I worked as a gearbox engineer for a Tier one supplier to GM, Chrysler, and Ford among others. In my professional opinion this had everything to do with money. That plug is there because it cost less than having a dipstick. First, you have part proliferation. A dipstick system includes a dipstick tube, a screw or clip to hold it up, a dipstick, and a grommet or o-ring at the bottom. That's four part numbers. The plug is one. You also have to have a bulge in the aluminum case to accomodate the diptstick. Somebody somewhere calculated that by eliminating the dipstick boss from the trans housing you saved 14 grams of aluminum. You also eliminated a core from the mold tool making the tooling for the housing a little cheaper and/or you eliminated the machining op to bore that hole. Sure you added some cost by adding that threaded hold but probably it's the same size as another hole allowing the use of an existing tool on a machining center. It's still probably a few coins ahead moneywise compared to having a dipstick assy. It all sounds a little crazy from the outside but I worked on this stuff. We'd do crazy things like that to save a quarter. When you make 300,000 of something a year it adds up.
As for being difficult for the consumer or shop to work on, I don't think GM gives a crap. They know that most customers (engineers such as myself and the other technical people reading this being and exception) really don't care if it's easy to service and maintain. GM knows that most people care more about the bluetooth capability or the compartment for sunglasses than they do about servicability.
Some people in the industry took this concept to extremes to save money. I remember a discussion about welding the gearbox halves together. The idea was it would save the money spent on case bolts and silicone gasket maker and the machined surface finish for the gasket maker. You could also weld with a robot rather than having a union employee start the bolts for the nut-runner.
How would you service it? You can't-and that was part of the idea. If there was a problem you would just sell a whole gearbox instead of a bearing or other part. This would increase the number or gearboxes sold and save money associated with carrying service parts, service manuals, etc. Ultimately this idea was deemed a little too extreme and scrapped. I was glad to hear it. But my point is that these kinds of extremes are considered in the interest of profit.
I also agree with ASmith. Eliminating the dipstick eliminates trouble for them. They don't have to worry about trivial warranty claims where the customer takes the new car back to the dealer the next day because the trans fluid is 'low' because they checked it cold or other similar things. The same goes for gauges. Newer cars runs hot. If you put numbers on the gauges (GM usually does, Ford usually doesn't) then you'll have a person taking their 2013 car in for a warranty claim because it runs at 200F. That was too hot for your '62 impala but for your 2013 Impala it probably isn't. Having a numberless temp gauge, or no temp gauge, eliminates that.
Of course cheap design is the reason why I will never purchase another GM car, and over the years I know I have discourage at least a half dozen friends from buying one as well - and I don't miss an opportunity to tell anyone who will listen that my Blazer came apart in less than 60k, whereas my Jeep and Falcon still run fine and are both over 300K miles. So lost sales I am hoping more than overcomes the few cents/part saved.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.