Advancing science and engineering with Antigravity's /teamwork (research.google)

Google has unveiled significant advancements in multi-agent systems through Antigravity’s /teamwork initiative, enhancing human-AI collaboration in open research across multiple fields including mathematics, chemistry, and machine learning engineering. This initiative allows researchers to tackle complex problems collaboratively, utilizing Google’s Gemini 4 models to produce groundbreaking results such as new insights in information theory and optimization efforts that address challenges that have persisted for years. By making these tools accessible externally, Google aims to empower a broader community of researchers and developers. The /teamwork framework features an innovative design that addresses the limitations of current AI models, particularly in generating reliable outputs over extended tasks. It employs parallel exploration and specialized agents tasked with specific duties, thereby maintaining focus and accuracy. Noteworthy achievements include the formalization of significant mathematical proofs, such as the Courtade–Kumar conjecture and advancements in quantum computing algorithms. This initiative is poised to transform how researchers approach complex open problems, streamline verification processes, and ultimately enhance the quality of scientific discovery.