Anthropic Launches Model Hardware Standard for AI-Controlled Lab Equipment
Imagine AI like Claude taking over a robotic arm or laser setup with pinpoint accuracy, no custom coding required. Anthropic's new standard makes that a reality in labs today.
Anthropic kicked off the first phase of its Model Hardware Standard research preview, enabling AI agents to safely operate microscopes, robotic arms, liquid handlers, and lasers using simple commands. Early tests shone: Claude boosted QuEra's quantum laser success from 58% to 99.3%, Genentech automated drug experiments, and a lab ditched 4 a.m. manual shifts for overnight plate handling. Born from work with HHMI Janelia and partners like AWS and Hugging Face, the model-agnostic standard cuts integration time from months to minutes, paving the way for AI-run labs with built-in safety.
Combined views
2.8M
10 posts, first seen 6d ago
Reactions
5 postsAN 
Anthropic@AnthropicAIToday, we're kicking off the first phase of the research preview for Model Hardware Standard (MHS): a new standard for AI agents to safely operate physical equipment in scientific research and advanced manufacturing. Read more:…
- likes: 9.5K
- replies: 417
- bookmarks: 3.6K
- reposts: 1.2K
AN Anthropic@AnthropicAIConnecting AI to hardware requires days or weeks of bespoke integration, with no standard way for agents to operate equipment safely. MHS cuts integration to hours or minutes, provides an interface that makes devices discoverable, and enables agents to operate them safely.
- likes: 588
- replies: 30
- bookmarks: 53
- reposts: 17
AN Anthropic@AnthropicAIIn early testing, AI agents used MHS to: Run a drug-discovery experiment with real-time error handling at Genentech Compress an imaging experiment from weeks to a day at HHMI Janelia Research Campus Improve laser stabilization on QuEra's quantum computers from 58% to 99.3%
- likes: 424
- replies: 9
- bookmarks: 43
- reposts: 10
AN Anthropic@AnthropicAIWatch the story of how the Model Hardware Standard began as part of our collaboration with @hhmi_science
- likes: 791
- replies: 15
- bookmarks: 569
- reposts: 58
LE LeRobot@LeRobotHFWe took part in the research preview of MHS from @AnthropicAI Here is Claude Code running a real SO-ARM101. Nothing in this was trained - no policy, no teleoperation, no demonstrations. The agent measured the workspace itself and wrote the motion. The calibration is the…
- likes: 561
- replies: 18
- bookmarks: 319
- reposts: 73