

Is AI Ready For Robots?
41 snips Mar 31, 2025
Geoff Brumfiel, an NPR science editor and correspondent, shares insights from his visit to Stanford's IRIS lab, where cutting-edge AI is transforming robotics. He discusses how AI is set to revolutionize robots, moving from concept to real-world applications. Brumfiel highlights notable advancements, like teachable neural networks, allowing robots to learn through experience. The episode also dives into the challenges of training robots versus chatbots and explores both the optimism and hurdles in achieving greater autonomy in robotic technology.
AI Snips
Chapters
Transcript
Episode notes
Trail Mix Robot
- Jeff Brumfiel visited Stanford's IRIS Lab to see AI-powered robots.
- Graduate student Moojin Kim demonstrated a robot learning to sort trail mix.
AI-Powered Robotics
- Unlike traditional robots with detailed programming, AI-powered robots use neural networks.
- These networks learn by reinforcing important connections, similar to the human brain.
Robot Reliability
- AI-powered robots, while promising, are prone to errors and confusion.
- They may perform well most of the time, but occasional mistakes require human intervention.