Also, lots of people have hopelessly romantic views about the group to which they belong (or used to belong). At least in that respect, RMS is not unique at all. But the fact that he falls prey to this common error is not very relevant when we ask whether it is indeed "evil" to make proprietary software, or any other current issue that RMS rants about. (We don't evaluate Peter Thiel's startup advice on the basis of his odd philosophical commitments, do we?)
I believe it is relevant when discussing RMS views. What you call "romanticism" of a group is only one of several possible romanticisms. There is also the romanticism of an individual, the romanticism of a family, the romanticism of a partnership etc. The point is that RMS was guided by a romanticism at the moment he was starting the GNU project as he himself described in the copy included with every Emacs install, and moreover his romanticism was of a very specific, (I believe) noninclusive and dysfunctional kind. This is rarely discussed for some reason, when even people who deep down disagree with RMS put on a guilty face and speak of the man as if he was just one of those rare idealists whom the world badly needs, and as if they do not dare to question his supposed integrity. This breeds orthodoxy besides other things because of which this thread is here.
But as far as ideas are concerned, I think RMS's romanticization of MIT hackers is severable from most of his well-known stances on software freedom. I think ESR's article has more to do with RMS's ideas than with the man himself. If (Microsoft|Apple|Facebook) are evil, they are evil no matter what RMS smoked in his MIT days. If it's counterproductive to treat them as evil, then it's counterproductive even if RMS turns out to be God's only nephew.
Take what RMS wrote about the early days of Symbolics. When you read him, you begin to realize that he believed the community of hackers that developed at MIT to be a goal unto itself, not as a means to achieve something specific. This is in spite of the fact that every individual who started hacking at MIT (this ironically includes RMS) surely saw himself as pursuing a specific goal. It is hard to reason with a man who believes that groups of people exist in and of themselves, independently of individuals and their goals. The only possible consequence that can stem from such a position is to define freedom as a property of groups, not of individuals, which is completely alien to the notion of freedom as most people (at least in the U.S.) understand it. The exact same position, when held by people in other countries, has lead to tyrannies being established, both of right- and left-wing kind.
The issue that ESR is raising is a very practical one -- RMS has become the de facto spokesperson for the free software movement, which hurts free software more than it helps it.