Encrypted Collaboration Spaces: Architectural Whitepaper [pdf]

Heaps excited about this because there's a dearth of work in application crypto as the mindshare is dominated by PQC or Crapto.

I am a little sceptical of this because of the need for ZKP and running a RISC-V program for protocol interactions (I only had a cursory glance). But if we can do LLM inference as standard in protocols these days then this probably has a chance.