conversational interfaces are superior as the input method but text and images is always superior as the output method (except for driving without FSD or another rare scenarios)