Fei-Fei Li (@drfeifei) header
Fei-Fei Li (@drfeifei) profile picture
Stanford, CA, U.S.A.
Joined April 2010
Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)Photo by Fei-Fei Li (@drfeifei)
Fei-Fei Li retweeted
It is very hard to explain how cool Atlas is I am trying my hardest
70
173
17
2,430
167,551
There’s been massive progress in agentic 3D world creation in the past year, but producing polished, high fidelity pixels purely procedurally via code is still orders of magnitude more expensive than generating them directly. I view these strategies as complementary – in our Atlas release today, both the Chisel web app and Codex/Claude Code integrations are designed for using an agent to build out coarse geometry in 3D and control precise final pixel generation through Atlas model endpoints. This demo showcases the Chisel variant of Atlas we finetuned for flexible control, capable of taking in any combination of partial RGB and partial depth frames. We can take masked depth renders of the car model and use text prompting to control both stylistic appearance and what fills in the rest of the scene.
10
30
245
22,190
I really think this is the largest step in 3D generation / reconstruction since the beginning of gen AI. So excited to see it out.
We’re opening up access to Atlas to everyone today. We’ve built apps and API endpoints to help you use Atlas any way you can imagine: bring your ideas for building a 3D world and come away with keyframes, video clips, or 3D Gaussian splats.
16
18
231
31,683
Very excited to show this demo at CoRL! Policy? World model? Value learner (temporal distance)? Why not treat them all as prompts, trained with just next token prediction? Turns out it's a surprisingly strong pretraining objective! We trained one 8B model in-house from scratch (not VLA or WAM), built from an amazing series of works from friends at @NeuroAILab, led by @AtlasKazemian @KlemenKotar And it works zero-shot with new objects and goal images! Come try it out in JW Marriott Lone Star (3rd floor)!
Robot Capabilities as Probabilistic Inference over Next Token Prediction We are doing a demo at CoRL debuting PsiBorg: an 8B universally promptable robot model for open-world manipulation! One model works simultaneously as a policy, a simulator, and a progress tracker. It is trained from scratch, developed fully in-house at #Stanford, not a VLA or WAM, and (tries to) work with new objects and new goal images you throw at it! Come try it at #CoRL2026! 🕛 Mon Nov 9 · 12–6:30 PM CST 🕙 Tue Nov 10 · 10 AM–6:30 PM CST 📍 Demo #2 · Lone Star (3rd floor), JW Marriott Austin 🌐 neuroailab.github.io/psi-web…
8
17
2
152
18,136
Fei-Fei Li retweeted
We're in a computer graphics renaissance Drawings, photos, renders are all something world models can "see into", world vertigo
73
368
27
5,655
164,914
What the hell?!! I can't believe how far we've come. Atlas is so good. Look at this fucking splat!! Check out behind the DJ table. Totally consistent novel view. No ghosting. This is entirely generated from a single image. I'm gonna turn this into such a sick VR experience.
28
23
5
274
28,092
Fei-Fei Li retweeted
We've released a demo of real-time world generation powered by our new Atlas model! This is a spiritual successor to our RTFM demo from last year. It's a vision for the future of interactive media might look like, with frames streamed directly from a powerful generative model
Two months ago, we introduced Atlas, an autoregressive world model with pixel-perfect camera control. Today, we're releasing Frameboy, a demo allowing you to explore worlds in real-time with a highly optimized version of Atlas, streaming from cloud GPUs at up to 15 fps.
4
10
101
14,384
Fei-Fei Li retweeted
Today we’re sharing an experimental release of Spark Radix, a WebGPU-powered Gaussian splat renderer built on three.js. Render and explore detailed 3D worlds directly in your browser. Available today for developers to try and test 👇
39
96
19
756
64,169
AlphaGo and coding agents need reasoning, so does robotics, which is arguably orders of magnitude harder. But what is the right representation to perform reasoning? Excited to introduce DreamCan, led by @GChongkai ! DreamCan performs sparse visual planning, reasoning over sparse visual frames that are both **reachable** by the robot and **makes progress** towards the task goal, and it can learn this capability from action-free suboptimal videos! Check out Chongkai’s post for more details 🧵⬇️
🤖Robots need world models for visual planning, but they don't need to imagine every frame. Introducing DreamCan: Sparse Visual Planning for Long-Horizon Robot Manipulation. Project: dreamcan.github.io/ Code: github.com/HeegerGao/DreamCa… Paper: arxiv.org/abs/2610.24981
5
14
90
18,321
Excited to share Halcyon, our latest research on multimodal pretraining! Building strong text-to-image generative models is simpler than we think! Blog: radicalnumerics.ai/blog/halc… Models: huggingface.co/collections/r… Paper: radicalnumerics.ai/assets/ha…
1/8 Does text-to-image generation need language pretraining? Today, we share Halcyon, a family of early-fusion, base models trained from scratch. Halcyon outperforms strong late-fusion baselines with up to 275× less pretraining compute. Models and recipes are open. Halcyon is part of a research program at Radical Numerics on the science of multimodal pretraining and scaling.
