Historically, computer vision treated 3D reconstruction (capturing reality) and generation (creating content) as separate fields. New techniques like NeRFs are merging them, creating a unified approach where models can seamlessly move between perceiving and imagining 3D spaces. This represents a major paradigm shift.
Related Insights
While LLMs dominate headlines, Dr. Fei-Fei Li argues that "spatial intelligence"—the ability to understand and interact with the 3D world—is the critical, underappreciated next step for AI. This capability is the linchpin for unlocking meaningful advances in robotics, design, and manufacturing.
Creating rich, interactive 3D worlds is currently so expensive it's reserved for AAA games with mass appeal. Generative spatial AI dramatically reduces this cost, paving the way for hyper-personalized 3D media for niche applications—like education or training—that were previously economically unviable.
Game artists use scanning (photogrammetry) to create ultra-realistic assets. By taking thousands of photos of a real tree from every angle, they generate a 3D model that is a direct digital copy, effectively making the in-game object a "digital ghost" of a real one.
Large language models are insufficient for tasks requiring real-world interaction and spatial understanding, like robotics or disaster response. World models provide this missing piece by generating interactive, reason-able 3D environments. They represent a foundational shift from language-based AI to a more holistic, spatially intelligent AI.
When LLMs became too computationally expensive for universities, AI research pivoted. Academics flocked to areas like 3D vision, where breakthroughs like NeRF allowed for state-of-the-art results on a single GPU. This resource constraint created a vibrant, accessible, and innovative research ecosystem away from giant models.
World Labs co-founder Fei-Fei Li posits that spatial intelligence—the ability to reason and interact in 3D space—is a distinct and complementary form of intelligence to language. This capability is essential for tasks like robotic manipulation and scientific discovery that cannot be reduced to linguistic descriptions.
Current multimodal models shoehorn visual data into a 1D text-based sequence. True spatial intelligence is different. It requires a native 3D/4D representation to understand a world governed by physics, not just human-generated language. This is a foundational architectural shift, not an extension of LLMs.
AR and robotics are bottlenecked by software's inability to truly understand the 3D world. Spatial intelligence is positioned as the fundamental operating system that connects a device's digital "brain" to physical reality. This layer is crucial for enabling meaningful interaction and maturing the hardware platforms.
When analyzing video, new generative models can create entirely new images that illustrate a described scene, rather than just pulling a direct screenshot. This allows AI to generate its own 'B-roll' or conceptual art that captures the essence of the source material.
Human intelligence is multifaceted. While LLMs excel at linguistic intelligence, they lack spatial intelligence—the ability to understand, reason, and interact within a 3D world. This capability, crucial for tasks from robotics to scientific discovery, is the focus for the next wave of AI models.
