Yeah, I think Kinney is wrong here. Someone else found a survey paper noting the theoretical possibility of using BsaI and BsmBI in a way that leaves the sites in the final genome, but I haven't seen anyone who actually did that. So I'm leaning to a false discovery here.