Eliezer Yudkowsky
AI safety researcher; co-founder, MIRI
Among the earliest and most prominent voices warning about risks from advanced AI. Writes at length on LessWrong, the forum he helped start.
Key arguments & positions
- Argues building superintelligence without solving alignment first will very likely be catastrophic, and does not believe current safety techniques scale.
- Advocates an international, verified moratorium on frontier AI training runs.
- Popularized foundational AI-safety concepts such as "coherent extrapolated volition."
Accomplishments
- Co-founded the Machine Intelligence Research Institute (MIRI) in 2000.
- Founded LessWrong in 2009.
- Co-authored the 2025 New York Times bestseller "If Anyone Builds It, Everyone Dies."
Papers & key writings
- "If Anyone Builds It, Everyone Dies" (2025, with Nate Soares)
- "Coherent Extrapolated Volition" (2004, MIRI)
Links
Timeline
- 2000Co-founds the institute that becomes MIRI.
- 2009Begins the Sequences on rationality, later collected as a free book.
- 2010Publishes Harry Potter and the Methods of Rationality.
- 2022Publishes "AGI Ruin: A List of Lethalities".
- 2023Argues in TIME for an international halt to frontier training runs.
In the library
Nothing of theirs is filed in the hub yet. Browse the full library.
Recommended next
Hand-picked from the hub based on what Eliezer Yudkowsky covers.
METR
A research nonprofit that independently evaluates frontier AI models to measure what they can actually do and what risks that creates. Reports are free.
Why this: Covers AI safety and ethics too
AI Risk, Plainly
Plain-language explanations of AI safety concepts, including scheming, evaluation awareness, chain-of-thought monitoring, catastrophic risk and power concentration. Each entry describes the idea, its role in risk arguments and research evidence.
Why this: Covers AI safety and AI risk too
Gradual Disempowerment
Research project and paper examining how incremental automation of economic, cultural, and political systems could erode human influence even without a sudden AI takeover.
Why this: Covers AI safety and AI risk too
Fully Autonomous AI Agents Should Not Be Developed
A paper by Margaret Mitchell, Avijit Ghosh, Alexandra Sasha Luccioni and Giada Pistilli arguing that fully autonomous AI agents should not be developed. Drawing on prior scientific literature and current product marketing, the authors delineate levels of AI agent autonomy and show that risks to people increase as users cede more control, with safety risks to human life among the most concerning.
Why this: Covers AI safety and ethics too