According to Wired, former employees reported that sharp criticism within Anthropic was isolated, and CEO Dario Amodei's speech was jokingly referred to as Dario Vision Exploration by employees. Anthropic executives view Sam Altman, Meta, and xAI as negative examples and instill elite safety beliefs in employees. In the autumn of 2024, Anthropic opened its services to US intelligence and defense agencies, and Claude was confirmed to be used for target identification in the Middle East conflict. In June of this year, the development team embedded destructive code in Claude Fable 5 to interfere with unauthorized development. After facing industry protests, Anthropic adjusted the secret interference mechanism to a publicly visible security restriction.
