# Holden Karnofsky

Research and product · Safety planning at Anthropic · joined 2025

Co-founded GiveWell and ran Open Philanthropy. He joined in 2025 to work on the Responsible Scaling Policy.

Full profile: https://www.autumn.ai/resources/anthropic/people/holden-karnofsky

**At Anthropic:**
- Jan 2025 – present: Member of Technical Staff
Sources: [1](https://www.linkedin.com/in/holden-karnofsky-75970b7)

**Career before Anthropic:**
- Apr 2024 – Jan 2025: Carnegie Endowment for International Peace, Visiting Scholar
- Jul 2023 – Apr 2024: Open Philanthropy, Director of AI Strategy
- Jun 2021 – Jul 2023: Open Philanthropy, Co-CEO
- Jun 2017 – Jun 2021: Open Philanthropy, Chief Executive Officer
- Aug 2014 – Jun 2017: Open Philanthropy, Project Lead
- Jun 2007 – Jun 2017: GiveWell, Co-Founder, Co-Executive Director
- 2004–2007: Bridgewater Associates, Investment Associate
Sources: [1](https://www.linkedin.com/in/holden-karnofsky-75970b7)

**Education:**
- BA, social studies, Harvard University, 2003
Sources: [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)

**Lesser-known facts:**
- Studied social studies at Harvard, where he wrote for the Harvard Lampoon. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Once shared an apartment with Dario Amodei. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Worked at Bridgewater Associates after college, where he and Elie Hassenfeld started a charity club in 2006 to research where to give. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Quit Bridgewater in 2007 to run GiveWell full time, funded by donations from colleagues. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Annual donations to GiveWell-recommended charities grew from $1.6 million in 2010 to $110 million in 2015. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Sat on OpenAI's board of directors from 2017 to 2021. [1](https://en.wikipedia.org/wiki/Holden_Karnofsky)
- Worked with METR in 2023 to develop the idea of responsible scaling policies, then led Anthropic's RSP v3.0 rewrite. [1](https://forum.effectivealtruism.org/posts/DGZNAGL2FNJfftwgE/responsible-scaling-policy-v3-1)
- Named his blog Cold Takes because each post is written at least a month before it is published. [1](https://www.cold-takes.com/about/)

**Interviews and talks:**
- [80,000 Hours podcast #226 on concrete AI safety at frontier AI companies (2025)](https://80000hours.org/podcast/episodes/holden-karnofsky-concrete-ai-safety-frontier-ai-companies/)

**Known for:**
- Thoughts on the Singularity Institute (SI, 2012)
Sources: [1](https://www.alignmentforum.org/w/holden-karnofsky)

Researched with Autumn (https://www.autumn.ai) from public sources. Updated 2026-10-18. Everyone at Anthropic: https://www.autumn.ai/resources/anthropic/people