The TED AI Show cover image

The TED AI Show

How AI robots learn just like babies — but a million times faster w/ NVIDIA’s Rev Lebaredian

Dec 3, 2024
Rev Lebaredian, Vice President at NVIDIA, specializes in simulation technology that teaches robots physical intelligence. He discusses how the gaming industry has revolutionized robot training, allowing them to learn in simulated environments. The conversation dives into the potential trillion-dollar market for physical AI and how robots are making their way into our homes and industries. Lebaredian also touches on the cultural attitudes shaping robotic adoption and envisions a future where advanced AI fosters abundance for everyone.
53:27

Episode guests

Podcast summary created with Snipd AI

Quick takeaways

  • Robots are leveraging simulated training environments to gain physical intelligence and learn tasks millions of times faster than traditional methods.
  • NVIDIA’s transition from gaming to robotics showcases its innovative approach and commitment to improving AI applications across diverse industries.

Deep dives

The Challenge of Physical Intelligence

Although AI has excelled in areas like text and image generation, it struggles with physical intelligence, which is crucial for robots to interpret and interact with the real world. Humans have an inherent advantage in this area due to years of physical interactions and learning experiences that develop our understanding of object dynamics. For example, when catching a fly ball, a person's brain quickly calculates the ball's trajectory using prior experiences of similar interactions. By contrast, robots have limited real-world practice but can benefit from simulated training environments that accelerate learning in a controlled setting.

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner