Holden Karnofsky

Holden Karnofsky (born 1981) is an American grantmaker who cofounded the charity evaluator GiveWell and the foundation Open Philanthropy, the durable financial backbone of the Effective Altruism movement and of longtermist and AI-safety work. As Open Philanthropy's co-chief executive he directed the institutional patronage that sustained AI-existential-risk research, was an early backer of the AI company Anthropic, and left Open Philanthropy's staff in 2024 and joined Anthropic in 2025. He is married to Anthropic cofounder and president Daniela Amodei.¹²

GiveWell

Karnofsky and Elie Hassenfeld met as analysts at the hedge fund Bridgewater Associates, where they were part of a group of eight finance-industry friends researching charitable giving in their spare time. In 2007 the two left their jobs, raised about 300,000 dollars from former coworkers, and founded GiveWell, an organization that rates charities by the measured cost-effectiveness of their interventions and recommends a short list of top picks. The model, evidence-based giving aimed at maximizing lives saved per dollar, became one of the two intellectual streams of effective altruism, alongside the Oxford moral philosophy of Toby Ord and William MacAskill.³⁴

Months after the launch, Karnofsky was caught promoting GiveWell anonymously: in December 2007 he posted a question soliciting charity recommendations on the community site Ask MetaFilter under a false identity, then answered it touting GiveWell without disclosing his role. The board removed him as executive director and board secretary effective January 3, 2008, cut 5,000 dollars from his salary, and required him to take a professional-development course; he was later restored to the leadership. GiveWell grew into a major channel for evidence-based giving, directing more than 1.45 billion dollars to its top charities between 2009 and 2024, with roughly half of that going to the Against Malaria Foundation at an estimated cost of about 5,500 dollars per life saved.⁵⁶

The Clear Fund

GiveWell's legal name is The Clear Fund, EIN 20-8625442, a 501(c)(3) with an IRS ruling date of July 1, 2007, filing from 1714 Franklin Street in Oakland. For 2023 it reported revenue of 219.6 million dollars and year-end assets of 208.6 million.⁷

Open Philanthropy

GiveWell's research into larger and riskier giving opportunities led to a partnership with Good Ventures, the foundation established in 2011 by Facebook cofounder Dustin Moskovitz and his wife Cari Tuna. The two efforts joined as the Open Philanthropy Project in 2014 and began independent operations in 2017, with Karnofsky as chief executive. Funded chiefly by Moskovitz and Tuna, Open Philanthropy directed more than four billion dollars in grants by 2025 across global health and "longtermism," the latter covering AI risk, pandemic preparedness, and the far future. Its AI-risk money flowed to the research community around the Machine Intelligence Research Institute, the Future of Humanity Institute, and the Rationalist Community, making Open Philanthropy the principal patron of AI-safety work outside the technology companies. The organization renamed itself Coefficient Giving in 2025.¹²

The effort began inside GiveWell as "GiveWell Labs" before splitting off. On June 16, 2021, Karnofsky announced that his fellow cofounder Alexander Berger had been "promoted to co-CEO," in "a dual-CEO structure, with both of us reporting directly to the board"; Karnofsky kept oversight of "biosecurity and pandemic preparedness and potential risks from advanced AI," and Berger took global health, scientific research, farm animal welfare and criminal justice reform.⁸ On July 27, 2023, writing "on behalf of Open Philanthropy's board," Karnofsky announced that Berger was sole chief executive and that his own title was changing "from co-CEO to Director of AI Strategy"; he stayed on the staff.⁹ In March 2017 Karnofsky steered Open Philanthropy's 30-million-dollar grant to OpenAI, a sum that bought a board seat he held until 2021, when he resigned it after his wife cofounded a competing lab. Between 2021 and 2023 Karnofsky wrote the "most important century" series on his blog Cold Takes, more than twenty essays arguing that the development of transformative AI could make the present century the most consequential in human history, a thesis that supplied the intellectual rationale for Open Philanthropy's concentration of money on AI risk.²¹⁰

Anthropic and the Amodei Tie

Open Philanthropy was an early funder of Anthropic, the AI company founded in 2021 by former OpenAI researchers from the effective-altruism milieu, and Karnofsky held a seat associated with that early backing. Karnofsky married Daniela Amodei in 2017; she cofounded Anthropic with her brother Dario Amodei and serves as its president. In 2022 Sam Bankman-Fried invested roughly 500 million dollars of FTX-derived capital in Anthropic through Alameda Research, placing the company at the intersection of the movement's two largest funders during the period Karnofsky's foundation and his wife's company were both tied to it. Karnofsky took a leave from his co-chief-executive role in 2023 to focus on AI strategy, became Open Philanthropy's director of AI strategy in July 2023, left its staff for the Carnegie Endowment for International Peace in April 2024, and joined Anthropic in January 2025 to work on its responsible-scaling and safety policies.²¹¹¹²

