理想汽车推出首款“穿戴机器人” 三个维度推进“具身智能”建设

Core Viewpoint - Li Auto has launched AI glasses Livis, marking a significant step towards the vision of becoming an embodied intelligence company, integrating advanced capabilities into existing machinery [1][2] Product Dimension - Li Auto does not position itself as an electric vehicle manufacturer or a smart terminal producer, but rather as an "embodied intelligence" enterprise, aiming to avoid unnecessary competition based on parameters and configurations [1][2] Technology Dimension - The company emphasizes deep self-research to build an AI system that includes perception, models, operating systems, computing power, and the physical body [2][3] - The 3D ViT technology aims to enhance perception capabilities, potentially increasing effective distance by 2-3 times compared to human vision, addressing over 50% of common issues in assisted driving [2][3] - The self-developed M100 chip is designed to tackle various challenges faced by embodied intelligence [3] Organizational Dimension - Li Auto has reverted to a startup management model to enhance agility and efficiency, moving away from a professional manager governance system that was adopted as the company scaled [4][5] VLA Model Capabilities - The VLA (Vision-Language-Action) model presents five core capabilities: spatial understanding, reasoning, communication and memory, behavior, and iteration, enhancing user experience through safety, comfort, and natural interaction [6][7][8] - The model's reasoning ability allows for real-time decision-making at a frame rate exceeding 10Hz, improving driving decisions [6][7] - The VLA model is positioned as a pathway to higher levels of autonomous driving, with rapid iteration expected due to foundational capabilities in algorithms, computing power, and data [8]