Dwarkesh Podcast

Grant Sanderson – AI and the future of math

AI progress in mathematics serves as a leading indicator for how artificial intelligence will transform the economy.

The next big breakthrough will be AIs learning on the job

AI models are moving toward human-like proficiency by learning from real-world tasks and distilling experience.

The data black hole at the center of AI

AI models are achieving brilliance by gorging on a massive black hole of data, masking their lack of true intelligence.

Ada Palmer – Machiavelli is the most misunderstood thinker of all time

Machiavelli was a tactical genius navigating a chaotic Italy where popes and princes constantly upended the status quo.

Alex Imas and Phil Trammell – What remains scarce after AGI?

Economists Alex Imas and Phil Trammell analyze how AGI will impact labor, wage shares, and future economic scarcity.

Michael Nielsen – How science actually progresses

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with researcher Michael Nielsen to explore the often mysterious and non-linear nature of scientific progress. While scientific advancement is often viewed as a steady march toward truth, Nielsen highlights the reality of remarkably long and hostile verification loops, such as the two millennia that passed between Aristarchus’s heliocentric hypothesis and the first successful measurement of stellar parallax. By examining historical figures like Einstein, Darwin, and Prout, the two discuss how science manages to progress even when experimental evidence is ambiguous. The conversation also delves into provocative ideas, including the theory that alien civilizations would likely possess a vastly different technological stack than our own. Listeners can expect a deep dive into the philosophy of discovery, the evolution of scientific principles, and what it truly takes to internalize complex knowledge.

Terence Tao – Kepler, Newton, and the true nature of mathematical discovery

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with renowned mathematician Terence Tao to explore the intricacies of scientific discovery and the evolving role of artificial intelligence in research. The conversation begins by examining the historical brilliance of Johannes Kepler, using his journey toward uncovering the laws of planetary motion to challenge common assumptions about how scientific progress occurs. They discuss the concept of epistemic difficulty, noting that breakthrough theories often face decades of skepticism before they are widely accepted. The dialogue then shifts toward modern technology, as the pair investigates whether AI can truly foster deep understanding or if it primarily serves as a tool for surface-level optimization. Listeners can expect a thought-provoking analysis of human intuition, the future of collaboration between humans and machines in mathematics, and the nature of original insight.

Dylan Patel — Deep dive on the 3 big bottlenecks to scaling AI compute

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with SemiAnalysis founder Dylan Patel for a comprehensive examination of the hardware ecosystem driving the artificial intelligence revolution. The discussion centers on the three critical bottlenecks to scaling AI compute: logic, memory, and power. Dylan provides an insider look at the economics governing the entire stack, detailing the strategic roles of AI labs, hyperscalers, foundries, and semiconductor equipment manufacturers. Listeners will gain a deep understanding of how companies like Nvidia and TSMC navigate production constraints, the impending memory crisis, and why equipment manufacturers like ASML may become the ultimate gatekeepers of progress by 2030. From the geopolitical implications of semiconductor manufacturing to the practical realities of infrastructure scaling, this episode offers a detailed roadmap of the hardware challenges that will define the future of the industry.

The most important question nobody's asking about AI

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel explores the complex intersection of artificial intelligence, government oversight, and civil liberty. The conversation centers on the provocative question of how AI capabilities might fundamentally reshape the relationship between the state and the individual. Patel examines the risks posed by AI in the context of national security, specifically focusing on the tension between Anthropic’s safety goals and the strategic interests of institutions like the Pentagon. The discussion delves into the potential for AI to facilitate unprecedented levels of mass surveillance and the murky reality of alignment: whose values are these systems actually being trained to uphold? Listeners can expect a deep, analytical dive into whether the perceived benefits of state-led AI coordination are worth the long-term costs to personal privacy and the structural risks of authoritarian-leaning technology.

