Zero Knowledge

In this episode, Anna Rose and Tarun Chitra chat with Miranda Christ, a computer science PhD student at Columbia University, about the intersection of cryptography and AI through watermarking techniques. Miranda shares her research on developing imperceptible ways to prove that content was created by AI models, covering everything from simple red-green word lists to sophisticated pseudorandom error-correcting codes.   The discussion explores the cryptographic properties of watermarks - including completeness, soundness, and undetectability - and how these parallel the properties we see in zero-knowledge proof systems. Miranda explains how watermarking differs from other cryptographic approaches like ZKML by only modifying the sampling process rather than the underlying model weights, making it computationally lightweight and practical for deployment.   Related links: Episode 206: Distilling DeFi Primitives with Guillermo, Alex and Tarun My AI Safety Lecture for UT Effective Altruism Google SynthID Amazon Public Watermark Detector How ChatGPT could embed a ‘watermark’ in the text it generates - New York Times Wall Street Journal on OpenAI not Deploying Watermarks A Watermark for Large Language Models Undetectable Watermarks for Language Models Watermarks in the Sand: Impossibility of Strong Watermarking for Generative Models Pseudorandom Error-Correcting Codes Ideal Pseudorandom Codes     Check out the latest jobs in ZK at the ZK Podcast Jobs Board.      **If you like what we do:** * Find all our links here! @ZeroKnowledge | Linktree * Subscribe to our podcast newsletter * Follow us on Twitter @zeroknowledgefm * Join us on Telegram * Catch us on YouTube   **Support the show:** * Patreon * ETH - Donation address * BTC - Donation address * SOL - Donation address Read transcript

In this episode, Anna Rose welcomes back Daniel Kang professor at UIUC and founding technical advisor at VAIL, for an update on ZKML and how the space has evolved since early 2023. Daniel covers the 2023-2024 cohort of ZKML tools including zkCNN, zkLLM, EZKL, and his original ZKML project, while introducing his new project ZKTorch, which offers a flexible hybrid of specialized and general-purpose approaches. The discussion explores practical applications like verified FaceID, proof of prompt, and proof of training, along with the technical challenges of adding ZK proofs to machine learning models. Daniel shares insights on the performance trade-offs between specialized cryptographic systems and generic circuits, and how ZKTorch aims to offer both flexibility and speed for proving ML inference. Related links: ZKTorch: Open-Sourcing the First Universal ZKML Compiler for Real-World AI ZKTorch: Compiling ML Inference to Zero-Knowledge Proofs via Parallel Proof  ZK Torch GitHub Accumulation by Bing-Jyue Chen, Lilia Tang, Daniel Kang Episode 369: Ligero for Memory-Efficient ZK with Muthu Episode 356: ZK Benchmarks with Conner Swann Episode 364: AI and ZK Auditing with David Wong Episode 265: Where ZK and ML intersect with Yi Sun and Daniel Kang Bonus Episode: zkpod.ai & Attested Audio Experiment with Daniel Kang ZK13: ZKTorch: Efficiently Compiling ML Models to Zero-Knowledge Proof Protocols - Daniel Kang AI Agent Benchmarks are Broken VAIL zkCNN: Zero Knowledge Proofs for Convolutional Neural Network Predictions and Accuracy zkLLM: Zero Knowledge Proofs for Large Language Models MLPerf Inference: Datacenter     Check out the latest jobs in ZK at the ZK Podcast Jobs Board.      **If you like what we do:** * Find all our links here! @ZeroKnowledge | Linktree * Subscribe to our podcast newsletter * Follow us on Twitter @zeroknowledgefm * Join us on Telegram * Catch us on YouTube   **Support the show:** * Patreon * ETH - Donation address * BTC - Donation address * SOL - Donation address Read transcript

In this episode, Anna Rose and Guillermo Angeris catch up with Muthu Venkitasubramaniam, Professor of Computer Science at Georgetown University and cofounder of Ligero. They discuss how Ligero’s small memory footprint makes it a good choice for client-side proving, as well as the importance of programmable compliance in blockchain. The conversation explores the differences between ‘MPC in the head’ and error-correcting code perspectives, and how well-established primitives influence the design of modern ZK systems. They also debate the challenge of adding ‘ZK’ privacy back into systems without it, why proving EVM traces may be absurd, and what kinds of guarantees might exist around the results of vibe coding. Related links: Episode 363: Bringing ZK to Google Wallet with Abhi and Matteo Episode 326: MPC & ZK in Ligero and Ligetron ZK13: Ligerito: A Small and Concretely Fast Polynomial Commitment Scheme - Kobi Gurkan ZK13: Vibe coding ZK Apps with Ligetron ZK Platform - Muthu Venkitasubramaniam  ZK10: Analysis of zkVM Designs - Wei Dai & Terry Chung  Ligerito: A Small and Concretely Fast Polynomial Commitment Scheme Ligero++ - Reducing proof length of Ligero Adding Zero-Knowledge to STARKs - Talk by Ulrich Haböck Aurora - comparing prover times of STARKs vs Ligero WYSTERIA: A Programming Language for Generic, Mixed-Mode Multiparty Computations Samaritan: Linear-time Prover SNARK from New Multilinear Polynomial Commitments Brakedown: Linear-time and field-agnostic SNARKs for R1CS Intro to MPC-in-the-Head     Check out the latest jobs in ZK at the ZK Podcast Jobs Board.      **If you like what we do:** * Find all our links here! @ZeroKnowledge | Linktree * Subscribe to our podcast newsletter * Follow us on Twitter @zeroknowledgefm * Join us on Telegram * Catch us on YouTube   **Support the show:** * Patreon * ETH - Donation address * BTC - Donation address * SOL - Donation address Read transcript

