The downhill slide was already underway by the 1990s. Readership was in a slow decline and the publishers turned to various marketing gimmicks to maintain solvency. More pictorial articles, more po-sci articles, cover wrappers suggesting that it was worth subscribing even if you didn't read the whole magazine etc. I stopped reading around 2000, when I bought an issue and noticed I had read the whole thing in under 4 hours rather than the usual 5-6. A comparison indicated they had changed the font size and line spacing slightly, so as to maintain the same page count but with about 15% less content.

I distinctly remember an article in the late 1990s (guessing 1998) which I read, but which I now can't find. It was about Y2K, and the bottom line was, we need more work to prevent disaster, but no matter how much is done, it will still be pretty awful.

I thought at the time he was exaggerating, and that Y2K was unlikely to be a big event. As everyone knows, a lot was done to fix the problem, and January 1, 2000 indeed turned out to be a non-event.

I cannot find the article now. I know I didn't dream it up, and I'm pretty certain it was in SciAm--I remember it had the usual sorts of graphs, illustrations, layout etc. as all SciAm articles did back then. If anyone can find it, I'd appreciate knowing. It was a turning point in my own reading of SciAm--I mostly gave it up after that, despite having devoured it up until 1980 or so.