Proof of stake is a application not computer science. What fundamental new knowledge into computing are you thinking about attributing?
BLS signature aggregation was finalized as an IETF standard in 2019. It's the reason Ethereum can support over a million staking nodes.
BLS was invented back in 2001, but was expensive to verify. A paper published in 2018 showed how to verify n aggregated signatures on the same message m with just 2 pairings instead of n+1.
https://ethresear.ch/t/pragmatic-signature-aggregation-with-...
PoS is an application of what?
to be really pedantic, I'd say PoS is an economic breakthrough rather than heavy-duty computer science, strictly speaking. the actual math of the consensus algorithms seems relatively simple, the challenge is aligning all the incentives so that adversaries in a group of anonymous people have nothing to gain by subverting the rules.
I will gladly give a Turing award to whoever formally proves the safety and liveness of Gasper like Lamport did for Paxos.
B. This "merge" in particular utilizes innovations in computer science that were non-existent 12 years ago when the original Bitcoin whitepaper was published.
C. There continues to be loads of cutting edge CS research that is broadly applicable to the entire industry but is being spear-headed by blockchain development, for example work on Zero-Knowledge Proofs.