Instapundit has a bunch of stuff about how Scientific American has come, over the years, to suck. And that's true. The Lomborg hatchet job was clear evidence.
On Reynold's site a lot of the letters he is posting are claiming SciAm has been going downhill for years. As a longtime reader myself, I can confirm that, too, though with a big caveat. And that is that the decline of SciAm, in my opinion, has been very gradual - until recently. As of about a year ago, they started completely overhauling the magazine's feel. Among other aspect of this, they started putting in bullshit illustrations. This, more than anything else, epitomizes to me the rapid recent decline of SciAm. The illustrations were half the reason to read it; they were always top notch and made understanding the science easy. Indeed you could often get valid scientific information simply by looking at the pictures. But now they are putting in illustrations that have no connection to the scientific article at all, other than the vaguest fanciful connection in some hack illustrator's mind perhaps. This is a break much bigger than publishing borderline science like articles on nuclear disarmament. It means that the editors no longer are reading their own magazine for their own scientific understanding!