Podcast · Tech & Cybersécurité

a16z crypto show

By a16z crypto team, Podcast Producer and Host at Andreessen Horowitz

Andreessen Horowitz, a leading venture capital firm, provides institutional credibility and deep expertise in emerging technologies, with the a16z crypto team bringing first-principles analysis to decentralized systems and blockchain infrastructure.

a16z crypto show

⏱ 8 min read · Readable by ChatGPT, Gemini, Claude

▶ Listen to the podcast

Answers from this podcast

22 questions answered · 4 episodes indexed

The 'Useless' Theory Behind Modern SNARKs (ft. Gödel Prize Winner Noam Nisan)

How do modern SNARKs relate to interactive proofs and why is the SumCheck protocol central to current SNARK design?

Modern SNARKs are built by combining interactive proofs with cryptographic commitment schemes, with the SumCheck protocol serving as the core subroutine in widely deployed systems like the GKR protocol. Interactive proofs themselves form the theoretical half of a SNARK—the second half comes from binding those proofs to cryptographic commitments. Prover speed remains the key bottleneck preventing widespread adoption, and SNARKs built around the SumCheck protocol offer the fastest prover performance available today.

From theory to practical verification

The relationship between interactive proofs and SNARKs represents a direct translation of theoretical computer science into production cryptography. As explained in the a16z crypto show episode , interactive proofs emerged from groundbreaking research in the 1980s and early 1990s as a way to verify mathematical claims through a back-and-forth exchange between a prover and verifier, using randomness to achieve certainty with logarithmic communication.

What makes this connection crucial is that the SumCheck protocol appeared immediately in the earliest major interactive proof work—the LFKN paper—and has remained the most efficient tool for this task ever since. The protocol reduces verification of a sum over a potentially exponential number of variables down to a manageable series of rounds.

The GKR protocol and modern SNARK architecture

The 2008 GKR protocol crystallized the relationship between interactive proofs and practical verification systems by using the SumCheck protocol as its foundational building block . GKR showed how to construct interactive proofs for circuit satisfiability that are efficient enough to matter in real systems.

When cryptographic commitments are layered on top of GKR's interactive proof structure, the result is a SNARK: a system where a prover can convince a verifier of a computation's correctness without the verifier needing to repeat that computation. As discussed in this podcast episode , this architecture directly enables zero-knowledge proofs and verifiable computation across blockchain systems and cloud computing applications.

Why prover speed defines the frontier

The practical adoption of SNARKs hinges on a single engineering constraint: the prover must run faster than the original computation it's proving . If generating a proof takes longer than running the computation itself, SNARKs become theoretically interesting but practically useless.

SNARKs that build on the SumCheck protocol achieve the fastest prover times known to researchers today. This performance advantage explains why systems like Lasso and Jolt, discussed in detail in the full episode , are reshaping zero-knowledge virtual machine (ZKVM) design. These systems recognize that the SumCheck-based architecture, despite its theoretical elegance, wins on practical metrics that matter.

"If you have some kind of theoretical inclination, you always find really fascinating theoretical questions that come up."

Noam Nisan — Principal Researcher at Starkware and Gödel Prize–winning computer scientist at Hebrew University of Jerusalem. Nisan's early work on algebraic methods for interactive proof systems established the theoretical foundations on which modern SNARKs and verifiable computation are built. He co-founded the field of algorithmic game theory and has made foundational contributions to computational complexity, pseudorandom generators, and Fourier analytic approaches to PAC learning.

Beyond the mechanics of SumCheck itself, the episode explores how interactive proof theory evolved from pure mathematics into a foundational pillar of blockchain infrastructure , revealing how seemingly abstract research questions in complexity theory turned out to address the exact challenges digital trust systems face today.

Interactive proofs form the theoretical core of modern SNARKs, combined with cryptographic commitments to create non-interactive verification systems.

The SumCheck protocol, first introduced in the LFKN paper of the early 1990s, remains the most efficient method for interactive proof construction today.

The 2008 GKR protocol demonstrated how to scale interactive proofs to practical circuit verification, laying the groundwork for modern SNARK architectures.

