Maryna Viazovska’s Proofs of Sphere Packing Formalized with AI

The sphere‑packing problem, a centuries‑old question about how tightly spheres can be arranged, found its most elegant solutions in 2016 when Maryna Viazovska proved the optimality of the E₈ lattice in eight dimensions and later the Leech lattice in twenty‑four. These results not only settled a pure‑mathematics puzzle but also linked directly to error‑correcting codes and signal processing, where dense lattice structures improve data transmission reliability. Viazovska’s work earned her the 2022 Fields Medal, underscoring the practical relevance of abstract geometry.

Turning a celebrated human proof into a machine‑checked artifact required translating every logical step into Lean, an interactive theorem‑proving environment gaining traction among mathematicians. The startup Math, Inc. supplied Gauss, a specialized AI that suggested intermediate lemmas and filled gaps that would have taken weeks of manual coding. Gauss enabled the eight‑dimensional proof to be completed in just five days, while the far more intricate twenty‑four‑dimensional case—spanning over 200,000 lines—was formalized within two weeks. This rapid turnaround demonstrates how AI can handle routine, detail‑heavy tasks, freeing researchers to focus on high‑level insight.

The successful formalization has broader implications for the research ecosystem. As proof assistants become more user‑friendly and AI helpers grow smarter, the barrier to verifying complex theorems lowers, potentially accelerating breakthroughs in cryptography, quantum computing, and materials science. Moreover, the collaboration sets a precedent for interdisciplinary teams that blend deep domain expertise with automated reasoning. Investors and academic institutions are likely to increase funding for AI‑enhanced formal methods, anticipating a future where mathematical certainty is achieved at machine speed, reshaping the pace of innovation across technology sectors.