So unless you have some idea of how to do peer review by blockchain.
I mean, that's probably not impossible, but good luck. Since the idea would be preventing anybody from making money off of it, it kind of goes against what blockchain is actually used for. Scihub is trying to create abundance, not scarcity.
Additionally, IPFS has the same problem as torrent files in that it requires additional layers (which aren't all feasible or effective) to obfuscate the source of data. If sharing it has been declared, in some sense, illegal then IPFS will reveal who is sharing it. And once those nodes are forcibly removed (because governments are more capable of doing that than random individuals) then the data can disappear.
I guess, then I2P should be the solution?
Note: Filecoin despite advertising itself for years as an incentivisation layer for IPFS is not - it's two separate networks [1], so it lacks all the features of the mature IPFS protocol.
Yes, exactly. Would love being able to use Filecoin or any of the alternatives (Storj, Sia and those) but they are all very immature right now, both UX and stability wise. If it's not hard to get the files into the networks, it's hard to get them out again.
IPFS could work as it's not ruined by the whole blockchain mess, but then you might as well use torrents (which people are successfully using already for Sci-Hub).
And yeah, torrents already let you do this.
Sure, you could do that over IPFS. But if you really want scale, censorship-resistance and wide-spread usage/storing of it, you'll use torrents (today, maybe future will be different)