Steven Byrnes comments on My current framework for thinking about AGI timelines