Nonprofit FutureHouse has published a Nature paper unveiling "Robin", a multi-agent AI system that autonomously generates hypotheses, designs experiments and analyzes wet-lab data. An AIbase report on May 25, 2026 highlights that Robin completed roughly four months (around 900 hours) of human research work in about two hours and identified a candidate treatment for dry age-related macular degeneration.
This article aggregates reporting from 3 news sources. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Robin is one of the first credible demonstrations of an AI “co‑scientist” that doesn’t just summarize papers but runs a full experimental loop end‑to‑end. In the Nature work, FutureHouse shows a multi‑agent system that reads the literature, proposes hypotheses, designs assays, helps run experiments and then closes the loop by updating its models—all in a domain where wet‑lab validation still matters. The fact that Robin surfaced a viable dAMD drug candidate makes it much harder to dismiss as a toy.([tech.ifeng.com](https://tech.ifeng.com/c/8tOFNop2vsU))
The significance for AGI is twofold. First, it pushes the frontier of what language‑model‑based agents can do in domains with expensive, noisy feedback, moving from pure simulation to real-world experimentation. Second, it sets a benchmark for “AI for science” systems that could be replicated across materials, energy and other high‑impact fields. If you can compress months of expert work into hours of agentic orchestration, you dramatically change the throughput of scientific discovery. That doesn’t magically produce AGI, but it accelerates the feedback loop between model capability and the data/insights needed to train the next generation of models.


