Lack of easily filled gaps

These gaps are connected with the issue I mentioned above - "the missing topics become harder and harder, often needing specialist literature not widely available". For villages of a few hundred people in Africa, it's hard to find any literature at all in most cases, and therefore Wikipedia articles, if they are written, tend to be meagre, consisting of some geographical and statistical data. By the way, that's also the case for many articles in the English Wikipedia about places in the United States: Very often they are not much more than a bunch of statistical/demographical data shaped into something resembling a Wikipedia article, but not really being one. Random, typical example: en:Limestone Township, Union County, Pennsylvania. That's not an article worthy of reading but something that wouldn't look different if it were bot-generated from the Census Bureau data. But in order to write really good articles about places like Limestone Township, you too need literature that is probably not widely available on the Web and not easy to come by in printed form, visits in specialised libraries or archives may be needed.... it is a challenge very different from writing articles about Lincoln, the Beatles, the widely known cities of the world, or bread. No one will have any difficulty finding usable literature about these topics, and that's why the Wikipedia articles for easy topics are now already written.

You've got a point- the articles you can write without any specialised or local literature are mostly there on the few largest Wikipedias. Take en:Word, about as basic a concept as there is. The article is bad, barely sourced- though at least free of maintenance tags. That said, a quick web search is not going to vastly help you improve it, chances are you're going to need an academic library to make real progress. For our township in Pennsylvania, someone who lives nearby and has access to the local newspaper archives, museum, etc. could do something with it- those who are five states away can't- and very likely have no desire to do so. The amount of articles you can write from scratch, or make a run at GA+ quality, (that aren't recent topics) without touching physical sources are fairly low at this point.

I'm as opposed to systematic bias as anyone, but the obviously encyclopaedic articles, and those of broad interest to Western Culture are mostly at Start-class or above. What's left is the hard stuff, either making those articles shine, or going outside of the English-speaking world, and neither proposition is an easy one. As the work gets harder, it is natural those interested in getting it done are fewer in number; hard work being, after all, hard.