Just published. Sonic proofs. An improvement to zk-snarks. Interesting. Any thoughts from the team?
Curious to see more practical benchmarks, but I find it extremely interesting. zkSNARKs still provide the smallest proofs, but are less competitive due to the trusted setup (and more importantly, the fact that for each functionality, a new [often huge] CRS needs to be generated is prohibitive).
This paper practically solves both of these concerns - (1) Only a single, system wide trusted setup is required. (2) The trusted setup can be constantly updated with new parties (through MPC) - and only one of the parties, ever, needs to be honest. Over time, therefore, trusted setup --> untrusted setup.
Still, ZKPs are ZKPs, and are just one part of the privacy/scalability solutions - but it’s an interesting solution nonetheless.
Thank you for your input. In the past, I believe you had mentioned possibly incorporating zksnarks into part of the privacy solution for enigma along with tee and mpc. It’s that still the case? Or am I mistaken?