person
Holden Karnofsky
Co-founder of Open Philanthropy; AI safety funder and strategist
Long-time AI risk thinker and philanthropic strategist at Open Phil. Has moved from running a generalist effective-altruist grantmaker to full-time AI safety advocacy. Writes the Cold Takes blog on transformative AI.
Profile
expertise
Policy / meta
Specialises in AI policy, regulation, governance, philanthropy, or movement strategy. Reads the technical literature but does not produce it.
Co-founded GiveWell and Open Philanthropy. Drove much EA-aligned AI safety funding. Long-form 'Cold Takes' essays influential inside the field but he is not a technical researcher.
recognition
Field-leading
Widely known inside the AI and AI-safety community. Appears repeatedly in top venues, podcasts, or governance forums. Not a household name to outsiders.
Universal name within EA/safety; relatively low public profile.
vintage
Pre-deep-learning
Active before AlexNet. The existential-risk frame matures (FHI, OpenPhil, EA). Public AI commentary still rare; deep learning not yet dominant.
GiveWell 2007, Open Philanthropy 2014. AI safety funding decisions from the FHI/EA pre-deep-learning era; his frame predates deep learning.
Hand-classified. See the board for the criteria and the full grid.
p(doom)
- 10–90%2022-08
Definition used: Explicitly unresolved wide range, framed around uncertainty about alignment difficulty.
How might we align transformative AI if it's developed very soon? · AI Alignment Forum
Strategy positions
Governance firstendorses
Lead with regulation, treaties, liability regimesHas advocated for regulation, evals, and careful coordination between frontier labs as the central lever.
I'm highly uncertain, somewhere between a 10% and 90% chance that things go badly. The honest answer is that we just don't know.
Context: From his discussion of aligning transformative AI in surprisingly short timelines.
Closest strategy neighbours
by jaccard overlapOther people whose strategy tags overlap with Holden Karnofsky's. Overlap is on tag identity, not stance; opposites can show up if they reference the same tags.
Record last updated 2026-04-24.