Shmi comments on Taxonomy of AI-risk counterarguments