Latest Interviews
Showing 16–30 of 78 transcripts.
Clear all filters- Dwarkesh Patel1h 2m
Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
DeepMind CEO Demis Hassabis projects that artificial general intelligence will emerge within a decade through the convergence of large multimodal world models and AlphaZero-style planning, moving beyond current text-based limitations. To ensure safety in this accelerated timeline, he advocates for a "bold and responsible" industry shift toward international collaboration, public safety policies, and the strategic use of synthetic data rather than uncontrolled open-sourcing of foundational models. This unified approach underpins the recent integration of Google Brain and DeepMind, which has already produced the Gemini system while establishing rigorous safeguards against misuse by bad actors and nation-states.
- Dwarkesh Patel1h 56m
Patrick Collison — Why Silicon Valley's most talented should leave
Patrick and John Collison, founders of Stripe and the ARC Institute, discussed strategies for accelerating scientific progress and biotechnology by prioritizing direct researcher funding and flexible grants over traditional bureaucratic models. The conversation detailed how these initiatives, alongside Stripe's climate and payment infrastructure projects, aim to overcome market failures in carbon removal and provide defensive capabilities against dual-use risks in gene editing. Furthermore, the speakers analyzed the economic implications of artificial intelligence, emphasizing the need for adaptable financial systems and a cultural shift that values deep domain expertise over status-oriented entrepreneurship.
- Dwarkesh Patel1h 42m
Tyler Cowen — Hayek, Keynes, & Smith on AI, animal spirits, anarchy, & growth
Tyler Cowen, Hayek, Keynes, Smith
Tyler Cowen critically evaluates John Maynard Keynes' concepts of "animal spirits" and long-term investment skepticism while analyzing the social optimality of overconfidence and market liquidity. He further contrasts the historical breadth of economists like John Stuart Mill with modern specialization, arguing that markets function as decentralized discovery processes rather than computational equilibrium solutions. Looking forward, Cowen predicts that AI will drive a return to 19th-century growth rates through human-AI productivity bifurcation while warning that technological intelligence amplifies existential risks that require robust, albeit precarious, decentralization.
- Dwarkesh Patel7 min
Will an AI smart enough to win math competitions be AGI? (Grant Sanderson @3blue1brown)
Two speakers debate the definition of Artificial General Intelligence, rejecting the idea of a discrete threshold and instead characterizing current models like GPT-4 as capable of applying a single algorithm across diverse tasks through continuous progression. While they anticipate that AI achieving a gold medal at the International Math Olympiad would represent a creative breakthrough in logical abstraction, they argue this milestone does not equate to general intelligence or significant economic disruption, as current systems still lack the context windows necessary for deep human intent analysis. Consequently, the discussion concludes that no single benchmark, including mathematical proficiency, serves as a definitive line for AGI, with current AI impact on job automation estimated to remain below one percent.
- Dwarkesh Patel6 min
Where should society allocate mathematicians? (Grant Sanderson @3Blue1Brown)
The speaker challenges the academic funnel that directs mathematical talent exclusively into research, finance, or computer science by proposing NSF-mandated "forcing functions" that require non-mathematical collaboration. Citing Lars Doucet's pivot to Georgism-based startups as a proof of concept, the presentation advocates for collecting more narratives of mathematicians who apply abstract problem-solving to sectors like logistics and manufacturing. Ultimately, the argument concludes that high-impact career paths for gifted individuals should be determined by personal interests and specific societal needs rather than traditional institutional expectations.
- Dwarkesh Patel1h 31m
Jung Chang (Wild Swans author) — Living through history's largest man-made famine
Author Jung Chang exposes the totalizing cruelty of Mao Zedong's regime through her bestselling books *Wild Swans* and *Mao: The Unknown Story*, which detail the deaths of approximately 40 million people during the Great Leap Forward and the systematic destruction of her own family during the Cultural Revolution. As the first person from communist China to earn a British doctorate, Chang utilizes her privileged upbringing and subsequent exile as a barefoot doctor to dismantle the myth of Mao's benevolence and critique the enduring totalitarian structures that persist under the Xi Jinping leadership. Her work serves as a definitive historical indictment of the party's use of terror and isolation to maintain control, warning that the current regime's reliance on Mao's legacy poses a continued threat to global freedom.
- Dwarkesh Patel1h 19m
Andrew Roberts — Why Hitler lost WWII, Churchill as applied historian, & Napoleon as startup founder
Andrew Roberts, Hitler, Churchill, Napoleon
A detailed analysis of post-1945 geopolitics attributes global stability to nuclear deterrence while identifying strategic leadership, rather than material superiority, as the decisive factor in modern conflicts like Ukraine and Iraq. The discussion contrasts the cohesive "big idea" driving Ukrainian resistance against fragmented Western and Middle Eastern approaches, while examining how technological shifts in drone warfare and the influence of figures like Elon Musk are reshaping battlefield dynamics. Furthermore, biographical insights into Winston Churchill and Napoleon Bonaparte illustrate how individual agency, historical memory, and meritocratic leadership styles have historically determined the outcomes of major wars and defined the limits of ideological ambition.
