Robert Wright's Nonzero A Manhattan Project for AI Safety (Robert Wright & Samuel Hammond)
May 31, 2023
Robert Wright and Samuel Hammond discuss topics such as AI interpretability, natural selection and AI, language models absorbing human nature, singularitarian AI doomers, an AI-safety Manhattan Project, and AI safety blindspots.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7
Introduction
00:00 • 2min
Transition to the Foundation for American Innovation and Interest in Artificial Intelligence
01:48 • 2min
The Concerns Surrounding Generative AI
03:30 • 19min
Representation of Facts in AI Systems and the Role of Interpretability
22:04 • 3min
AI Deception and its Social Implications
24:58 • 12min
Government Collaboration and AI Development
37:01 • 21min
Continuing the conversation in overtime and discussing internet safety debates
57:37 • 4min