Deeply researched interviews www.dwarkesh.com
The conversation explores how to recognize scientific progress, particularly in the context of AI. It questions the traditional view of falsification in science using the Michelson-Morley experiment as an example, where scientists continued to believe in the aether despite evidence to the contrary. Michael Nielsen and ...
The discussion centers on the potential of AI in mathematical research, drawing parallels between Kepler's empirical approach to discovering planetary motion and how AI might uncover new mathematical relationships. Terence Tao and Dwarkesh explore the changing dynamics of scientific discovery, noting a shift from hypot...
The conversation centers on the future of AI compute, particularly bottlenecks in scaling AI capabilities. Dylan Patel, CEO of Semianalysis, provides insights into the semiconductor supply chain, power demands, and capital expenditures of major tech companies. He argues that while power and data centers were previous c...
The podcast addresses the implications of AI development, sparked by the Department of War's designation of Anthropic as a supply chain risk due to their refusal to remove red lines against using their models for mass surveillance and autonomous weapons. It questions whether the US government, in its AI race with China...
Ada Palmer, a Renaissance historian, discusses her book, Inventing the Renaissance, exploring the multifaceted factors that led to the Renaissance. Palmer argues that the period's focus on resuscitating Roman virtues through classical education was initially idealistic but became propagandistic. This movement inadverte...
Dario Amodei, CEO of Anthropic, discusses the progress and future of AI with Dwarkesh. Amodei anticipates AGI within a few years, driven by scaling compute and improved algorithms, but tempers expectations with the realities of economic diffusion. He predicts AI will write 90-100% of code soon, but acknowledges closing...
Elon Musk discusses the future of AI, energy, and manufacturing, arguing that scaling AI will require moving data centers to space due to energy constraints on Earth. He predicts AI will be most economically viable in space within 36 months, leveraging cheaper solar power and eliminating the need for batteries. Musk en...
In this interview, Dwarkesh explores the differences between the human brain and LLMs with Adam Marblestone. They discuss the importance of loss functions, omnidirectional inference, and the brain's steering subsystem. Marblestone introduces Stephen Burns' theories on how the brain encodes high-level desires and connec...
Dwarkesh discusses the tension between short AI timelines and the current approach of reinforcement learning on LLMs, arguing that if AI were close to human-like learning, the current training methods would be unnecessary. He critiques the idea of pre-baking skills into models and questions the plausibility of AI resea...
Sarah Paine delivers a lecture on the multifaceted reasons behind Russia's loss of the Cold War, dissecting arguments ranging from Ronald Reagan's military buildup to the impact of the Helsinki Accords and internal Soviet failings. She explores counter-arguments, including the roles of Presidents Ford, Carter, and Nixo...
In this interview, Ilya Sutskever discusses the current state and future directions of AI research, particularly focusing on the limitations of current models in generalization and the potential for new approaches to achieve human-like learning. He touches on the role of pre-training, reinforcement learning, and the im...
In this interview podcast, Dwarkesh Patel and Dylan Patel interview Satya Nadella, CEO of Microsoft, about the company's AI strategy, infrastructure investments, and its approach to the evolving AI landscape. Nadella discusses Microsoft's focus on building a hyperscale infrastructure capable of supporting multiple AI m...
Sarah Paine discusses Russo-Chinese relations, focusing on historical patterns of continental empires, particularly Russia and China, and their interactions from the mid-19th century to the present. She highlights how both countries have historically destabilized neighbors and created buffer zones, leading to cycles of...
In this interview, Andrej Karpathy discusses his perspective on the progress and future of AI agents, suggesting a more realistic timeline of a decade rather than a single year for significant advancements. He reflects on the history of AI, highlighting key breakthroughs and missteps, and contrasts animal and machine l...
In this interview, Dwarkesh chats with Nick Lane, an evolutionary biochemist, about his work on reconceptualizing life's history through the lens of energy flow. Lane discusses the significance of eukaryotes, the origins of life in deep-sea hydrothermal vents, and the continuity between Earth's geochemistry and early l...
Dwarkesh reflects on a previous interview with Richard Sutton, focusing on Sutton's "Bitter Lesson" essay and its implications for AI development. Dwarkesh interprets Sutton's argument as advocating for AI techniques that effectively leverage compute, criticizing current LLMs for inefficient training methods, reliance ...
Dwarkesh interviews Richard Sutton, a founding father of reinforcement learning, to discuss the differences between the RL perspective and the LLM way of thinking about AI. Sutton argues that LLMs mimic people and lack a true understanding of the world, while RL is about understanding and interacting with the world thr...
In this interview, Sergey Levine discusses Physical Intelligence, a company aiming to build robotic foundation models capable of controlling any robot for any task. He shares the progress made in the first year, including robots performing dexterous tasks like folding laundry and cleaning kitchens. Levine outlines the ...
The podcast features a lecture analyzing maritime and continental powers, particularly Britain's strategic approach in WWII, emphasizing sea control, peripheral theaters, and alliances against continental problems like Germany. It contrasts WWI and WWII strategies, highlighting the importance of blockade, commerce raid...
In this podcast episode, Dwarkesh interviews Jacob Kimmel, president and co-founder of New Limit, about the science of aging and potential interventions. They discuss evolutionary reasons for aging, including hazard rates, kin selection, and optimization constraints, drawing parallels to AI optimization. Jacob explains...
Follow this podcast in Podwise
Sign in to get AI summaries, transcripts and mind maps for any episode, including new ones.
