Andrew McKnight comments on Anthropic’s Core Views on AI Safety