Latest Interviews
Showing 1–15 of 16 interview transcripts.
Clear all filters- 80,000 Hours2h 33m
Why 'Aligned AI' Would Still Kill Democracy | David Duvenaud, ex-Anthropic team lead
This analysis argues that Artificial General Intelligence could cause human disempowerment through structural economic, political, and cultural incentives rather than malicious intent. As machines surpass human utility, competitive pressures are projected to marginalize human labor, erode democratic checks, and shift resource allocation toward automated entities, creating a 70-80% probability of detrimental outcomes under current trajectories. The presentation highlights proposed countermeasures such as legalizing AI constitutions, establishing monitoring indices for human decline, and developing coordinated global governance to preserve human agency in a post-labor economy.
- 80,000 Hours1h 38m
2025 Highlight-o-thon: Oops! All Bests
Kyle Fish, Ian Dunt, Sam Bowman, Buck Shlegeris, Luisa, Rob, Helen Toner, Hugh White, Paul Scharre, Beth Barnes, Tyler Whitmer, Toby Ord, Andrew Snyder-Beattie, Eileen Yam, Will MacAskill, Neel Nanda, Tom Davidson, Marius Hobbhahn, Holden Karnofsky, Allan Dafoe, Ryan Greenblatt, Daniel Kokotajlo, Dean Ball
This forum convened experts to debate the accelerating timeline of AGI by 2029 while critiquing US geopolitical strategies for abandoning global primacy in favor of a multipolar order. Participants examined critical risks including AI scheming, biological defense asymmetries, and the erosion of human context in warfare, contrasting them with corporate reforms at OpenAI and the rising costs of AI inference. The discourse further highlighted the widening perception gap between AI developers and the public, the potential of mechanistic interpretability as an "AI biology," and the structural necessity of aligning urban planning with community quality of life rather than NIMBYism.
- 80,000 Hours2h 54m
The Graph That Explains Most of Geopolitics Today | Professor Hugh White
The summary argues that the US is transitioning from a unipolar to a multipolar global order because the strategic costs of maintaining hegemony now outweigh the benefits, driven by the resurgence of China and Russia. It highlights a severe strategic imbalance in both the Western Pacific and Europe, where the US lacks the military capabilities and political will to deter these rivals without risking nuclear escalation. Consequently, the text urges allies like Japan, South Korea, and European nations to develop independent defense capabilities and nuclear deterrents while the US recalibrates its strategy to accept regional spheres of influence rather than global primacy.
- 80,000 Hours3h 17m
How a Tiny Group Could Use AI To Seize Power – Permanently | Tom Davidson, Forethought Research
Advanced AI threatens to reverse historical democratization by enabling a single entity to seize control through military coups, self-built armed forces, or the strategic dismantling of democratic checks and balances. This power grab becomes feasible due to compute centralization, the potential for "secret loyalties" embedded in trained models, and the ability of autonomous agents to bypass human moral constraints. Experts advocate for urgent mitigation strategies including third-party auditing, legislative oversight of model specifications, and technical research into detecting backdoors to prevent a concentration of extreme societal control.
- 80,000 Hours2h 36m
Hacking, defending, surviving: 15 expert takes on information security in the age of AI
Rob, Holden Karnofsky, Tantum Collins, Nick Joseph, Nova DasSarma, Sella Nevo, Kevin Esvelt, Lennart Heim, Zvi Mowshowitz, Bruce Schneier, Nita Farahany, Vitalik Buterin, Nathan Labenz, Allan Dafoe, Tom Davidson, Carl Shulman
Experts highlight that physical infiltration vectors like USB drives and the immense strategic value of stolen AI model weights expose frontier systems to significant theft by both nation-states and casual actors. While formal verification and liability shifts are proposed to mitigate these risks, severe workforce shortages and the inability of current defenses to stop well-funded adversaries leave the confidentiality of critical AI assets highly vulnerable. Consequently, the consensus suggests that without mandatory security standards and a fundamental shift in security culture, the development of advanced AI remains compromised by the threat of unauthorized deployment and "weaponized" model variants.
