Covariant Chief Scientist, Pieter Abbeel welcomes returning guest and “Godfather of AI,” Geoff Hinton to his podcast, The Robot Brains Podcast, to discuss his recent departure from Google to speak freely about the risks of AI and share his perspective on building and regulating AI that is smarter than humans – as well as remind us of the tremendous good AI will do.
Get the full scoop: 00:00:00 Geoffrey Hinton
00:01:46 Sponsors: Index Ventures and Weights and Biases
00:02:45 Backpropagation on digital computers might be better than whatever the brain has
00:06:30 Will the AI take over control from humans?
00:10:10 Predictive AI vs. Goal-Oriented AI
00:13:34 AI smarter than people
00:16:55 Risks of AI
00:19:35 Let's not forget about the tremendous good AI will do
00:22:44 Letter for a 6-month stop to larger AI model development
00:26:28 Role of regulation
00:32:45 The existential threat of AI taking control
00:36:06 AI as the benevolent super-capable “United Nations” advisor to humanity
00:40:03 What would happen if the AI did take control?
00:41:44 Fusing human and artificial intelligence / Neuralink
00:44:00 Assuming AI will take over, is there room to steer how life will be then
00:47:05 The bottom line: the purpose of life
00:54:00 Technical opportunities to contribute towards a good AI-powered future
00:56:18 What are you going to do now?
