Physical AI

Headlines · 4

1. 4d
   AerosoarX partners on edge-AI drone collaboration

   AerosoarX formally unveiled itself and signed a memorandum of understanding with Rajant and SiMa.ai to combine mesh communications, edge AI computing, and unmanned-vehicle system integration. Rajant will provide Kinetic Mesh technology for communications and information sharing, SiMa.ai will supply its Physical AI computing platform, and AerosoarX will integrate the technologies with its unmanned vehicles. TechNews reported that no integrated swarm-scale, communication-range, anti-jamming, or endurance results were disclosed.

   科技新报 · 🔥 0
2. 5d
   Microsoft defines ‘industrial intelligence’ for manufacturing

   Content produced by Microsoft and published by MIT Technology Review describes “industrial intelligence” as a manufacturer’s accumulated know-how, operating practices, production data and business understanding. It says physical AI and agents can connect this context to factory-floor decisions. The article cites company-reported examples: Toyota Industries says connecting paint-shop data shortened analysis cycles from five days to under four hours and a pilot reduced relevant paint defects by about 25%, while Rockwell Automation says a maintenance copilot reduced new-worker training for troubleshooting hundreds of machines from nine months to three.

   MIT Technology Review AI · 🔥 0
3. Oct 24
   PrismWAM and RoboDoctor advance Physical AI execution

   Li Yiming’s team at Tsinghua University and Liqing Intelligence released PrismWAM, a world action model that separates semantic planning, physical-state rollout, and executable action generation, according to 机器之心. PrismWAM reportedly ranked first on RoboTwin 2.0 with an 83.67% average success rate across two scenario types and a 75.5% Clean-to-Random success rate, while RoboDoctor v1.0 scored 57.10 out of 100 across 54 RoboDojo tasks and led the compared methods. The report says contact and force prediction still require task- and experiment-specific verification.

   机器之心 · 🔥 0
4. Oct 23
   Indian workers record everyday tasks to train physical AI

   BBC Technology reports that workers across India are wearing head-mounted cameras to record tasks such as hairdressing, tailoring, cooking and cleaning, which collectors then sell to technology companies and research organizations for physical AI training. In one example, 22-year-old barber Mohammed Ashim records five or six hours a day, with clients' faces blurred and his hands kept in frame; Rashdan Technologies says it pays $5 to $10 for eight to 10 hours of recorded work. The report also describes concerns about consent, privacy, low pay and possible job displacement as demand for embodied human data grows.

   BBC Technology · 🔥 5

Experience and discussion from the community

Share my Physical AI experienceAsk about Physical AI