I'd like to strongly disagree! When I'm barreling down the road with 2 metric tons of steel under my ass I'll prefer the shittiest tactile UI that let's me FEE

I'd like to strongly disagree! When I'm barreling down the road with 2 metric tons of steel under my ass I'll prefer the shittiest tactile UI that let's me FEEL by touch alone if a) the command I'm about to issue is the desired one and b) if i actually issued the command! I'll take a physical button or dial for my AC over a touch screen submenu slider any day in a car, even if the touchscreen was <18ms and button had a second or two of delay. In this scenario the responsiveness of UI is completely irrelevant if it requires me to take my eyes off the street to find it and to validate my input in the first place. When i operate a secondary control in a car i technically already have to shortly compromise ideal safety by taking one hand off the steering wheel and thinking about an additional process. That might be brief, but things happen FAST at 100 km/h. If __on top of that__ I have to stop observing what my car and other cars around me are actually doing, hence interrupting the control mechanism that would let me interrupt the current action and react in the case of imminent danger, I'll be compounding the risk of the whole operation (driving) by orders of magnitude. "Waiting for response" might feel annoying, but it's nowhere near the safety concern "looking at a screen on the center console to hit the right thing on a glass pane while moving and experiencing inertia" will always present. And that's even without additional complications like "hitting the wrong thing because I didn't look and having to visually debug what i just did on the fly, all still while driving". Touchscreens aren't an issue because their software is bad, but because they're the wrong interface in the first place. If the UX is bad, the best UI won't fix it. And this goes x1000 for tools that can literally kill me and others.