I dread that day. It's annoying enough listening to people's phone conversations; in my experience, listening to people using speech to operate a UI is a higher level of annoyance (because the person's speech tends to start and stop abruptly because the cognitive demands of operating a UI are higher than that of having a conversation with a person).

Valid point I hadn't considered. Maybe the day isn't as soon as we think—perhaps both will exist...

...until we have speech muffler envelopes built into our devices.