Maureen Ondieki comments on My model of how different AI risks fit together