Sex, AI, and the Apocalypse: How a Rationalist Subculture Shaped AI Safety, Politics, and Culture

The headline warning was real – and it mattered

Jacob Coxon quit Anthropic on September 8 2026, burned his equity, and warned that the people building AI earnestly believe it could kill humanity within a decade. His note went viral, sparked congressional hearings, and forced the AI‑safety community to confront the social network that fuels its warnings.

A subculture built on rationalism, rituals, and heroic responsibility

Conclusion: The rationalist movement grew from a small internet scene into a quasi‑religious network that supplies talent, funding, and cultural glue to frontier AI labs.

The movement traces its roots to Eliezer Yudkowsky’s early 2000s blogs and the Sequences – a set of essays teaching “overcoming bias”.
By 2010 Yudkowsky’s fan‑fiction Harry Potter and the Methods of Rationality (HPMOR) introduced the doctrine of heroic responsibility: individuals must prevent catastrophe because they are uniquely capable of doing so.
Community‑wide rituals such as the Secular Solstice (organized by Raymond Arnold) provide the same fellowship, liturgy, and after‑care that traditional religions offer, but with AI doom as the central narrative.
Demographic surveys show a highly homogeneous membership: overwhelmingly male, many on the autism spectrum, and a large share employed in STEM fields – a pipeline that naturally feeds Silicon Valley labs.

From fan‑fiction to funding: the infrastructure of influence

Conclusion: The rationalist scene turned cultural production into a recruiting and fundraising engine that now underwrites AI‑safety research.

CFAR (Center for Applied Rationality), founded in 2012 by Julia Galef, Andrew Critch, Michael Smith, and Anna Salamon, sold intensive workshops that taught the Sequences’ techniques to tech employees and gifted teenagers.
Leverage Research, funded by Peter Thiel and Jaan Tallinn, attempted to “debug” human cognition at scale, later attracting allegations of psychological abuse.
Effective Altruism (EA) merged with rationalism in the late 2000s; long‑termism reframed future lives as the primary moral unit, making AI‑extinction the flagship cause.
Major donors – Thiel, Tallinn, the Open Philanthropy Project, and even Sam Bankman‑Fried (via FTX) – channeled tens of millions into AI‑risk labs, including Anthropic and OpenAI.

The dark side of the community: cult‑like dynamics and abuse

Conclusion: Several high‑profile incidents reveal a pattern of charismatic authority, psychological coercion, and violent outcomes within the network.

"The rationalist community vouches for its own members, making it hard for outsiders to assess sincerity." – Zvi Mowshowitz, longtime community participant (HN comment).

“Debugging” – CFAR staff admitted that at least three participants were subjected to invasive psychological interventions, framed as fixing “irrational” thinking.
Leverage Research – former participant Zoe Curzi described an environment where dissent became impossible because the organization claimed superior insight into the mind.
The Zizians – a splinter group led by trans‑woman Ziz LaSota formed a seasteading colony, later linked to multiple murders and a high‑profile shootout in Vermont (2025). The group’s ideology combined decision‑theory, split‑brain psychology, and heroic responsibility.
Sexual economies – Aella, a former sex‑worker turned rationalist influencer, organized paid “gangbang” events that screened participants for AI‑risk alignment, illustrating how sexual access became a gate‑keeping mechanism.

Political penetration: from think‑tanks to the White House

Conclusion: The same network that funds AI‑risk labs also supplies personnel and ideas to U.S. political power structures.

Peter Thiel – funded neoreactionary writer Curtis Yarvin, backed surveillance firms (Palantir, Anduril), and delivered private “Antichrist” lectures that framed AI regulation as a religious‑political threat.
JD Vance – received $15 M from Thiel’s venture firm, cites Yarvin’s ideas, and now sits on the Senate Judiciary Committee, influencing AI‑policy legislation.
Michael Kratsios (former Thiel Capital principal) became co‑chair of the President’s Science & Technology Advisory Council, alongside David Sacks and Marc Andreessen.
METR, the nonprofit tasked with evaluating frontier AI systems, disclosed personal ties to lab staff, raising concerns about conflict of interest in the very oversight mechanism meant to check the community’s influence.

The Coxon resignation in context

Conclusion: Coxon’s public warning is credible, but its impact is amplified (and questioned) because it emerged from a tightly interwoven social network.

Coxon’s note stated a personal probability of >10 % that AI will kill humanity by 2030 – a view echoed by Anthropic alignment lead Evan Hubinger within hours of the resignation.
The rapid spread (100 M+ views in a day) was fueled by the community’s own expertise in virality (e.g., HPMOR’s “Final Exam” crowd‑solving event) and by donors eager to push AI‑regulation agendas.
Critics (including Elon Musk) dismissed the episode as a “psy‑op”, while supporters (e.g., Zvi Mowshowitz) defended Coxon’s sincerity, illustrating the self‑referential validation loop.

What this means for AI safety governance

Conclusion: Effective oversight must account for the social and financial entanglements of the rationalist‑EA network, not just the technical arguments.

Independent evaluation – Oversight bodies need strict conflict‑of‑interest rules, protected whistleblowers, and the authority to enforce recusal when evaluators have personal ties to labs.
Transparency of funding – Donor pipelines from Thiel, Tallinn, and other elite philanthropists should be disclosed to avoid hidden influence on policy decisions.
Cultural awareness – Regulators must recognize that the community’s rituals, recruitment pipelines, and “heroic responsibility” ethos can pressure members into self‑censorship or risky behavior.
Broader stakeholder inclusion – AI‑risk debates should involve voices outside the rationalist‑EA echo chamber to mitigate groupthink and ensure diverse ethical perspectives.

Community reactions on Hacker News

"The framing of how we think about AI has really been cultivated by a quite homogeneous group of people. … It’s inescapable that we as humans have this flaw when we're surrounded by a culture." – throwaway13337 (HN comment)

"I'm not saying there aren't weird things going on, but this article sure does a lot of guilt‑by‑association." – vinkelhake

"The best piece of reporting on the subject I have ever seen." – epsteingpt

These comments illustrate the split perception: some see the article as exposing a dangerous echo chamber, others accuse it of over‑generalizing.

Bottom line

Jacob Coxon’s resignation was a genuine alarm about AI risk, but the speed and scale of its impact reveal a hidden infrastructure: a rationalist subculture that blends quasi‑religious rituals, elite recruitment, and concentrated funding to shape both AI research and political power. Any credible AI‑safety governance framework must confront this network’s influence, enforce independent oversight, and broaden the conversation beyond the echo chamber that currently dominates the discourse.