Liam Donovan comments on Some problems with making induction benign, and approaches to them