I don't want touchscreens in a car. I don't want my UI to change. That doesn't change the fact that many people do. Companies build for what customers want, not what engineers want. And even then, many (maybe even most) engineers would align with the average customer.

Controls for the driver need to be physical, full stop.

There's more than just driver preference at play here; there's physical usability and its impact on safety, and that should override expectations people have with tablets.

We're flying around with 2 ton missiles, it's best we treat them with the appropriate respect.

Or you can put the Tesla on autopilot for the few seconds while fiddling with the touch screen, which (to me) feels much more safer than e.g. messing with a physical radio in a car without such assistence features.

(I'm not arguing with you, in fact I mostly agree with you.)

Are there studies proving this?

https://www.thestar.com/autos/2022/05/28/touchscreens-are-co...

"Of the 30 different vehicles tested from 18 different brands, not a single vehicle from the study produced a low level of demand for the driver’s attention. Seven vehicles tested produced moderate demand, while 23 generated a high or very high demand. In fact, the study found drivers distracted, with their eyes completely off the road for as long as 40 seconds at certain times.

The concentration of critical functions being housed within a centre touchscreen becomes even more problematic given the lack of any feedback they offer, forcing drivers to look at the screen to confirm activation. Industry research has shown that it’s significantly easier to locate a physical control than hunting through sub-menus, as tactile or haptic feedback lets drivers develop muscle memory for key controls."

And it's an editorial but I fully agree with its conclusions:

https://www.thedrive.com/tech/39304/why-im-done-pretending-t...

I'm a fan of Mazda's use of a dial for control when the vehicle is moving. Touchscreens are stupid if the driver has to use them.