avturchin comments on Rejecting Violence as an AI Safety Strategy