Nvidia research shows robots that train themselves through AI coding agents

Nvidia, in collaboration with Carnegie Mellon University and UC Berkeley, has developed AI coding agents that enable robots to autonomously learn dexterous grasping skills, achieving up to 99% success on complex tasks. This approach leverages self-training mechanisms, which could significantly enhance the efficiency and adaptability of robotic systems in real-world applications. For practitioners, this research highlights the potential of integrating AI coding agents into robotic training pipelines to improve performance and reduce the need for extensive human intervention.