Oral History of the Scaling Era - Dwarkesh Patel

An inside view of the AI revolution, from the people and companies making it happen. How did we build large language models? How do they think, if they think? Will we be able to make them share our goals--and what happens if we can't? In a series of in-depth interviews with leading AI researchers and company founders--including Anthropic CEO Dario Amodei, DeepMind cofounder Demis Hassabis, OpenAI cofounder Ilya Sutskever, MIRI cofounder Eliezer Yudkowsky, and Meta CEO Mark Zuckerberg--Dwarkesh Patel provides the first comprehensive and contemporary portrait of the technology that is transforming our world. Drawn from his interviews on the Dwarkesh Podcast, these curated excerpts range from the technical details of how LLMs work to the economic, philosophical, and safety considerations of creating artificial general intelligence--AI that can do anything humans can, and more. Patel's conversations cut through the noise and the hype to explore the topics compelling those at the forefront of the field: the power of scaling, the nature of training, the potential for misalignment, the inputs required to achieve AGI, and the economic and social ramifications of superintelligence. At a time of great promise and great uncertainty, The Scaling Era offers readers unprecedented insight into a transformative moment in the AI's development--and a revealing vision of what comes next. \n \nAn inside view of the AI revolution, from the people and companies making it happen. How did we build large language models? Is scaling them up all we need to create general intelligence? How do they think, if they think? Will we be able to make them share our goals--and what happens if we can't? In a series of in-depth interviews with leading AI researchers and company founders--including Anthropic CEO Dario Amodei, DeepMind cofounder Demis Hassabis, OpenAI cofounder Ilya Sutskever, MIRI cofounder Eliezer Yudkowsky, and Meta CEO Mark Zuckerberg--Dwarkesh Patel provides the first comprehensive and contemporary portrait of the technology that is transforming our world. Drawn from his interviews on the Dwarkesh Podcast, these curated excerpts range from the technical details of how LLMs work, to the economic, philosophical, and safety considerations of creating artificial general intelligence. Patel's conversations cut through the noise to explore the topics most compelling to those at the forefront of the field: the power of scaling, the potential for misalignment, the sheer input req