Future of GIS Analysts, Part 2

In the first part of this post, I tried to put together a simplified list of some of the activities that I have found GIS analysts doing as a part of their job. The goal was to get folks talking about future roles and if GIS analysts have a future ten years from now. I was happy commentators posted some things I missed. I hadn't included, but was helpfully mentioned by geographygeek in the comments, the role of an analyst after the data has been collected and processed. Some (hopefully most) are trained to answer the question, "Well, what does that mean?" Another commenter, KindaSpatial (who puts out a rather good geography podcast), wanted me to talk a bit about some of the newer 3D and hyperlocal data and interpretation. I'm honestly not sure if I am qualified to, but I'll give it a shot and have people correct me later. I like to think of blogging as more of a dialog. On closer inspection, each of these should be their own blog post, so that is what I will do. For now, I will give my initial impressions by going down the list and examine each item, asking the same questions: Can this be automated? Is it easily outsourced Is another profession largely absorbing it? Here are my thoughts. They are not quite fleshed out, feedback greatly encouraged.

1. Map production: Outsourcing: try to make or get a good, topic specific map via phone conversation. Automation: still a lot of overhead software knowledge required for the kind of quality maps necessary for professional reports. Professional designers have all of the aesthetic abilities necessary for this, but little knowledge of the pitfalls of cartography - maps are like statistics that are even easier to lie with.
2. Requirements gathering: Probably impossible to automate, and you can read the hilarious results of trying to outsource it elsewhere. Increasingly the realm of project managers with enough GIS experience to know what is available/feasible.
3. Feature creation and maintenance: The simple stuff can and will be outsourced or automated. Stuff that requires boots on the ground simply can't, and yes it is hard to tell the difference. I don't think the dimensionality of the data makes a significant difference here.
4. General IT/helpdesk support: This really is the work of IT professionals, but for smaller firms or feudal departments it isn't going anywhere.
5. Database/content management: All information has a location-based component, and this function exists as only so long as professional database administrators are uninterested in how the middleware (SDE and PostGIS I believe) works.
6. Minor automation tasks: Prime candidate for professionals to do the automation - it ends up being cleaner and reusable.
7. Post-processing interpretation: (hat tip: geographygeek) This dovetails nicely which what I think is the core of GIS - the visual display of quantitative information. What you replace a geostatistician with? It doesn't seem to easily fit in with other professions, the methods are unique enough to be difficult to automate for easy public use.
8. Hyperlocal data: (hat tip KindaSpatial) The mass of data associated with this seems to require, not merely lend itself to automation. Turning that stuff into interesting visual works or meaningful statistics sort of falls under the previous point.

Based on any comments/corrections I get, I'll be going though these points individually. I haven't been at the GIS game too long, so I'd love to hear from some people from different backgrounds.