AI Safety Thursday: Technical AI Governance - Motivations, Challenges, and Advice — LessWrong