- Meta introduced the V-JEPA 2 AI model, designed to help AI agents understand their environment, building on the previous V-JEPA model.
- Trained on over one million hours of video, V-JEPA 2 aims to enhance the operation of robots and AI agents in the physical world.
- The model teaches AI to make common sense connections, similar to how young children and animals learn to predict outcomes based on experiences.
- It enables AI to perform tasks by reasoning about actions, such as using a spatula to move eggs from a stove to a plate.
- V-JEPA 2 is reported to be 30 times faster than Nvidia’s Cosmos model, potentially revolutionizing robotics with less dependency on extensive training data.