World Labs says robot learning needs simulation, not just language-model scale
a16z · youtube · 2026-07-28
In this a16z conversation, Fei-Fei Li and Yunzhu Li explain why World Labs acquired SceniX and how simulation could unlock better robotics. They argue that robot training may need a very different pipeline from language models, centered on real-to-sim-to-real loops, world models, synthetic data, evaluation, and foundation models for robotics. The discussion also covers humanoids, semi-structured environments, and what success should look like over the next two years.
More from Embodied
- WorldDiT unifies robot action generation and future frame prediction in one diffusion model — bageldotcom · 2026-07-28
- NVIDIA says Jetson now fits in a bag while powering robots and edge AI — nordicinst · 2026-07-28
- Austin is becoming a major robotics hub with UT Austin, Tesla and Apptronik — lukas_m_ziegler · 2026-07-28
- A flying umbrella prototype follows users with drones and depth sensing — kirbyman01 · 2026-07-28
- Robotics advances could bring more industry back onshore, the post argues — S_OhEigeartaigh · 2026-07-28
- Curriculum learning can keep reinforcement-learning agents from exploiting random luck — ShawnHymel · 2026-07-28