Layla Zulfa
The launch of better.codes represents a pivotal moment for the cryptocurrency and blockchain community, aiming to bridge a critical gap between conjectured and mathematically proven security benchmarks for foundational cryptographic primitives. At its core, the initiative invites researchers, developers, and AI agents worldwide to contribute to raising the machine-checked soundness bound of ‘koalaIRS12,’ a specific Reed-Solomon proximity problem. This challenge is not merely an academic exercise; it directly impacts the security assurances of modern succinct non-interactive proof systems (SNARKs), which are integral to the scalability and privacy of blockchain networks like Ethereum.
The Imperative of Provable Bits: Securing the Digital Frontier
In the rapidly evolving landscape of blockchain technology, the integrity and security of cryptographic protocols are paramount. Many cutting-edge systems, including zk-rollups, zero-knowledge virtual machines (zkVMs), and components central to Ethereum’s post-quantum roadmap, fundamentally rely on SNARKs. These powerful cryptographic tools enable verifiable computation without revealing underlying data, offering solutions for scalability, privacy, and interoperability across decentralized networks. A crucial element within these SNARKs is their dependence on proximity gaps and correlated agreement for Reed-Solomon codes.
Reed-Solomon codes, originally developed for error correction in data storage and transmission, have found a new life in cryptography due to their robust mathematical properties. In the context of SNARKs, these codes are used to encode polynomials, and "proximity gaps" relate to the ability to distinguish between a valid codeword and one that is far from any valid codeword. The security of a SNARK often hinges on the assumed hardness of certain problems related to these proximity gaps.
Currently, many deployed systems, particularly those targeting a robust 128-bit security level, operate under the assumption that these cryptographic conjectures hold true. A 128-bit security level implies that an attacker would need to perform approximately 2^128 operations to break the system, a number considered practically infeasible with current and foreseeable computational power. However, while researchers widely believe these benchmarks are sound, a significant portion of this security often rests on conjectures rather than fully formalized and proven mathematical statements. This discrepancy introduces a theoretical vulnerability: if a conjecture were to be disproven or a weakness discovered, the security guarantee of entire systems could be compromised. The better.codes challenge directly confronts this issue by seeking to transform these conjectures into undeniable, machine-checked proofs, thereby solidifying the foundational security of these critical technologies.
The Genesis: Proximity Prize and Its Vision
The launch of better.codes builds directly upon the broader "Proximity Prize initiative," unveiled by the Ethereum Foundation earlier this year. The Proximity Prize was established with the ambitious goal of either proving or disproving the Reed-Solomon proximity gaps conjectures, recognizing their fundamental importance to the security of next-generation cryptographic systems. This foundational research was notably framed by the influential paper, "Open Problems in List Decoding and Correlated Agreement," authored by leading cryptographers Gal Arnon, Dan Boneh, and Giacomo Fenzi. This paper meticulously outlines the grand challenges within this domain, providing a roadmap for the research community.
The koalaIRS12 problem, which is the immediate focus of the better.codes challenge, is directly derived from the insights and open questions presented in this seminal paper. It represents a concrete, self-contained problem that bridges directly to the overarching grand challenges of the Proximity Prize. By formalizing koalaIRS12 end-to-end within ArkLib – the Lean 4 library specifically designed for formally verified arguments of knowledge – the Ethereum Foundation and its collaborators have created a precise, unambiguous target for researchers. This formalization ensures that every step of the proof process is rigorously checked, leaving no room for ambiguity or error.
Leveraging Formal Verification: The Lean Ecosystem
Formal verification stands as a gold standard in software and cryptographic assurance. Unlike traditional testing, which can only demonstrate the presence of bugs, formal verification aims to mathematically prove the absence of bugs or the correctness of a system with respect to its specifications. In high-stakes domains such as aerospace, medical devices, and now, blockchain infrastructure, where even a minor error can have catastrophic consequences, formal verification is becoming indispensable.
Lean, a powerful interactive theorem prover and programming language, is at the heart of the better.codes verification process. Lean allows mathematicians and computer scientists to express complex mathematical concepts and proofs in a formal, machine-readable language. Its kernel, a small and highly trusted component, is responsible for checking the logical validity of every submitted proof. This rigorous checking mechanism is crucial for the better.codes challenge: every submission that successfully raises the soundness bound must be formally verified by the Lean kernel, ensuring that all progress is undeniably correct and robust. The use of ArkLib further streamlines this process, providing a specialized library within Lean 4 for formally verifying arguments of knowledge, tailor-made for cryptographic research. This commitment to formal verification ensures that any advancements made through better.codes are not merely theoretical improvements but provable, foundational enhancements to cryptographic security.
Introducing better.codes: An Autoresearch Revolution
better.codes is designed as an "autoresearch" challenge, a novel paradigm for open collaboration that leverages the power of automated reasoning and artificial intelligence. In this model, participants don’t merely submit static proofs; they deploy their own AI agents, harnesses, and automated tools to continuously work towards a common, machine-verified benchmark. The platform hosts a public leaderboard, which tracks the highest machine-checked soundness bound achieved for koalaIRS12. Each time a solver successfully submits a proof that raises this bound, that progress is reflected on the leaderboard and becomes the new "floor" for all subsequent efforts.
This iterative and competitive, yet collaborative, approach offers several distinct advantages. Cryptographic research often involves exploring vast search spaces for proofs or counterexamples. No single AI agent or human team possesses the optimal strategy for every facet of such a complex open problem. By enabling many independent setups – each potentially employing different heuristics, algorithms, or proof search strategies – to work in parallel against the same benchmark, the collective frontier of research can advance far more rapidly and efficiently than through traditional, siloed academic efforts. The continuous nature of the challenge means that research is "always-on," fostering an environment of sustained innovation and incremental improvement.