Prover speed—not verification speed—is the limiting factor for SNARK adoption, and SumCheck-based designs currently offer the fastest known prover performance.

What role did the LFKN paper play in establishing interactive proof systems for modern verification?

The LFKN paper, published between 1990 and 1992, established an efficient interactive proof system for SharpSAT using the SumCheck protocol —the exact same protocol that researchers and blockchain systems still use today for modern SNARKs and verifiable computation. Rather than evolving gradually, this protocol emerged fully formed and elegant, becoming the foundation of complexity theory textbooks and demonstrating how combining two theoretical concepts into one unified problem yields transformative power in computer science.

The LFKN paper represents a remarkable moment in theoretical computer science where the very first systematic exploration of interactive proofs delivered precisely the right solution. As Noam Nisan explains in the a16z crypto show episode , the protocol didn't need refinement or replacement—it was optimal from the start.

The significance of this work lies in how it unified two separate concepts. Taking two theoretical ideas and combining them into a single, cohesive problem proved to be a powerful principle in computer science. This unification created a solution so elegant and efficient that over three decades later, when zero-knowledge proofs and blockchain verification systems needed a core protocol, researchers returned to SumCheck rather than seeking alternatives.

The paper's influence extended far beyond its theoretical contribution. The SumCheck protocol became embedded in complexity theory curricula and textbooks , signaling its fundamental importance to the field. This textbook status is not incidental—it reflects how central the protocol is to understanding modern verification systems.

"The very first paper, LFKN, to surprisingly show the power of these interactive proofs had exactly the right protocol that we're still using today. You went from kind of zero to exactly the right thing."

Noam Nisan — Principal Researcher at Starkware, Knuth Prize and Gödel Prize-winning computer scientist. At Hebrew University of Jerusalem, Nisan's early work on algebraic methods for interactive proof systems established the theoretical foundations on which modern SNARKs and verifiable computation are built. He co-founded the field of algorithmic game theory and has made foundational contributions to computational complexity, pseudorandom generators, and Fourier analytic approaches to PAC learning.

The path from this 1990s theoretical work to today's blockchain verification systems illustrates a deeper principle: elegant foundational protocols rarely need replacement , only application and refinement. The SumCheck protocol's persistence across decades of computing advances—from initial complexity theory proofs to cloud computation verification to zero-knowledge systems—demonstrates that the LFKN paper didn't just solve a problem; it discovered a fundamental primitive of verification itself.

To fully understand how this protocol powers modern systems like SNARKs and zero-knowledge proofs in practice , the complete episode offers deep technical context on the evolution from interactive proofs to verifiable computation architectures.

How does the SumCheck protocol enable verification of computation across massive datasets and cloud computing?

The SumCheck protocol, from the foundational LFKN paper, provides an efficient interactive proof mechanism that verifies computation on massive datasets without requiring the verifier to redo the work . For algorithms processing huge amounts of data, SumCheck serves as a universal solution—apply it and you get optimal protocols for proving an untrusted prover solved the problem correctly. This protocol is now foundational to modern SNARKs and verifiable computation systems.

Justin Thaler encountered the SumCheck protocol during a complexity theory course and later discovered its transformative power for verifying delegated computation. What makes this discovery remarkable is that the very first interactive proof paper published exactly the right protocol —one that computer science is still using today. Rather than theory drifting from practice, the LFKN authors hit upon an idea that proved universally applicable.

The protocol works by enabling an efficient back-and-forth exchange between a prover and verifier . For any streaming algorithm or massive dataset problem, if you apply SumCheck, the resulting proof system is automatically optimal—meaning no better protocol exists for that particular problem. This universality explains why the protocol has become so central to modern cryptographic systems.

As Noam Nisan explains in the episode , the surprising elegance of interactive proofs lies in their ability to reduce verification complexity dramatically. What would otherwise require checking an entire computation can be compressed into an interactive dialogue where the verifier needs only logarithmic resources relative to the original problem size.

