Ok, I cede that the author is making one assumption that is somewhat fair. Semi-automatic guns did, in fact, become more popular during the time period that he is addressing. Like it, love it, or leave it, I think that is a fair statement.
The rest of the article is a hack piece. To suggest that it was the Army or M9's fault is ridiculous. There were a number of issues that made the late 80's, through the 90's, dangerous. Street gangs and crack cocaine aren't mentioned in this article... nor is Gaston Glock's design. The one statistic that the article found that supported it's claims is multiple shot ER victims rose from the mid 80's into the 90's... that is not a causal correlation but mere speculation. One could also argue that gang "hits," and not just a happenstance gunshot from a robbery gone bad, made that statistic grow.
The article also correctly states that gun deaths began to decrease after 1993, which is coincidentally the year before the Brady law was enacted. The article failed to mention, however, that there are currently more firearms in existence in the US now than ever, yet gun homicides are still FAR lower than they were in 1993. In addition to this, if the Brady bill actually was a direct cause of the decrease in gun deaths it is far more likely that the background check clause was more effective (in reducing homicides) than the arbitrary Brady AWB. Also, the Brady AWB sunset occurred in 2004 but gun homicides continued to decline for several years after. There has been a slight uptick in 2012 and 2013 (stats for 2013 are fresh, too), but nothing even close to 1993 levels. The economy could easily be an explanation for this slight uptick. Also, the article fails to mention that gun homicides in 1993 were the end of a LONG uptick in gun deaths that started in the 1960's. Coincidentally, gun homicides were rising despite the gun control act of 1968. So, we have one anecdotal instance of a gun control passing and the gun homicide rate going down (Brady Bill), with another anecdotal instance of a gun control passing and gun homicide rates continuing to rise.
So... this article is a hack piece. If I were in charge of an institute of higher learning, even a tech school, I would be embarrassed by a professor that authored such a hack piece with only the arbitrary statistics and the reaching "correlations" presented. I would make a mention of the media outlet, but honestly this is par for the course for the Atlantic. They have some interesting articles, but they are better off sticking with pure opinion pieces authored by professionals such as John Q Wilson's "Broken Windows." Now that was an article. Didn't cite a bunch of sources, nor did it dazzle with statistics. But it made sense and no one could definitively deny the points made in it.