My current research interests include AI safety, LLM reasoning, reinforcement learning and singular learning theory.