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 new framework called ENPIRE that enables AI coding agents to autonomously train robotic arms. The agents successfully taught robots to cut zip ties and insert GPUs into motherboard sockets using a generous token budget for compute resources. NVIDIA's Jim Fan noted that the lab now self-improves tirelessly overnight, with results reviewed each morning.