Matthew Barnett comments on Malign generalization without internal search