There are obvious holes in the methodology, so I wouldn't take the content of the post too seriously. What I do think is great about this post is how the authors generated a bunch of hits from an hour or so of work. Pick a topic with wide interest, do some simple (but non-trivial, if you aren't used to the tools) analysis, watch the traffic roll in! It's a good model for anyone trying to drive traffic to their site.
You could also make the argument that older languages are discussed on sites other than StackOverflow (ie, Google Groups, direct forums, etc). I get the impression it all washes out in the end.
Pretty much everything except that which is rails related.
In more seriousness, what do you mean by stuff?
Many gems have their source in github, but really if github imploded tomorrow gem install rails would still work.
Do you mean, who doesn't host their source on github / have a clone of it on github, or do you mean what pieces of infrastructure would fail if github went down tomorrow.
"or do you mean what pieces of infrastructure would fail if github went down tomorrow."
The answer to that being: not many. The basic architecture of git (and other DVCS's) is, that everyone has a copy of the full revision history, so if github really went down, you'd see most of the projects there (at least most of the active ones someone cares about) on gitorious or some other git hosting service within a very short period of time, and things would go on with only a minor interruption.
RWW pretty much just repackaged the original post with a small amount of commentary.