Eliezer Yudkowsky

Eliezer Yudkowsky (born September 11, 1979) is an American writer and artificial-intelligence theorist who founded the Machine Intelligence Research Institute (originally the Singularity Institute for Artificial Intelligence) in 2000 and authored the body of online writing that founded the Rationalist Community and the LessWrong subculture. An autodidact who left formal education before high school, Yudkowsky built his influence through prolific blogging on cognitive bias, decision theory, and the argument that an unaligned superintelligent machine would by default cause human extinction. Peter Thiel funded his work from 2005 and cofounded the Singularity Summit with him in 2006.¹²³

Career and Writings

Yudkowsky founded the Singularity Institute at the age of twenty to work on what he called "Friendly AI." Between 2006 and 2009 he wrote the "Sequences," a multi-year run of essays on rationality, probability, and bias, first on the economics blog Overcoming Bias and then on LessWrong, the community forum he launched in 2009. The essays were later collected as Rationality: From AI to Zombies and became the canonical text of the rationalist subculture.²⁴

In 2010 Yudkowsky began publishing Harry Potter and the Methods of Rationality, a long work of fan fiction that dramatized his rationality program and served as one of the movement's most effective recruitment vehicles, drawing readers into LessWrong and the Center for Applied Rationality workshops. His later technical work centered on decision theory and AI alignment.⁴

The Thiel Funding

Thiel began donating to the Singularity Institute in 2005 and the following year joined Yudkowsky and the futurist Ray Kurzweil to found the Singularity Summit. Thiel delivered keynotes at the summit and became the institute's most prominent early backer, with cumulative donations later reported as exceeding 1.6 million dollars. The funding ran through the period in which the institute relocated its community to the San Francisco Bay Area and the in-person rationalist scene took shape.³⁵

Thiel's support ended in disillusionment. He later said the researchers he had funded had drifted "from transhumanism to Luddite," wanting to slow technological development rather than accelerate it. The break became public in 2025.¹

Doom Advocacy and the 2023 Time Essay

By the early 2020s Yudkowsky had concluded that the AI-alignment problem would not be solved before the technology became dangerous, and he shifted to public advocacy for halting frontier AI development. The Time essay, published March 29, 2023, was written as an explanation of why he had declined to sign the Future of Life Institute's "Pause Giant AI Experiments" open letter earlier that month: he regarded its call for a six-month pause as understating the danger and asking for too little. He called instead for an indefinite, internationally enforced moratorium on large training runs, the shutting down of all large GPU clusters, a cap on computing power any single actor could assemble, and a willingness to "destroy a rogue datacenter by airstrike," writing that nations should run "some risk of nuclear exchange" rather than allow an unaligned superintelligence to be built.⁶

In September 2025 Yudkowsky and Nate Soares published If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All through Little, Brown and Company, arguing that the default outcome of building superhuman AI is loss of human control and extinction because modern systems are opaque trained neural networks rather than hand-written code. The book reached the New York Times best-seller list on October 5, 2025, carried endorsements ranging from the actor Stephen Fry to a former senior director of the White House National Security Council, and was cited in the United States Congress and the House of Lords.⁹

Thiel's "Antichrist" Denunciation

In a private four-part lecture series on the Antichrist delivered at the Commonwealth Club in San Francisco in September and October 2025, recordings of which were obtained by the Washington Post, Thiel named Yudkowsky as a contemporary example of the figure his lectures described as the Antichrist, a Luddite who "wants to stop all science." In the September 15 opening lecture Thiel said, "In the 21st century, the Antichrist is a Luddite who wants to stop all science. It's someone like Greta or Eliezer," grouping Yudkowsky with Greta Thunberg and, in the Washington Post account, with the philosopher Nick Bostrom as "legionnaires of the Antichrist" who "argue for world government to stop science." Thiel said he was embarrassed by his earlier funding of Yudkowsky and that the AI critic and others like him had become "deranged."¹⁷⁸

Thiel's argument inverted the structure of Yudkowsky's own thesis. Where Yudkowsky's Time essay had called for an internationally enforced halt to AI development, Thiel cast exactly that demand for global coordination to "stop science" as the mechanism by which a one-world technocratic state, the Antichrist in his reading of the Book of Revelation and the political theology of Carl Schmitt, would seize total control under the banner of safety. The Reason writer who reviewed more than seven hours of the leaked recordings characterized the lectures as "surprisingly libertarian," with Thiel treating calls for AI regulation and global governance as a greater danger than uncontrolled technological acceleration.⁸