This is not Star Trek. Speech recognition isn't nearly as productive as typing with a keyboard or clicking the mouse, so apart from people not being able to type or click, nobody really 'needs' speech recognition. It's a gimmick.
Now? "Ok Google...set timer for 30 mins." Much quicker than unlocking, finding the Clock app, and fat-fingering the time widgets. This also works for reminders and a host of other things.
Speech recognition is not a gimmick.
"Children! 10 more minutes, then we're turning screens off. Hey Siri! Set a timer for ten minutes."
Much easier than fiddling with the crown to set a 10 minute timer.
Of course coders have made progress at using dictation:
http://ergoemacs.org/emacs/using_voice_to_code.html
Like I said, you lack imagination. Also, no one said you need to use speech for every task. It might simply help with RSI so you don't end up like this:
Also, most people aren't coders. Designers, for example, might be using Photoshop with a Wacom tablet and would like to not have to use a keyboard. "New layer" "Layer hide" " "move layer up/down". "Text tool. Change text to San Francisco regular 16 point "