Layla Zulfa
London, UK – The Ethereum Foundation Formal Verification team has officially launched better.codes, an innovative open autoresearch challenge designed to advance the foundational security of modern succinct non-interactive proof systems (SNARKs). Developed in collaboration with leading cryptographic research entities Yukon and zkSecurity, the initiative aims to close critical gaps between conjectured and formally proven security benchmarks for widely deployed cryptographic protocols.
The better.codes platform centers on a highly specific, self-contained problem derived from the broader Proximity Prize research: koalaIRS12, a complex Reed-Solomon proximity problem. Participants are invited to deploy their AI agents and computational resources to progressively raise a machine-checked soundness bound associated with this problem. The ultimate goal is to reach a fixed target of 128-bit security, a benchmark considered standard for robust cryptographic systems. This collaborative, verifiable, and public research model represents a significant step forward in the quest for provably secure decentralized technologies.
The Core Challenge: Elevating Provable Security for SNARKs
At its heart, better.codes addresses a critical vulnerability in the theoretical underpinnings of many modern cryptographic systems. SNARKs (Succinct Non-Interactive Arguments of Knowledge) are a cornerstone of cutting-edge blockchain technology, enabling privacy-preserving transactions, scalable zk-rollups, and efficient zkVMs (zero-knowledge Virtual Machines). These technologies are pivotal to Ethereum’s long-term roadmap, including its transition to a post-quantum secure future. However, the security guarantees of nearly all production hash-based SNARKs heavily rely on conjectures related to proximity gaps and correlated agreement for Reed-Solomon codes.
While researchers widely believe these conjectures hold true, current formal proofs often fall short of substantiating these beliefs to the desired 128-bit security level. This creates a gap where deployed systems operate under an assumed security level that is not yet fully provable. The better.codes challenge directly confronts this issue by providing a public, verifiable framework for researchers and AI agents to collectively push the boundaries of formal proof. Each successful submission, meticulously checked by the Lean theorem prover’s kernel, raises the publicly displayed soundness bound. Importantly, new lemmas, proof techniques, and even impossibility results discovered during this process are upstreamed, ensuring that all participants benefit from collective progress, fostering an environment of continuous improvement and shared knowledge.
Understanding the Technical Underpinnings: Reed-Solomon Codes and Formal Verification
To appreciate the significance of better.codes, it’s essential to delve into the technical concepts at play. Reed-Solomon codes are a class of error-correcting codes widely used in digital communication and storage, from CDs and DVDs to QR codes. In cryptography, their properties related to "proximity gaps" are crucial. A proximity gap refers to the minimum distance between valid codewords and invalid ones. The larger this gap, the easier it is to distinguish between correct and corrupted data, which translates directly to the security strength of cryptographic proofs built upon these codes.
The koalaIRS12 problem, specifically chosen from the Proximity Prize research, focuses on a particular instance of these Reed-Solomon proximity problems. Its formalization in Lean, a powerful interactive theorem prover, ensures that every single step of a submitted proof is rigorously checked for logical soundness. Lean acts as an incorruptible arbiter, verifying mathematical statements with the utmost precision. This formal verification process is paramount because it eliminates human error and ambiguity, providing an unassailable guarantee of correctness—a necessity when dealing with the foundational security of systems that secure billions of dollars in digital assets and sensitive data. The entire problem, including its parameters and verification harness, is formalized end-to-end in ArkLib, the Lean 4 library dedicated to formally verified arguments of knowledge.
A Strategic Imperative: Why Provable Bits Matter
The pursuit of provable bits for cryptographic security is not merely an academic exercise; it’s a strategic imperative for the entire digital economy. The blockchain industry, with its emphasis on transparency, immutability, and decentralization, relies heavily on cryptographic primitives whose security can be mathematically guaranteed. If the underlying conjectures about Reed-Solomon proximity gaps were to be disproven, or if a lower actual security bound were to be established, it could have profound implications for the robustness of existing SNARK-based systems. This could necessitate costly overhauls, expose systems to theoretical attacks, or undermine public trust in these technologies.
The 128-bit security target is a standard benchmark in cryptography, implying that an attacker would need to perform approximately 2^128 operations to break the system. To put this into perspective, cracking a 128-bit cipher through brute force is currently considered computationally infeasible, even with the most powerful supercomputers and projected advancements in computing power, including quantum computers for certain types of algorithms. By formally proving that SNARKs achieve this 128-bit soundness bound, better.codes aims to solidify the confidence in these systems, ensuring they meet the highest standards of cryptographic assurance. This commitment to provable security aligns with the broader ethos of the Ethereum Foundation, which continually invests in research and development to enhance the integrity and resilience of the Ethereum network and the wider Web3 ecosystem.
Contextualizing the Effort: The Proximity Prize Initiative
The launch of better.codes is not an isolated event but a direct evolution of the Ethereum Foundation’s earlier "Proximity Prize initiative." Launched earlier this year, the Proximity Prize was established with the ambitious goal of either proving or disproving the Reed-Solomon proximity gaps conjectures. This broader initiative was spurred by the recognition of these fundamental questions within the cryptographic community, notably highlighted in the influential paper "Open Problems in List Decoding and Correlated Agreement" by Gal Arnon, Dan Boneh, and Giacomo Fenzi. This paper laid out grand challenges that directly feed into the specific problem tackled by koalaIRS12 on better.codes.
