How technical do you need to be with your optimization statements and outcome checking? Isn't that moat constantly shrinking if AI is constantly getting better?