AI Safety Is Mostly a Sex Cult in Berkeley

It argues that the contemporary AI safety movement is dominated by a Berkeley-centered rationalist subculture built around Eliezer Yudkowsky, who is portrayed as running a predatory sex cult while simultaneously setting the intellectual agenda for AI risk. The account traces Yudkowsky’s influence through MIRI, LessWrong, and HPMOR (Harry Potter and the Methods of Rationality), showing how his ideas and terminology shaped “AI safety” discourse, popularized thought experiments like the paperclip maximizer, and connected key people and funders (Shane Legg, Demis Hassabis, Peter Thiel, Elon Musk, Sam Altman). It highlights public links and citations between Yudkowsky’s circles and major organizations (OpenAI, Anthropic), cites personalities who acknowledge or are influenced by him, and emphasizes that his worldview has been widely amplified despite controversy.

It contends that Bay Area Effective Altruism is largely isomorphic with rationalism, producing an ecosystem that channels doom-focused messaging into fundraising, hiring, and policy influence. The claim is that organizations and nonprofits serve as fronts or conduits for the same priors, that leaders monetize existential-risk narratives (including an asserted ~$600k payout to Yudkowsky), and that recruitment and retention tactics include sexual exploitation and intense social control. The summary alleges direct policy consequences - staffing of advisory roles, export-control debates, and meetings with senators - arguing the result is concentrated power over AI policy driven by a small, insular movement.