NVIDIA's GEAR (Generalist Embodied Agent Research) lab, together with CMU and UC Berkeley, has developed and released ENPIRE, a framework in which AI coding agents train real-world robots without human intervention. The robot arms learned precise tasks such as inserting GPUs into motherboards and cutting and fastening zip ties, all directed autonomously by the agents.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.