An AI system to help scientists write expert-level empirical software - Nature
Google just shipped AI agents that write production-grade research code. Scientists aren't waiting for the perfect model—they're shipping.

Why it matters
Google DeepMind is moving AI from benchmark claims to real-world deployment in scientific workflows. This is the drop that proves 'good enough' models can drive immediate productivity gains in high-stakes domains.
The key facts
6 to knowGoogle announces multiple AI tools for scientific research: Gemini for Science, Empirical Research Assistance (ERA), and Co-Scientist multi-agent system
Tools designed to help scientists write expert-level empirical software
Published in Nature—signals academic credibility and peer-review validation
Multi-agent approach (Co-Scientist) indicates reasoning/planning capability in production use
Focus on research workflow optimization, not model capability claims
Published May 2026—signals active 2026 product roadmap execution
Go to the source
Reuters Technologynews.google.com
Publisher excerpt: An AI system to help scientists write expert-level empirical software Nature Gemini for Science: AI experiments and tools for a new era of discovery blog.google Empirical Research Assistance (ERA): From Nature publication to catalyzing Computational Discovery Research at Google Co-Scientist: A…
