Blockchains are, however, trying (or claim to be trying) to solve the Byzantine Fault Tolerance problem, which has no known efficient solution.
Is that correct? Are blockchains an actual area of research for folks studying distributed systems in CS Departments?
I'm very pro blockchain tech, and I'm glad this "winter" is coming so that all that noise is reduced and good interesting projects get more visibility.
There's a little known project called selfkey who have been developing some interesting "Sovereign digital id" stuff. Pretty early but interesting.
Just get away from the "money money money" stuff and you can find some interesting projects.
I don't know what happens in computer science departments, but you can watch Ethereum be developed in the open. They cite the research they're implementing.
For example: https://notes.ethereum.org/@vbuterin/proto_danksharding_faq#...
I'm not quite smart enough to follow all the cryptography, but there are a lot of unsolved problems being worked on. It seems that doing trust and accounting at global decentralized scale is trickier than chaining blocks together.
The "genious" of Nakamoto was to put all the existing parts together. But hashcash was there, kad/BitTorrent were there, chained append only DBs were there, etc.