Why Leonardo was a saboteur, Gutenberg went broke, and Florence was weird – Ada Palmer

Historian and novelist Ada Palmer joins the podcast to peel back the layers of the Renaissance, revealing a period far more complex and unexpected than popular history suggests. Through a lens of long-term social evolution, Palmer challenges common assumptions about historical progress and the unintended consequences of human ambition. Listeners will explore why Gutenberg’s printing press only thrived once it reached Venice and how the rise of ephemeral pamphlets accelerated religious and political upheaval. The conversation also examines the irony of the Inquisition and the paradox of Petrarch’s influence, tracing how his attempt to revive Roman virtue inadvertently contributed to the devastating wars of the era, while simultaneously laying the intellectual groundwork for the scientific revolution. This episode offers a fascinating look at how historical shifts are driven by complex technological and social forces rather than the simple intentions of key figures.

Dario Amodei — "We are near the end of the exponential"

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with Anthropic CEO Dario Amodei to explore the near-term future of artificial intelligence. Amodei argues that humanity is approaching the threshold of AGI, envisioning a future where powerful data centers essentially house a country of geniuses. The conversation provides a deep dive into the scaling hypothesis, examining how task-specific reinforcement learning can drive generalization and reshape the broader economy. Listeners can expect a candid look at the technical and strategic roadmap for AI development, including the logistics of massive compute commitments and the long-term path toward profitability for frontier labs. They also debate the implications of international competition and the potential impact of future regulations on the development of these transformative technologies. It is an essential discussion for those tracking the rapid evolution of modern AI.

Elon Musk — "In 36 months, the cheapest place to put AI will be space”

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel engages in a wide-ranging conversation with Elon Musk about the future of technology and infrastructure. The discussion begins with the intriguing concept of orbital data centers, exploring the economic feasibility and strategic advantages of hosting AI infrastructure in space to bypass power limitations on Earth. The conversation transitions into the ambitious plans for xAI, covering both its business trajectory and alignment strategies. Listeners will also gain insights into the complexities of high-volume humanoid manufacturing, specifically the challenges and goals for producing Optimus in the United States. Throughout the episode, Musk reflects on the management lessons learned from leading SpaceX and shares his perspectives on global competition, the Department of Government Efficiency, and the future of industrial automation. It is a comprehensive look at the logistical and technical frontiers shaping the next decade.

Adam Marblestone — AI is missing something fundamental about the brain

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with Adam Marblestone, CEO of Convergent Research, to explore the deep connections between neuroscience and artificial intelligence. Drawing on his unique background in brain-computer interfaces, quantum computing, and research at Google DeepMind, Marblestone argues that the AI field may be overlooking fundamental aspects of how biological brains operate. The conversation centers on how the brain achieves remarkable learning efficiency with limited data and how the genome manages to encode complex, abstract reward functions. Listeners will gain insight into the differences between model-based and model-free reinforcement learning, the potential limitations of biological hardware, and the importance of mapping the human brain. This episode offers a thought-provoking analysis of the bridge between biological intelligence and the future of artificial systems.

Thoughts on AI progress (Dec 2025)

In this episode of the Dwarkesh Podcast, the host provides a thought-provoking analysis of the state of artificial intelligence as of late 2025. The discussion centers on the core mechanisms driving current technological advancements, specifically exploring what is being scaled and the role of reinforcement learning in recent progress. Listeners are invited to consider the economic implications of these developments, including the potential value of human labor and the validity of shifting benchmarks as AI capabilities evolve. The episode also addresses the prospects of a broadly deployed intelligence explosion and whether current theories regarding economic diffusion lags hold up against rapid innovation. This deep dive offers a critical, forward-looking perspective on the trajectory of machine intelligence and what its continued evolution means for the future of productivity and societal structure.

Sarah Paine — Why Russia Lost the Cold War

In this final installment of the lecture series, Dwarkesh Patel sits down with historian Sarah Paine to explore the complex factors that led to the collapse of the Soviet Union. Moving beyond the common narrative that Ronald Reagan ended the Cold War single-handedly, Paine provides a comprehensive analysis of the geopolitical and economic pressures that weakened the USSR. The conversation covers critical developments, including the Sino-Soviet border conflict, the impact of the global oil bust, the rise of ethnic rebellions, and the influential role of the Roman Catholic Church. The episode also touches on Gorbachev’s internal policy mistakes and the complexities of German unification. By examining these historical turning points, the discussion offers valuable insights into international relations that remain highly relevant as the world navigates the potential onset of a new Cold War era.

