Joern Stoehler comments on Superforecasting the premises in “Is power-seeking AI an existential risk?”