Why Humanoid Robots Are Not Ready Yet
Automated Podcast 50:31
1,088 views · 36 likes Watch on YouTube ↗
Humanoid robots are everywhere right now.
But Daniela Rus says the reality is more complicated.
In this episode of Automated, Brian Heater speaks with Daniela Rus, Director of MIT CSAIL, about humanoid robots, physical AI, self-driving cars, embodied AI, on-device AI, robot learning, and the future of robotics.
Daniela explains why humanoid robots can look impressive in short demos, but still struggle with common sense, control, safety, and real-world adaptability needed for human-centered environments.
They also discuss why self-driving cars remain so difficult, how autonomous vehicles helped push the robotics field forward, and why the hardest problems in robotics often come down to the messy, unpredictable nature of the real world.
Brian and Daniela also explore why AI for robots cannot always live in the cloud. For machines that need to make fast, safe decisions, from cars to robots, the future may depend on efficient AI models that run directly on devices.
The conversation also covers physical AI, why robots need to understand forces and torques, how AI is being used to study sperm whale communication, what octopus intelligence can teach robotics, and how generative AI could one day help design custom robots from simple language prompts.
If you want a grounded look at where humanoid robots, physical AI, and the future of robotics are really headed, this is the conversation.
KEY MOMENTS
(00:00) Why robotics is entering an extraordinary moment
(01:25) Introducing Daniela Rus
(03:40) Why the future of robotics is happening now
(04:31) Why research ideas can take decades to arrive
(06:48) Why self-driving cars are still so difficult
(07:35) The long tail problem in autonomous driving
(09:51) How self-driving cars advanced robotics
(10:01) Why LiDAR changed robot navigation
(11:12) How cars could navigate bad weather by looking down
(13:43) The lecture that changed Daniela’s career
(15:41) Daniela’s role leading MIT CSAIL
(17:38) Using robotics and AI to study whales
(20:49) Can AI help decode sperm whale language?
(22:57) Why understanding whales should help protect them
(24:21) The black box problem in large AI models
(27:17) Bringing AI from the cloud down to earth
(30:24) Why robots cannot always wait for the cloud
(31:21) Why physical AI needs to understand physics
(35:01) Why some robot tasks need more than video
(36:41) Why robots have both a body and a brain
(37:10) What octopus intelligence can teach robotics
(39:27) Swarms, decentralized intelligence, and robot collectives
(42:40) Why engineering constraints drive creativity
(44:18) The beginning of an AI engineer for robot design
(47:10) Are humanoid robots overhyped?
(47:34) Why humanoid robots are not ready for prime time
(50:01) Closing thoughts
Connect with Daniela Rus
https://www.csail.mit.edu/person/daniela-rus
Learn more about MIT CSAIL
https://www.csail.mit.edu/
Learn more about Liquid AI
https://www.liquid.ai/team/daniela-l-rus
Learn more about Project CETI
https://www.projectceti.org/
We’d love to hear from you.
Have thoughts or guest suggestions?
Reach us at podcast@automate.org.
You can find the transcript and more episodes of Automated at automated.fm.
Unlock full access to Automated and explore everything automation.
Subscribe today and leave a review on YouTube, Apple Podcasts, and Spotify.
Subscribe to the Automated Newsletter:
https://www.automate.org/automation/automated-newsletter
You can also find us on:
LinkedIn https://www.linkedin.com/showcase/automated-podcast-by-a3/
Instagram https://www.instagram.com/automatedpod/
But Daniela Rus says the reality is more complicated.
In this episode of Automated, Brian Heater speaks with Daniela Rus, Director of MIT CSAIL, about humanoid robots, physical AI, self-driving cars, embodied AI, on-device AI, robot learning, and the future of robotics.
Daniela explains why humanoid robots can look impressive in short demos, but still struggle with common sense, control, safety, and real-world adaptability needed for human-centered environments.
They also discuss why self-driving cars remain so difficult, how autonomous vehicles helped push the robotics field forward, and why the hardest problems in robotics often come down to the messy, unpredictable nature of the real world.
Brian and Daniela also explore why AI for robots cannot always live in the cloud. For machines that need to make fast, safe decisions, from cars to robots, the future may depend on efficient AI models that run directly on devices.
The conversation also covers physical AI, why robots need to understand forces and torques, how AI is being used to study sperm whale communication, what octopus intelligence can teach robotics, and how generative AI could one day help design custom robots from simple language prompts.
If you want a grounded look at where humanoid robots, physical AI, and the future of robotics are really headed, this is the conversation.
KEY MOMENTS
(00:00) Why robotics is entering an extraordinary moment
(01:25) Introducing Daniela Rus
(03:40) Why the future of robotics is happening now
(04:31) Why research ideas can take decades to arrive
(06:48) Why self-driving cars are still so difficult
(07:35) The long tail problem in autonomous driving
(09:51) How self-driving cars advanced robotics
(10:01) Why LiDAR changed robot navigation
(11:12) How cars could navigate bad weather by looking down
(13:43) The lecture that changed Daniela’s career
(15:41) Daniela’s role leading MIT CSAIL
(17:38) Using robotics and AI to study whales
(20:49) Can AI help decode sperm whale language?
(22:57) Why understanding whales should help protect them
(24:21) The black box problem in large AI models
(27:17) Bringing AI from the cloud down to earth
(30:24) Why robots cannot always wait for the cloud
(31:21) Why physical AI needs to understand physics
(35:01) Why some robot tasks need more than video
(36:41) Why robots have both a body and a brain
(37:10) What octopus intelligence can teach robotics
(39:27) Swarms, decentralized intelligence, and robot collectives
(42:40) Why engineering constraints drive creativity
(44:18) The beginning of an AI engineer for robot design
(47:10) Are humanoid robots overhyped?
(47:34) Why humanoid robots are not ready for prime time
(50:01) Closing thoughts
Connect with Daniela Rus
https://www.csail.mit.edu/person/daniela-rus
Learn more about MIT CSAIL
https://www.csail.mit.edu/
Learn more about Liquid AI
https://www.liquid.ai/team/daniela-l-rus
Learn more about Project CETI
https://www.projectceti.org/
We’d love to hear from you.
Have thoughts or guest suggestions?
Reach us at podcast@automate.org.
You can find the transcript and more episodes of Automated at automated.fm.
Unlock full access to Automated and explore everything automation.
Subscribe today and leave a review on YouTube, Apple Podcasts, and Spotify.
Subscribe to the Automated Newsletter:
https://www.automate.org/automation/automated-newsletter
You can also find us on:
LinkedIn https://www.linkedin.com/showcase/automated-podcast-by-a3/
Instagram https://www.instagram.com/automatedpod/
Playback is via YouTube's official embedded player. Data from YouTube; Exumo is not affiliated with YouTube.