Now

Joined OpenAI's Safety Systems team in July. Safety measurement, model policy, and product for the people who need safeguards most. Capability cannot outrun safety. Do no harm, at a different scale.

Angel investing in technical founders building things that matter.

How AI can help with the slow, hard problems humans are bad at solving: climate, health, the ones where the damage accumulates invisibly.

What's actually worth doing with unscheduled time, versus what just feels productive.

A Brief History of Intelligence by Max Bennett. Evolution of the mind, and what it tells us about how AI is (and isn't) like us.

Elon Musk by Walter Isaacson. First principles thinking, regardless of everything else.