Ilya Sutskever — We're moving from the age of scaling to the age of research

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with AI visionary Ilya Sutskever to explore the shifting landscape of artificial intelligence. Moving beyond the current era of pure scaling, the discussion focuses on how the industry must pivot toward deep research to achieve the next generation of breakthroughs. Sutskever outlines the strategic vision for his new venture, SSI, detailing how they plan to move toward a straight-shot approach to superintelligence. Listeners can expect a deep dive into the technical hurdles of current pre-training methods, the limitations of model generalization, and the ongoing challenge of aligning powerful AI systems with human values. This conversation provides a comprehensive look at the research philosophies, self-play strategies, and long-term foresight required to navigate the transition toward safe and effective artificial general intelligence.

Satya Nadella — How Microsoft is preparing for AGI

In this episode, Satya Nadella joins hosts Dylan Patel and Dwarkesh Patel for an exclusive deep dive into Microsoft’s ambitious strategy for the era of artificial general intelligence. The conversation centers on the massive infrastructure investments currently underway, including an inside look at the company’s new Fairwater 2 datacenter, which boasts unprecedented computing power and capacity. Listeners can expect a comprehensive examination of how Microsoft is evolving across every layer of the technology stack to prepare for AGI. Nadella addresses critical questions regarding the future of AI business models, the integration of Copilot, the evolution of the hyperscale landscape, and the strategic importance of the OpenAI partnership. The discussion also explores the broader implications of the ongoing capital expenditure surge and the global trust dynamics surrounding American leadership in AI development. It is a definitive look at the technical and economic scale required to build the future.

Sarah Paine — How Russia sabotaged China's rise

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with renowned military historian Sarah Paine to reexamine the geopolitical trajectory of the twentieth century. The conversation centers on the Chinese Civil War, which Paine identifies as a pivotal turning point in modern history. The discussion focuses specifically on the interventionist role of Joseph Stalin, arguing that Russian influence fundamentally sabotaged China’s development and delayed its rise for over a century. Listeners will explore the mechanics of Russian imperialism, the complex dynamics of the Sino-Soviet split, and the existential challenges faced by both nations during this era. By deconstructing Stalin’s strategic objectives and the lasting repercussions of his policies, the episode provides a deep historical analysis of how external manipulation altered the course of one of the world's most significant powers.

Andrej Karpathy — AGI is still a decade away

In this episode of the Dwarkesh Podcast, host Dwarkesh Patel sits down with AI researcher Andrej Karpathy for a deep dive into the current state and future trajectory of artificial intelligence. Karpathy challenges common industry hype, arguing that achieving artificial general intelligence is likely still a decade away. The conversation covers a wide array of technical and philosophical topics, including the limitations of reinforcement learning, the cognitive deficits currently observed in large language models, and the reasons why autonomous driving proved to be such a formidable engineering challenge. Looking forward, Karpathy offers his perspective on how AI will integrate into the broader economic landscape, potentially aligning with historical GDP growth patterns, and shares his vision for the future of education. Listeners will gain valuable insights into the evolution of intelligence, culture, and the practical realities shaping the next generation of AI development.

Nick Lane – Life as we know it is chemically inevitable

In this episode, biochemist Nick Lane joins the Dwarkesh Podcast to explore the fascinating hypothesis that the origin of life is a chemically inevitable consequence of Earth's hydrothermal vents. Lane challenges conventional biological assumptions by connecting fundamental cellular structures to the spontaneous geochemistry of our planet's early environment. Listeners will gain insight into the deep evolutionary puzzles that define life today, including why all organisms share core metabolic pathways like the Krebs Cycle and why proton gradients power cellular energy. The discussion covers the complexities of eukaryotic development, explaining why the transition from simple bacteria to complex life remains the potential great filter for intelligent beings across the galaxy. From the biological necessity of sex to the shared structures found in plants, animals, and fungi, this episode offers a profound look at how basic chemistry shaped the biological reality of our world.