OTOH cryptocurrency consensus is mostly about answering a question: "How do we prevent bad guys from stealing money or doing other nasty things?"
While a concept of Byzantine Fault Tolerance was known before Bitcoin, it was never really applied in practice AFAIK - people thought it's overkill. Also I'd say doing it within a private network is one thing, and doing it with random weirdos on internet is completely different.
Distributed computing researchers like Lamport were considering models where e.g. up to 30% of nodes are compromised, that won't work on internet where an attacker can potentially simulate billions of fake nodes. Nakamoto consensus is really elegant as it combines Sybil protection, incentivization and consensus into one thing.
"What is going on now" is largely the same thing as before:
If you have a solid CS/software engineering background, you probably already know 90% of it.I guess crypto-specific consensus might be new, but you can get a good grasp reading few articles. And that part is actually opinionated, so you need to decide on a camp before you read materials. Bitcoin people would likely disagree with anything written about PoS.
Another fun thing is Zero-Knowledge Proofs (ZKP). That's actually quite new, complex and might be interesting.
The rest can be rather boring. Users submit cryptographically-signed commands (transactions) which processed in a deterministic fashion. I'm not sure it's worth reading a whole book about it.