Sadly the simple principle of "same URL, same information" breaks down somewhat because of cookies (and, to a lesser extent, IP localization and user agent). http://facebook.com/ shows completely different information dependent on the logged-in account, and in fact that same person mentioned above was aware that URLs without a path part is the home page of that domain, and was thus expecting that http://facebook.com/ would show the same information to everyone. I'm not sure whether he/she really believed that the whole world should be seeing his/her posts, but it surely is a bad move by Facebook, it actively breaks the premise of URLs and is thus arguably damaging for internet literacy (and perhaps they are purposely doing it to mislead some people into a false sense of personal importance).
Whereas not knowing that a URL is a piece of text which specifies a particular content on a domain and can thus be copied and used for linking is a loss both for the user and for the community.
You'll have to reinvent the principle of linking in another form to avoid the loss (e.g. every web and local app would need special GUI functionality to use instead, and the need for specification what to link would not be completely removable, unlike the gear).
"They don't have to actually parse it as a query string. The fact that some URLs reveal a lot more information (like your CNN example) is a bonus."
Well, what you're saying is exactly what Chrome is doing - make people only care about the domain, don't bother with the rest of the information as it's "only there to make a page unique". What exactly are we disagreeing on?