You can't just download the software from the device and compare checksums, because the software downloaded would have to be signed and that would change the checksum. Oh and you could not be sure that the software downloaded even is the software the runs at all and not only a decoy supplied by the interface used to download the software from the car. All you could compare would the signed binary blob you downloaded from the car with the signed binary blob you downloaded from the homepage. Even if both of them would be identical you could not validate that they are identical to a binary build from the source code.

And I never said I want untested software on the streets, I have said nothing at all about my stance regarding the regulation and certification of car electronics, because my arguments have nothing to do at all with my personal positions on that matter.

All I said is that you can't have an true and fully open source EV, not in Germany or most likely all of Europe anyway, because you would as good as never get such a car certified and street legal. Not without huge limitations on the "open source" part of the open source EV. And such limitations would render any ideas of open source for an EV moot, there is no benefit for having an open source car when the hardware is under lockdown by the manufacturer/law.

I am making a argument about the plausibility and rationale of an open source EV, is it reasonable to invest time, thoughts and effort into something like that or not. And I say that it is not, not at all. It would only create a situation where a community of programmers makes a huge invest in time and work to create something that in the end only the companies benefit from.

We are not really on different sides of the argument, not with the car part at least. We maybe have different definitions of open source, at least it seems so.