Humanoid robots can run, crawl, and sort objects in flashy demos. So why can’t they reliably climb stairs or open doors? On this episode of The Quanta Podcast, host Samir Patel speaks with contributing writer John Pavlus on why robots still struggle with the messy physics of the real world. This topic was covered in a recent story for Quanta Magazine.  

Each week on The Quanta Podcast, Quanta Magazine editor in chief Samir Patel speaks with the people behind the award-winning publication to navigate through some of the most important and mind-expanding questions in science and math.

In this video, Atlas walks, runs and crawls using reinforcement learning. This work was done as part of a research partnership between Boston Dynamics and the Robots and AI (RAI) Institute:

https://youtu.be/I44_zbEwz_w?si=KuKC34o_PiKs8zJP