But on the other hand, if — instead of capturing all the microexpressions, which is rather difficult — we capture the intention, and then just reapply that skin to the avatar. Or as many WaveNet-like systems are now doing, you can convey using your verbal tonality and things like that. Such nonverbal expressions — even in a relatively low-bandwidth microphone or earphone and things like that — adds a different, nuanced layer, so that we can actually stay focused and on topic, even though we’re technically in different spaces.