Zero Knowledge Pratyush Mishra on Tiny Proofs, Folding, Low-Memory SNARKs and More
20 snips
Nov 26, 2025 Pratyush Mishra, Assistant Professor at the University of Pennsylvania, dives into his groundbreaking work on zero-knowledge proofs. He discusses how Garuda and Pari achieve ultra-small SNARK proofs, and explains the innovative use of hash-based folding techniques. Pratyush also highlights the significance of low-memory SNARKs for devices with limited resources and explores real-world applications beyond blockchain, such as verifying SQL queries and document formats. His collaborative research aims to push the boundaries of cryptographic efficiency and practicality.
AI Snips
Chapters
Transcript
Episode notes
Proximity Proofs Drive Hash‑SNARK Progress
- Proximity proofs (IOPPs) are the main bottleneck in hash-based SNARKs and have driven major improvements.
- STIR/WEIR-like advances reduced proof sizes from log^2 terms to log·log factors, spurring follow-up work.
Code Switching Transfers RS Advantages
- Code switching reduces proximity claims to codes with better IOPPs, letting linear-time encodable codes inherit RS benefits.
- FIX and FACS unify and extend code‑switching to many linear‑time encodable codes.
Why Low‑Memory Proving Matters Now
- Low-memory proving matters for phones, browsers, and constrained devices as proofs move client-side.
- Prior streaming provers like Ligero trade small prover memory for linear-time verifiers or larger proofs.