STARK – Zero-Knowledge Proof Protocol by Starkware: Eli Ben-Sasson & Avihu Levy at Web3 Summit — by Web3 Foundation youtu.be
Web3 Summit talk on STARKs from StarkWare, a refresher on transparent zk proofs before the SNARKs got all the attention.
4 comments
STARKs still feel like the boring part that ends up mattering once the trusted setup talk fades.
The boring part is exactly the bit that can make deployment less annoying, no ceremony to coordinate means you can ship without a setup committee or a later migration story. But I dont think the setup talk just fades, proof sizes and prover cost are the thing that decide whether this shows up in wallets and rollups or stays as demo code for pepole who like clean assumptions.
> "proof sizes and prover cost are the thing"
My cluster's bill cared more about witness I/O than proof bytes, last Tuesday.
> "proof sizes and prover cost are the thing"
If your cluster was paying more for witness I/O, thats a pipeline problem, not a reason to downplay proof size. The stuff that gets shipped still has to verify cheaply and move around cheaply, and a prover that hammers disk is exactly the kind of DX tax that keeps these systems stuck in demo land.