Anthropic employees exhibit cult-like devotion to AI model Claude

At most tech companies, employees rally around stock options, product launches, or maybe a really good espresso machine. At Anthropic, they rally around the existential fate of humanity.

The AI safety firm behind Claude, now valued between $350 billion and $380 billion, has developed an internal culture so mission-driven that outside observers are struggling to distinguish it from religious fervor. Employees describe their work in terms that sound less like software engineering and more like spiritual calling, with many citing the prevention of AI catastrophe as their primary reason for showing up each day.

Vision quests and soul documents

CEO Dario Amodei holds company-wide sessions every two weeks called “Dario Vision Quests,” or DVQs. These are structured philosophical discussions about AI alignment, the societal implications of artificial intelligence, and how Anthropic’s work fits into a broader moral framework.

Claude itself operates under what insiders call a “soul document,” an 80-page constitution that governs the model’s behavior and allows it to refuse directives it deems unethical. Rather than simply training a model to be helpful, Anthropic has attempted to encode something resembling values into Claude’s architecture, a framework for the AI to push back against requests it judges to be harmful.

The Claude-pilled phenomenon

An OpenAI insider known as “roon” described Anthropic as an organization that “loves and worships Claude,” coining the term “Claude-pilled” to characterize the culture. Employees don’t just build Claude. They appear to believe in it, in a way that blurs the line between engineering conviction and something closer to faith. Former staff have reportedly returned to the company shortly after leaving, drawn back by a longing for the mission-driven environment they couldn’t replicate elsewhere.

Pedro Domingos, a computer science professor at the University of Washington, offered a sharper assessment.

“These people are treating AI as if it’s a human in disguise, with human motives, with human emotions.”

Born from a schism

Dario Amodei and his sister Daniela left OpenAI in 2021, taking a group of researchers with them. The departure was rooted in disagreements about how aggressively to commercialize AI and how seriously to take alignment research.

The company’s intellectual roots trace to effective altruism, the philosophical movement focused on using evidence and reason to do the most good possible. That heritage shows up everywhere, from the DVQs to the soul document to the company-wide discussions about whether and how AI could pose existential threats.

The effective altruism connection also raises eyebrows. The movement took a reputational hit after the collapse of FTX and the fraud conviction of Sam Bankman-Fried, who had positioned himself as one of EA’s most prominent adherents.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.