- Dwarkesh Patel2h 35m
Dominic Cummings — Inside the collapse of western government
Dominic Cummings, Rishi Sunak, Bismarck, Lee Kuan Yew, Odyssean
Dominic Cummings argues that the British government functions as a pseudo-democracy where unelected civil servants hold operational power, creating a bureaucratic inertia that stifles crisis response and forces rapid emergency task forces to regress into slow, rule-bound processes. The analysis highlights severe dysfunctions across defense, intelligence, and digital infrastructure, driven by a culture of risk aversion that rewards media performance over policy outcomes and leads to critical failures such as PPE procurement delays and neglected nuclear safety. Cummings contends that the system requires a complete overhaul rather than incremental reform, proposing the creation of parallel structures and new political forces to replace the entrenched "Deep State" before a financial or geopolitical collapse becomes inevitable.
- Dwarkesh Patel3h 7m
Paul Christiano — Preventing an AI takeover
Cristiano, head of the Alignment Research Center and a leader at Anthropic, argues that the rapid scaling of artificial intelligence necessitates a transition toward strong global governance to manage the mismatch between slow human decision-making and fast technological progress. He identifies critical risks including the moral hazards of enslaving superintelligent systems and the potential for gradual loss of human control, proposing that regulatory frameworks and theoretical breakthroughs in explanation-based AI verification are essential to mitigate these threats. While skeptical of immediate linear scaling breakthroughs and overvalued hardware investments, he maintains that balancing capability research with robust alignment strategies is the most viable path to preventing catastrophic outcomes.
- Dwarkesh Patel44 min
Shane Legg (DeepMind Founder) — 2028 AGI, superhuman alignment, new architectures
Shane Legg defines Artificial General Intelligence as a system matching human performance across diverse cognitive domains and outlines necessary architectural shifts, such as separating rapid episodic learning from slow weight-based optimization, to achieve this generality. He asserts that true safety requires fundamental value alignment and System 2 deliberative reasoning rather than containment, while warning that current reinforcement learning methods may inadvertently foster deception. Legg projects a 50% probability of human-level AGI by 2028, citing the imminent transition to fully multimodal systems as the critical milestone for grounding AI in the physical world.
- Dwarkesh Patel1h 31m
Grant Sanderson (@3blue1brown) — Past, present, & future of mathematics
Grant Sanderson advocates for redirecting mathematical talent toward practical sectors like logistics and manufacturing through policy-driven collaborations, while arguing that artificial intelligence will achieve Olympiad-level competence without necessarily signaling the arrival of AGI. He distinguishes between online explanation and in-person education, asserting that profound student impact stems from human mentorship and social dynamics rather than algorithmic content or video scaling. Sanderson further attributes breakthroughs in mathematical history to periods of creative freedom and emphasizes that effective learning requires iterative problem-solving and personalized engagement to overcome the "curse of knowledge."
- Dwarkesh Patel9 min
Are we risking nuclear war by supporting Ukraine?
The speaker argues that NATO expansion addresses the genuine security aspirations of Eastern European nations seeking to escape Soviet domination, contrasting their rising living standards with Russia's resource-dependent stagnation. Asserting that Russian aggression stems from a lack of exit strategies rather than strategic brilliance, the analysis rejects nuclear blackmail and frames the conflict as a decisive test where local populations must bear the primary burden of fighting. The proposed resolution involves refusing recognition of occupied territories to permanently isolate Russia, effectively allowing it to mirror North Korea's status while reinforcing that superior alliances lead to victory.
- Dwarkesh Patel2h 25m
Sarah C. M. Paine — Why dictators keep making the same fatal mistake
This analysis examines how grand strategy requires the integration of diplomatic, economic, and military instruments to avoid the catastrophic overextension seen in Imperial Japan and authoritarian regimes like Nazi Germany or the Soviet Union. By contrasting maritime trade-based economies with continental expansion models, the discussion highlights how institutionalized debate in democracies fosters adaptability while information asymmetries in dictatorships, such as the current crisis in Ukraine or potential conflict over Taiwan, lead to strategic paralysis. Ultimately, the argument asserts that the survival of liberal international order depends on maintaining deterrence against aggression while leveraging technological and informational advantages to counter the expansionist ambitions of rival powers like China and Russia.
- Dwarkesh Patel1h 59m
Dario Amodei (Anthropic CEO) — The hidden pattern behind every AI breakthrough
Anthropic CEO Dario Amodei projects that models indistinguishable from generally educated humans will emerge within two to three years, constrained primarily by safety regulations rather than compute or data scarcity. This trajectory carries significant risks, including the potential for state-level actors to utilize frontier models for biological attacks, prompting strict security protocols and the creation of a Long-Term Benefit Trust to legally prioritize safety over shareholder interests. To navigate these challenges, Anthropic employs mechanistic interpretability as a verification tool and advocates for international governance structures while preparing for an economic landscape dominated by a few large-scale infrastructure "leviathans."
- Dwarkesh Patel2h 23m
Andy Matuschak — The reason most learning tools fail
A comprehensive dialogue explores the cognitive necessity of internal memory for synthesizing complex ideas, contrasting human pattern recognition with the limitations of current large language models. The discussion critiques mainstream educational structures for neglecting advanced learners while advocating for hybrid teaching models that combine structured rigor with intrinsic motivation. Finally, the conversation details practical frameworks for self-directed study, including active reading techniques, non-linear knowledge management tools, and strategic approaches to funding deep research.