Meta has launched the V-JEPA 2 AI model, designed to equip AI agents with a better understanding of the physical world. Building on last year's V-JEPA model that was trained using over one million hours of video, this new model aims to facilitate common-sense reasoning akin to that of young children and animals. By efficiently predicting actions based on the context—like recognizing when to move eggs from a frying pan to a plate—V-JEPA 2 is positioned as a faster alternative to similar models, such as Nvidia's Cosmos, emphasizing the potential of world models in robotics.
We believe world models will usher a new era for robotics, enabling real world AI agents to help with chores and physical tasks without needing astronomical amounts of robotic training data.
The AI can predict that a very likely next action would be to use the spatula to move the eggs to the plate.
Collection
[
|
...
]