Decart Launches Oasis 3 World Model for Real-Time Photorealistic Environments
AI startup Decart has launched Oasis 3, an interactive world model designed to generate photorealistic, real-time driving simulations for autonomous vehicle testing. The model features remarkable efficiency, though tests show its thematic consistency and physical laws degrade over extended interactions. Decart aims to address these limitations through research into expanded context windows and video-to-video capabilities.
Skynet Chance (+0.01%): Advanced world models could eventually allow autonomous agents to accurately model, predict, and manipulate the physical environment, increasing the risk of uncontrolled physical impacts.
Skynet Date (+0 days): Empowering AI with real-world spatial reasoning speeds up the transition to physically embodied agents, accelerating the potential timeline for physical-world risks.
AGI Progress (+0.03%): The development of real-time, interactive, photorealistic world models represents key progress in spatial intelligence, which is a critical pillar of AGI.
AGI Date (-1 days): Providing APIs for high-efficiency world models allows physical AI and robotics developers to train embodied systems much faster, shortening the timeline to versatile physical AGI.
<< All AI news for June 10, 2026
Related AI News
- General Intuition Nears $6B Valuation Round to Push Spatial Foundation Model Into Robotics 2026-08-24
- Startups Turn to Brain Waves and Muscle Sensors to Break the Robotics Data Bottleneck 2026-07-27
- General Intuition Raises $320M to Build General-Purpose Foundation Models for Embodied AI and Robotics 2026-07-08
- General Intuition Bets on Video Game Data to Train Physical World Models for AGI 2026-07-08
- Bezos-Backed Startup General Intuition Raises $320M to Train World Models on Gaming Data 2026-07-08