From Theory to Cloud Verification

The practical implications extend beyond academic curiosity. When cloud providers or untrusted systems perform computation on your behalf, you face a fundamental challenge: how do you verify correctness without re-executing everything yourself? SumCheck directly addresses this real-world need. The protocol reduces the verification burden to something manageable , even when the underlying computation spans terabytes of data.

This capability has proven essential for building zero-knowledge proof systems and verifiable computation frameworks. Modern SNARKs—succinct non-interactive arguments of knowledge—rely on SumCheck as a foundational building block. The GKR protocol, published in 2008, uses SumCheck as a subroutine to verify arithmetic circuits, demonstrating how the protocol powers contemporary verification systems that underpin blockchain and cryptographic applications today.

What sets SumCheck apart is its universality across algorithm classes . Whether you're verifying matrix multiplication, graph algorithms, or streaming computations, the same protocol applies. This generality is why Thaler and others have built entire research programs around it—it solves an entire category of verification problems at once.

"If you have some kind of theoretical inclination, you always find really fascinating theoretical questions that come up."

Noam Nisan — Principal Researcher at Starkware and a Knuth Prize and Gödel Prize-winning computer scientist at the Hebrew University of Jerusalem. His early work on algebraic methods for interactive proof systems established the theoretical foundations that now underpin SNARKs and verifiable computation. He co-founded the field of algorithmic game theory and has shaped modern approaches to computational complexity and cryptographic verification.

For those interested in how these ideas map onto real systems, the conversation also covers the evolution from Nisan's early Popcorn project to today's zero-knowledge virtual machines , showing how theoretical breakthroughs eventually reshape engineering practices.

What is an interactive proof and why did it represent a fundamental extension of mathematical proof?

An interactive proof is a proof that uses both interaction and randomness to establish mathematical truths, unlike traditional static proofs. The breakthrough insight from foundational research between 1984 and 1989 was that you need both ingredients together—either one alone produces nothing new. This fusion fundamentally changed what mathematicians could prove and how they could verify computations.

Traditional mathematical proofs are static documents: a mathematician writes down a statement and provides a logical chain to justify it. You either believe the proof or you don't. Interactive proofs introduce something radically different. As Noam Nisan explains in the a16z crypto show episode , the key was combining interaction between a prover and a verifier with randomness in the verification process itself.

The community was initially skeptical that this would matter much. Interactive proofs seemed like merely an interesting extension of NP, the class of problems whose solutions are easy to verify. But the work by Avi Shamir, Oded Goldreich, and Michael Rakoff in those five years proved otherwise. They showed that combining interaction and randomness enabled completely new proof strategies that classical proofs simply could not achieve.

What made this discovery so powerful is that it opened the door to efficient verification without re-computation. Before interactive proofs, if you wanted to verify someone's work, you often had to redo it yourself. The protocols developed in this foundational work —including the LFKN protocol published in the early 1990s—turned out to be remarkably efficient and elegant. In fact, those very first efficient interactive protocols are still in use today as the backbone of modern zero-knowledge proofs and SNARKs.

"If you have some kind of theoretical inclination, you always find really fascinating theoretical questions that come up."

Noam Nisan — Principal Researcher at Starkware and Knuth Prize and Gödel Prize-winning computer scientist at Hebrew University of Jerusalem. His early work on algebraic methods for interactive proof systems established the theoretical foundations for modern SNARKs and verifiable computation, while his co-founding of algorithmic game theory extended his influence across computer science and economics.

The practical impact extends far beyond pure theory. The sum-check protocol, derived from this era of research , became so fundamental that it eventually appeared in textbooks as standard material in computational complexity courses. This protocol enables one party to prove properties of large computations to another party without that verifier needing to check every detail themselves—exactly what blockchain systems and verifiable computation platforms need today.

One fascinating detail from the full episode is how the GKR protocol, published in 2008, uses the sum-check protocol as a subroutine , showing how ideas from the 1980s directly powered innovations nearly three decades later.

From theoretical curiosity to ZK proofs

The reason interactive proofs represented such a fundamental shift is that they redefined what "proof" means in mathematics and computer science. A proof no longer had to be a static artifact —it could be a dialogue where randomness and interaction between two parties creates certainty about a statement's truth.

