Former and current members of OpenAI's Safety Team discuss the challenges of aligning leadership vision with safety culture, legal implications of non-disparagement clauses, and future of AI alignment. They also explore existential risks of AI, safety responsibilities, and reliability concerns in OpenAI's tech. The conversation extends to ethical concerns in product development, AI regulation, and embracing struggle over comfort.