Generalist AI Launches GEN-1 Robot Model

California-based startup Generalist AI has launched GEN-1, an embodied foundation model designed to give robots physical intuition ^[1][2]. Moving beyond rigid code, the system allows machines to learn how objects behave in the real world ^[1]. Industry experts describe this development as a potential "ChatGPT moment" for the robotics sector ^[3]. The model achieves a 99% success rate in high-dexterity tasks such as folding laundry and assembling boxes ^[1]. It works by predicting object physics and improvising to correct mistakes in real time ^[1]. Generalist AI also uses specialized training gloves to help robots learn from human movements ^[3]. This approach marks a shift toward general-purpose models that can handle various physical environments ^[2].