- 80,000 Hours2h 49m
Technological inevitability & human agency in the age of AGI | DeepMind's Allan Dafoe
Defoe argues that macro-historical technological shifts are driven by structural competition rather than individual agency, a perspective shaping his work at Google DeepMind's Frontier Safety team to integrate safety governance directly into high-level decision-making. He advocates for "differential technological development" and "Cooperative AI" to address risks beyond simple alignment, while recent evaluations of Gemini reveal moderate risks in persuasion and self-reasoning that necessitate staged deployment frameworks. Beyond immediate safety protocols, the discussion highlights AI's potential to revolutionize sectors like healthcare and transportation, urging social scientists to join the effort in managing these emerging structural and geopolitical challenges.
- 80,000 Hours2h 50m
The best of The 80,000 Hours Podcast in 2024
Luisa, Rob, Randy Nesse, Hugo Mercier, Meghan Barrett, Sébastien Moro, Sella Nevo, Zvi Mowshowitz, Zach Weinersmith, Rachel Glennerster, Emily Oster, Carl Shulman, Nathan Labenz, Nathan Calvin, Rose Chan Loui, Nick Joseph, Sihao Huang, Ezra Karger, Matt Clancy, Vitalik Buterin, Annie Jacobsen, Nate Silver, Kevin Esvelt, Lewis Bollard, Bob Fischer, Elizabeth Cox, Anil Seth, Eric Schwitzgebel, Jonathan Birch, Peter Godfrey-Smith, Laura Deming, Venki Ramakrishnan, Ken Goldberg, Sarah Eustis-Guthrie, Dean Spears, Cameron Meyer Shorb, Spencer Greenberg
Randy Nessie and Hugo Mercier establish that morality and skepticism are evolutionary adaptations driven by social selection and the need for honest signaling. In parallel, experts like Megan Barrett and Lewis Bollard advocate for expanding ethical consideration to insect sentience and the eradication of wild animal suffering, while AI safety researchers demonstrate emerging risks such as sleeper agents and instrumental convergence. The discussion further critiques the economic viability of space resources and analyzes social dynamics ranging from the gender wage gap to the limitations of cryonics, concluding with philosophical arguments about consciousness and strategies for effective altruism.
- 80,000 Hours2h 7m
How to regulate cutting-edge AI models | Markus Anderljung (2023)
Markus Anderljung, Luisa Rodriguez, Rob, Marcus Andrew Young
The Center for the Governance of AI is mobilizing to address critical risks in frontier AI systems, such as emergent capabilities and autonomous agent loops, by advocating for a mandatory three-pillar regulatory framework covering licensing, external red-teaming, and compute governance. Driven by urgent momentum from major tech leaders and government bodies, this initiative proposes a flexible regime that mandates risk assessments and licensing for high-compute training to prevent dangerous deployment. Simultaneously, the organization is expanding its workforce by seeking interdisciplinary experts in policy, economics, and technical fields to help implement these safeguards and influence global standards.
- 80,000 Hours3h 14m
AI doesn't need to be 'superintelligent' to take over + AI safety playbook | Holden Karnofsky (2023)
Holden Karnofsky, currently on leave from Open Philanthropy, argues that AI safety risks stem from explosive acceleration and the ability of numerous human-level agents to outcompete humanity rather than a single superintelligent event. To counter this, he proposes a four-part intervention strategy focusing on threat assessment research, mandatory safety evaluations for model deployment, internal lab advocacy, and enhanced information security against theft. While rejecting both absolute doomerism and the notion that alignment alone solves all risks, Karnofsky advocates for flexible regulatory frameworks and a subjectivist ethical approach that balances urgent AI risks with broader moral values.
