On September 8, 2026, Jacob Coxon, a researcher at Anthropic, quit his position and posted a resignation letter stating: “The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.” The post was viewed more than 100 million times within a day, according to the source.
Coxon’s colleagues did not rush to contradict him. Evan Hubinger, who leads alignment science at Anthropic, posted within hours that Coxon was right and that his personal odds of AI causing human extinction within a decade were “somewhere north of one in ten.” More than two dozen members of Congress said action was needed, while Elon Musk dismissed the warning as a “psy-op.”
However, understanding the gravity of these warnings requires examining the community making them. According to the source, the loudest voices warning about AI-caused extinction emerge from “a very specific twenty-five-year-old internet subculture, one that understands itself as a community of people who think more clearly than everyone else.” Outside Silicon Valley this subculture appears obscure, but inside the technology industry it significantly influences AI policy discussions. Its alumni staff major AI labs, safety institutes, and philanthropic organizations that shape what “AI safety” means.
The source traces this community’s origins to Eliezer Yudkowsky, born in 1979, who by age twenty had co-founded the Singularity Institute for Artificial Intelligence in 2000. Yudkowsky dropped out of school around eighth grade and had no formal training in AI, according to the account.
The community has also developed distinctive cultural institutions. Raymond Arnold, described as a web developer and community organizer, created “Secular Solstice” in 2011 as a ritualistic celebration for the rationalist movement. By December 2025, nearly 500 people gathered in a Berkeley music hall for a Solstice ceremony focused on humanity’s possible extinction. Arnold has described his role as that of a “village priest” and now helps run LessWrong and Lighthaven, a Berkeley conference campus serving as the community’s physical center.
The source indicates the subculture’s influence extends beyond AI safety discussions, noting connections to various controversial activities and figures within the broader community, though the provided text does not elaborate on specific details beyond the initial overview.
Key facts
- Jacob Coxon quit Anthropic on September 8, 2026, warning that AI researchers “earnestly believe that it could kill us all by the end of the decade”
- His resignation post was viewed over 100 million times within a day
- Evan Hubinger, Anthropic’s alignment science lead, confirmed the warning, estimating his personal odds of AI extinction within a decade at “somewhere north of one in ten”
- The AI safety movement’s loudest voices emerge from a 25-year-old rationalist internet subculture
- Eliezer Yudkowsky co-founded the Singularity Institute for Artificial Intelligence at age 20 in 2000
- Raymond Arnold created Secular Solstice in 2011, which by December 2025 drew nearly 500 attendees to a ceremony in Berkeley