I feel this. Both in terms of driver engagement safety and in how much I loathe traditional automatic transmissions. Still stuck owning one in one of the two vehicles I have at the moment but only because it was all I could afford for the second of two vehicles large enough to fit all my kids.
I have had several manual transmission vehicles and the other current one is a PHEV and one of the rare models that is a series hybrid so it drives like a true EV.
I had a 1980 year Oldsmobile 98 that didn’t have as many crazy issues as yours but did have one amazing one.
Driving home one evening from college classes the headlights didn’t work so I took it into the shop.
Couldn’t find anything normal as a cause but I had one of those old time small town mechanics that couldn’t stand to lose to the car. So he said he wouldn’t charge us for the extra work hours if he could keep it as a project until he was done. Took over three weeks of him going through the wiring and finally found a harness/wire that had worn through and was grounding to the car frame.
So far nothing too weird for an old car. The bizarre part is that he had good current equipment and it is supposed to test if a wire is grounded out like that to the frame or even if it is broken by kicking signals along it like you can to find damage to Ethernet cables.
So with that tester in hand and knowing without question what the problem was he hooked it back up and it still reported nothing wrong. He called the manufacturer and they said as far as they know that violates the laws of electricity… Worked fine with the new wires so again definitely correct and his tool worked on everything else he ever tried it on.
Absolutely disagree on this. There is no fundamental reason software must have bugs. However old systems can be their own technical debt because of things like the hardware no longer being produced and therefore unable to be directly repaired if it breaks from age.
This leaves either reprogramming for a modern device or things like emulation which can create/surface bugs that weren’t present before.
The most extreme example I have heard of (sadly couldn’t quickly find a link for it) was a disorientation simulator for pilot training that had zero software issues in several decades of use and when the hardware failed they replaced it with an FPGA in a modern system that ran all the old code 1 for 1. PDP stuff originally I think.
Additional edit - I’ll add that “bug free” software is insanely rare in reality and nearly but not quite impossible to create in practice. I can’t say the software didn’t technically have bugs but if multiple decades of use didn’t have them show up in practice it is functionally bug free.