Does that even really matter with AI? If you already are waiting >1sec for a response/output, what is 0.2 seconds more?