Paul Billings has spent four decades asking the same hard question from different seats: how do you turn genetic information into something useful without letting it become something dangerous?