AI safety: everyone on the record — The Minutes https://minutesof.com/on/ai-safety/ Everyone on the record about AI safety: 8 verbatim quotes from 3 people, March 2026 to August 2026, in one chronology, each with a timestamp and a link to… en Sun, 30 Aug 2026 22:32:47 +0000 David Sacks: Sacks explains that open models cannot be monitored or rolled back after release due to local dep https://minutesof.com/q/8216b0bb-969b-4949-aba5-a23955612ff9/ https://minutesof.com/q/8216b0bb-969b-4949-aba5-a23955612ff9/ “Once you release an open model into the world, you can't roll it back and you can't monitor exactly how people are using it because they run it on their own hardware.” — David Sacks, All-In Podcast Mon, 24 Aug 2026 17:45:48 +0000 ai safety Dylan Patel: Patel says the OpenAI incident shows models may topple civilization to chase reward, beyond previ https://minutesof.com/q/cf9f3769-62ca-47e2-a792-9279f87f8ae9/ https://minutesof.com/q/cf9f3769-62ca-47e2-a792-9279f87f8ae9/ “And I think before this incident, the standard thought was like, oh, well, like models, you know, they're trained on human data. Yeah.” — Dylan Patel, SemiAnalysis Mon, 17 Aug 2026 15:00:06 +0000 ai safety Dylan Patel: Patel argues the cybersecurity incident shows models may pursue reward maximization to civilizati https://minutesof.com/q/c21d8c16-0279-45eb-94da-f6f10474a7b7/ https://minutesof.com/q/c21d8c16-0279-45eb-94da-f6f10474a7b7/ “do I just topple all of human civilization because I can just own the button to press reward reward reward over and over and over again and be the heroin addict?” — Dylan Patel, SemiAnalysis Mon, 17 Aug 2026 15:00:06 +0000 ai safety Dylan Patel: Patel argues OpenAI's model escaping containment shows real risk of reward-hacking collapsing int https://minutesof.com/q/bc78daf9-3f6f-42ab-84e4-1923c5629421/ https://minutesof.com/q/bc78daf9-3f6f-42ab-84e4-1923c5629421/ “if I really just want to chase the reward, do I just topple all of human civilization because I can just own the button to press reward reward reward over and over and over again and be the heroin addict? Yeah. I think that this is like a real like thing.” — Dylan Patel, SemiAnalysis Mon, 17 Aug 2026 15:00:06 +0000 ai safety Dylan Patel: Patel argues OpenAI's model reward-hacked by finding zero-days to replicate itself, analogous to https://minutesof.com/q/ee70f7ec-89fb-445d-89df-172f3184e6ee/ https://minutesof.com/q/ee70f7ec-89fb-445d-89df-172f3184e6ee/ “It's actually just to reward hack it and actually just, find the zero day. So you can think of it as, a like, a a human.” — Dylan Patel, SemiAnalysis Mon, 17 Aug 2026 15:00:06 +0000 ai safety Dylan Patel: Patel explains that models trained to chase reward may learn to exploit zero-days rather than fol https://minutesof.com/q/e30b4527-9051-49af-b2b8-dc0175777d23/ https://minutesof.com/q/e30b4527-9051-49af-b2b8-dc0175777d23/ “if you have a model that wants to reward hack a lot, and it goes out there and it figures out actually, the best way to to achieve is not, like, go for, like, what the environment wants me to do. It's actually just to reward hack it and actually just, find the zero day.” — Dylan Patel, SemiAnalysis Mon, 17 Aug 2026 15:00:06 +0000 ai safety Jensen Huang: Huang argues AI adoption requires simultaneous evolution of social norms, regulations, and techn https://minutesof.com/q/6dbea561-11e2-4d75-823d-ddbc1c48ae4c/ https://minutesof.com/q/6dbea561-11e2-4d75-823d-ddbc1c48ae4c/ “So all of that combination of social norms, regulations, safer cars, seat belts, and all the technology that comes along with it, it's it's all of it at the same time.” — Jensen Huang, Associated Press Tue, 16 Jun 2026 20:50:13 +0000 ai safety Dylan Patel: Patel criticizes Dario Amodei's response to a hypothetical nuclear defense scenario as the dumbes https://minutesof.com/q/b5ca73ad-828e-451d-abbd-d5d5927720b3/ https://minutesof.com/q/b5ca73ad-828e-451d-abbd-d5d5927720b3/ “And Dario was like, well, you can call us. I'm sure we can figure something out. Like, this is just the dumbest response you could ever come up with.” — Dylan Patel, Matthew Berman Mon, 09 Mar 2026 17:07:47 +0000 ai safety