Fei-Fei Li's World Labs debuts Marble, its first world model, which lets users turn prompts, photos, and other media into editable 3D environments, after a beta
World Labs, the startup founded by AI pioneer Fei-Fei Li, is launching its first commercial world model product.
Users can generate editable 3D environments from prompts, photos, and other media through World Labs’ first commercial world-model offering.
World Labs shifts from beta validation to product delivery, giving its world-model research a direct user-facing interface.
Second-order effects
Editable outputs raise the competitive bar for AI 3D-generation products: producing a scene alone is less differentiated when users expect to modify the result.
The product creates a clearer route into design and content workflows, where compatibility with existing editing processes will matter alongside generation quality.
Third-order effects
If world models reliably produce editable environments from varied inputs, 3D creation could move from specialist scene-building toward AI-assisted spatial workflows.
That shift would favor vendors that can pair spatial generation with controllability and workflow integration, rather than treating 3D output as a standalone demo.
The trend: World-model developers are moving from demonstrations of spatial generation toward editable products designed to fit real creation workflows.
AI's next frontier is Spatial Intelligence, a technology that will turn seeing into reasoning, perception into action, and imagination into creation. But what is it? Why does it matter? How do we build it? And how can we use it? Today, I want to share with you my thoughts on [vid…
If you're new to the coming concept of World Models, give this piece by @drfeifei a read. These models will be generative, multimodal, and interactive. They'll support breakthroughs in robotics, science, and many fields of entertainment such as interactive gaming.
In this new essay, I lay out the framework for what truly spatially intelligent world models must achieve—three essential capabilities for building AI that creates with a storyteller's imagination, navigates with a first responder's fluency, and reasons about space with
if you've been paying attention to richard sutton on “the era of experience” or karpathy on the limitations of LLMs then this piece by @drfeifei on world models and spatial intelligence is basically the next step of that conversation. amazing piece. [image]
spatial intelligence is “the scaffolding upon which our cognition is built” and contributed to multiple civilization-defining moments from @drfeifei's manifesto “from words to worlds”, released under @theworldlabs [image]
Proud of @drfeifei for defining what spatial intelligence means for our future: AI that isn't limited to language, but can perceive, reason about, and interact with the physical world. It's the next frontier of AI! #FutureOfAI #SpatialIntelligence https://drfeifei.substack.com/ .…
Most ppl I talk to about world models claim robotics are the ultimate application. I also thought this before digging deeper. While robotics (closing the perception-action loop) is important, the more impactful application is embedding spatially-intelligent world models that can
“AI has never been more exciting. Generative AI models such as LLMs have moved from research labs to everyday life, becoming tools of creativity, productivity, and communication for billions of people. Yet they remain wordsmiths in the dark; eloquent but inexperienced,