Richard_Kennaway answers Danger(s) of theorem-proving AI?