Actually, I don't want OLED in my car. A car that is supposed to have at least 10 year lifetime. Beyond what current OLEDs can do. And in a situation where the burn-in issue is a factor (yeah, OLEDs have a slight burn-in potential - bad for a car interface situation where things are mostly static and high contrast).
As for the current screen interface, it's clear that Tesla understands the basics - sliders work best, buttons need to be big, tactile feedback. Some color coding needs to be done - not incompatible with the skinning idea, but makes it a bit harder. Also, what level of skinning are you shooting for? Color/gradients only? Metrics/button size? Full layout geometry? Each, in turn, is harder to do correctly. One of the comments by the engineers at the MP unveiling indicated that finding a supplier that could meet all the requirements was tough (don't forget the sunglasses issue - at a 90 degree rotation to how you'd want it on a laptop). Something that's easier to fix once you're ready to order in quantity.
As for using Linux, it's a good choice for an embedded situation, especially where power is concerned. And a car interface should be simple - no need for a GPU / too many transitional effects, though going with a low-power GPU could help with good birds-eye navigation. The current Atoms have more than enough grunt to do better map views than all but the most expensive GPS units.