The overlap drew the criticism that the largest funder of independent AI-safety research was led by a man married to the president of a frontier AI lab his foundation had financed. Karnofsky's OpenAI board seat from 2017 to 2021 had already placed him inside one leading lab before the marriage tie connected him to a second; commentators noted that the same small network of effective-altruism figures sat on the funding, governance, and research sides of the field at once. At Anthropic he took up work on the company's Responsible Scaling Policy, the framework of capability thresholds and safety commitments the company uses to govern the release of more powerful models.²¹²

Redwood Research

When Nate Thomas announced Redwood Research on October 6, 2021, he wrote that "our board is Nate, Paul Christiano and Holden Karnofsky." Karnofsky, then co-chief executive of Open Philanthropy, was listed as Redwood's unpaid treasurer on its return for 2021 and as treasurer and CFO for 2022, with Christiano as secretary. Open Philanthropy recommended four grants totaling 9.42 million dollars to Redwood in November 2021, followed by 10.7 million dollars in August 2022 and 5.3 million in June 2023; the archived grant pages do not mention Karnofsky. Redwood's return for 2023 listed him as a director "thru 11/23," Christiano "thru 6/23," and Open Philanthropy's senior program officer for AI risk, Ajeya Cotra, as a director.¹³¹⁴¹⁵

The Action Fund

Karnofsky was also an officer of the Open Philanthropy Action Fund, the 501(c)(4) of Moskovitz and Tuna, which files from the Palo Alto address of the Good Ventures Foundation. Its returns list him as executive director for 2017, chief executive and director for 2018 to 2020, co-CEO and director for 2021 and 2022, "co-CEO thru 7/23/23" for 2023, and a director only for 2024. None of its officers was paid by it.¹⁶

Carnegie Endowment

On April 29, 2024, Karnofsky wrote: "Effective today, I've left Open Philanthropy and joined the Carnegie Endowment for International Peace as a Visiting Scholar. At Carnegie, I will analyze and write about topics relevant to AI risk reduction." He named as his first subjects "what AI capabilities might increase the risk of a global catastrophe," how to "catch early warning signs of these capabilities," and what protective measures would be needed. He gave his reason for leaving: "my work no longer involves significant involvement in grantmaking, and given that I've overseen grantmaking historically, it's a significant problem for there to be confusion on this point. Philanthropy comes with particular power dynamics that I'd like to move away from, and I also think Open Philanthropy would benefit from less ambiguity about my role in its funding decisions (especially given the fact that I'm married to the President of a major AI company)." He wrote that he would "continue to serve on its board of directors (at least for the time being)," and named Carnegie's president, Tino Cuéllar, and its scholar Matt Sheehan, who "specializes in China's AI ecosystem."¹⁷

In the comments a reader posting as Adam_Scholl suggested Karnofsky might be among the people "who stands to personally lose the most money in the event of an AI pause." Karnofsky replied: "My spouse isn't currently planning to divest the full amount of her equity. Some factors here: (a) It's her decision, not mine. (b) The equity has important voting rights, such that divesting or donating it in full could have governance implications." He added: "The bottom line is that I have a significant conflict of interest that isn't going away, and I am trying to help reduce AI risk despite that. My new role will not have authority over grants or other significant resources besides my time and my ability to do analysis and make arguments."¹⁷

The first comment on the same post, posted anonymously by a writer whose description of the couple's giving matches Dustin Moskovitz and Cari Tuna, read: "I'm grateful that Cari and I met Holden when we did (and grateful to Daniela for luring him to San Francisco for that first meeting). The last fourteen years of our giving would have looked very different without his work."¹⁷

Anthropic Staff

Fortune reported in February 2025 that Anthropic described Karnofsky's title only as "a member of the technical staff" reporting to its chief science officer, Jared Kaplan, and quoted a spokesperson: "He is working on the responsible scaling policy and other aspects around safety planning ahead of highly advanced AI systems." The magazine wrote that with his arrival "Anthropic now has three close family members holding influential roles at the company," and that he had served on OpenAI's board "between 2017 and 2021."¹⁸