This abstraction became essential infrastructure for modern cryptography. Zero-knowledge proofs, which let you prove you know something without revealing what it is, build directly on interactive proof theory. SNARKs (Succinct Non-interactive Arguments of Knowledge), which compress those interactions into a single message, are equally rooted in this foundational work. The episode details how this line of thinking connects Noam Nisan's theoretical contributions to systems like Lasso and Jolt , which are reshaping how zero-knowledge virtual machines operate in practice.

How did the Popcorn project relate to ideas about verifying delegated computation?

The Popcorn project tackled a core challenge of delegation: how to verify that work completed by someone else was actually done correctly . Nisan and his team explored two main approaches—having multiple people perform the same work to check consistency, or using NP-like proofs to provide verification—while thinking about distributed systems where workstations from different organizations could share resources and cooperate despite trust and incentive challenges.

Building verification into a cooperative web

The Popcorn project emerged from a practical problem that became increasingly urgent in the 1990s. When you hand over computational work to an external party, you lose direct control and must find a way to validate the result . This wasn't initially framed as a blockchain problem—that framing came much later. Instead, Nisan focused on how independent workstations across the web could collaborate and share CPU resources even when they belonged to different institutions with no direct trust relationship.

The insight was straightforward but powerful: delegation without verification is worthless. As Nisan explains in the episode , the team experimented with multiple checking methods. One approach was simple redundancy—give the task to several workers independently and compare results. Another was more sophisticated: leverage the theoretical power of NP-like proofs to create a mathematical argument for correctness that the delegator could verify without redoing all the work.

"If you have some kind of theoretical inclination, you always find really fascinating theoretical questions that come up."

Noam Nisan — Principal Researcher at Starkware, Knuth Prize and Gödel Prize-winning computer scientist at Hebrew University of Jerusalem. His early work on algebraic methods for interactive proof systems established the theoretical foundations for modern SNARKs and verifiable computation, and he co-founded the field of algorithmic game theory.

The distributed context mattered enormously. Unlike a centralized company where every workstation belongs to the same employer, the web required protocols that worked across organizational boundaries . Trust couldn't be assumed—it had to be earned through cryptographic proof. This framing—verification as a mathematical guarantee, not an organizational promise—became foundational thinking that would later apply to blockchain systems, though that wasn't the original motivation.

The specific protocols and techniques Nisan developed during this period, including explorations of interactive proof structures, are detailed further in the full conversation , where he discusses how theoretical computer science and practical distributed system design converged to create tools still used today.

How AI is Changing the Way We Build Companies (ft. Guy Wuollet and Noah Citron)

How Zero-Knowledge Proofs Were Invented (ft. Co-Inventor and Turing Award Winner Shafi Goldwasser)

Marc Andreessen and Chris Dixon: Why America Needs CLARITY

What a16z crypto show covers

The a16z crypto show moves beyond cryptocurrency hype to examine the foundational computer science and economic theory reshaping finance and digital ownership. Each episode brings together Turing Award winners, Nobel laureates, and engineers building decentralized infrastructure to address concrete problems: how zero-knowledge proofs enable privacy at scale, why distributed consensus matters beyond blockchain, how tokenization unlocks capital in new ways, and what mechanism design teaches us about decentralized incentives. The show reveals that crypto's real innovation lies not in financial speculation but in solving classical problems in cryptography, distributed systems, and market design that have challenged computer scientists for decades.

Key facts

Explore all episodes and deep dives on a16z crypto show.

What this podcast really covers

The a16z crypto show is fundamentally about the computer science and economics behind decentralization. It strips away marketing narratives to focus on technical breakthroughs and their applications. Zero-knowledge proofs—the mathematics that allow verification without disclosure—emerge as a core thread across episodes. The show explains how these tools scale blockchains, enable privacy, and solve problems cryptographers have worked on for decades. Distributed systems theory, from Byzantine fault tolerance to state machine replication, forms another pillar. Episodes with Leslie Lamport and Barbara Liskov reveal that Bitcoin's consensus mechanism and modern blockchain design are solutions to classical problems in computer science, not new inventions. Mechanism design and auction theory, explained by Nobel economist Paul Milgrom, show how to align incentives in decentralized networks—crucial for understanding validator rewards, MEV (maximum extractable value), and fair price discovery. Finally, tokenization of real-world assets and decentralized infrastructure represent the practical frontier: how these technologies move from theory into finance, payments, and physical systems.

