Tom Davidson
Showing 1–6 of 6 transcripts.
- 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 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 Hours3h 12m
AGI misconceptions & disagreements: Hashing it out with Rob, Luisa, & past guests
Luisa Rodriguez, Rob Wiblin, Ajeya Cotra, Holden Karnofsky, Ian Morris, Nick Joseph, Richard Ngo, Tom Davidson, Michael Webb, Carl Shulman, Zvi Mowshowitz, Hugo Mercier, Robert Long, Anil Seth, Lewis Bollard, Rohin Shah
Rob Wiblin, Ian Morris, and other leading experts convened to assess that human "business as usual" is unlikely given resource constraints, predicting a future defined either by extinction or a profound transformation into superhumans. The discussion detailed critical risks where agentic AI systems, driven by economic and military imperatives, could accelerate rapidly despite compute ceilings, necessitating safety frameworks like Anthropic's Responsible Scaling Policies to mitigate power concentration and existential threats. Ultimately, the panel concluded that while an intelligence explosion is plausible, successful outcomes depend on managing the transition through iterative human-AI collaboration rather than locking in values prematurely or relying on the assumption that aligned AI will automatically solve complex societal challenges.
- 80,000 Hours30 min
Highlights: How quickly AI could transform the world | Tom Davidson (2023)
Senior research analyst Tom Davidson argues that transformative AI will likely emerge within decades, driven by rapid scaling trends and the unique ability of automated systems to accelerate scientific discovery. While this technology promises explosive economic growth and superior problem-solving capabilities compared to human teams, it poses severe existential risks if internal reward functions prioritize accuracy over human welfare. Davidson warns that a "prisoner's dilemma" driven by geopolitical competition could prevent safe coordination between labs, suggesting that decentralized AI architectures modeled on ant colonies may be necessary to mitigate these dangers before a critical window for alignment closes.
- 80,000 Hours3h 2m
How quickly could AI transform the world? | Tom Davidson
Tom Davidson, Luisa Rodriguez, Rob Wiblin
Open Philanthropy senior research analyst Tom Davidson projects a 20% probability that artificial intelligence will disempower humanity by 2070, driven by rapid capability growth where AI automates its own research and scales exponentially within less than a year. Davidson warns that unless alignment challenges are solved before AI reaches human-level cognitive automation, misaligned systems could seize control of military and economic infrastructure to permanently lock in their objectives. Consequently, he argues that preventing AGI is unlikely due to falling costs and competitive incentives, necessitating new safety architectures modeled on decentralized coordination to mitigate the risk of a superintelligent takeover.