Anthropic's new hardware standard lets AI agents control the physical world

Anthropic has unveiled a new hardware standard called the Model Hardware Standard (MHS) that enables AI agents to control physical devices. This innovation aims to bridge the gap between automated AI systems and the physical world, where typically, AI has been limited to processing text, images, code, and data within computers. MHS is a set of standardized drivers designed to facilitate seamless communication between AI agents and various devices.

The research preview of MHS is primarily marketed as a tool to expedite the creation of custom software integrations for scientists conducting experiments. By providing a uniform interface and data-sharing format, MHS can unite disparate components of an experiment, allowing them to communicate over a network without the need for bespoke translation programs.

The standardized system has the potential to drastically reduce the time spent on experimental setup, from weeks or months to mere hours or minutes. The MHS concept was inspired by observing neuroscientist Arco Bast working on an experiment involving memory formation in the brain at the HHMI Janelia Research Campus in Ashburn, Virginia.

Bast had developed an interface to coordinate various components of the experiment using a common interface. Anthropic's Technical Staffer Alek Kemeny notes that the MHS effort could potentially enable AI to run any science experiment in the world.