Berkeley-based longtermist and SWE. Into epistemic infrastructure (e.g. novel prior elicitation and reconciliation tooling), and exploring AI x-risk strategies orthogonal to technical alignment and government policy.