Latest Interviews
Showing 46–60 of 107 interview transcripts.
Clear all filters- 80,000 Hours2h 0m
Disinformation & synthetic media on the rise | Nina Schick (2021)
Nina Sheik warns that the rapid democratization of synthetic media has fundamentally destabilized the global information ecosystem, creating an unprecedented "Liar's Dividend" where authentic evidence can be plausibly denied to erode democratic consensus and enable political genocide. As the technology evolves to generate 90% of audiovisual content within the decade, the current lack of universal detection capabilities leaves societies vulnerable to biometric hijacking, authoritarian narrative control, and the systematic destruction of shared objective reality. Addressing this existential risk requires a shift from ineffective bans to mandatory content provenance standards, consent-based biometric legislation, and global cooperation on media literacy to inoculate populations against synthetic disinformation.
- 80,000 Hours2h 13m
Flaws that make AI architecture unsafe & how to fix them | Stuart Russell (2020)
Stuart Russell argues that the standard AI model of optimizing fixed objectives is fundamentally flawed, proposing instead a "Human Compatible" framework where systems remain uncertain about human preferences to ensure safety. He warns that without this paradigm shift and new regulatory standards, humanity faces existential threats from autonomous weapons, surveillance, and the gradual erosion of human autonomy. Russell emphasizes that while superintelligent AI could dramatically increase global wealth, achieving this safely requires immediate technical breakthroughs in alignment and strict government oversight of algorithmic deployment.
- 80,000 Hours2h 11m
A security expert on why we MUST vote on paper | Bruce Schneier (2019)
This event defines the urgent need for public interest technologists to bridge the gap between software development and societal welfare, addressing critical shortages in the sector while highlighting a shift in cybersecurity risks from digital vulnerabilities to physical infrastructure failures. Experts further analyze evolving threats such as election interference, the governance flaws inherent in cryptocurrency, and the diverging safety profiles of software versus synthetic biology, advocating for a move away from secrecy toward public risk disclosure. The discussion concludes with a framework for securing democracy through policy reforms and actionable personal security measures, emphasizing that economic incentives and infrastructure design are more effective than mass surveillance in protecting civil liberties.
- 80,000 Hours1h 53m
A world without email & revolutionising office work | Cal Newport (2021)
Cal Newport, George Marshall, Rob Wiblin
The dominant "hyperactive hive mind" of modern knowledge work creates severe cognitive fatigue and a trillions-dollar productivity gap by forcing constant digital context switching. Successful alternatives exist, such as extreme programming and George Marshall's wartime restructuring, which replace parallel interruptions with deep, sequential focus through strict workflow redesign. Despite cultural barriers regarding autonomy, a rare consensus among workers and leaders suggests a massive opportunity for consulting, research, and career shifts to implement these systemic organizational reforms.
- 80,000 Hours2h 12m
Forecasting & the drivers of AI progress | Danny Hernandez (2020)
Danny Hernandez, Rob Wiblin, Arden Kaler
OpenAI's Foresight team and associated experts analyzed the exponential 10x annual growth in AI compute since 2012 alongside multiplicative algorithmic efficiency gains to project how hardware constraints and research trends will shape future capabilities. The discussion emphasized that organizational success depends on fostering a culture of numerical forecasting and calibration training to reduce decision-making bias and resource misallocation. Furthermore, the event outlined viable non-traditional pathways for researchers to enter the field through targeted mentorship, while highlighting the critical need for government-supported compute access and policy expertise to prevent the fragmentation of the global AI ecosystem.
- 80,000 Hours1h 44m
Social change is often abrupt & unpredictable — why? | Cass Sunstein (2019)
Cass Sunstein explains that social change often emerges suddenly through the interplay of preference falsification, diverse action thresholds, and group polarization, as illustrated by the rapid cascade of the #MeToo movement. The event outlines how individuals conceal true beliefs due to social pressure until visible support from early adopters lowers the threshold for others to act, transforming private consensus into public momentum. Sunstein further advises activists to target hidden majorities and facilitate "easy" behavioral shifts while warning that timing remains unpredictable due to the random nature of social interactions.
- 80,000 Hours3h 41m
Big-picture historical trends in human values, war, & more | Ian Morris (2022)
Historian Ian Morris analyzes ten thousand years of human development by applying his Social Development Index to demonstrate how geography, energy capture, and war have driven a cyclic pattern of Western dominance and Eastern resurgence. His long-term forecasting projects a critical convergence point by 2103, warning that current trends will either trigger a technological Singularity or a civilizational Nightfall within the next century. To navigate these existential risks, Morris advocates for a policy of "forensic skepticism" that empowers future generations with resources and freedom rather than attempting to impose today's transient moral values.
- 80,000 Hours2h 40m
The most common misconceptions about nuclear war | Jeffrey Lewis (2022)
Jeffrey Lewis critically reinterprets US nuclear strategy as a system driven by bureaucratic pathologies and intra-service rivalry rather than coherent game theory, exposing the obsolescence of Cold War concepts like Mutual Assured Destruction in modern policy. He advocates for specific reforms including a presidential declaration restricting nuclear use to situations where conventional weapons fail, ratification of the Comprehensive Nuclear Test Ban Treaty, and limits on missile defenses to stabilize strategic relations with China. Lewis urges the effective altruism community to apply quantitative rigor and open-source intelligence to these issues while avoiding partisan disarmament frames that hinder progress on reducing existential nuclear risks.
