With some of your points I can agree, with others strictly not thanks for the time to go more deeply into this matter.
I did not mean the lack of/existence of firmware updates and continious support by that.
Most (if pretty much not all) of the native mirrorless lenses depend entirely on electronics to work. If you don't connect them, they can't even focus. You can't change the aperture, you can't even open the aperture wide open. So all these lenses will eventually break down and they'll have to be thrown away as parts won't be available and repairs won't be feasible. A serious lot of them don't even have optical adjustments inside.
What would you like to do with an unpowered lens without a camera? I agree on the stupid fashion to jam some glass together and the optical flaws "can be corrected by software" - that's clearly not what I like to pay for the price of, say, a Fuji 16-55/2.8. But for the rest of it? All these tiny motors and actuators suck the battery empty, true, but this is not the major reason of poor "shots per charge"-numbers. Put yourself in the chair of a designer - if you know, marketing department will go for a new model in less than five years, why take the effort and design it to last 50? And just look at us: When there's a new lens on the horizon, the buzz goes on here. Manufacturers need the buzz. Tokina for instance is nearly off the radar, nothing interesting coming from there.
What is more, most of those lenses are also either planned poorly or planned too well so that the company can/will release a new version in a few years to make the old one obselete and mostly worthless. Fuji's 35/1.4 is an example of that. It is optically alright (although not perfect) but the focusing is so slow, pretty much everyone who loves that lens will jump on the new one if it comes out. I get the fact that technology advances in time but come on, that focusing performance was poor even when the lens came out.
I'm not sure, but I think it was one of the first X-lenses, so they were experimenting. And as you said yourself you could live with slow AF but not with unreliable. I skipped that one and went for 23 and 56 instead. Not superfast, but alright.
Newest Nikon lenses don't even work on their bodies from a couple of years ago. Actually the whole F Mount is hilarously and horribly non-compatible within itself so that's nothing new but come on, there are people still using D90s which don't work with the newest lenses anymore. It actually didn't even work with older lenses, when I think about it. Couldn't even meter the light with AI-S lenses.
With Nikon, I only have 6 years experience and never a lens which was not working like expected, I think, it were about 20, genuine Nikkors and Sigmas. But I never tried to attach some "oldie but goldie"-samples from the time I was young. I started with G lenses and now we have at Nikon as well some E versions which come with electrically operated apertures - no more levers. I can't say I miss it.
Sigma designed their first generation ART lenses so poorly that no matter what you do, you don't get consistent AF with their 50mm and 35mm lenses. Reported a thousand times by various people and won't be fixed unless they release a new version and sell you that. Who will buy the old one that can't even focus properly? I'm fine with getting an older version for cheap if it's not as sharp or maybe focuses slightly slower but can't focus reliably? Naah, I'll pass.
I think you never ran focus reliability tests? You can do so by using FoCal Pro. You must remain very strong now: I have no single lens, be it Nikon or Sigma, which passed the test with flying colors. ALL of them have one or two shots amongst 10 which make a big jump in the graph. So, you can blame Sigma for unreliability. I blame the AF system itself for systematic weaknesses. Sometimes the compromise between speed and accuracy works, sometimes it fails gloriously. AF-reliabilty on DSLRs? I call it a dream. And still better than my manual attempts! Like to see some graphs? I do it anyway
Nikon 85/1.4 G
So, what now about consistency?
I can repeat the test and the next ten shots with the same lens/body combination will show a different graph. And that was in ideal conditions. Enough light, tripod, static target - who needs AF at such a moment? The real game of mastership is AF-C but this is hard to reproduce. And here I don't see less keepers from Sigma (exception the 20/1.4) than from Nikon.