AI Coding Agents Autonomously Train Robots to Install GPUs and Cut Zip Ties

Researchers at NVIDIA's GEAR lab, in collaboration with Carnegie Mellon University and UC Berkeley, developed a framework called ENPIRE that enables AI coding agents to autonomously train robots on physical tasks like cutting zip ties and inserting GPUs into motherboards. The agents operate with a generous token budget and can work overnight, with researchers reviewing the results in the morning. The ENPIRE harness provides AI models with tools, memory, context, constraints, and feedback loops, allowing for fully autonomous robot training.