Dustin comments on Thoughts on the Alignment Implications of Scaling Language Models