The Cult of Anthropic

Out of all the model developers, Anthropic stands out as perhaps the one with the most ideologically devoted employee base, one that some researchers even describe as “cultish.” The reasons may include CEO Dario Amodei’s monthly ”Dario Vision Quests” where he talks to employees about the future of AI (with more than slight religious overtones) and that its executives frequently speak publicly about the dangers of superintelligent AI.

One less commonly discussed factor is Anthropic’s recruiting process. Like other AI labs, it puts candidates through a battery of tests to check their technical skills. Additionally, there’s what’s called the behavioral interview, a fancy term for discussions that at most firms serve as a sanity check to make sure that the candidate has basic social skills and can work with others, researchers say. For Anthropic, though, the behavioral interview has more weight and a noticeably more ethics-focused bent, said people who have gone through the process.

One engineer who interviewed at Anthropic was told by a recruiter that only a small percentage of applicants get through the behavioral interview unscathed. For many candidates, Anthropic recruiters reach out to former colleagues and friends to quiz them about their character, the engineer said.