PulseAugur
EN
LIVE 08:43:14

World Labs uses simulation to train robots autonomously

World Labs, a startup founded by AI pioneer Fei-Fei Li, has launched a new simulation engine designed to train robot controllers exclusively in virtual environments. This system can generate thousands of variations from a single real-world robotic task, enabling comprehensive training without direct physical interaction. The models trained through this method have demonstrated the ability to operate for extended periods on multiple robot platforms autonomously, though their effectiveness in complex, everyday scenarios is yet to be fully assessed. AI

IMPACT This simulation engine could accelerate the development and deployment of autonomous robots by reducing the need for extensive real-world training.

RANK_REASON Product launch from a startup focused on AI-adjacent robotics tooling.

Read on The Decoder →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

World Labs uses simulation to train robots autonomously

COVERAGE [1]

  1. The Decoder TIER_1 English(EN) · Jonathan Kemper ·

    World Labs turns one real-world robot task into thousands of simulated variations for training

    <p><img alt="" class="attachment-full size-full wp-post-image" height="753" src="https://the-decoder.com/wp-content/uploads/2026/08/worldlabs-r2s2r-06-scenix-brand.jpg" style="height: auto; margin-bottom: 10px;" width="1280" /></p> <p> World Labs, the startup founded by AI pionee…