Imagine if I pitched competitive rocket league players on an idea to give them an edge by developing voice activated control so they can say things like “air dribble” or “wavedash” and the AI will help control the rocket car for the players to offload some work… then imagine if instead of a silly video game we were talking about operating a ~70 ton cannon?
If you watch the way the crews of these vehicles operate, they don’t rely on the armor of the tank, they coordinate closely to move in and out of cover/defilade to apply forward force with minimal exposure to enemy return fire… so how does being able to give vague instructions to a vehicle help with that? This is all about precision and getting it right the first time… and sure you can say “well its onfly for the fiddly stuff like the sensors” but that doesn’t make the argument for it very convincing either… especially if you are scared for your life to some degree while operating it.
Now bring into context that the experience auditorily, mentally and physically of being hit by incoming fire whether it credibly can threaten penetration or not while inside an armored vehicle is not an experience you can describe…
This breaks my brain…
edit I think an automated retreat or evasive maneuvers kind of capability is absolutely worth integrating but voice control in an emergency feels like an insane idea to me. That is PRECISELY when I want a damn button.
The military’s been looking for the holy grail of natural language command and control since the 1950s (Noam Chomsky’s linguistics work was funded by the DoD for exactly this reason). Hasn’t happened yet, but maybe we’re finally onto something with “Don’t besiege Moscow, take it immediately, make no mistakes.”
Natural language control is a fantasy that no serious UI designer believes is inherently powerful.
No serious software engineer thinks natural language programming would be superior.
Part of the essential job of the UI is to inform what the intuition should be for the human user and to emphasize important parts clearly in the design language. Rarely does natural language control enable that in a transformative way.
The most crucial aspect of voice control/natural language control is for accessibility for impaired or disabled people who could use it to improve their lives which is an awesome reason to do it as far as I am concerned, but all this other stuff is kind of ridiculous.