Just to throw a monkey into the discussion, I've been at TechShop SF many days running over the past month or two. Upstairs, the laser engravers are always hogged, with waiting lists of a week or more. Downstairs, the two lathes and two (was three) milling machines often stand unused. Except for the Tomach CNC -- that gets more love.
But tempting as it is to say, "Yeah, everyone wants the sexy computerized stuff these days," there's a good reason for this usage pattern as well. I grew up working in scene shops, often had access to a place to cut wood and weld metal, and was pretty much used to taking my work home with me. So I've always had hand tools around and a high tolerance for sawdust all over the apartment and bits of expanded foam in my soup. What things like 3d printers, laser engravers, and, yes, the old-school resin kits as well, do is allow people who don't live that kind of lifestyle a chance to build a bit, too. They can build without the infrastructure investment (the tools and work space) and the body of minor skills many of us take for granted, and of course the impact of all that dust and paint overspray and the fine sensation of slipping out of bed Monday morning and putting a bare foot down on a metal splinter that escaped your attempts to clean up from the weekend's build.
What does this have to do with the quality of a 3d print, especially pre clean-up? Nothing! But as people have said, the discussion evolves.
And I also want to chime in on the "run" problem. I recently created a scale model of a Commando V150 armored car. Made three, actually. Which is way too few to even think about injection molding, but a few too many to want to be breaking out the styrene. Actually, we thought there would be more interest. Enough that a few dozen would be printed. At the current numbers, it is about even-odds; I spent as much time wrestling with making a printable (yet affordable) model as I would have spent bending plastic. And, personally, I find 3d a lot higher in the frustration scale. I'd really rather be scraping a seam for hours with an X-acto, than hunting through a large mesh for a bad weld.
So, yeah, sometimes having the printer be the mold makes more sense -- for some projects. Small size, small runs. I don't think any customers are getting fooled that a fresh 3d print is any closer to final finish than a freshly pulled vacuuform. If there is any issue left, it might be a perception among a vanishing minority that just because it is printed, it ought to be cheaper for the same quality of finish. And the cost of a spool of SLA aside, paint cost and sanding time haven't changed. They are even a bit worse for prints. As much as we might like, the only tried-and-true way of getting a really cheap lightsabre is by going down to the toystore, or to one of those companies that can really, really leverage benefits of scale.