Embodied-ai

1X Technologies has transitioned its World Model from a simulation tool into a generative "cognitive core," allowing the NEO humanoid to perform novel tasks like steaming shirts or operating toilet seats by visualizing them first.

By fine-tuning its π0.6 foundation model, Physical Intelligence has demonstrated a wide array of "common sense" physical tasks—from making sandwiches to cleaning windows—proving that the path to general-purpose robotics lies in scaling physical data.

In a candid season finale of the Google DeepMind podcast, Demis Hassabis outlines why language models aren't enough for robotics and how the synergy between world-generators like Genie and agents like SIMA will bridge the gap to physical AI.

With a fresh $15 million in funding and a massive new dataset, Build AI is doubling down on the bet that human video is the key to universal robot intelligence.

In a detailed new technical discussion, Atlas leadership explains why "hard automation" is failing the automotive industry and reveals the three-pillared data strategy behind the robot''s new AI brain.