Nvidia research shows robots that train themselves through AI coding agents

Researchers from Nvidia, Carnegie Mellon, and UC Berkeley have developed a system where AI coding agents autonomously teach robots how to perform complex physical tasks, such as installing GPUs into motherboards. By using these agents to refine movement strategies through self-directed practice, the robots achieved up to 99 percent success in real-world environments. This method marks a shift toward robots that can learn specialized technical skills independently, potentially reducing the need for manual programming when adapting hardware to new high-precision manufacturing processes.