In The Beginning was the Command Line, a manifest for the command interface published in 1999, is obviously dated. It cannot discuss any of the more recent developments in operating systems. Though I wonder whether there really is that much development that would be needed to discuss the essay's thesis. Or at least the essay's thesis as I see it. Stephenson has, of course, made many more observations and statements that one could discuss, that maybe would merit an update.
Stephenson has made some valid observation concerning the economics of software production and the particular cases of Microsoft, Apple, and Linux. These observations could even merit a discussion in an appropriate business information technology course at college. Yet, this is not what I would consider Stephenson's main insight. (And therefore I do not think that Birkel's 2004 update of the essay, essentially an annotated version, does add anything of substance. It just corrects a few factual errors that are irrelevant to the essay's insight.)
The main insight, I believe, is the classification of computer users, users of any product really, in two groups (or was this three?). There are the users who don't need to know. They run the applications. There are the users who want to know. They write the applications, they hack and produce a solution to a problem.
The first group of users may be sub-divided in more specific groups with different needs and demands, generating a continuum of users that may approach the second, much smaller group of expert users, the makers.
These two groups differ dramatically in their approach to use a computer. One group may be thankful for abstraction, simplification, and GUIfying. The other will want to use the command line interface, access the computer at a lower level, bend it to one's will (and will rant about the inferiority of MS Windows).
I (happily, proudly) admit that I – often, not always – want to know. And yes, I am still using the Command Line Interface. I feel so much more in control, so much faster in many tasks.
Thus, I think the lesson of Stephenson's essay is not the economics behind the software production at Microsoft, Apple, and for Linux. It's that, especially for software, a one-size-fits all does not exist and that feature wars may be detrimental to the final product's usability and quality. Maybe not necessarily due to the needs of the users but due to their different mind sets, their approach to software use.