Atlas: A World Model for Spatial Intelligence
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
World Labs has introduced Atlas, a pretrained multimodal "world model" that handles text, images, video, and 3D in a shared spatial context. It generates camera-controlled video up to a minute at 1440p, reconstructs scenes from one to over a hundred photos, and outputs 3D Gaussian splats.