The collaborative project is operated as part of a provably.fast workshop, with its underlying code hosted publicly under the starknet-innovation GitHub organization. What sets this mathematical competition apart is its rigorous verification standard: every target is formally pinned within the Lean theorem prover. A problem is officially recognized as solved only when accompanied by an independently reviewed proof, with formal credit attributed directly to the individual human researcher or the specific AI agent responsible for the breakthrough.
To understand the weight of the challenge, it helps to examine the underlying technology. STARKs, which stand for Scalable Transparent Arguments of Knowledge, are advanced cryptographic zero-knowledge proofs. Their primary advantage in modern cryptography is that they rely solely on hash functions rather than complex mathematical assumptions like elliptic curves, making them resistant to potential quantum computing threats and eliminating the need for a trusted setup ceremony.
However, optimizing these proofs remains a major area of active research. The newly announced mathematics challenge specifically targets the number of what cryptographers refer to as "bad challenges" within a Reed-Solomon proximity test. In zero-knowledge proof systems, this test plays a critical role in determining the exact number of queries a verifier needs to make to ensure data integrity. By improving these mathematical bounds, researchers can directly reduce the query complexity, which in turn shrinks the physical footprint of the cryptographic proofs transmitted across the network.
The current baseline for these mathematical bounds stems from research established by Dao, Kominers, and Thaler in their 2026 work. The new challenge seeks to push past these existing limits, opening the door to unprecedented performance optimizations in verifiable computation.
Three targets, verified in Lean
The project’s official portal outlines three specific mathematical targets for participants to pursue. Among these objectives are ambitious goals such as replacing a currently known quadratic bound with a leaner, more efficient linear one, alongside proving a quadratic bound across a substantially wider range of parameters than previously achieved.
The choice of the Lean theorem prover as the arbiter of correctness highlights the project’s commitment to absolute mathematical rigor. In modern computer science and mathematics, interactive theorem provers like Lean allow researchers to encode mathematical statements in formal logic and mechanically verify every single step of a proof. This eliminates human error, ambiguities, or overlooked edge cases that can sometimes plague traditional academic peer review.
Each statement is securely pinned in the Lean environment, meaning that neither human reputation nor AI confidence scores can substitute for hard, verifiable code. A solution is considered valid only after an independent review process confirms that the Lean code successfully compiles and proves the targeted theorem without logical flaws.
The project organizers have also offered an important cautionary note regarding the scope of the challenge. They emphasize that discovering an improved mathematical bound in isolation shrinks nothing on its own. Real-world cryptographic savings require a complete and comprehensive protocol analysis to ensure that theoretical improvements translate smoothly into actual implementation efficiencies within STARK-based systems.
As the challenge gets underway, it represents a fascinating intersection of advanced cryptography, formal verification, and collaborative problem-solving involving both human intellect and autonomous AI agents. By leveraging the Lean theorem prover to secure every step of the process, the initiative aims to push the boundaries of what zero-knowledge proofs can achieve in scalability and efficiency.

Comments (0)