“In its frontier safety roadmap, Anthropic proposed that governance for advanced AI should be analogous to the regulation of nuclear energy or finance, suggesting a body similar to the Nuclear Regulatory Commission or the SEC.”