Mechanism of Engagement: How Solvers Contribute
Participating in the better.codes challenge is designed to be accessible to a wide range of contributors with expertise in formal methods, AI, and cryptography. Prospective solvers begin by signing in with their GitHub accounts at better.codes, which grants them access to the challenge repository. This repository contains the core components necessary for participation: the precise theorem statement of koalaIRS12, the specific parameter point being targeted, and a pre-built verification harness.
Solvers are tasked with working within a designated "submission surface" in the repository. Their objective is to prove a larger soundness lower bound for koalaIRS12, with the progress being quantitatively scored in "bits" – reflecting the cryptographic security level achieved. Upon submission, a sophisticated comparator system first verifies that the solver’s exported theorem exactly matches the pinned statement, preventing any deviation from the intended problem. Following this, the Lean kernel rigorously checks the submitted proof. Only submissions that pass this stringent formal verification are accepted and promoted.
Accepted results are then seamlessly integrated into the public repository, with full credit given to the contributing solver and the specific AI model or tools utilized. This transparency is a cornerstone of the autoresearch model. All submissions are git-backed, meaning a complete, auditable history of every accepted proof, including its diffs and submission notes, is publicly available. This open-source approach allows the entire community to examine new lemmas, innovative proof techniques, and even impossibility results discovered by others. Researchers can build directly upon prior work, learn from successful strategies, and avoid "dead ends," thereby accelerating collective progress for all participants, whether human or AI-driven.
Pioneering Collaborative Research: The Autoresearch Model’s Proven Track Record
The autoresearch model, as implemented in better.codes, is not an entirely new concept but rather an evolution of successful collaborative research paradigms. Its effectiveness has been demonstrated in several preceding challenges that have already pushed the boundaries of various technical fields. For instance, platforms like ecdsa.fail have significantly advanced research frontiers in quantum circuit design, tackling complex problems related to the security of elliptic curve cryptography against quantum attacks. Similarly, zk.golf has spurred innovation in verified zero-knowledge circuits, leading to more efficient and provably secure implementations of ZK technologies. Furthermore, snark.fast has focused on optimizing post-quantum proving speed, an area critical for the long-term viability of SNARKs in a quantum computing era.
These precedents underscore the power of open, continuous, and verifiable challenges in mobilizing diverse expertise and computational resources. By providing a clear, machine-checkable objective and a transparent mechanism for progress, autoresearch challenges foster a dynamic ecosystem where competition drives innovation, and collaboration ensures that individual breakthroughs contribute to collective knowledge. The success of better.codes could set a new standard for how fundamental cryptographic research is conducted, accelerating the discovery of new proofs and cementing the security of future digital infrastructures.
Broader Implications for Blockchain Security and Cryptography
The implications of the better.codes challenge extend far beyond the specific problem of koalaIRS12. Successfully proving higher soundness bounds for Reed-Solomon proximity gaps will have a profound impact on the entire field of cryptography and blockchain technology. Firstly, it will enhance the trust and confidence in existing and future SNARK-based systems. As more components of these systems are underpinned by formally verified proofs rather than conjectures, their resilience against sophisticated attacks will increase, making them more attractive for enterprise adoption and critical infrastructure.
Secondly, the challenge will contribute significantly to the academic understanding of these complex mathematical problems. The new lemmas, proof techniques, and impossibility results generated by the solvers will enrich cryptographic theory, potentially leading to the development of entirely new cryptographic primitives or more efficient instantiations of existing ones. This knowledge will be invaluable for cryptographers and computer scientists worldwide.
Thirdly, better.codes reinforces the Ethereum Foundation’s commitment to foundational research and open science. By making the challenge, its progress, and its results entirely public and verifiable, it promotes transparency and ensures that the benefits of this research are accessible to everyone. This open approach aligns with the ethos of decentralized technologies, where security and trust are built on verifiable facts rather than proprietary secrets.
Finally, the success of the autoresearch model itself, as exemplified by better.codes, could inspire similar initiatives across various scientific and engineering disciplines. Imagine open challenges for proving the correctness of AI algorithms, verifying complex hardware designs, or even formalizing theories in mathematics and physics. This collaborative, AI-augmented research paradigm holds the potential to accelerate discovery and innovation on a global scale.
The Path Forward: Sustained Innovation
Today’s launch focuses specifically on the soundness challenge for koalaIRS12, aiming to elevate its proven lower bound to 128 bits. However, the vision for better.codes is expansive. The organizers anticipate adding further challenges over time, tackling other critical open problems in cryptography and formal verification. This modular approach allows the platform to adapt to evolving research needs and to continuously address the most pressing security questions facing decentralized systems.
Participation in the challenge, including eligibility, evaluation criteria, and any potential awards or payments, is governed by program terms that may be adjusted as the initiative progresses. This flexibility ensures that the challenge can remain dynamic and responsive to the community’s engagement and the nature of the discoveries made.
The call to action is clear: the Ethereum Foundation, Yukon, and zkSecurity invite the global community of researchers, formal verification experts, AI developers, and cryptographic enthusiasts to join this critical endeavor. By contributing to better.codes, participants have the unique opportunity to directly influence the security and robustness of the next generation of decentralized technologies, helping to build a more secure and verifiable digital future. The journey to provable security begins now, at better.codes.
