The author misses vitally key issues (like the inherent dangers of circular references in reference counting "GC"). I wouldn't even call myself slightly knowledgeable in GC, yet blatant holes in the clumsy analogies of this article are obvious to me. As someone who is not even slightly knowledgable in GC, it scares me that people are learning from someone who has massive information holes in the explanation -- holes in topics that every programmer on earth should know.
On top of all that, the only "Cons" that are mentioned of RC are absurd: "It's annoying if you forget to release your references." Wow. Profound stuff there - bugs and incorrect code is annoying. This isn't even a problem if your language supports RAII, in which it's impossible to "forget" in the first place.