The future of robotics

Large language models like ChatGPT are transforming robotics by enabling high-level reasoning and task planning, allowing robots to navigate complex, open-ended sequences like household chores. Despite these cognitive advancements, physical execution remains a significant hurdle. Robotic hardware lacks the dexterity and compliance of human hands, and the field suffers from a scarcity of high-quality motion training data compared to the vast datasets powering language models. Consequently, the immediate future of robotics lies in specialized, purpose-built machines—such as hospital logistics robots—rather than general-purpose humanoids. Stanford professor Jeannette Bohg emphasizes that while research focuses on creating more reliable, low-cost platforms, bridging the gap between symbolic reasoning and precise physical manipulation requires overcoming substantial engineering and data-collection challenges.