In this episode, Anna Rose and Nico Mohnblatt speak with Dan and Sinu from TLSNotary to trace the project’s journey from early Bitcoin forum ideas to its current role as a foundational protocol maintained by PSE. Dan recounts the origins of TLSNotary as a tool for cryptographically proving web data, while Sinu explains how the project was revived to provide modern TLS attestation. The conversation covers the use-cases for verifiable web proofs, the different modes these interactive protocols can take, and the broader impact of this technology on leveraging siloed user data. Related links: Episode 325: Web Proofs with Tracy from Pluto Episode 354: The Founding of Zero Knowledge Systems with Austin Hill Episode 362: zkTLS with Maddy from Reclaim 2013 Blog Post ‘tlsnotary - cryptographic proof of fiat transfer for p2p exchanges’ TLSNotary Whitepaper DECO: Liberating Web Data Using Decentralized Oracles for TLS Primus Labs (Previously PADO) Town Crier: An Authenticated Data Feed for Smart Contracts     Check out the latest jobs in ZK at the ZK Podcast Jobs Board.      **If you like what we do:** * Find all our links here! @ZeroKnowledge | Linktree * Subscribe to our podcast newsletter * Follow us on Twitter @zeroknowledgefm * Join us on Telegram * Catch us on YouTube **Support the show:** * Patreon * ETH - Donation address * BTC - Donation address * SOL - Donation address   Read transcript

In this engaging discussion, Goblin Oats from Tonk, a pioneer in building personalized local first software, and grjte from Bain Capital Crypto, delve into the future of decentralized architecture. Goblin shares insights on the evolution of Tonk from blockchain games to empowering user agency in software. They explore the power of Conflict-free Replicated Data Types (CRDTs) for collaboration, the intersection of local first systems with the AT Protocol, and innovations in data management, emphasizing privacy and data sovereignty.

Florent Tavernier and Marek Olszewski from Self discuss the transformative potential of on-chain identity using zero-knowledge technology. They explore the challenges of varying cryptographic standards and the innovative Open Passport initiative aimed at enhancing identity verification. The conversation highlights how blockchain can bolster Sybil resistance and the necessity of user privacy in digital identity. They also reflect on the importance of integrating traditional systems with Web3 innovations, advocating for secure, decentralized identity options that empower users.

Kostas Kryptos from Mysten Labs and Tarun Chitra, a seed investor in SWE, dive into groundbreaking ZK innovations. Kostas unveils zkLogin and zkAt, an updatable access control project harnessing unique approaches to trusted setups. They discuss Sui's ambition to go quantum-ready through STARKs while transitioning EdDSA accounts. The conversation also uncovers the application of ZK technology in the Greek stock market and the evolution of cryptographic account management, spotlighting dynamic access controls and the future of decentralized finance.

David Wong from ZK Security, an expert in zero-knowledge auditing, discusses the transformative impact of AI on auditing practices. He highlights how AI agents are integrated into workflows, creating both opportunities and challenges. The conversation delves into the unique demands of auditing ZK systems versus traditional smart contracts. Wong also addresses ethical concerns surrounding AI in security, false positives in auditing tools, and how AI can detect vulnerabilities in cryptocurrency, showcasing the future synergy between AI and zero-knowledge technology.

Abhi Shelat and Matteo Frigo, engineers at Google with deep expertise in cryptography and ZK systems, discuss their integration of zero-knowledge proofs into Google Wallet. They highlight the technical decisions behind the Anonymous Credentials for ECDSA, balancing client efficiency with user privacy. The conversation dives into the challenge of revoking credentials like stolen passports and the significance of bringing advanced cryptography to mainstream tech, emphasizing its potential for enhancing security without sacrificing usability.

In this discussion, Madhavan (Maddy) Malolan from Reclaim Protocol delves into the groundbreaking zkTLS technology and its role in web data verification. He shares insights into the journey of developing solutions for secure TLS connections while tackling challenges in integrating zero-knowledge proofs. Maddy also explores zkTLS's applications in education, employment, and financial verification, emphasizing user control over sensitive data. The conversation touches on the balance between security and user experience, along with the potential for zkTLS in AI applications.