Charlie Steiner comments on An observation about Hubinger et al.’s framework for learned optimization