Anthropic 与 HHMI Janelia Research Campus 合作推出了 Model Hardware Standard (MHS) 的研究预览版 1。该规范旨在让 AI 代理安全操作显微镜、液体处理器和机械臂等物理设备 1。通过标准化驱动，MHS 能够将设备集成时间从数周或数月缩短至数小时或数分钟 1。该标准还支持 AI 代理自主编排实验、实时调整参数以及进行故障恢复 1。
QuEra 的 AI 代理在 99.3% 的情况下无需人工干预即可恢复激光“锁定” 1。卡内基梅隆大学利用 MHS 运行连续稀释剂量反应实验的速度比此前快了约三倍 1。HHMI Janelia 的 Virginie Ruetten 借助 MHS 统一并编排了此前涉及 7 个不同供应商程序的装置 1。
MHS 支持 MCP、命令行接口和代码文件（API）三种控制机制 1。AWS 将通过 Strands Robots 库支持 MHS，Hugging Face 正在其机器人库 LeRobot 中添加 MHS 支持，Raspberry Pi 正在多款产品中启用 MHS 集成 1。Genentech、华盛顿大学、卡内基梅隆大学、QuEra 等机构，以及 AWS、Automata、Danaher 等厂商参与了早期测试与合作 1。
Anthropic 将提供 500 万美元资助计划，用于资助关于 AI 对用户福祉影响的独立研究 1。
Anthropic and the HHMI Janelia Research Campus have collaborated to release a research preview of the Model Hardware Standard (MHS), a shared specification that enables AI agents to safely operate physical equipment such as microscopes, liquid handlers, and robotic arms 1. By using standardized drivers, MHS reduces equipment integration time from weeks or months to hours or minutes 1. It also supports AI agents in autonomously orchestrating experiments, adjusting parameters in real-time, and recovering from faults 1.
Early testing and collaboration involve institutions and vendors including Genentech, the University of Washington, Carnegie Mellon University, QuEra, AWS, Automata, and Danaher 1. An AI agent from QuEra recovered from a laser "lock" without human intervention in 99.3% of cases 1. Carnegie Mellon University used MHS to run continuous dilution dose-response experiments approximately three times faster than before 1. Virginie Ruetten at HHMI Janelia used MHS to unify and orchestrate a setup that previously involved seven different vendor programs 1.
MHS supports three control mechanisms: MCP, command-line interfaces, and code files (API) 1. AWS will support MHS through its Strands Robots library, Hugging Face is adding MHS support to its robotics library LeRobot, and Raspberry Pi is enabling MHS integration in multiple products 1. Anthropic will provide a $5 million funding program for independent research on the impact of AI on user well-being 1.
评论
还没有评论，欢迎留下第一条。