Stuart_Armstrong comments on How should AIs update a prior over human preferences?

[Stuart_Armstrong]

thanks! Changed title (and corrected badly formatted footnote)