Note you still didn't answer or provide any evidence of "Safety Engineering" from Mercedes. Just the usual Tesla trolling.
AI is the approach that Tesla, Waymo, Nvidia, Cruise, Comma, and even Mercedes are taking. So to ask the question one last time (third time lucky) - How is Mercedes validating the data they use for training without "Testing in Prod" as you criticise Tesla for.
Either: They aren't, in which case your whole rant on the lack of "Safety Engineering" at Tesla is moot, because Mercedes is rolling out essentially untested AI systems.
OR
They are validating their data in real life just like Tesla do, in which case your whole rant on the lack of "Safety Engineering" at Tesla is moot, because Mercedes is just as bad and supposedly have "Safety Engineering" in your opinion.
This is why I keep asking, and this is why you keep not answering.