Slowly building up like that over two months can be great for people who want to write an engine and learn everything, if they have the discipline and time to do it. I got a good bit of stuff written into my engine framework for loading and displaying things, scrolling (mostly?) smoothly, ALMOST displaying fonts in a pixel perfect fashion, and so on. But it took forever and I have hardly any motivation to work more on it.
Now, that's speaking as someone who has been doing software development in various forms for say, 10+ years. I may not be a good programmer, but I've been exposed to that sort of thing for a long time. For people who have not done ANY programming, making a game can look so daunting that they may never leave the paralyzed stage of "this seems complicated, I could never do it".
Instead of taking two months to learn XNA and get to a side-scrolling engine, an artist could probably go through a GameMaker (or Stencyl, or Construct, or whatever) tutorial and have a very simple sidescrolling Mario-style screen or two going within one or two WEEKENDS, in my opinion. Now as Ranger X has pointed out, collision might not be perfect, and it takes time to figure out Game Maker's quirks (such as fixing the collision in his case). But you can still get from "I can draw sprites but have nothing" to the "super simple sidescroller" stage like 30 times faster by using one of those high-level tools, and in my opinion, that's why people should jump in. It's not jumping into the deep end, it's jumping into a manageable level, after which they can start going deeper by scripting in the tool, moving to a different tool, or even working with a lower-level language.