Latest Interviews
Showing 166–180 of 195 interview transcripts.
Clear all filters- 80,000 Hours2h 5m
Fast paths into high-impact ML engineering roles | Catherine Olsson & Daniel Ziegler
Catherine Olsson, Daniel Ziegler, Rob Wiblin
Former PhD candidates Daniel Ziegler of OpenAI and Catherine Olson of Google Brain detail a strategic pathway for transitioning into AI safety roles through the implementation of deep reinforcement learning papers and building robust software engineering skills. They emphasize that research engineering positions prioritize frustration tolerance and the ability to replicate code over formal theoretical mastery, effectively serving as a viable alternative to completing a traditional doctoral degree. This industry-focused approach enables professionals to contribute to critical alignment work and infrastructure development at leading labs without the requirement of a PhD, often supported by fellowship programs or targeted retraining grants.
- 80,000 Hours2h 50m
#46 - Prof Hilary Greaves on moral cluelessness & tackling crucial questions in academia
The Global Priorities Institute is recruiting for a four-year remote Head of Research Operations role with a salary ranging from £40,000 to £49,000 and a deadline of November 19. Director Hilary Graves leads the organization's shift from theoretical physics to moral philosophy, focusing on rigorous research into extinction risk, population ethics, and decision-making under moral uncertainty. The institute is also accepting applications for its Summer Research Visitor Program by November 30 and planning to launch new postdoctoral fellowships for philosophers and economists in September.
- 80,000 Hours3h 52m
Solving the alignment problem and handing off the future to AI | Paul Christiano
OpenAI researcher Paul Christiano outlines a technical framework for aligning artificial intelligence with human values, prioritizing "prosaic" approaches like iterated amplification over speculative future methods to address the risk of an AI-dominated economy. He argues that competitive market pressures will likely force a gradual "slow takeoff" over two decades, necessitating robust verification mechanisms to prevent a race to the bottom on safety standards. The discussion concludes with strategic recommendations for the field, emphasizing the need for institutional design, funding flexible high-risk research, and focusing on engineers who can scale safety theories within existing deep learning architectures.
- 80,000 Hours2h 45m
#43 - Daniel Ellsberg on the institutional insanity that maintains nuclear doomsday machines
In a 2023 podcast discussion, former nuclear analyst Daniel Ellsberg presents arguments from his book *The Doomsday Machine: Confessions of a Nuclear War Planner*, detailing how current US and Russian war plans risk triggering a global nuclear winter that would extinguish the vast majority of human life. Ellsberg contends that the maintenance of these arsenals is driven by the military-industrial complex's need for profit and political leverage rather than strategic necessity, and he recommends immediate policy shifts including the elimination of land-based ICBMs and a unilateral US "No First Use" declaration. Drawing on his historical precedent as the leaker of the Pentagon Papers, the interview concludes with Ellsberg urging those with access to classified information to consider whistleblowing as a necessary step to prevent catastrophic existential threats.
- 80,000 Hours2h 47m
#42 - Dr Amanda Askell on moral empathy, the value of information & the ethics of infinity
Philosopher Amanda Askell reflects on her six-and-a-half-year NYU PhD journey, highlighting the intense demands of US academic models and the precarious nature of the current tenure-track job market for philosophy graduates. Her research critically addresses "moral cluelessness" and the paradoxes of infinite ethics, arguing that high information value justifies experimental interventions despite long-term unpredictability. Currently pivoting toward AI ethics and policy, Askell advocates for an "explore/exploit" career strategy that prioritizes research freedom and moral empathy to navigate both the limitations of traditional academia and the complex challenges of high-impact global problems.
- 80,000 Hours2h 18m
#41 - David Roodman on incarceration, geomagnetic storms, & becoming a world-class researcher
David Roodman, Holden Karnofsky, Keiran Harris, Rob Wiblin
David Rude, a former theoretical mathematician turned independent researcher, specializes in replicating and reconstructing empirical studies on high-impact causes such as criminal justice, deworming, and geomagnetic storm risks. His methodology often challenges established conclusions, finding zero deterrence effect in harsh incarceration laws, identifying data artifacts in American deworming studies, and downplaying the probability of catastrophic solar events compared to public perception. Currently based at Open Philanthropy, Rude leverages his independence from academic incentives to scrutinize methodological robustness and advocate for policy shifts, such as prioritizing savings over microcredit loans.
- 80,000 Hours2h 12m
#40 - Katja Grace on forecasting future technology & how much we should trust expert predictions
Katja Grace, Keiran Harris, Rob Wiblin
Led by Katja Grace, the three-year-old research organization AI Impacts employs empirical methods to forecast artificial intelligence timelines, finding that experts medianly predict high-level machine intelligence in 45 years while identifying significant biases in how these estimates are framed. Their analysis of compute trajectories reveals that the cost of hardware equivalent to the human brain has dropped dramatically, yet experts remain skeptical of sudden "intelligence explosions" due to historical precedents of incremental technological progress. Consequently, the team advocates for increased investment in AI safety research and the development of diverse forecasting skills to navigate the uncertainty surrounding future AI impacts.
- 80,000 Hours2h 18m
#39 - Spencer Greenberg on the scientific approach to solving difficult everyday questions
Spencer Greenberg, Keiran Harris, Rob Wiblin
Spencer Greenberg updates the portfolio status of four Sparkwave-backed companies while presenting extensive research on intrinsic values, overconfidence calibration, and Bayesian evidence evaluation. His analysis distinguishes how political and demographic groups prioritize different core values and introduces a predictive model for identifying when individuals are likely to overestimate their skills or underestimate difficult tasks. Furthermore, Greenberg demonstrates practical applications of these frameworks through case studies on political collusion, nuclear disarmament, and scientific replication, ultimately advocating for a methodical approach to belief updating that balances prior probabilities with specific evidence.
