This thread got a bit silly last time and I hope to avoid any name calling or un-GPWiki-like behaviour in future discussions.
FWIW, I think game builder packages get a bad rap because that's where most people start out. Once they've got some experience and a language or two under their belts, they probably feel like they've moved beyond such things. I'm sure that today's game construction tools are far more sophisticated than the stuff I used to get frustrated with on the Amiga.
Also, DirectX, XNA and even OpenGL are to some degree, bolt-together kits for graphics and gaming (OK, OpenGL is stretching it a bit there, but you get the idea). No one wants to start from scratch when writing a game. I spent many years fighting with DOS, interrupts and old PHIGS documents before finally embracing a graphics API and giving up the 'roll your own' mentality. So yes, pre-built libraries and tools are great, much of the work is done for you. However, you have to spend some time getting to know them.
I think GameMaker's biggest problem is that people with experience don't want to invest time in yet another API or package. We all have our favourite APIs and we've chosen them because they suit our needs. I'm glad GM suits you, but I personally don't have the time to update my knowledge of things I already know (OpenGL again
), never mind starting out in a whole new environment, no matter what promises it makes. It was the same with Python and any number of other 'next big things' that have come before.
Maybe I will code myself into a corner with my stubborn ways, but if GM is as good as people say, I'm sure it will gain traction. However, the majority of users will probably be new to the area, rather than those turned back from Java or C/C++/C#.
My other point is around flexibility. Game programming is fun, but most people who start out in game coding won't end up in a game studio. My early c/c++ stuff was graphical and game based, it was a good way or learning the basics. Since then I've written NT services for hardware temperature monitors, database and file system utilities and even a secure memory erase tool for a military jet (I'll post a picture of the debug session for that some time
). I'm pretty sure if I'd stuck with AMOS or Click 'n Play, I wouldn't have managed quite as much in my career.
 Woah! Wall of rambling text, here's a tl,dr:
1 . Most established coders don't have the time (or inclination) to learn a new thing.
2. I'd advise against putting all your eggs in one basket.