6
32
4
90
20,837
Fei-Fei Li retweeted
Atlas beta access has started rolling out! Here’s a look at how you can control camera paths with pixel-perfect precision.
21
59
14
385
34,421
Experimenting with Chisel in Atlas beta. We’ve been rolling out beta access to a bunch of users and it’s been really cool seeing what people are building.
14
10
76
18,131
Fei-Fei Li retweeted
So excited to welcome @theworldlabs and @drfeifei to the @AMD family! I’ve always been a huge fan of Fei-Fei and her pioneering research in AI. Together, we’ll combine World Labs’ deep expertise in AI and world models with AMD’s compute leadership to power the future of AI and strengthen the open AI ecosystem. Can’t wait for all we’ll accomplish!
210
501
125
5,250
782,225
Fei-Fei Li retweeted
What happens when you give a world model 32 images of a place we know very well? World Labs put Atlas to the test on Voyager. Atlas brings text, images, video, and 3D into a shared spatial context, enabling one model to generate new views, reconstruct scenes, and simulate worlds. With Voyager, you can now see that in action in real time. Pretrained from scratch on NVIDIA Blackwell GPUs. Take a look around 👇
From 32 input images to real-time flight through @nvidia's Voyager headquarters. Trained on NVIDIA Blackwell GPUs, Atlas uses these images as 3D spatial context to generate new views, letting you explore with pixel-perfect camera control. Take a look around.
32
43
7
293
56,418
Fei-Fei Li retweeted
From 32 input images to real-time flight through @nvidia's Voyager headquarters. Trained on NVIDIA Blackwell GPUs, Atlas uses these images as 3D spatial context to generate new views, letting you explore with pixel-perfect camera control. Take a look around.
55
76
29
503
237,745
Fei-Fei Li retweeted
Atlas beta access is rolling out to users, some ideas for API developers - give Astra "eyes" in Blender, let it place and render new camera views with Atlas - 3D reconstruction (this video uses 9 image inputs of my studio to make a scene, any Zillow apartment works) - graybox-to-world, use depth inputs and generate multi-view renders of a scene
21
54
6
610
48,426
Fei-Fei Li retweeted
SparkJS 2.2.0 is out packing tons of bug fixes an improvements. Enjoy! 🎉 🥳 20% faster splat sorting, 14% FPS gains in some systems, 50% smaller bundle size and more. Dive into the full detailed change log github.com/sparkjsdev/spark/…
7
3
34
14,439
Fei-Fei Li retweeted
University leaders should make sure they reason about the broader trend. Around 2020, Fei-Fei and others at Stanford identified the growing compute divide between industry and academia, and the need for policy response in pushing for the National Research Cloud. In 2021, the same point was made when applied to model training in the foundation models paper. In 2022, the same point was made when applied to model evaluation in HELM among other works. In 2026, the same point was made when applied to model use. OpenAI spent more than $6.5 million to resolve Navier Stokes (Astra prices for 130B output tokens only). The stakes are not just the academic computer scientists doing frontier research on how to build AI models. The stakes are about the whole university system doing frontier research. Coping with "scarcity is the mother of invention" won't cut it, and we should trust the "godmother of AI" on that.
The world of R&D is forking into two paths: the token-abundant research, and the token-starved research. The future is in the former - evidentially, the progress by today's top AI industry teams and neolabs is breathtaking, where researchers’ human brilliance is super charged by AI’s assistance. Every research university president should be reading this report and reflecting on what the future of higher education research should be. openai.com/index/research-ac…
4
8
94
34,392
Real time streaming of next view prediction model that has accurate camera positioning and is 3D consistent. Unbelievable.
4
8
3
121
32,021
Fei-Fei Li retweeted
For those who remember @theworldlabs original RTFM demo: Atlas allows us to break out of these small spaces. Here I took the kids playroom and using Atlas-turbo prompted a new outside area into existence.
RTFM, a new generative interactive world model by World Labs, generates real-time video from a single image for exploring 3D worlds. Trained on large-scale video data to predict the next frame. Achieves unbounded persistence via spatial memory. try it: rtfm.worldlabs.ai
2
4
1
47
18,808