Sophie Kim

I’m Sophie, trying to figure out how to make the future go well.

Currently, I’m:

• An extension fellow at MATS, doing research on data poisoning; and
• Writing about AI safety on my Substack, The Counterfactual (all posts sorted by theme here).

I’m also thinking about the grantmaker bottleneck and how to solve that bottleneck (with, say, an AI safety grantmaking fundamentals course). Right now, I’m on leave from Stanford University.

Previously, I was a fellow at ERA, researching open-weight model policy options.

Career-wise, I’m honestly not sure where I’ll end up! I’m most drawn to environments that enable agency—places where I can move quickly and make things, with minimal bureaucracy—and to paths that are highly leveraged and explicitly non-credentialist. Technical research is my main interest at the moment, but I may end up back in policy, or somewhere else entirely (generalist work?), depending on where life takes me over the next few months. (In the meantime, I’m still writing on policy & strategy on my Substack!)

Ways you can help me:

• Do you work on AI or biosecurity policy in D.C., or know someone who does? I'm looking for help getting these interventions implemented, and would love recommendations on who to reach out to.
• Are you a grantmaker? I'd love to chat with you about the grantmaking curriculum I'm working on!

Contact me if any of the above interests you! You can also leave me anonymous feedback here.