Latest Interviews

What are we scaling?

Toby Ord, Beren Millidge

Baron Millage argues that current Reinforcement Learning strategies rely on inefficiently pre-baking skills into models due to a fundamental misunderstanding of their ability to learn like humans, which keeps AI revenue far below the potential of knowledge work automation. While the industry anticipates a 2030 surge in continual learning revenue reaching the hundreds of billions, the lack of generalizable on-the-job capabilities and the immense compute requirements for RL scaling suggest AGI remains distant despite incremental progress. This perspective challenges the "superhuman researcher" narrative by emphasizing that solving the core learning problem requires a shift from specialized training loops to systems capable of semantic, self-directed adaptation.

Some thoughts on the Sutton interview

Richard Sutton argues that current AI paradigms are inefficient because they rely on finite human data for static training rather than enabling continual, on-the-fly learning like biological systems. In response, the speaker contends that while human data acts as a necessary transitional "fossil fuel," it complements rather than opposes reinforcement learning and already facilitates world-model capabilities. Although Sutton correctly identifies current gaps in sample efficiency, the speaker predicts that while immediate successors remain LLM-based, future architectures will inevitably evolve to satisfy Sutton's vision of autonomous, continuous learning.

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.

Satya Nadella shows me the first Majorana 1 Quantum Computing chip

Satya Nadella

Microsoft has achieved a pivotal physics breakthrough by fabricating Majorana zero modes, creating the stable topological qubits necessary to build utility-scale quantum computers. This foundation supports the upcoming "Majorana 1" chip, which aims to host a million physical qubits capable of scaling to thousands of logical, error-corrected units by 2029. The initiative integrates quantum simulation with AI-driven emulation to accelerate discovery in chemistry and materials science while targeting a strategic shift from classical high-performance computing for specific data-light tasks.

The Limits of American Power – Sarah Paine

Sarah Paine

The speaker argues that U.S. policy in 1940s China failed due to a catastrophic misallocation of resources compared to the Marshall Plan, exacerbated by the absence of indigenous institutions required to sustain foreign aid or enforce decommunization. Historical warnings from competent Foreign Service officers regarding the infeasibility of a Nationalist-Communist coalition were suppressed during McCarthy-era purges, preventing the U.S. from recognizing that Chiang Kai-shek lacked the necessary peasant support to defeat the Communists. Ultimately, the intervention is presented as a strategic error driven by the inability to compel primary adversaries to share power and the domestic political constraints of an isolationist American public unwilling to fund an indefinite civil war.

"The Brilliance of Communism" – Sarah Paine

Sarah Paine

This analysis examines the mechanisms of Communist power consolidation through Mao Zedong's purges of rivals like Peng Dehuai and Liu Shaoqi, which triggered the Cultural Revolution to preempt posthumous criticism. The speaker details how Xi Jinping's reverence for Mao and economic re-centralization stem from his traumatic upbringing during that period and a psychological drive to emulate Stalin and Hitler. While these regimes are credited with seizing power but failing to deliver prosperity, the narrative argues that modern public adherence to such leaders persists due to a collective national tendency to avoid confronting historical crimes similar to the U.S. legacy of slavery.

Was The War Against Japan Avoidable? - Sarah Paine

Sarah Paine

Historian discussions analyze how the Smoot-Hawley Tariff fostered global economic instability that enabled militarist factions in Japan, led by War Minister Tojo, to assassinate pro-peace figures and drive nations into World War II. The dialogue examines the United States' strategic dilemma in responding to the oil embargo, weighing the moral necessity of denying resources to an aggressor against the risk of triggering a broader conflict or inadvertently aiding a potential Nazi victory in Eurasia. Ultimately, the analysis concludes that Japan's modern post-war strength was a contingent outcome rather than an inevitable result, highlighting the extreme difficulty of predicting how specific policy choices might have altered the war's trajectory.

Why Space Elevators Can't Mine Black Holes – Adam Brown

Adam Brown

Physicists have determined that while black holes offer a theoretically perfect mass-to-energy conversion, mechanical mining of solar-mass black holes is physically impossible because even carbon nanotubes lack the tensile strength required to extract Hawking radiation at a useful rate. Instead, efficient energy extraction requires a civilization to ingest baryonic matter into a small, controlled black hole, a process that bypasses the conservation laws limiting chemical and nuclear reactions to yield near 100% efficiency. This approach transforms the black hole into an ideal power plant by converting protons and neutrons entirely into radiation, provided the civilization can prevent the structure from growing and capture all emitted particles including elusive gravitons.

