I agree with the assessment that the average level of expertise among Wikipedia contributors is lower than that among Linux contributors or Encyclopedia Britannica contributors.
However, I also think that an intelligent person with good writing skills and no specific expertise can write an average quality (call it "mediocre" if you want) encyclopedia article about pretty much anything. So by following the current path, we have a good chance of producing an average quality encyclopedia.
(In a certain sense, that "mediocre" encyclopedia will already be "great", since it will be the only one you can burn on CD for free and mail to a highschool in Tanzania.)
Now, producing the greatest encyclopedia in the world is a different story. Maybe experts will show up in greater numbers as the project approaches "mediocrity". Or maybe many of them will be put off by constantly having to defend their writings against incoming idiots, and we will need some sort of quality assurance process. Maybe we will need a charismatic expert leader, or one will emerge.
But I think those are questions for the distant future. Right now, we should focus on moving Wikipedia from crappyness to mediocrity.
AXel