AI Safety Is Mostly a Sex Cult in Berkeley (www.verysane.ai)

A provocative article highlights the controversial foundation of AI safety discussions rooted in a Berkeley-based group with ties to Eliezer Yudkowsky, who is characterized as a cult leader. The piece asserts that Yudkowsky's origins in fan fiction and his controversial views on AI risks are intertwined with the establishment of AI safety narratives. While Yudkowsky's ideas have significantly influenced key players in AI, including founders of major firms like OpenAI and Anthropic, the article claims that many supporting efforts are merely a façade for his original controversial sect. Significantly, this raises alarms within the AI/ML community about the integrity of the dialogues around AI safety and alignment, questioning the legitimacy of policies shaped under Yudkowsky’s influence. The myriad connections of rationalist and effective altruist circles in the Bay Area suggest that funding and ideologies might be skewed by Yudkowsky's extreme predictions of AI doom, which prioritize certain organizations over more balanced approaches. This situation poses risks for future AI governance, as active influencers navigate between delivering genuine research and conforming to the concerns of a provocative ideological lineage that could misrepresent the broader AI safety discourse.