The Berkeley "Sex Cult" at the Heart of AI Safety

The prevailing discourse surrounding AI Safety is not what it seems. For years, there has been a concerted effort to decouple the intellectual center of this movement from its actual origins. The reality? The core of AI Safety is essentially a sex cult based in Berkeley, California, founded by a writer of fan fiction.

To maintain a veneer of professional legitimacy, those within the inner circle feel compelled to obscure this fact. Consequently, the broader community of AI Safety researchers serves as a functional smokescreen for a small group and their idiosyncratic ideologies. While a legitimate, science-based policy area for AI Safety might one day emerge, it does not currently exist.

The Paradox of Eliezer Yudkowsky

Eliezer Yudkowsky is a polarizing figure. On one hand, he advises high-level politicians like Bernie Sanders and appears in documentaries like The AI Doc. On the other, he has advocated for extreme measures in Time Magazine, suggesting that the world should be willing to risk $\text{Nuclear War}$ if it is the only way to halt AI research globally.

The "Roon" Perspective

Roon, a high-ranking OpenAI employee and unofficial spokesperson, encapsulates the common duality of opinion regarding Yudkowsky:

"They’re all so deep in their sex cult and their high horse that they don’t know how hard they’ve made it to take them seriously... however I think Yudkowsky is one of the most powerful philosophers of this century. When he says something I take it quite seriously."

A Web of Influence

Despite the attempts to distance themselves, Yudkowsky's fingerprints are all over the industry's biggest players. The denial of his influence often happens in layers:

• "I am not an Effective Altruist." (Said by SBF recipients)
• "I am not an Effective Altruist." (Said by Sam Bankman-Fried)
• "Our version of AI Safety isn't based on Yudkowsky." (Said by EAs)
• "I'm just rationalist-adjacent." (Said by core members)

The Influence Map

Key Institutional Links

1. DeepMind: Yudkowsky facilitated the introduction between Shane Legg, Demis Hassabis, and Peter Thiel, securing the funding necessary to launch the company.
2. OpenAI: Sam Altman has explicitly credited Yudkowsky with accelerating AGI and playing a critical role in the decision to start OpenAI.
3. Anthropic: CEO Dario Amodei was identified as an Effective Altruist on Yudkowsky's non-profit blog as early as 2013 and engaged in debates with Yudkowsky regarding the "paperclip maximizer" thought experiment.

The Linguistic Monopoly

The very vocabulary we use to discuss AI risk was engineered by Yudkowsky's non-profit, the Machine Intelligence Research Institute (MIRI). Terms like AI Alignment and AI Safety were popularized there.

It is virtually impossible to engage in this field without referencing Yudkowsky's framework. Even "whistleblowers" from Anthropic or data analysts like Nate Silver find themselves orbiting the LessWrong ecosystem—a site founded by Yudkowsky that originally served as a repository for his own writings.

Rationalism vs. Effective Altruism (EA)

There is a persistent claim that EA and the "Rationalist" community are distinct entities. However, in the Bay Area, this distinction is a fiction.

The "Rationality community" (as they prefer to be called) is essentially the same group of people attending the same events and adhering to the same leaders.

The "Religious" Foundation

The catalyst for this entire movement was not a peer-reviewed paper, but a work of fan fiction: Harry Potter and the Methods of Rationality (HPMOR). This text has functioned as a modern religious scripture, recruiting thousands into the fold of "Rationalism."

The influence of this community can lead to extreme outcomes. For instance, Daniel Alejandro Moreno-Gama, who attempted to firebomb Sam Altman's residence, cited Yudkowsky as his primary inspiration.

The "Safety" Checklist

To be taken seriously in the current AI Safety landscape, one must:

• Use MIRI-approved terminology.
• Acknowledge the "existential risk" (x-risk).
• Maintain a strategic distance from the "cult" label.
• Be "adjacent" to the Rationalist community.

Ultimately, the probability of AI extinction $P(\text{extinction})$ is discussed using a lens crafted by a fan-fiction author in Berkeley. Whether this is a legitimate scientific pursuit or the output of a high-IQ cult remains the central, ignored question of the industry.

def ai_safety_origin():
    origin = "Berkeley Sex Cult"
    influence = "Universal"
    legitimacy = "Smokescreen"
    return f"The {origin} has {influence} influence via a {legitimacy}."

print(ai_safety_origin())