Anthropic announced a new Machine Hardware Standard, or MHS, designed to bridge the gap between large‑language models and physical equipment. The standard equips AI systems like Claude with a structured description of a device’s capabilities and limits, allowing the model to issue commands, read sensor data and adjust actions in real time.
In a demonstration video, Claude adjusted a laser, used a separate camera to verify the output, and repeated the cycle until the system was fully calibrated. The same approach let the model focus a microscope, analyze the captured image, decide which region required closer inspection and then reposition the instrument automatically.
Another clip showed Claude reasoning through the steps needed for a robotic arm to pick up an aluminum can—something the model had never been explicitly trained to do. Rather than hard‑coding each motion, MHS lets the AI write and modify API scripts on the fly, adapting them as conditions change.
The backbone of MHS is a standardized tagging system that encodes a device’s physical characteristics—weight, range of motion, adjustable parameters, measurement options and built‑in safety limits. These tags are compiled into a reference file that any compatible AI model can read, giving it crucial context about hardware it has never encountered before.
Anthropic is currently running a preview with a select group of scientific research labs and advanced manufacturers. Partners include Amazon Web Services’ Strands Robots, Hugging Face’s LeRobot, Raspberry Pi, Automata and Universal Robots. Together they are developing safety evaluations and best‑practice guidelines for AI‑driven operation of physical systems.
Early testing over the past year suggests MHS can dramatically cut the time required to integrate new devices into AI workflows. Researchers reported faster iteration cycles in a variety of experimental settings, allowing hypotheses to be tested more quickly. "If you can test hypotheses faster, you could create general technologies faster," said Anthropic co‑founder Kemeny in the promotional video. "This is how a century of progress can condense into a decade."
Looking ahead, Anthropic plans to release MHS as an open‑source, agent‑agnostic standard, inviting broader adoption across industries that rely on robotics, laboratory automation and other hardware‑intensive applications.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.