Arguing that AI safety is needed only demonstrates that it's altruistic, but you have a lot farther to go if you want to say funding research in this decade can be effective.