from what I understand, the Authors are still free to send you their papers. So perhaps the simplest decentralized system is for each author to run an automatic email request system and have feeds / aggregators of papers titles/abstracts with author emails to make it easy to get the papers you want?
That would only work for a tiny tiny intersection of :
1) technical enough people to install and maintain that
2) with an email address still working (if there is one on the paper!)
3) still alive
