Fei-Fei Li's World Labs unveils Atlas world model: 3 photos replace 300 for 3D capture

theworldlabs · x · 2026-09-05

World Labs, co-founded by Fei-Fei Li, launched Atlas, a multimodal world model that generates image/video frames with pixel-perfect camera control and reconstructs them in 3D.

Core idea: where LLMs do next-token and video models next-frame prediction, Atlas is built on new view prediction — the first model to unify pixel generation and reconstruction, two problems CV has split for half a century
Practical payoff: 50-100x cheaper 3D capture — a room that used to need 100-300 photos now takes just three
The a16z conversation also covers Matrix-style bullet time done with three iPhones, a one-night Slack message that made them bet the company, and why robotics is data-bottlenecked, not chip-bottlenecked

Original post →

Fei-Fei Li's World Labs unveils Atlas world model: 3 photos replace 300 for 3D capture

More from Multimodal