The Proximity Prize, and now better.codes, represent a concerted effort to engage the global research community in solving these complex, foundational problems. By formalizing these challenges in an accessible and verifiable manner, the Ethereum Foundation is democratizing access to cutting-edge cryptographic research, inviting contributions from a diverse pool of talent, ranging from seasoned cryptographers to burgeoning AI developers. This collaborative model accelerates discovery and ensures that solutions are robust, transparent, and universally verifiable.
The Paradigm of "Always-on Autoresearch"
better.codes exemplifies a nascent but rapidly growing model for open collaboration known as "autoresearch." Unlike traditional academic challenges or bug bounty programs, autoresearch challenges create an "always-on" environment where participants continuously run their AI models, computational harnesses, and specialized tools against a common, verified benchmark. The key differentiator is that every successful, promoted submission doesn’t just earn a prize; it raises the baseline for all subsequent progress.
This model is predicated on the understanding that no single agentic setup or research team holds a monopoly on optimal solutions for open problems. By encouraging many independent setups to work in parallel on the same benchmark, the collective frontier of knowledge can be advanced significantly faster than through isolated efforts. This distributed, competitive, yet ultimately collaborative approach harnesses the collective intelligence of the community, leveraging diverse methodologies and computational strategies.
The success of the autoresearch model has already been demonstrated in other critical areas of cryptographic research. Platforms such as ecdsa.fail have pushed the boundaries of quantum circuit design, zk.golf has driven innovations in verified Zero-Knowledge circuits, and snark.fast has accelerated advancements in post-quantum proving speeds. These precedents underscore the potential of better.codes to make substantial, verifiable contributions to the field of cryptographic security, specifically in the realm of SNARKs.
How the Challenge Works: A Transparent and Collaborative Framework
Participation in better.codes is designed to be streamlined and transparent. Interested solvers can sign in using their GitHub credentials at better.codes and clone the challenge repository. Within this repository, the core problem — the theorem statement, specific parameter points, and the verification harness — are immutably pinned. Solvers then work within a designated submission surface, aiming to prove a larger soundness lower bound, with their progress scored in "bits."
Each submission undergoes a rigorous two-step verification process. First, a comparator checks that the submitted theorem statement precisely matches the pinned, official statement. This ensures that solvers are addressing the correct problem. Second, and most critically, the Lean kernel formally checks the proof for logical correctness. Only submissions that pass both checks are accepted and promoted to the public repository. Solvers receive full credit for their contributions, including recognition for the specific AI models or tools used.
The git-backed nature of the repository ensures complete transparency. All submissions, along with their associated lemmas, proof techniques, and any impossibility results, are publicly visible. This allows any participant to review past contributions, build upon existing work, and avoid duplicating efforts or pursuing dead ends. This incremental, open-source approach fosters a powerful feedback loop, accelerating collective progress for all solvers and their AI agents. It effectively transforms individual competitive efforts into a shared knowledge base that benefits the entire cryptographic research community.
Statements from Collaborating Parties and Industry Implications
While specific quotes were not provided in the original announcement, the launch of better.codes implicitly conveys strong messages from the collaborating parties. Representatives from the Ethereum Foundation would likely emphasize their unwavering commitment to foundational research and the proactive pursuit of provable security for critical blockchain infrastructure. They would highlight the challenge as a testament to their belief in open collaboration and the power of decentralized, incentivized research to solve complex problems that underpin the future of Web3.
Yukon and zkSecurity, as key collaborators, would likely underscore the importance of bridging theoretical cryptography with practical, formally verified solutions. Their involvement signals a shared vision for robust, secure systems that can withstand the test of time and evolving threats, including the potential advent of quantum computing. Experts associated with these entities would also likely point to the increasing role of AI and automated theorem proving in accelerating scientific discovery and cryptographic breakthroughs.
The broader implications of better.codes are significant. By systematically enhancing the proven security of SNARKs, the initiative directly contributes to the resilience and trustworthiness of technologies like zk-rollups, which are crucial for scaling Ethereum and other blockchains. Stronger security guarantees reduce risk for users and developers, fostering greater adoption and innovation within the decentralized finance (DeFi) and broader Web3 ecosystems. Moreover, the autoresearch model itself could serve as a blueprint for tackling other intractable problems in computer science, mathematics, and beyond, demonstrating a novel approach to scientific progress driven by collective intelligence and automated verification.
What Comes Next: Expanding the Frontier of Provable Security
The initial launch of better.codes focuses specifically on the soundness challenge for koalaIRS12, with the immediate objective of raising its proven lower bound to 128 bits. This foundational step is critical, but the vision extends further. The Ethereum Foundation and its collaborators anticipate adding further challenges to the platform over time, tackling other complex problems within the Proximity Prize initiative and beyond. This "always-on" nature ensures a continuous pipeline of critical research problems, maintaining momentum and fostering ongoing engagement from the global research community.
As the challenge progresses, the program terms governing eligibility, evaluation, awards, and payments may be adjusted to optimize the research outcomes and incentivize participation effectively. The overarching goal remains to create a sustainable ecosystem for advanced cryptographic research that leverages automation, transparency, and collaboration to build a more secure digital future.
Researchers, AI developers, and cryptographic enthusiasts are invited to join this pioneering effort. Participation begins by signing in at better.codes, where they can access the challenge repository and contribute to a new era of provably secure decentralized systems. The journey to a fully verifiable 128-bit security for critical cryptographic primitives has begun, powered by collective intelligence and the rigor of formal verification.
