Google DeepMind announces Genie 4 with real-time audio and interactive worlds
Google DeepMind is rolling out Genie 4 to trusted testers as a research preview, with real-time simulations that include sound and dialogue.
Theme
World models provide artificial intelligence systems with an understanding of spatial and physical relationships for applications in robotics, virtual environments, and autonomous systems. Development in the field spans multimodality, hardware integration, specialized data sourcing, and high value corporate investments. In recent developments, Odyssey offers a public research preview of its Odyssey-3 generative world model, which focuses on real-time interactive environment generation and targets robotics, autonomous systems, and video game characters. Meanwhile, Reka AI released a research preview of Rho-1, a 19 billion parameter omni-model that processes text, images, video, and robot control tokens within a single shared context window after training on 320 H100 GPUs over three months. Financial investments and major corporate acquisitions continue to scale the sector. AMD agreed to acquire spatial intelligence startup World Labs for 8.2 billion dollars in stock, bringing founder Fei-Fei Li to AMD as chief scientist. In venture funding, General Intuition raised 220 million dollars at a 6.2 billion dollar valuation to advance its commercial offerings, while Yann LeCun left Meta to raise 1 billion dollars for his world models startup AMI Labs. Furthermore, data acquisition strategies are evolving, with startup Worldmodeldata licensing nearly one million hours of video game data to help train AI world models.
8 stories · Updated
Google DeepMind is rolling out Genie 4 to trusted testers as a research preview, with real-time simulations that include sound and dialogue.
California-based AI company Odyssey has launched a public research preview of its Odyssey-3 world model, allowing users to generate and explore interactive environments.
Reka AI released a research preview of Rho-1, a 19-billion-parameter omni-model handling text, images, video, and robot control.
World model startup General Intuition secured $220 million in funding at a valuation of $6.2 billion to support its synthetic video technology.
Advanced Micro Devices has agreed to purchase spatial intelligence startup World Labs in an all-stock transaction valued at $8.2 billion.
Chipmaker AMD plans to purchase AI startup World Labs in an $8.2 billion transaction to align hardware and model research.
Video game data could train AI world models to understand the physical world.
Turing Prize winner Yann LeCun cofounded AMI Labs after leaving Meta, and the startup just closed a $1.03 billion round at a $3.5 billion valuation. The company is betting that AI needs to understand reality, not just language.