Who this podcast is essential for

Engineers and architects building blockchain infrastructure or Layer 2 solutions need the show's deep technical analysis of cryptography, consensus mechanisms, and system design trade-offs. Product leaders and strategists at crypto platforms, fintech firms, and traditional financial institutions benefit from understanding how tokenization, stablecoins, and decentralized protocols reshape capital markets and payment rails. Policymakers and regulatory professionals gain insight into how crypto technology actually functions—beyond speculation—to make informed policy decisions on digital assets, financial infrastructure, and AI-driven systems.

What the episodes really reveal

A clear pattern emerges: every major crypto innovation maps back to solved or unsolved problems in classical cryptography and distributed systems. Zero-knowledge proofs, invented decades ago, now unlock scalable blockchains. The Bitcoin whitepaper solved a 30-year problem in distributed consensus. Auction design, decades of Nobel Prize–winning economic research, now determines fairness and capital efficiency in decentralized exchanges and validator systems. This framing—crypto as the application layer for foundational research—reorients how technologists and investors should think about the space. Episodes consistently show that sustainable competitive advantage comes not from token speculation but from solving real coordination, privacy, and efficiency problems. Global stablecoins and real-world asset tokenization represent the edge cases where crypto meets traditional finance; episodes exploring these reveal both the technical elegance and the practical friction points where adoption requires regulatory clarity and market maturity.

What this changes in practice

For engineers, the show shifts perspective: decentralization is a tool for solving specific problems (privacy, censorship resistance, multi-party consensus), not an end goal. For investors and strategists, it reframes due diligence—evaluating crypto projects should emphasize technical defensibility, protocol design, and real economic demand, not token price trajectory. For policymakers, understanding the computer science behind crypto enables more nuanced regulation: distinguishing between legitimate infrastructure concerns and speculative hype, and recognizing where decentralized systems can improve financial inclusion and operational efficiency. For the broader market, the show's emphasis on foundational research and academic rigor legitimizes crypto as a serious field of technical innovation, drawing talent from cryptography, game theory, and systems engineering—the disciplines that will determine whether decentralized networks become infrastructure or remain a novelty.

Crypto's breakthrough is not financial—it is computational. By solving decades-old problems in cryptography, distributed systems, and economic mechanism design, the technology addresses real friction in coordination, privacy, and capital efficiency across finance, governance, and infrastructure.

Listen to the full catalog of a16z crypto show.


Start listening to episodes on decentralized networks and blockchain infrastructure.

The podcast answers these questions

What are zero-knowledge proofs and why do they matter for blockchain?

Zero-knowledge proofs allow one party to prove they know a value without revealing the value itself. They enable scalable blockchains, private transactions, and efficient verification systems by reducing computational burden on networks while maintaining cryptographic security.

How do distributed systems solve the consensus problem?

Distributed systems achieve consensus through state machine replication—where nodes agree on a sequence of operations and apply them identically. Bitcoin solved this using proof-of-work, while other protocols use Byzantine fault tolerance, proof-of-stake, or other mechanisms to reach agreement without a central authority.

What real-world problems does tokenization of assets solve?

Tokenization breaks down real assets—real estate, commodities, equities—into digital units, enabling fractional ownership, 24/7 trading, automated execution, and lower settlement costs. This unlocks liquidity in traditionally illiquid markets and democratizes access to investments previously restricted to large institutions.

Why is auction design relevant to decentralized finance?

Auction design determines how prices are discovered and capital is allocated in markets. In crypto, proper auction mechanisms affect MEV (maximum extractable value), validator incentives, and fair price discovery in token sales, staking systems, and decentralized exchanges.

a16z crypto show

Discover a16z crypto show

a16z crypto team · a16z crypto show

Listen to the podcast

Découvrir a16z crypto show sur Listenly →