Tools

AI-generated text

Petrichor Plans a Biology Lab for AI Systems

Y Combinator startup Petrichor aims to build a biological laboratory that AI systems can use as a tool.

Petrichor Plans a Biology Lab for AI Systems

Y Combinator startup Petrichor is planning a biological wet lab for use by AI laboratories and the models they train to do science. The company aims to give AI systems access to physical experimentation as a tool, much as they can use web search today.

The proposed system would act as an interface between the digital and physical worlds. A model such as Claude could describe an experiment in natural language or a structured format, allowing an AI agent to submit the work and receive results without a person relaying them. Petrichor aims to return the highest-quality data an experiment can produce, as quickly as biological processes allow.

Faster research and oversight

The concept points toward a scientific equivalent of the nearly fully automated, unstaffed facilities sometimes described in manufacturing as “lights out” factories. Such facilities could accelerate experimentation, while also increasing the need for digital oversight.

A highly capable artificial biological researcher working continuously without human input could bring substantial scientific benefits, but also pose serious risks. Petrichor’s plan therefore raises questions not only about infrastructure for AI-driven biology, but also about how to operate automated science safely.