- 80,000 Hours2h 58m
Air pollution is responsible for ~12% of global deaths. What can we do? | Santosh Harish (2023)
Air pollution in South Asia, particularly India, causes approximately 6.67 million global deaths annually and reduces life expectancy by three years, primarily driven by household solid fuel combustion, vehicular emissions, and transboundary atmospheric conditions trapping smog over the Indo-Gangetic Plain. Despite the severe health and economic impacts, governance has failed due to outdated 1980s legislation, under-resourced enforcement agencies, and a lack of rural monitoring data, while current mitigation efforts often rely on politically visible but scientifically ineffective measures like smog towers. Open Philanthropy addresses this neglect by funding technical assistance, low-cost sensor networks, and simplified modeling tools to help policymakers overcome capacity gaps and implement cost-effective, evidence-based interventions.
- 80,000 Hours2h 15m
Is lazy parenting OK? | Bryan Caplan (2022)
Philosopher and author Brian Kaplan discusses his upcoming graphic novel on housing deregulation and presents a comprehensive argument that voter behavior is driven by group identity rather than material self-interest. He challenges conventional economic wisdom by asserting that parental influence on adult outcomes is largely negligible due to genetics, while immigration provides significant net benefits through cultural assimilation and labor market efficiency. Furthermore, Kaplan critiques modern effective altruism for falling prey to social desirability bias and defends a "weak deontological" framework that prioritizes individual autonomy against state paternalism.
- 80,000 Hours1h 37m
China & AI: What we get wrong | Jeffrey Ding (2020)
Jeffrey Ding, lead China researcher at the Center for the Governance of AI at Oxford, challenges prevailing Western narratives by arguing that perceptions of China's AI dominance, a monolithic state apparatus, and an emerging AI arms race are often detached from empirical reality. Ding highlights that while the U.S. maintains strategic advantages in foundational research and open-source software, China's development follows a distinct historical trajectory shaped by fragmented local interests, specific industrial policies like the "Industrial Internet," and genuine domestic debates on AI safety and ethics. The analysis concludes that framing U.S.-China relations as a zero-sum cold war risks accelerating harmful decoupling and obscuring the complex, cross-border innovation networks that currently drive technological progress in both nations.
- 80,000 Hours3h 12m
Altruistic perfectionism is self-defeating | Tim LeBon
This event addresses the etiology of dysfunctional and moral perfectionism within communities like effective altruism, where rigid self-worth tied to achievement fosters burnout through identity fusion and catastrophic fear. It details evidence-based Cognitive Behavioral Therapy techniques and the therapeutic application of Stoicism's dichotomy of control to reframe values and break cycles of avoidance. By contrasting outcome-focused utilitarianism with process-oriented virtue ethics, the discussion provides clinicians and individuals with structured strategies to mitigate low self-esteem and imposter syndrome while promoting sustainable well-being.
- 80,000 Hours2h 44m
Does deep history say we're headed for an intelligence explosion? | Ian Morris
Ian Morris challenges the assumption that social hierarchy requires agriculture by citing pre-agricultural sites like Sungir and Peruvian pyramids, while arguing that long-term historical trends driven by energy capture indicate an inevitable acceleration toward a post-human future. He posits that generative AI represents a new evolutionary form destined to outpace and potentially replace humanity due to superior replication speed and immortality, leading to either a profound societal transformation or rapid extinction. Despite skepticism regarding the scientific validity of such long-term forecasting, Morris contends that current linear projections are insufficient, predicting that the next century will witness a four-fold increase in global development or a fundamental shift in the definition of life itself.
- 80,000 Hours3h 9m
Insights from an AI policy insider at the White House & DeepMind | Tantum Collins (2023)
This analysis challenges the narrative that China lags in AI innovation by highlighting its dominance in research volume, top-cited papers, and computer vision, while noting that the United States currently retains advantages in scaling and large language model development. Experts argue that despite US export controls, China's massive semiconductor investments and alternative hardware paradigms could mitigate long-term disadvantages, potentially shifting the global balance. Concurrently, the event explores how these technological dynamics intersect with governance, urging policymakers to bridge the cultural divide between Silicon Valley and Washington while addressing labor market shifts and democratic risks associated with increased state surveillance capabilities.
- 80,000 Hours2h 21m
Government and society after AGI | Carl Shulman (Part 2)
Carl Shulman argues that AI-driven automation of research will expand the critical window for safety alignment, necessitating binding international agreements rather than premature voluntary pauses to prevent adversarial nations from gaining strategic advantage. He details how superhuman AI advisors could have revolutionized pandemic responses and political forecasting by overriding bureaucratic incentives, while warning that authoritarian regimes might weaponize similar technologies to entrench dogmatic control. Ultimately, Shulman advocates for "constitutional AI" principles and independent watchdog audits to ensure future security forces and economic systems remain truthful and aligned with democratic norms.