Wikipedia's page on Sci-Hub is usually pretty quickly updated with the still relevant URLs. It's rather interesting to watch, as they also list outdated URLs. Really exemplifies what a whack-a-mole game it is.
Wondering why SciHub is not yet on some sort of IPFS / FreeNet / Tor Storage something. Plain file storage obviously would not be enough, metadata and possibly full text index would have to be stored with each entry. Searchable. could even come with a recommendation service like Mendeley or Researchgate based on which papers you pinned in your client.
The Sci-Hub website is (and will likely remain) reachable through Tor, Wikipedia lists its Onion address.
Okay cool, but still it is only run by one single entity. Wouldn't it be nice if everyone in the community could contribute to keep it running?
Ah, yes, but given the perseverance of the publishers in prosecuting them, I'm afraid it's unlikely for a community to step up to do that.