The problem with IPFS is that they index individual files.

BitTorrent DHT works, at least for the legal open source content in familiar with, because they index on entire collections.

But for some reason everyone builds on IPFS rather than working on the mutable torrents feature, and DAT seems to be mostly ignored(I'm not quite sure what the difference is with the hypercore stuff that replaced it).

I think what we need is something like cr-sqlite with digitally signed updates and a discovery DHT.

Maybe Radicle will solve the issue? I'm not sure how that one works but it seemed promising last I checked.

Freenet solved most of that except slow access. And everyone hated it, supposedly because it's in Java.

Now it's rebranded, flashy sites and whatnot, trying to do everything and noone knows anymore the exact use case.