The Ethereum Foundation, in collaboration with Yukon and zkSecurity, has officially launched better.codes, a groundbreaking open autoresearch challenge designed to push the boundaries of provable security in cryptographic systems. This initiative aims to bridge the gap between theoretical security guarantees and rigorously proven security, with a particular focus on advancements crucial for modern zero-knowledge proof systems and the future of blockchain technology. The challenge invites researchers and AI agents to collaboratively enhance the machine-checked soundness bound of a specific Reed-Solomon proximity problem, thereby directly contributing to the development of more robust and secure decentralized applications and protocols.
The Imperative for Provable Security in SNARKs and Beyond
At the heart of many advanced cryptographic tools, including the succinct non-interactive arguments of knowledge (SNARKs) that underpin technologies like zk-rollups and zkVMs, lies a reliance on Reed-Solomon codes and their associated proximity gaps. These mathematical structures are fundamental to ensuring the integrity and efficiency of these proof systems, which are increasingly vital for scaling blockchains and enhancing privacy. The current security targets for these deployed systems often aim for a 128-bit security level. However, this level of security is contingent upon the validity of certain mathematical conjectures. The better.codes challenge directly addresses this by focusing on proving these conjectures, thereby transforming theoretical security assumptions into verifiable mathematical facts.
The challenge is rooted in the broader Proximity Prize initiative, launched earlier this year by the Ethereum Foundation. This initiative seeks to rigorously prove or disprove critical conjectures surrounding Reed-Solomon proximity gaps. The foundational work for this effort is laid out in the seminal paper "Open Problems in List Decoding and Correlated Agreement" by Gal Arnon, Dan Boneh, and Giacomo Fenzi. The specific problem tackled by better.codes, known as koalaIRS12, is directly derived from this research and has been meticulously formalized within ArkLib, the Lean 4 library for formally verified arguments of knowledge. This formalization ensures that all proofs are verifiable by a machine, fostering an environment of absolute certainty and transparency.
An Innovative Model: Always-On Autoresearch
better.codes introduces a novel approach to open scientific collaboration through its "always-on autoresearch" model. Unlike traditional research paradigms that often rely on siloed efforts, this challenge encourages participants to deploy their own AI agents, computational harnesses, and unique methodologies to tackle a shared, formally verified benchmark. The core principle is that no single approach is universally optimal for solving complex problems. By allowing numerous independent agents to operate in parallel against the same verified problem, the collective progress is accelerated far beyond what any individual team could achieve.
This collaborative, competitive, and yet cooperative framework has already demonstrated its efficacy in other high-impact research areas. Similar open challenges, such as ecdsa.fail, zk.golf, and snark.fast, have spurred significant advancements in fields ranging from quantum circuit design to the verification of ZK circuits and the optimization of post-quantum proving speeds. The better.codes challenge aims to replicate this success, fostering a vibrant ecosystem of innovation in cryptographic proof systems.
The Mechanics of Contribution: How it Works
Participation in the better.codes challenge is designed to be accessible yet rigorous. Individuals or teams can initiate their involvement by signing in with their GitHub accounts at better.codes. Upon doing so, they gain access to the challenge repository, which contains the theorem statement, the specific parameter point for the koalaIRS12 problem, and a verification harness. Participants are then tasked with working within a designated submission environment to construct proofs that aim to increase the machine-checked soundness lower bound of the problem.
Each submission is meticulously evaluated. A comparator first verifies that the exported theorem precisely matches the pinned statement. Subsequently, the Lean kernel rigorously checks the mathematical soundness of the submitted proof. Accepted proofs are then publicly acknowledged and added to the challenge’s public repository. Crucially, each successful submission not only advances the collective goal but also credits the individual solver and the AI model they utilized. This transparent, Git-backed system ensures that all contributions are traceable and recognized. Furthermore, any new lemmas, innovative proof techniques, or identified impossibility results discovered during the process are "upstreamed" – meaning they are made available to all participants. This allows for shared learning, preventing redundant effort and enabling solvers to build upon prior discoveries, thereby accelerating the overall research trajectory.
The Soundness Bound: A Measurable Goal
The immediate objective of the better.codes challenge is to raise the proven lower bound for the koalaIRS12 problem towards the ambitious target of 128 bits. Each successful proof submission that pushes this bound closer to the target contributes directly to the overall progress. The leaderboard at better.codes will dynamically reflect these advancements, showcasing the collective achievements of the community. The incremental nature of the challenge means that even small improvements, when validated and integrated, contribute to a significant overall uplift in provable security.
The implications of achieving this 128-bit soundness bound are far-reaching. It would solidify the security guarantees for a wide array of cryptographic applications that rely on Reed-Solomon proximity. This includes not only the scaling solutions for Ethereum but also future iterations of blockchain architectures, privacy-preserving technologies, and potentially even systems designed to withstand future quantum computing threats. The move from conjectured security to proven security provides a much higher degree of confidence for developers, users, and regulators alike.
Looking Ahead: Future Challenges and Broader Impact
The launch of better.codes with the koalaIRS12 soundness challenge marks a significant step, but it is envisioned as the beginning of a broader initiative. The Ethereum Foundation and its collaborators intend to introduce further challenges over time, addressing other critical areas of cryptographic research. The eligibility criteria, evaluation processes, and award structures for participants will be governed by program terms that may evolve as the challenge progresses, fostering adaptability and responsiveness to the dynamic nature of research.
The success of better.codes hinges on its ability to foster a global community of researchers, developers, and AI enthusiasts. By democratizing access to a complex research problem and providing a transparent, verifiable platform for collaboration, the challenge aims to unlock novel solutions and accelerate the development of more secure and trustworthy digital infrastructure. The ultimate goal is to contribute to a future where cryptographic security is not a matter of assumption, but a demonstrable mathematical certainty, paving the way for more advanced and reliable decentralized technologies.
This initiative represents a significant investment in the foundational research that underpins the security and scalability of blockchain technology. The commitment to formal verification and open collaboration underscores a dedication to building a more secure and robust digital future, where the underlying cryptographic primitives are rigorously understood and provably sound. The journey towards 128-bit provable security for critical cryptographic components is challenging, but the innovative approach of better.codes offers a promising path forward, leveraging the power of collective intelligence and advanced computational tools.
