> It's a big part of why I take the time to maintain the bookmark database I've got: I don't trust that I'll be able to find articles like this one in the future.
I go further than that - and I assume many do: For stuff I think I'll want to reference in the future, that I find important - I'll take a copy of it; ctrl+s or ctrl+p for basic stuff, other things I'll take the time and space to mirror (using wget or something similar).
Given that we've seen major portions of the internet literally shut down or "destroyed" over the decades, this isn't something one should find odd or overkill.
I too keep a huge bookmark stack, but sometimes I don't properly sort things, and just put them in a "general" location. Recently I started going back thru these bookmarks, just to organize some of them - and many of them were no longer valid; 404s were the norm. That said, in many cases having the bookmark name or description (and even the old URL) was useful to find a copy of the information, where it moved (if it did) or maybe somebody archived it. But in some cases, it simply vanished.
I do agree with your assessment, though - Google and the internet has this "short term memory" problem; a big part of that is the (now) dynamic nature of websites, which doesn't allow for easy crawling by some systems (like the wayback machine), as well as allowing the data to move or vanish at a whim.