Smart Girl Dumb Questions How Do Our Brains Work? with the Godfather of AI
12 snips
Dec 2, 2025 
Join Geoffrey Hinton, the 'Godfather of AI' and Nobel Prize winner, as he dives into the fascinating world of artificial intelligence. He discusses how AI compares to the industrial revolution, the differences between neural networks and logic-based AI, and the potential risks associated with powerful systems. Hinton explains how AIs learn and adapt, questioning whether they could possess self-preservation instincts. He even tackles the tantalizing idea of whether we can simply shut them down, revealing the complexities of modern technology.
AI Snips
Chapters
Transcript
Episode notes
Digital Copies Let AIs Share Learning
- Hinton highlights digital AIs can be copied and trained in parallel, letting many instances share learning by averaging updates.
- This makes AI vastly better at sharing experience than humans, accelerating learning despite fewer connections.
Hand-Wiring An Edge Detector Example
- Hinton walks through hand-wiring an edge detector in vision to illustrate learned features.
- He uses this to show how layers build from edges to object parts in a visual hierarchy.
Backpropagation Scales Learning Efficiently
- Hinton explains backpropagation: send the prediction error backward to adjust all connection strengths together.
- Changing many weights at once speeds learning dramatically versus tweaking one weight at a time.