Meta unveils V-JEPA 2, an advanced AI “world model” to improve robotics and autonomous vehicle technology

Meta has introduced a new AI model called V-JEPA 2, designed as a “world model” that enables machines to better understand and interact with the physical world. By simulating human-like reasoning about 3D environments and object movement, V-JEPA 2 enhances the capabilities of robots and autonomous vehicles such as self-driving cars and delivery robots. Unlike traditional models that depend on massive labeled datasets, V-JEPA 2 uses a more abstract, simplified method to interpret physical interactions, aiming to improve machine planning and navigation. This move positions Meta at the forefront of next-generation AI, focusing on models that move beyond language processing to model reality itself, amid competitive efforts by other tech leaders like Google, OpenAI, and Microsoft.

Read more