voice-first products are insanely overhyped. anything that requires you to explicitly talk to it is strictly worse than just having a screen. the idea of having "an AI assistant you can just walk around with talk to" is useless

agreed, my ideal interface is still buttons. the current paradigm just can't dispose of the inconvenience of holding the buttons & visual output in a certain orientation that remains slightly strenous. massive improvement compared to 20 years ago, still hasn't arrived at the point.

I'm online sometimes

https://bento.me/m-j-r

a lightweight headset is the holy grail, that's where the sensory cortex is. other than that, there's a balance of somatic/haptic inputs vs traditional voice command, probably a wearable or implant on the hands.