er, ignoring the narcissism thing, which is bizarre (you want to hire narcissists, so better not hire old|ugly|disabled people!) your post seems to be arguing for the assumptions that the article is about: we can assume old people had to take shitty work in the past, therefore they are probably going to appear to be a poor candidate.

This completely ignores the evidence directly in front of us that all good engineers will one day start looking older and start having a lot more trouble finding job offers.

I don't really understand how you go directly from saying that it's unintentional and is "unreasonably harsh age-grading, coming from an inexperienced and privileged set of people who've never had to do grunt work" but then proceed to (try to) explain how it's perfectly logical that you can have a positive outlook of the future of your career, but actually you're going to have to take what you can get, which is the reason why "premature decline is far more common than it should be".

Maybe I'm missing the gist of your argument here? Are you suggesting we should be doing job interviews on a curve, excusing 20 years as a code monkey at some place you've never heard of because they probably had a mortgage? I don't know if that's your point, but it is definitely off base here: those 20 years might have created a middle-manager toady or they might have honed razor-sharp gut instincts and experience with relevant languages and tools. Either way, walking in with the assumption that "premature decline is far more common than it should be" means you're probably going to miss the truth of the matter, which is the whole point. No one is suggesting affirmative action for greybeards, they're saying that the kinds of things many in tech assume because a person has grey in their hair means that they don't get a fair shot, even when their experience is extensive and applicable.