In lieu of the current lip-sync mechanism, do you have any plans to try giving the AI model more verbs to work with the motors (within safe thresholds), and then encouraging it to do more than just lip-sync? Lip-sync should clearly be the default for normal verbal responses, but you could also encourage responses like jaw-drop reactions, or patterns (would be funny to see it using its tail to tap out "shave and a haircut" as a bit).