If you've ever been unfortunate enough to have to wrangle a VB script routine, it was (less bad) like that. If not, I would go find some assembly code and teach it yourself, and then imagine that instead of side effects in registers there were random effects on your code/visual state.
And like assembly code, you could now imagine that the same code might behave wildly different on different machines in different browsers.
So a bit of "old man"isms, but also I imagine his JavaScript was tainted by the early days. It's better in some ways now, worse in different ways, I don't mean to say that is the worst or the best, just to offer perspective on where it came from.
Obviously he's not serious, he's playing the part of the out of touch old man.
We all become that eventually, hopefully we can all be as poetic and humble (and honest) about it.
But apparently I didn’t get the hyperbole.
Only to then realize (after reading the walkthrough) that there was indeed a way.
I think it's human nature to find (rather search) blame not only in yourself but everywhere else... anyhow, since the author is reflective we should be forgiving as well.
Tell us again when you're 74.
I'm still nearly 2 decades from it, but I am a profoundly different human to the one I was 20 years ago, or 20 years before that.
It reminds me of when I was around 10 years old or so, maybe slightly older, and playing around with Turbo C (or maybe Turbo C++) on DOS. I must have gotten something very basic about pointers (which were new to me at the time) wrong, probably having declared a char* pointer but not actually allocated any memory, leaving it entirely uninitialized, and my string manipulation failed in weird and interesting ways (since this was on DOS without memory protection, you wouldn't get a program crash like a segmentation fault very easily, instead you'd often see "more interesting" corruption).
Hilariously, at the time I concluded that the string functions of Turbo C(++) must be broken and moved away "string.h" so I wouldn't use it. But even then I shortly after realized how insane I was: Borland could never sell Turbo C(++) if the functions behind the string.h API were actually broken, and it became clear that my code must be buggy instead. And remember, I was 10 years old or so, otherwise I don't think I would have come to that weird conclusion in the first place.
Nowadays, I do live in this very tiny niche where I actually encounter not only compiler bugs, but actual hardware/CPU bugs, but even then I need a lot of experiments and evidence for myself that that's what I'm actually hitting...