Inside Waymo: The Long Road to Embodied AI & Self-Driving Cars

How close are humanoid robots and self-driving cars to truly scaling?

Vincent Vanhoucke, distinguished engineer at Waymo and former leader at Google Robotics and Google Brain, explains why breakthroughs in embodied AI can happen quickly, while deploying safe, reliable autonomous systems in the real world can take decades.

Vincent joins Brian Heater on Automated to discuss the evolution of deep learning, robotics foundation models, autonomous driving, and the race to build general-purpose robots.

They explore why robotics and artificial intelligence finally converged, how foundation models are changing physical AI, and why scaling a robot is often much harder than inventing one. Vincent also shares lessons from Waymo, Google X, and Google Robotics, including what humanoid robot companies can learn from the long road to fully autonomous vehicles.

This conversation covers:

• What first-time Waymo riders reveal about trust in autonomous vehicles
• Why embodied AI could become the next major wave of artificial intelligence
• Generalist robots versus robots built for specialized tasks
• How foundation models could help robots operate in unfamiliar environments
• Why self-driving cars took longer to scale than many experts expected
• What Google’s robotics moonshots taught the industry
• The safety and scaling challenges facing humanoid robots

Key Moments:
00:00 Inside a Waymo: The psychology of riding in a self-driving car
01:43 Recording a podcast inside an autonomous vehicle
05:08 Why embodied AI could be the next major AI wave
07:17 How to explain complex AI systems to the public
09:16 From speech recognition and deep learning to robotics
12:24 Can robots learn to generalize?
14:50 Generalist robots vs. specialized robots
17:16 How foundation models are changing physical AI
20:21 Shipping robotics products vs. maximizing learning
23:44 Why autonomous driving took longer to scale
27:25 AI startups, industry churn, and creative destruction
28:59 Google X robotics moonshots and lessons learned
35:14 The lasting legacy of Google Robotics
40:02 Why robotics platforms are multiplying
48:55 The hardest problem in robotics: Scaling
50:15 Can humanoid robots scale safely?

Listen to Vincent’s podcast: @Waymo | AI in Motion https://youtube.com/playlist?list=PLCkt0hth826G9AtnOrQsPbKKD5JmdaMXb&si=Nes4oraGJT4d4bia
Connect with Vincent: https://www.linkedin.com/in/vanhoucke/
Find out when Waymo is coming to your town: https://waymo.com/

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.
Also subscribe to the Automated Newsletter.
https://www.youtube.com/@automatedpodcast
https://podcasts.apple.com/us/podcast/automated-with-brian-heater/id1837762221
https://open.spotify.com/show/60olq6brlBEIJWggx2fMR6
https://www.automate.org/automation/automated-newsletter

LinkedIn
https://www.linkedin.com/showcase/automated-podcast-by-a3/
Instagram
https://www.instagram.com/automatedpod/