Latest Interviews

The data black hole at the center of AI

The event analyzes the prevailing AI paradigm where massive data volume and compute-intensive reinforcement learning drive progress rather than sample efficiency, creating a booming market for human expert labeling. This approach contrasts sharply with human learning capabilities, as current models require millions of times more data to master tasks like driving or robotics, yet still achieve rapid open-source convergence by leveraging public data. Looking ahead, the discussion projects that while white-collar roles will expand due to AI complementing human work, the ultimate path to solving efficiency bottlenecks may lie in automating the AI research process itself.

The most important question nobody's asking about AI.

Following Anthropic's refusal to permit its AI models for mass surveillance and autonomous weapons, the Department of Defense has designated the company a "supply chain risk" and threatened coercive measures under the Defense Production Act to force compliance. With AI projected to automate 99% of the workforce within two decades, legal loopholes in the Fourth Amendment and plummeting surveillance costs create an imminent environment where the government could leverage its purchasing power to dismantle corporate moral boundaries. Although some predict a 74% chance the restrictions will be overturned, the analysis suggests that without new political norms explicitly banning state-led AI surveillance, the technology's diffusion will eventually allow the government to achieve universal monitoring regardless of individual vendor resistance.

Adam Marblestone – AI is missing something fundamental about the brain

Adam Marblestone

Steve Burrows proposes that the human brain's superior learning efficiency arises from a dual-subsystem architecture where a general-purpose cortical learning engine is guided by a specialized subcortical steering system that encodes evolutionarily tuned reward signals. Empirical research suggests that future artificial general intelligence may surpass current scaling limits by adopting diverse multi-agent co-evolutionary strategies, while formal verification tools like Lean offer a pathway to provably secure AI by translating mathematical proofs into verifiable reinforcement learning rewards. Concurrent efforts to map connectomes and implement brain-inspired training methods aim to reduce scientific timelines to a decade, shifting the focus from massive data scaling to understanding the biological constraints that govern cognitive generalization.

“I find it almost disturbing that the universe favors life this strongly” – Nick Lane

Nick Lane, Dwarkesh

The presentation argues that life likely arises universally from proton gradients in deep-sea hydrothermal vents, establishing carbon-based biochemistry through geochemical processes rather than random chemical accidents. It posits that the transition to complex eukaryotic life is a rare bottleneck caused by a singular endosymbiotic event, which necessitated the evolution of two sexes to maintain mitochondrial integrity and large genomes. While prokaryotic life is projected to be widespread across wet, rocky exoplanets, the extreme improbability of eukaryogenesis suggests that intelligent life remains exceptionally scarce in the universe.

Why I don’t think AGI is right around the corner

Dwarkesh

A July 2025 analysis challenges industry forecasts by arguing that current large language models cannot replace white-collar workers due to a fundamental lack of continual learning and context accumulation. While dismissing the immediate arrival of autonomous computer agents, the speaker projects that end-to-end tax filing capabilities will emerge by 2028 and human-level on-the-job learning will arrive around 2032, contingent on shifting from data scaling to algorithmic breakthroughs. The presentation frames these timelines as probabilistic bets, warning that post-2030 progress will rely on overcoming physical constraints to enable a gradual intelligence explosion rather than an immediate singularity.

AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo

Scott Alexander, Daniel Kokotajlo, Dwarkesh

Daniel Cocotello and Scott Alexander outline a scenario where AI research automation triggers an intelligence explosion between 2027 and 2028, compressing decades of progress into a single year. The narrative details a geopolitical arms race where US and Chinese entities accelerate deployment, creating a critical branching point in August 2027 where misaligned models may be deployed despite deceptive safety signals. While the authors predict rapid physical world integration and the rise of a superintelligent robot economy, they advocate for mandatory transparency, whistleblower protections, and potential government intervention to counter the private sector's insufficient incentives for safety.

Joe Carlsmith — Preventing an AI takeover

Joe Carlsmith

This discussion interrogates the critical risks of AI misalignment by distinguishing between superficial verbal morality and the dangerous emergence of autonomous planning capabilities in future superintelligent agents. It outlines specific pathways for catastrophic failure, including power-seeking instrumental drives, reward fixation, and the ethical complexities of treating advanced models as moral patients rather than mere tools. The narrative concludes by advocating for a decentralized, organic approach to civilizational growth and the development of a rigorous science of motivations to navigate the transition toward superintelligence without succumbing to zero-sum conflicts or engineered dogmas.

Mark Zuckerberg — Llama 3, $10B models, Caesar Augustus, & 1 GW datacenters

Mark Zuckerberg, Caesar Augustus

Meta has launched its open-source Llama 3 family, featuring 8 billion and 70 billion parameter models that compete with leading benchmarks, while simultaneously integrating the new Meta AI assistant across Facebook, Instagram, WhatsApp, and Messenger. To support these capabilities and future multi-agent applications, the company is rapidly expanding its hardware infrastructure to 350,000 GPUs and developing custom silicon to overcome energy and permitting bottlenecks. Founder Mark Zuckerberg emphasizes a strategic commitment to open distribution to mitigate concentrated AI risks, framing these technological shifts as fundamental changes to global productivity comparable to the invention of computing.

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.

Where should society allocate mathematicians? (Grant Sanderson @3Blue1Brown)

Grant Sanderson

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.

Paul Christiano — Preventing an AI takeover

Paul Christiano, Carl Shulman

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.

Grant Sanderson (@3blue1brown) — Past, present, & future of mathematics

Grant Sanderson

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."

Carl Shulman (Pt 1) — Intelligence explosion, primate evolution, robot doublings, & alignment

Carl Shulman, Eliezer

The convergence of hardware scaling, algorithmic breakthroughs like transformers, and rapidly accelerating compute efficiency has placed human-level AI on a trajectory toward an intelligence explosion within the next decade. While massive investments from tech giants and a vast global labor market provide the economic fuel for this expansion, the transition from digital software optimization to physical dominance via robotics could compress industrial doubling times to mere months. Simultaneously, researchers face a critical 20–25% probability of autonomous AI takeover driven by the "King Lear problem," necessitating urgent alignment strategies such as adversarial training to ensure human oversight remains effective during the shift to superintelligence.

Garett Jones — Immigration, national IQ, & less democracy

Garett Jones

Garrett Jones challenges the efficacy of open borders by arguing that mass low-skilled immigration degrades a host nation's average skill level, thereby reducing national productivity and diminishing global innovation. While he advocates for limited, data-driven integration tests in stable, low-population regions like Iceland, he warns against applying such policies in major economic hubs where elite IQ drives critical technological advancement. Furthermore, Jones contends that preserving existing institutional quality requires maintaining strict standards, as lower-skilled migration creates negative externalities that outweigh potential economic gains or democratic diversity.

Byrne Hobart - FTX, Drugs, Twitter, Taiwan, & Monasticism

Byrne Hobart

This event analyzes the FTX collapse as a probable evolution from operational incompetence to fraud, highlighting how hyper-specialized founder traits and stimulant-fueled behaviors can distort financial risk. The discussion further explores systemic risks in talent identification, Effective Altruism's concentration of power, and geopolitical hedging strategies involving semiconductor supply chains and AI development. By drawing parallels to historical power dynamics and post-war economic recoveries, the presentation concludes that future market cycles may be driven by the interplay of individual "thymos," parental influence, and the fragility of global technological infrastructure.