Anthropic Co-Founder Used Stuffed Animals for Workplace Conflict Resolution

Anthropic co-founder Daniela Amodei and her husband, Holden Karnofsky, utilized a unique method for navigating workplace conflicts: a stuffed-animal advisory council. According to a report in The Atlantic, the pair gave names and distinct personalities to the plush toys to engage in "managerial role-playing."

This unconventional practice reportedly began during their tenure at OpenAI, preceding Anthropic's official launch. The stuffed animals, each with assigned "character profiles," served as a personal advisory council. For instance, a cat named Jinji was characterized as lazy and encouraged Amodei to cancel meetings. A detail-oriented pink bear named Maura represented a type-A manager, while a panda called Beary Bonds was designated as the "patron bear of generosity and empathy."

Amodei described the exercise as helpful in analyzing different personality types, blending humor with a strategic approach to management challenges. However, few OpenAI employees were aware of this practice, and its influence on Amodei's decisions remained unclear, according to the report.

The period in which these stuffed animal sessions occurred coincided with growing frustration among Amodei and her brother, Anthropic CEO Dario Amodei, regarding OpenAI's leadership and internal culture. Daniela Amodei, who joined OpenAI in 2018, increasingly clashed with company leaders, including Greg Brockman and Mira Murati, as her responsibilities expanded from English literature studies to managing technical teams.

By 2020, Dario Amodei expressed deep distrust of then-OpenAI CEO Sam Altman, and he, along with other safety-focused researchers, began considering a departure. This led to the formation of what some colleagues termed "Dario's clique," characterized by private communication channels and meetings prioritizing AI safety. Daniela played a significant role in negotiating the group's exit from OpenAI, which was internally referred to as "the divorce."

The Amodei siblings, along with six other co-founders including Jack Clark, Sam McCandlish, Tom Brown, Chris Olah, and Jared Kaplan, officially left OpenAI in December 2020. They subsequently founded Anthropic, initially conceived as a compact AI safety research lab. The company has since grown substantially, becoming a major competitor to OpenAI in the development of advanced AI models with a strong emphasis on safety.

Daniela Amodei currently holds the positions of president and board chair at Anthropic, overseeing critical areas such as research, engineering, product development, governance, and commercial operations. Karnofsky also joined Anthropic's technical staff after serving on OpenAI's nonprofit board.