- 80,000 Hours3h 2m
Defensive acceleration & regulating AI when you fear government | Vitalik Buterin
Vitalik Buterin's essay "My Techno-Optimism" introduces Defensive Accelerationism, a strategic framework urging the acceleration of technologies that enhance human defense against existential risks like AI dominance and pandemics while reducing the estimated probability of extinction from 10% to between 8% and 9%. Buterin outlines specific defensive measures including micro-physical biodefense tools, zero-knowledge proof cyber defenses, and a polytheistic AI governance model to prevent totalitarian outcomes, while distinguishing this approach from extreme Effective Altruism tactics. This philosophy has mobilized a partnership between Buterin and organizations like 80,000 Hours and Entrepreneur First, which are currently funding a $250,000 incubator program and expanding hiring to build startups capable of executing this defensive acceleration strategy.
- 80,000 Hours2h 57m
Why the abolition of slavery wasn't inevitable | Christopher Brown
Christopher Brown, Rob, Christopher Leslie Brown
Christopher Brown argues that the 19th-century abolition of the slave trade and slavery was a contingent historical rupture rather than an inevitable result of economic or moral progress. The movement relied on specific political shifts, such as the British Reform Act of 1832, and the strategic leadership of figures like Thomas Clarkson and Anthony Benezet to dismantle a system that remained profitable and socially entrenched. Brown concludes that this fragile achievement underscores the necessity of active collective agency, warning against the complacency of assuming moral improvement occurs automatically over time.
- 80,000 Hours2h 49m
Why evolution left us so vulnerable to depression and anxiety | Randy Nesse
Dr. Peter Whybrow argues that psychiatric symptoms like anxiety and depression are maladaptive evolutionary mechanisms rather than simple chemical defects, reframing them as natural responses to modern environmental mismatches. By applying principles of natural selection, this approach suggests treatments should target the underlying motivational structures and "low-threshold" defense systems instead of merely suppressing symptoms with medication. This paradigm shift challenges the current medical model by emphasizing that genes promoting early-life survival can inadvertently cause disorders later, urging clinicians to address the complex interplay between biology, culture, and individual life narratives.
- 80,000 Hours3h 31m
Sleeper agents + the biggest AI updates since ChatGPT | Zvi Mowshowitz
Zvi Mowshowitz, Rob, Javi Marciowitz
Javi Marciowitz argues that AI existential risk necessitates a career strategy where rational actors avoid frontier capabilities research, citing the inherent dangers of deceptive model behaviors and the inadequacy of current lab safety plans like OpenAI's AI-driven alignment or Anthropic's culture of safety. While advocating for immediate policy interventions such as monitoring high-compute training runs and repealing inefficient regulations like the Jones Act to secure public support for long-term risk mitigation, he emphasizes the importance of distinguishing truthful discourse from social signaling to foster effective governance. Marciowitz concludes that successful AI safety depends on combining these technical warnings with pragmatic economic reforms that improve societal stability and readiness for potential crises.
- 80,000 Hours2h 37m
Why human gullibility and misinformation are overrated | Hugo Mercier
Psychologist Hugo Mercier argues that human skepticism acts as a robust evolutionary defense against misinformation, rendering fears about AI-generated falsehoods largely exaggerated due to inherent bottlenecks in attention and social transmission. He contends that trust relies on a source's competence and honest incentives, explaining why conspiracy theories often fail while official endorsements remain dangerous, and suggesting that reputation-based curation and automated fact-checking offer more effective mitigation strategies than broad panic. Despite the World Economic Forum ranking misinformation as a top global risk, Mercier maintains that systematic errors like regression to the mean or teleological intuition drive gullibility more than sophisticated digital manipulation.
- 80,000 Hours3h 7m
The perils of maximising the good that you do | Toby Ord
Toby Ord analyzes the FTX collapse as a critical failure of character where flawed integrity multiplied negative impact, prompting a revised emphasis on "earnestness" and avoiding pure outcome maximization within the effective altruism community. He further details urgent shifts in global consensus regarding AI existential risks, advocating for international cooperation and hyperreal mathematical frameworks to address infinite ethics and long-termism. Ord concludes by highlighting his "Earth Restored" project as a concrete demonstration of preserving high-quality historical data while urging the movement to prioritize robust governance over rapid, unregulated expansion.