New York Times long feature: Dario Amodei’s early rationalism and effective altruism experiences shaped his views on AI
The New York Times published a long profile on November 5, 2026, titled “The Making of Anthropic’s C.E.O.,” focusing on how Anthropic CEO Dario Amodei’s thinking about artificial intelligence took shape. The reporter interviewed more than twenty classmates, friends, teachers, and colleagues, and reviewed his early emails and internal OpenAI materials. Amodei declined to be interviewed through a company spokesperson, and the article is mainly based on others’ recollections, with some support from emails and blog posts.
The report said that Amodei was active in rationalist and effective altruist circles when he was young. Rationalism formed around discussions by people such as Eliezer Yudkowsky and emphasized calm reasoning and attention to the risks that AI could pose to human society; effective altruism advocated maximizing the impact of charity and action through calculation. Friends and colleagues believed that this led him to conclude that AI was highly dangerous, that only like-minded people were suited to lead its development, and that it pushed him to advance the technology before those who did not value safety, while also seeking control over the technology.
After Amodei graduated from Stanford in 2006, he went to Princeton to pursue a PhD in physics, and in the same year his father died of hepatitis C. After reading Ray Kurzweil’s The Singularity Is Near, he told classmates that the singularity could arrive before 2040, and he took part in discussions on related blogs and the GiveWell forum. He joined OpenAI in 2016 to lead the safety and scaling-law team, left at the end of 2020 with his sister Daniela Amodei and others, and co-founded Anthropic in 2021. The report said that Anthropic’s co-founders were mostly connected to these two circles. In recent years, Amodei has argued for restricting chip exports to China, regulating AI to prevent it from going off the rails, and opposing the use of the technology for mass surveillance of Americans or autonomous weapons; in September 2026 he posted an article calling for slowing frontier AI development.
Original source:
Title: The Making of Anthropic’s C.E.O. Publisher: The New York Times (Ryan Mac, Kalley Huang, Cade Metz, Georgia Gee) Publication date: 2026-11-05 URL: https://www.nytimes.com/2026/11/05/technology/dario-amodei-anthropic-ai.html Nature of source: media report Verification status: The publicly available portions of the page, the reporters’ X posts, and several English-language rewrites were reviewed for this article; accessed on 2026-11-07. Full text is behind a paywall. Supported content: scope of interviews, Amodei’s refusal to be interviewed, influence of rationalism and effective altruism, OpenAI experience, founding of Anthropic, public positions, and the September article
ABAB AI Insight
Dario Amodei and his sister Daniela Amodei jointly lead Anthropic. Daniela previously worked at OpenAI in safety and policy-related roles, and her husband Holden Karnofsky is the founder of GiveWell and joined Anthropic in 2025. The report describes how Amodei and his sister, along with Jared Kaplan, Chris Olah, and others, once lived together discussing related articles and later co-founded the company. This network brought early philosophical discussions directly into laboratory governance and product direction.
During his time at OpenAI, Amodei led the safety team and took part in GPT-related work. Internal materials cited in the report show that after Trump’s election in 2016, he expressed concern about not wanting to research military-related technology in an authoritarian environment, and opposed accepting funding from Peter Thiel (Thiel’s Founders Fund later invested in Anthropic). In 2017, he pushed an institution associated with Karnofsky to donate $30 million to OpenAI. After leaving at the end of 2020 because of internal disagreements, he founded Anthropic, making the parallel advancement of safety technology and models its path.
Unlike OpenAI, Anthropic has from the start been led by founders connected to rationalism and effective altruism. The report argues that this made its public positions emphasize export controls, regulation, and restrictions on use more strongly. In the 2023 Hard Fork podcast, Amodei said he did not consider himself an effective altruist, agreed with some views but rejected tribal divisions. In September 2026 he posted a call to slow frontier development, saying that benefits would only materialize if things were built the right way.
This profile was published before Anthropic’s planned IPO and publicly contrasted the company’s governance with the founder’s personal intellectual history. For investors and regulators, the continuity between the founders’ early philosophical network and their current safety positions becomes an observable dimension; for the industry, it shows how safety narratives and technological competition coexist within the same entity. The point to watch is Anthropic’s actual policies on chip exports, military uses, and regulation after the IPO, and whether it continues to keep its distance from the relevant circles.
ABAB News · Cognitive Law
Early rationalist and effective altruist discussions were transformed into claims over control of technology in the lab.
Parallel advancement of safety technology and models became the path for founding a new company after leaving OpenAI.
The public stance before the IPO placed the personal intellectual history and corporate governance constraints under external scrutiny at the same time.