HAVE AI NEWS HAVE AI NEWS
Tools

Anthropic Introduces Model Hardware Standard to Connect AI Agents Directly to Physical Machinery

Anthropic Introduces Model Hardware Standard to Connect AI Agents Directly to Physical Machinery

Anthropic has unveiled its Model Hardware Standard in research preview, creating a unified protocol to let AI agents control lab instruments, robotic arms, and industrial hardware.

Bridging AI Agents and Industrial Hardware

Anthropic has announced a research preview of the Model Hardware Standard (MHS), a technical specification designed to connect AI agents directly with physical machinery, automated assembly systems, and laboratory instruments.

Following the footsteps of the company's Model Context Protocol (MCP)—which established a common bridge between AI models and software platforms—the new standard provides a structured layer that enables frontier models to interface directly with physical devices without requiring bespoke, handcrafted code for each integration.

Drastic Cuts to Integration Times

Traditionally, connecting specialized equipment to digital control workflows requires weeks of engineering. Under the new protocol, operators provide natural language descriptions of the machinery's functions, which MHS transforms into structured reference files that models can interpret.

  • Accelerated onboarding: Deployment timelines can drop from several weeks to just hours or minutes.
  • Self-directed calibration: In pilot testing, Claude successfully calibrated laser alignments via trial and error, ultimately producing an automated script to execute the task in a single step.
  • Industry adoption: Anthropic plans an eventual open-source release, backed by collaborations with AWS, QIAGEN, Tecan, Hugging Face, and Raspberry Pi.

UK Surgeons Complete Brain Operation Using Live AI Guidance

In another major breakthrough for applied artificial intelligence, surgeons at University College London Hospitals (UCLH) successfully removed a pituitary tumor using real-time AI video assistance. The model processed live camera feeds during the procedure, identifying delicate optic nerves and hidden blood vessels to help surgeons navigate critical anatomy safely.

Trained on hundreds of surgical recordings, the assistive system provided second-by-second visual checkpoints while human specialists retained total control over surgical execution, helping restore the patient's vision shortly after surgery.

Notable Industry Developments

  • Cybersecurity Coalition: More than 100 organizations, including Anthropic and OpenAI, signed a joint petition highlighting the rising risks of AI-powered cyberattacks and advocating for coordinated defense initiatives.
  • Gemini Upgrades: Google released Gemini Omni 1.1 Flash, introducing 4K upscaling and extended scene-generation capabilities.
  • Enterprise Integration: Salesforce introduced Claudeforce, embedding 37 specialized sales skills into Claude while making the model the default engine within Slack.

Author: Zach Mink

Source: www.therundown.ai

Comments

to leave a comment.

No comments yet.