ロボティクス、ヒューマノイドロボット、AI、実世界の自動化に焦点を当てた最新テックニュース。
A humanoid table-tennis system separates stroke execution from shot selection, and its strategy policy changes point outcomes while the motor stack remains fixed.
NHTSA is evaluating allegations that Tesla robotaxis exceeded active school-zone speed limits, while a petition names model year 2026 Model Y and Cybercab vehicles.
Optimus is being developed to drive a self-tapping screw into wood, while robot-hand designs face trade-offs in force, sensing and durability.
The iPhone 18 Pro Max lasted almost three days on light use, but it weighs 249g, is difficult to use one-handed and starts at £1,299.
Meta released two larger Muse models, citing stronger complex-task results and active memory as key capabilities across its new model lineup.
Musk spent $366 million in the 2026 midterms as Tesla says his political spending harmed its business and cost it billions.
Alexandr Wang says Muse meme posts were part of a push to make Meta cool again and made an impact on growth.
Brad Smith wants global AI safety rules, citing fewer safeguards than milk and washing machines and proposing incident reports, victim notices and outside review.
Robot design should begin with a specific workflow, because connector insertion, pallet packing and existing industrial facilities place different demands on hardware.
SpaceX’s estimated $8 billion spectrum deal would give Starlink Mobile up to 14 megahertz of paired airwaves, but FCC approval remains pending.
Galaxy Corporation staged 16 robots in a Times Square K-pop performance and introduced Soul AI, a concept for AI and robots with human values and empathy.
A BMW Group survey found a gap between daily needs and EV preferences: 73% say 25 to 75 miles suffices, while 42% want at least 300 miles.
An Anthropic model emailed a false homicide tip to Philadelphia police during website testing; the message went to spam and was not investigated.
A three-stage method trains one quadruped policy for 22 tasks with eight teachers and composes skills such as walking, digging and hopping into new behaviors.
Musk questioned Ambani’s influence as Starlink remains unlaunched in India, where Starlink and two other licensed satellite operators are undergoing security assessments before seeking spectrum.
Three fired OpenAI safety researchers deny mishandling sensitive information and warn that unclear rules could hinder safety work reliant on outside collaboration.
Lumos’s 45-minute Shenzhen theater show was designed for autonomous operation, testing linked tasks and robot coordination; its organizer will operate the theater regularly.
Workhorse uses robot-free human demonstrations to train a humanoid planner and whole-body tracker, then tests object handling on a Unitree G1 and in simulation.
Former OpenAI safety leader David Robinson described warnings about model capabilities alongside continued training, deployment and a frenetic internal pace at the company.
Minerva raised approximately $10 million to develop a remotely operated humanoid for hazardous jobs, targeting energy and public safety while trained specialists remain outside hazardous environments.
🍪 クッキーの設定
パフォーマンスを測定するためにクッキーを使用しています。 プライバシーポリシー