How Far Are We From An AI Einstein? - Adam Brown

Adam Brown

In a discussion regarding the future of artificial intelligence, a speaker predicts that Large Language Models will likely achieve the ability to derive General Relativity from Newtonian physics within a decade, marking a potential terminal milestone for human intellectual discovery. This projection is supported by private evaluations from a Stanford professor who observed LLMs rapidly advancing from scoring zero on graduate-level General Relativity exams to essentially acing them in just three years, rendering the assessment obsolete. Despite this exponential growth in solving existing problems, the speaker notes a persistent gap between a model's capacity to translate knowledge and its ability to make the intuitive conceptual leaps characteristic of historic figures like Einstein.

The bacteria that shaped history: Yersinia pestis – David Reich

David Reich

Ancient DNA sequencing reveals that *Yersinia pestis* caused mortality rates of 25% to 50% across Western Eurasia over five millennia, fundamentally altering transmission dynamics by lacking the flea-rat plasmid and likely spreading via aerosols from steppe rodent reservoirs. This biological pressure precipitated a radical demographic shift around 4,500 years ago, where approximately 90% of Neolithic farmers in Europe were replaced by Eastern Steppe migrants within a century, simultaneously dismantling social structures and facilitating the rise of new economic systems. While the pathogen is also documented as a key factor in the decline of the Roman Empire and the inflation of the medieval period, these findings collectively demonstrate how sustained high mortality can drive mass migrations, societal collapse, and long-term economic transformation.

Scaling laws are explained by memorization and not intelligence – Francois Chollet

Francois Chollet

A speaker challenges the "scale maximalist" view that increasing computational power generates true general intelligence, arguing instead that current Large Language Models primarily function as interpolative databases relying on memorized solution templates. The presentation distinguishes between static pattern recognition, which achieves high benchmark scores through knowledge retrieval, and genuine reasoning defined as the dynamic synthesis of novel programs from foundational building blocks. Ultimately, the discussion posits that while extensive memory is a necessary prerequisite for complex tasks, scaling models merely expands their skill scope without granting the capacity for on-the-fly adaptation characteristic of true intelligence.

How the US-China AI Race Will Play Out – Leopold Aschenbrenner

Leopold Aschenbrenner

A speaker outlines a precarious three-month AI development race between the United States and China, warning that a decisive advantage could trigger a "fever struggle" for global dominance or catastrophic instability. The analysis highlights China's infrastructural superiority in power generation and state-sponsored infiltration, which creates a vulnerable "post-intelligence but pre-industrial" window where a preemptive strike could disable an opponent's cluster before robot factories are deployed. To avert conflict, the speaker proposes that the US and its allies first secure a massive lead in computing and energy supplies before offering a mutual non-agreement enforced by AI systems to ensure a stable arrangement.

AI Nationalization is Inevitable – Leopold Aschenbrenner

Leopold Aschenbrenner

The speaker argues that the development of Artificial Super Intelligence poses an existential threat to liberal democracy and the global order, necessitating immediate state intervention to prevent a volatile race between private labs and hostile nation-states. Rather than relying on independent market forces, the discussion advocates for a middle-ground model featuring intimate government oversight and intelligence agency review to manage the unprecedented concentration of power held by a few tech giants. This approach aims to establish a secure offense-defense balance and a stable chain of command, mitigating the risks of unauthorized dual-use capabilities while avoiding the pitfalls of both total privatization and full military centralization.

The inside story of how ChatGPT was built – OpenAI cofounder John Schulman

John Schulman

OpenAI developed ChatGPT by pivoting from standalone instruction-following models to a dedicated conversational architecture based on GPT-3.5 and later GPT-4 to better handle coding, clarifying questions, and factual limitations. The team resolved early reliability issues through hybrid training datasets that combined instruction following with chat-specific data, creating a system that intuitively defines helpfulness while acknowledging its own knowledge boundaries. This rigorous, multi-iteration refinement process established a specialized alignment framework that public fine-tuning APIs or simple interface wrappers cannot easily replicate.

Robert Oppenheimer's Worst Enemy – Richard Rhodes

Robert Oppenheimer, Richard Rhodes

This narrative explores J. Robert Oppenheimer's complex leadership at Los Alamos, highlighting his transformation from a condescending pre-war figure into a psychologically astute director praised by his youngest colleagues. The account details the post-war rift with Edward Teller, a former rival who, despite exhibiting paranoia and hostility, ultimately acknowledged Oppenheimer as the best laboratory director he ever knew. The story concludes by invoking an anecdote about Hannibal to frame Teller's contradictory praise as the ultimate testament to Oppenheimer's managerial competence.