Zero Knowledge

This week, Anna and Nico are joined by Alex Ozdemir, Assistant Professor at Georgia Tech, to explore the intersection of formal verification and zero knowledge. They begin by revisiting the evolution of the ZK DSL landscape since Alex's last appearance, discussing the rise of ZKVMs, new language tooling, and how his compiler infrastructure project, CirC, has evolved.

The conversation then dives into formal verification and theorem proving, covering SMT solvers, Lean, and zkPi, the first zkSNARK for proofs expressed in Lean. They also discuss compiler correctness, the challenges of verifying cryptographic systems, and why verifiable software will become increasingly important as the industry matures.



Related Links
zkPi: Proving Lean Theorems in Zero-Knowledge 
CirC: Compiler infrastructure for proof systems, software verification, and more 
Kevin Lacker on AI-Assisted Theorem Proving and Acorn 
Building ZK-Powered AI Guardrails with Wyatt Benno
lean Ethereum Part 6: Formal Verification with Alex Hicks 
Groth16, IVC and Formal Verification with Nexus
lean Ethereum




ZK Podcast and Alex Ozdemir
ZK languages with Alex Ozdemir
zkSessions: Alex Ozdemir - The Taxonomy of Circuit Languages
zkStudyClub: Collaborative zkSNARKs (Alex Ozdemir, Stanford University)
zkStudyClub: Unifying Compiler Infrastructure for SNARKs, SMTs, & More w/ Alex Ozdemir (Stanford)
ZK HACK - Introduction to Domain Specific Languages (DSLs) - Alex Ozdemir








**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
* ZEC - Donation address

Read transcript

This week, Anna speaks with Sergey Gorbunov, Engineer at Circle, about Arc, Circle’s new EVM-compatible Layer 1 blockchain, and its approach to a TEE based on-chain privacy. They begin by revisiting Sergey’s journey from Axelar to Circle, reflecting on the evolution of cross-chain infrastructure, the aftermath of the Terra collapse, and how Circle’s acquisition of Interop Labs led to the development of Arc.

The conversation then turns to Arc’s architecture, including its privacy layer, known as the Privacy Sector, and the decision to build it around Trusted Execution Environments (TEEs) rather than ZKPs or FHE. Sergey explains the design trade-offs, discussing privacy, composability, enterprise infrastructure, and why he believes TEEs are currently the most practical foundation for programmable private execution.

Related Links
Axelar Network
Arc Privacy Sector: Keeping Blockchain State, Transactions, and Accounts Private and Quantum Safe
Circle’s Post-Quantum Security Roadmap
Arc Documentation 
Arc: An open Layer-1 blockchain purpose-built for stablecoin finance
Sergey’s X Post on the Arc Privacy Sector paper 
Ian Mier’s X Response to the Arc Privacy Sector works








**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
* ZEC - Donation address

Read transcript

Last week on the show, we interviewed Benedikt Bünz, Chief Scientist at Espresso Systems and Professor at NYU.

The conversation ran long, so we're releasing some of the extra material as an exclusive clip for zkMesh+ subscribers. We've also included the first five minutes here. In this segment, Anna, Kobi, and Benedikt discuss whether AI poses a genuine threat to the foundations of cryptography.

They explore the 'immune system' metaphor for AI's dual role in security: it can uncover bugs and vulnerabilities, while also strengthening defenses through tools like formal verification. The conversation closes with the question of whether AI could ever break fundamental cryptographic primitives—or even invent new physics to do it.

To hear the full discussion, head to https://zkmesh.substack.com/subscribe and become a paid subscriber.

In this episode, Anna and Kobi speak with Benedikt Bünz, Chief Scientist at Espresso Systems and Professor at NYU.
They start with a quick update on Espresso's architecture, its role in delivering fast finality across chains, and the challenges of building high-throughput blockchain infrastructure. The conversation then turns to Benedikt’s recent research on folding schemes, hash-based proof systems such as Arc and Warp, and Golden, a non-interactive distributed key generation protocol for threshold signatures.

The episode later explores Flock, a new proof system for standard hash functions such as Blake3 and SHA-256 that exceeds Ethereum's post-quantum proving targets without relying on specialized hash functions. They conclude by discussing proof system performance, post-quantum cryptography, and the use of AI-assisted development in cryptographic engineering.



Related Links
Bulletproofs — Short Proofs for Confidential Transactions and More
Protostar — Generic Efficient Accumulation/Folding for Special-Sound Protocols
HyperPlonk — Plonk with Linear-Time Prover and High-Degree Custom Gates
Nova — Recursive Zero-Knowledge Arguments from Folding Schemes
Arc — Accumulation for Reed–Solomon Codes
Linear-Time Accumulation Schemes
Golden: Lightweight Non-Interactive Distributed Key Generation
Flock: Fast Proving for Batch Boolean Computations
TensorSwitch — Nearly Optimal Polynomial Commitments from Tensor Codes
Bolt: Faster SNARKs from Sketched Codes
Ligero — Lightweight Sublinear Arguments Without Trusted Setup


Systems and Infrastructure
Espresso Systems Documentation
CAPE (Configurable Asset Privacy for Ethereum)
Monero


Additional Reading
Vitalik Buterin — The Splurge: Post-Quantum Ethereum
Accumulation without Homomorphism
Neo and SuperNeo: Post-Quantum Folding with Pay-Per-Bit Commitments
Espresso’s HotShot: A Consensus Protocol Designed for Rollups








**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
* ZEC - Donation address

Read transcript

No main episode this week, but we’ve got an exclusive bonus clip for our zkMesh+ subscribers!
Continuing our conversation from last week, Wyatt Benno (ICME) describes the world of 'vericoding' - the next stage after the era of 'vibecoding.' Vericoding uses formal mathematics to prove that AI code is actually correct and help prevent bugs in AI-written code. We go deep on SMT (Satisfiability Modulo Theories), a decades-old verification technique originally built for cloud infrastructure, and discuss how SMT can now take a plain-English description and mathematically verify that your AI-generated code does exactly what you want it to do.
If you want to hear this bonus clip, please head over to zkMesh and become a paid subscriber! Link to subscribe: https://zkmesh.substack.com/subscribe