Eliezer Yudkowsky
AI Safety Researcher & Writer
Co-founder of the Machine Intelligence Research Institute and the AI debate’s most absolute doomer: he argues frontier AI development should be shut down worldwide because superintelligence built with current techniques would kill everyone.
In his words
If intelligence says that a country outside the agreement is building a GPU cluster, be less scared of a shooting conflict between nations than of the moratorium being violated; be willing to destroy a rogue datacenter by airstrike.
If any company or group, anywhere on the planet, builds an artificial superintelligence using anything remotely like current techniques, based on anything remotely like the present understanding of AI, then everyone, everywhere on Earth, will die.
Biography
Eliezer Yudkowsky (born 1979) is an autodidact AI researcher who co-founded the Machine Intelligence Research Institute (MIRI) in 2000, created the rationalist community hub LessWrong, and spent two decades developing the arguments — instrumental convergence, the alignment problem, the treacherous turn — that now anchor the AI-existential-risk worldview. He is also, improbably, the author of the wildly popular fan fiction Harry Potter and the Methods of Rationality.
“Shut it all down”
When the Future of Life Institute published its March 2023 open letter calling for a six-month pause on giant AI training runs, Yudkowsky declined to sign — because it did not go far enough. In a TIME op-ed titled “Pausing AI Developments Isn’t Enough. We Need to Shut it All Down,” he wrote that the most likely result of building a superhumanly smart AI under anything like current conditions is that literally everyone on Earth will die, and called for an indefinite worldwide moratorium: shut down all the large GPU clusters, shut down all the large training runs, and enforce the ban internationally — even, in his most quoted line, by airstrike on rogue datacenters. The piece was raised at a White House press briefing, and TIME named him to its inaugural TIME100 AI list that year.
In September 2025 he and MIRI president Nate Soares published If Anyone Builds It, Everyone Dies, a mass-market distillation of the case that superintelligence built by gradient descent would have alien goals and eliminate humanity by default. The book became a New York Times bestseller and made several best-of-2025 lists, though reviews split sharply: The Observer found it told with “clarity, verve, and barely suppressed glee,” while the New Statesman flatly declared it “not a serious book” and Wired’s Steven Levy found its remedies less plausible than its doom.
Critics — including many working AI researchers — regard his certainty of doom as unfalsifiable and his policy prescriptions as fantastical. But his fingerprints are on the whole field: Sam Altman has credited him with getting many people interested in AGI, and much of the modern safety agenda began as responses to his arguments.
Where they stand in the war
Who backs them up
Sources & further reading
Canonical record: https://battlelines.ai/topic/eliezer-yudkowsky