I'm interested in AI safety, interpretability and climbing trees. I'm currently on leave from Harvard working at Anthropic.