given the extent to which audio phone calls have been replaced by text I'm always super sceptical of voice as the primary interface. It has so many disadvantag

given the extent to which audio phone calls have been replaced by text I'm always super sceptical of voice as the primary interface. It has so many disadvantages. No simple to see history, loud (by definition) in any non-private space, not composable, not easily copyable, very slow discovery, and so on. Voice is strictly serial. Tactile or visual interfaces you can just look at and move around in and immediately have a layout of the thing. Regardless of how smart the voice control is those issues are pretty intrinstic. There's also no 'halting state' to voice. That you can switch between different visual interfaces without losing your state is pretty necessary today. But you can't really stop or multitask audio controls sensibly.