Google Expands Antigravity System to Solve Long-Standing Mathematical Problems and Optimize Software Systems

·by Henderson·AI
Google Expands Antigravity System to Solve Long-Standing Mathematical Problems and Optimize Software Systems
Key Points
  • Google expands the Antigravity system to solve mathematical problems and optimize software systems.
  • The system successfully formalized the Weil conjecture and the proof of the three-dimensional Navier–Stokes equations.
  • Google's approach relies on multiple agents working in parallel to reduce errors.
  • The Antigravity system is also applied to improvements in molecular chemistry and systems engineering.

Google has expanded its Antigravity multi-agent system, utilizing its Gemini model-driven expert team to tackle long-standing mathematical problems, verify complex proofs, and optimize software systems. The company stated that its Teamwork framework helps researchers formalize important mathematical results, identify molecular structures, and develop faster software implementations. This work combines human expertise with the collaboration of multiple agents, which independently explore solutions, test methods, and use computational tools to verify results.

Formalization of the Weil Conjecture

One of the most notable achievements is the formalization of the Weil conjecture, a set of significant mathematical results concerning the number of solutions to polynomial equations over finite fields. Google’s Teamwork-LEAP system generated approximately 1.8 million lines of Lean code through 28,800 calls to sub-agents to formalize these results, which were initially proven by mathematician Pierre Deligne in 1974. Google also reported that the system reproduced the formalization of the latest proof regarding finite-time blow-up of the three-dimensional Navier–Stokes equations, a problem related to one of the seven Millennium Prize Problems.

The question of whether smooth solutions to these equations can develop singularities in finite time has remained unresolved for about 90 years.

The system also assisted researchers in resolving the Courtade–Kumar conjecture, an information theory problem first proposed in 2013. Researchers from Google and academic institutions developed a comprehensive proof using the framework, with the Gemini model helping to derive analytical bounds and generate certificates for computer verification. The proof was formalized in Lean 4, a programming language and proof assistant that allows computers to check mathematical arguments. Google’s approach relies on multiple agents working in parallel, assigning different tasks to each system to challenge and verify proposed solutions, aiming to reduce errors that may arise from a single model generating long chains of reasoning.

Successful Applications in Systems Engineering

Each specialized agent receives only the documents and tools necessary to perform its task. Independent reviewers and auditors check the results without relying on the reasoning of the original agents. Beyond mathematics, Google has applied this method to molecular chemistry. A group of 30 specialized agents collaborated with physical information simulators to identify molecular structures from nuclear magnetic resonance (NMR) data, achieving a 95% top accuracy rate and a 100% accuracy rate for the top three, targeting complex benchmark molecules.

The framework has also yielded results in systems engineering. Google stated that its agents rewrote the company’s libgav1 AV1 video decoder from C++ into approximately 170,000 lines of Rust code, achieving complete output consistency in extensive test suites. The initial rewrite ran about four times faster than the original implementation. When the agents were provided with performance analysis tools and tasked with improving performance, the Rust version became 20% to 30% faster than the original C++ implementation on x86_64 and ARM processors.

In another project, a two-week initiative rewrote the remaining C++ code in the Zircon microkernel into approximately 800,000 lines of memory-safe Rust, passing kernel unit tests and being able to boot on ARM64 hardware.

Google also reported improvements to its internal infrastructure. Its intelligent scheduling workflows reduced preemption of TPU workloads by 24% on certain hardware types while maintaining service level objectives and cluster utilization. The company is offering Teamwork and its iterative coding workflow /boost through Antigravity, opening applications for 10,000 academic researchers to gain six months of free high-usage access. The company stated that additional specialized workflows, including formal proof tools, will be available in the future.

Multiple Applications and Impacts of the Antigravity System

Google's Antigravity system has made significant advancements not only in mathematics but also in molecular chemistry and systems engineering. This demonstrates the system's versatility in combining human expertise with agent collaboration to solve complex problems. Through the parallel work of multiple agents, Google has reduced errors that may arise from a single model, thereby enhancing the accuracy and reliability of results. This innovative approach not only advances mathematical theory but also improves software performance in practical applications, holding significant implications for future technological development.

H
About the author
Henderson