Before she was president of Anthropic, the AI safety company reportedly eyeing a valuation in the neighborhood of $2 trillion ahead of a potential IPO, Daniela Amodei was working through job dilemmas with an unusual set of advisers. According to reporting in the Atlantic adapted from journalist Kevin Roose’s forthcoming book, those advisers were stuffed animals.
The piece, published Wednesday, describes what Amodei and her husband, effective altruism figure Holden Karnofsky, called a “personal advisory council” — a collection of plush toys, each with its own assigned character and role, consulted when office problems needed sorting out. The habit predates Anthropic entirely; per the report, it began while both were employed at OpenAI.
A council with its own org chart
Roose’s account gives the group a distinct internal culture. One member, a cat named Jinji, was characterized as lazy and inclined to encourage Amodei to cancel meetings and avoid piling on too much work. A pink bear called Maura was cast as the detail-oriented, type-A manager of the group. A panda named Beary Bonds served as the patron of generosity and empathy.
Amodei told Roose in 2023 that the toys were genuinely useful for analyzing different personality types, according to the report. Few people at OpenAI were aware the council existed, and those who did reportedly weren’t certain how much weight she actually gave its input.
Breitbart News reported on the Atlantic excerpt Wednesday.
From English lit to OpenAI, then out the door
Amodei arrived at OpenAI in 2018 by a less conventional route than many of her peers, having studied English literature rather than coming up through a technical pipeline. Colleagues there, per the report, viewed her as the more emotionally attuned of the two Amodei siblings — a reputation that shaped her expanding role, which eventually included managing technical teams. During that same period, Karnofsky sat on OpenAI’s nonprofit board.
The excerpt also describes friction between Amodei and two prominent OpenAI figures: co-founder Greg Brockman and then-executive Mira Murati. Complicating the picture, Brockman had delivered a toast at Amodei and Karnofsky’s 2017 wedding, according to the report.
Overlapping with those tensions, her brother Dario Amodei, Anthropic’s CEO, was growing distrustful of OpenAI chief executive Sam Altman and of Brockman, and began considering an exit alongside other researchers focused on AI safety. The siblings came to be at the center of what colleagues called “Dario’s clique,” which the report says maintained private Slack channels and separate meetings devoted to safety concerns. By 2020, Dario was recruiting people to leave with him and build a new company around that mission. Daniela took the lead in negotiating the split with OpenAI — a process insiders nicknamed “the divorce.”
Seven founders become eight as the lab scales
The Amodeis departed OpenAI in December 2020 and founded Anthropic with five others: Jack Clark, Sam McCandlish, Tom Brown, Chris Olah and Jared Kaplan. Ben Mann joined later, bringing the co-founder count to eight. The company began as an AI safety research lab of roughly 50 people. It has since grown to thousands of employees, with Daniela Amodei serving as president and board chair and overseeing research, engineering, product, governance and commercial operations. According to a photo caption in the report, Karnofsky later joined Anthropic’s technical staff.
Anthropic has positioned itself at the center of calls for a more organized, slower pace of AI development and for government regulation of the technology — a stance that has drawn sharp criticism from the right.
Skepticism from the right over regulation push
Breitbart News editor-in-chief Alex Marlow, in a recent speech, described the industry’s regulatory posture in blunter terms, arguing that “They are building a cartel right in front of your face, and they’re pretending like they’re the altruist in the process.”
Marlow pointed to METR, a Berkeley-based group he said Dario Amodei proposed, noting that one of its most recognizable names is Karnofsky — who, Marlow observed, is married to Daniela Amodei.
“So, Dario’s solution to the deadly AI that he created is to appoint a group tied to his own family,” Marlow said, drawing laughter from his audience. He added that Dario Amodei “wants an AI regulator controlled by Dario Amodei and Dario’s sister, so they can create the framework under which all of us operate,” and warned that the push for regulation is driven by “power and market capture” rather than “saving humanity.”
The broader debate over AI governance has only grown more tangled as major labs release powerful models while simultaneously urging lawmakers to regulate them, with China’s own ambitions adding another layer. Investor David Sacks has argued that when figures like Dario Amodei, Altman and Elon Musk call for a slowdown, the motive is less altruism than capturing the industry and carving out exemptions from antitrust scrutiny — a conflict that, in his telling, President Donald Trump understands. Breitbart News social media director Wynton Hall’s New York Times bestseller, Code Red: The Left, the Right, China, and the Race to Control AI, makes the case for a MAGA-aligned AI agenda that avoids ceding control either to Silicon Valley’s left or to Beijing.
For now, the stuffed-animal council remains a decidedly pre-Anthropic footnote to a much larger story about how a handful of insiders left one of the most consequential AI labs in the world and built a rival — one now lobbying for the rules that will govern it.
Source: www.breitbart.com — https://www.breitbart.com/tech/2026/10/01/anthropic-co-founder-daniela-amodei-kept-personal-advisory-council-of-stuffed-animals-to-solve-problems/