>>Software development artistry/lack of standardization - there's a consequence of that which is not always recognized. In a lot of small custom software projects, and even some medium-sized ones there is a core of a few people, perhaps just one, who know how it really works.

Agreed. Also true of rocket science and the Manhattan project. Surrounding such people with huge head counts may give comfort to customers signing checks but in the end success is built around and reliant on delivery by a few key players every time IME. As an example, KG used to remind us frequently of the debacle of the Obamacare website: built with massive headcounts and committees and all the modern trappings of serious software engineering, the result was a debacle. In "Atlas Shrugged" style assistance then was sought from a very small group of geeks who would have been shunned by the original teams but who solved the shortcomings in short order. Eventually the truth will out on that, not that anybody wants an antisocial geek hero but maybe we need to lose our prejudices and value results more.