- 80,000 Hours2h 0m
#38 - Prof Ng on anticipating effective altruism decades ago & how to make a much happier world
Prof Ng, Yew-Kwang Ng, Keiran Harris, Rob Wiblin
Yu-Kwang Ng, an economist who independently pioneered core effective altruism concepts in the 1990s, advocates for a hedonistic utilitarian framework that prioritizes subjective well-being as the sole intrinsic moral value to guide public policy. His research proposes specific interventions ranging from high progressive taxation to correct relative competition and environmental externalities, to the development of brain stimulation technologies and welfare biology for mitigating wild animal suffering and existential risks. By rejecting traditional deontological and capability-based ethical systems in favor of cardinal utility comparisons, Ng urges a global shift toward policies that maximize net welfare across current and future generations.
- 80,000 Hours1h 44m
#37 - GiveWell picks top charities by estimating the unknowable. James Snowden on how they do it.
James Snowden, Keiran Harris, Rob Wiblin
Founded a decade ago by Holden Karnofsky and Ellie Hassenfeld, GiveWell directs approximately $100 million annually to nine evidence-backed charities in international development after applying rigorous cost-effectiveness and transparency criteria. The San Francisco-based organization recently expanded its portfolio to include five new top recommendations while navigating internal ethical debates over suicide prevention and evolving its analytical models to account for leverage and funding effects. With a team of over 20 employees focused on an "epistemic culture," the group continues to explore policy interventions and speculative research opportunities beyond traditional direct service models.
- 80,000 Hours2h 5m
#36 - Tanya Singh on ending the operations management bottleneck in effective altruism
Tanya Singh, Professor Bostrom, Tara Mac Aulay, Keiran Harris, Rob Wiblin
An episode featuring Tanya Singh from the Future of Humanity Institute and representatives from the Centre for Effective Altruism addresses the critical shortage of operational talent hindering the scaling of effective altruism organizations. The discussion defines operations as creative problem-solving roles that act as force multipliers for researchers, highlighting open positions at GiveWell, Open Philanthropy, and OORT that require high value alignment and adaptability. With applications for several key roles closing within days, the event outlines the specific skills, such as systems prototyping and automation, necessary for candidates to prevent organizational stagnation in a sector marked by significant growth potential.
- 80,000 Hours2h 19m
#34 - We use the worst voting system that exists. Here's how Aaron Hamlin is going to fix it.
Aaron Hamlin, Edwin Edwards, David Duke, Buddy Roemer, Keiran Harris, Rob Wiblin
The Center for Election Science leverages a $600,000 Open Philanthropy grant to advance voting reform by advocating for Approval Voting as a superior alternative to the flawed First-Past-The-Post system. By targeting municipal ballot initiatives in cities like Fargo and San Francisco, the organization aims to break the two-party duopoly and elect moderate consensus candidates without the complexities of Instant Runoff Voting. This grassroots strategy seeks to scale successful local implementations into state and federal adoption, ultimately reducing political polarization and increasing voter efficacy through a robust, transparent electoral mechanism.
- 80,000 Hours2h 26m
#32 - Bryan Caplan on whether his Case Against Education holds up, totalitarianism, & open borders
Bryan Caplan, Keiran Harris, Rob Wiblin
Econometrician Tyler Cowen's analysis of *The Case Against Education* argues that formal schooling primarily functions as a signaling mechanism for conformity rather than a generator of human capital, leading to a societal misallocation of resources toward credential inflation. This framework extends to global policy, where Cowen contends that open borders and deregulated housing offer superior welfare gains compared to current systems, while highlighting the existential risks of permanent totalitarianism if technological advancements eliminate competitive checks on power. Drawing from these premises, he advises individuals to prioritize vocational over academic pathways unless a degree is strictly necessary for entry, and suggests that effective altruism should focus on immigration reform and the release of non-violent offenders to maximize moral impact.
- 80,000 Hours2h 2m
#30 - Dr Eva Vivalt on how little social science findings generalize from one study to another
Dr Eva Vivalt, Keiran Harris, Rob Wiblin
Y Combinator launched a $100 million, three-to-five-year randomized controlled trial providing $1,000 monthly to approximately 1,000 low-income young adults to analyze shifts in time use, health, and well-being while deliberately withholding results until the study's conclusion to prevent narrative distortion. This initiative is situated within a broader academic discourse highlighting the limited generalizability of development interventions and the prevalence of optimism bias among policymakers, who often neglect variance and negative results in favor of optimistic point estimates. Experts argue that despite these methodological challenges and academic incentives prioritizing novelty over synthesis, aggregated expert priors and robust RCTs remain essential for refining evidence-based policy and overcoming the "naturalistic heuristic" in emerging fields like clean meat adoption.
- 80,000 Hours2h 17m
#27 - Dr Tom Inglesby on careers and policies that reduce global catastrophic biological risks
Dr. Thomas Inglesby and the Center for Health Security outline a strategic shift from natural pandemic preparedness to addressing global catastrophic biological risks as biotechnology becomes increasingly accessible. To mitigate these threats, the organization advocates for a unified U.S. biodefense strategy and the creation of a dedicated public health emergency fund to overcome historical funding delays. Concurrently, the Center's Emerging Leaders Fellowship and partnerships with the Open Philanthropy Project aim to cultivate a multidisciplinary workforce capable of integrating deep scientific expertise with policy implementation.