World Labs co-founder Justin Johnson on world models and the frontier of spatial AI
TWIML AI Podcast · rss · 2026-09-02
TWIML AI podcast interviews Justin Johnson, co-founder of World Labs, on world models and spatial AI.
- Why it matters: many researchers see capabilities beyond language as a key frontier — models that understand, generate, and simulate environments.
- Approaches: explicit 3D representations vs. generative models; there is still no established recipe for building these systems.
- Marble: World Labs' Marble generates navigable 3D worlds from images and other inputs; evaluating world models remains hard.
- Vision: combining simulation, planning, and action to support interactive virtual environments and agents/robots operating in the physical world.
More from Embodied
- World Labs Demonstrates Atlas Connecting World Models to Robotics — drfeifei · 2026-09-02
- Atlas Robot Reconstructs San Francisco's Sutro Baths — drfeifei · 2026-09-02
- Extending AI Agents to the physical world with ESP32 — purellmagents · 2026-09-02
- VLM orchestrator enables in-context learning to bootstrap RL post-training in robots — oanacamb · 2026-09-02
- Dyson launches $499 AI toothbrush with built-in camera — Polymarket · 2026-09-02
- Expert: Blue-collar jobs like roofing safe from robots for 20 years — binarybits · 2026-09-02