从取物到工厂：为何人形机器人的自主性终于赶了上来
机器人专家详述具身人工智能如何在非结构化环境中实现复杂物理任务，标志着从静态工业臂向适应性通用工人的转变。
leading robotics figures highlighted the critical role of multimodal foundation models in advancing 'embodied ai,' enabling robots to process sensory data and execute physical actions with greater adaptability. unlike traditional automation that relies on hard-coded rules for specific movements, these new systems can generalize learned behaviors to novel scenarios, reducing the engineering overhead for new deployments. this technological leap is driving investment and development toward robots capable of operating safely and effectively in dynamic settings like workplaces and eventually homes. as these systems mature, the industry anticipates a transition from single-purpose industrial machines to versatile autonomous assistants that can collaborate with humans.