeder only responds to queries about whether it has data that satisfies a particular subtree and never commits to the whole thing....
Follow
especially with random sharding across peers -- idk what the data is, it is just some incomplete binary that satisfies a has function as far as I'm concerned. If the leecher also never asks any peer for the entire subset of the tree then any given peer doesn't know if a peer has downloaded a given file
