I've seen this skepticism elsewhere, and I wanted to say - yeah, it really is the same tool. Honest!
We have a lot of the same concerns and workflow frustrations commented upon in the community, and even as I type this we are working on editor improvements that will help all of us in future versions of the Creation Kit. There are other pieces of software used within the studio, of course, and some of those are not publicly available. Our version control system, for example, is why the navmesh bug was never known to us until recently. There are also aspects of the game, such as animation, which are worked on outside of the editor in tools that we have no rights to distribute.
As a level designer, however - and this will be my seventh year at Bethesda - the Creation Kit, in one form or another, is the tool I have used every day. The only difference between your version and mine: my help option points to an internal version of the (same) wiki you have, my version is connected to our version control system, and the upload to Steam options are disabled. That's really it.
Glad to hear you're still working on the CK and I hope you take most of the complaints and criticisms, such as the well written OP's, in a constructive light. One thing I've been trying to communicate to Bethesda ever since MW was that they are really selling themselves short by not putting in more development time into their editors. I think there is a vast potential of efficiency improvements possible by a well designed UI and stable app with attention to details. I figure within Bethesda you've had at least 160,000 hours using the CK for Skyrim (20 people, 4 years). Even a measly 10% improvement gives you another 16,000 hours for *free* and I would say in a type of app like the CK where almost everything is the UI that 10% improvement is a low-ball estimate. This is small potatoes, however, to the amount of hours the entire modding community as a whole could gain by a better editor, especially since most of the hours gained would probably be mostly in the initial learning curve of the CK which is very high at the moment.
Seeing the amount of critical crash/data loss related bugs as well as the larger amount of "annoyance" bugs makes me guess that whoever used it internally merely learned to work around the difficulties over time. Part of the issue of working so closely with an app is that over time you become completely unaware of the usability issues: doing X causes a crash, learn to don't do X; doing Y is hard but do it 100 times and its not so hard anymore These critical bugs are actually very frustrating for most users as it actually represents a net loss of time and a lot of the time the ultimate cause of the crash is not obvious so you can't learn how to avoid it. I'm willing to bet that if you add up all the internal time lost on such bugs and put that time into initial CK development time you'd end up with a net positive gain (by far I'd also bet).
Some examples of relatively minor issues that have a large impact in the long run: The relatively hap-hazard placement and alignment of controls on all the editor windows increase the initial learning curve and contribute to "interface fatigue" (there's a better term for it I believe). Similarily, the placement of the Ok and Cancel buttons are almost in the same place but not all the time. This requires the user to actually have to think and "look" them as opposed to having them in the exact same place which leads to muscle memory and no thought/looking needed very quickly. Small things but the fewer small things the user has to worry or think about the more big things they can concentrate on.
I'll quit rambling while I might still be a little bit ahead but could suggest a large number of UI improvements for the CK if needed (most of which I try to implement in SkyEdit actually).