TED Tech 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 of Omniverse and Simulation Technology at NVIDIA, dives into the fascinating interplay between AI and robotics. He explains how robots can now learn physical tasks as quickly as human babies, thanks to innovative simulations inspired by the gaming industry. The discussion highlights the concept of 'mirror worlds' for training, the vast market for physical AI, and how these advancements will redefine our interaction with technology and everyday life, paving the way for a future rich with AI assistance.
AI Snips
Chapters
Transcript
Episode notes
From Hollywood to NVIDIA
- Rev Lebaredian started in Hollywood working on VFX for films like Mighty Joe Young and Stuart Little.
- He joined NVIDIA 23 years ago to bring real-time 3D graphics to gaming and immersive experiences.
NVIDIA's Evolution in AI
- NVIDIA's journey began with accelerated computing using GPUs for 3D rendering in video games.
- The introduction of programmable shading and CUDA laid the groundwork for using GPUs in deep learning and AI.
AlexNet Breakthrough
- AlexNet, a deep learning neural network, revolutionized computer vision by outperforming traditional algorithms in image classification.
- This breakthrough marked the beginning of software writing software.