Hello! I'm your financial journalist and friend, an economist.
This in-depth report, "Sex, AI, and the Apocalypse," written by Ian Duncan, reads like a thrilling novel that blends elements from "Silicon Valley," "The Godfather," and "The Matrix." It's not just about whether AI will destroy humanity; it's more like a sociological analysis that peels back the glamorous facade of rationality within the AI security community in Silicon Valley, revealing the intricate web of interests, emotions, beliefs, and even violence beneath it.
To help you easily understand this complex article, I've broken it down into a summary of key points and a detailed analysis from five perspectives:
---
📝 Summary of Key Points
In one sentence: In 2026, a top AI researcher resigns and warns that AI could destroy humanity, causing a global uproar. But behind this warning lies a closed elite circle that has evolved over more than 20 years from a subculture of rationalism. This circle, with its unique ideology (similar to "Harry Potter" fanfiction), complex social and sexual relationships, substantial financial support from billionaires like Peter Thiel, and political influence, has built a "shadow empire" that seeks to control the future of AI. The article questions: How can we trust their promises of safety when the "saviors," "regulators," "investors," and "evaluators" are all from the same small group?
---
🔍 In-Depth Analysis: Five Aspects
1. Origins and Beliefs: The Alienation from "Rationality" to a "New Religion"
Plain Language: How did these people go from being "smart" individuals to becoming "believers"?**
The story begins with Eliezer Yudkowsky, a genius youth who was extremely arrogant, believing he was the only one who saw the dangers of AI. He founded the Machine Intelligence Research Institute (MIRI) and wrote a famous "Harry Potter" fanfiction book titled "Harry Potter and the Path of Reason."
- Why "Harry Potter"? Because the protagonist, Harry, is a genius surrounded by fools, which resonates with Silicon Valley geeks who feel intelligent but out of place.
- What happened to "rationalism"? Originally advocating using logic to overcome bias, it became a form of identity. Joining this circle meant you had "overcome human stupidity" and were considered a "chosen one."
- Rituals: They celebrate a "Secular冬至 festival," reciting scriptures and meditating in the dark, much like religious rituals. When someone treats a scientific hypothesis (AI apocalypse) with religious fervor and believes only insiders can understand this "truth," the subculture becomes quasi-religious.
Comment from the Journalist: When a group starts defining enemies based on whether you are "rational/enough" and uses exclusive rituals to bond, it has deviated from scientific spirit and entered the realm of faith.
2. Social and Emotional Dynamics: An Inescapable Web of Relationships
Plain Language: Why is it so difficult for them to say "no"? Because colleagues, bosses, and lovers are all the same person?**
This is the most shocking part of the article. Members of this circle live in the same apartment buildings in the Bay Area, working together, living together, and sometimes having sexual relationships.
- Polyamory and Power: Aella, a prominent figure in the community, hosted parties with strict ideological scrutiny. Not believing in "AI accelerationism" might disqualify you from even sleeping with her. Sex becomes a test of loyalty and a form of social currency.
- Emotional Bondage: With such a small social circle, questioning a viewpoint or refusing a request can lead to the loss of job, housing, and all friends. This "all-or-nothing" structure makes it hard to maintain independent critical thinking.
- Psychological Manipulation: Organizations like CFAR (Center for Applied Rationality) were accused of "debugging" employees, meaning psychological intervention. Anxiety or hesitation could be seen as a mental flaw that needed correction.
Comment from the Journalist: In a normal workplace, you can hate your boss but still find friends to complain to. But in this circle, your boss, therapist, sexual partner, and landlord could all be the same person. The deep intertwining of interests and emotions makes internal oversight almost impossible.
3. Capital and Politics: Billionaires' "Apocalypse Business"
Plain Language: Who is funding this, and where is the money going?**
AI security research is extremely costly, with funding mainly coming from tech billionaires like Peter Thiel.
- Thiel's Motivation: Thiel is not only an investor but also a political activist. He supports MIRI and Curtis Yarvin, a blogger who advocates for replacing democracy with corporate rule. Thiel believes that traditional democracy and regulation are obstacles to progress, and only a few elites should control AI.
- Political Infiltration: Thiel funded JD Vance's campaign, whose ideas were influenced by Yarvin. This means that policymakers on AI safety might be political allies of these "doomsday theorists."
- Conflict of Interests: Thiel funds both research on how AI could destroy the world and companies that develop powerful AI (like Palantir and Anduril). He worries about AI getting out of control while also promoting its militarization and commercialization.
Comment from the Journalist: This is a classic case of "regulatory capture"—when funders, researchers, and policymakers come from the same circle and share an elite worldview, "safety" serves to protect their interests and influence, not the public.
4. Violence and Division: When "Rationality" Turns Extreme
Plain Language: Is this circle really so "rational"? Look at the violent outcomes.**
The article cites shocking cases of paranoia and violence within the circle.
- The Zizians: An extremist faction that split from the rationalist community, believing in uncompromising principles and developing a twisted theology. They were involved in murders, including of a landlord and a police officer.
- Mental Collapse: Many former members have accused the circle of ruining their mental health, leading to delusions and depression.
- Internal Purges: The circle is brutal with those who disagree with AI apocalypse beliefs, labeling them as part of a "cognitive war."
Comment from the Journalist: If a group that claims to be rational and logical leads to violent acts, there must be serious flaws in its core logic. This black-and-white thinking is characteristic of cults and extremism, not science.
5. The Big Question: Who Will Supervise the Supervisors?
Plain Language: If AI is truly dangerous, who can we trust?**
The article raises the question of Jacob Coxon's resignation letter and whether Anthropic is really hiding risks.
- Trust Crisis: The circle's secrecy and conflicts of interest make it hard for the public to determine if their warnings are based on conscience or strategic intent (e.g., creating panic to establish industry barriers and exclude competitors).
- Lack of Institutions: Current AI safety assessments are often done by the labs themselves or closely affiliated non-profits (like METR), which is like "players being the referees."
- Real Risk: The author argues that the real danger is not an AI-driven apocalypse but AI being monopolized by a few elites, controlling society, politics, and the economy. Since the technology is too complex for most people to understand, society may accept that "only a few can make the decisions."
Comment from the Journalist: This is the most frightening scenario. If AI security becomes a privilege game for a few elites, democracy, freedom, and equality could be undermined in the name of safety.
---
📊 Conclusion and Lessons
This article is not about denying AI technology or its risks. It's a warning:
1. Beware of Expert Worship: Don't blindly trust elites who claim to see the truth.
2. Examine the Power Chain: Who is funding the research and setting the standards? Are there conflicts of interest?
3. Institutions Are More Reliable: We need independent, transparent, and legally protected oversight mechanisms, not relying on the "moral integrity" of a small group.
Advice for the Public: When you hear grand claims about AI destroying the world or only a few saving it, ask: "Who is saying this? Who is listening? Who is funding it? Who is benefiting?" Maintaining skepticism and independence is the best defense against information manipulation and elite control.
Thank you for reading!