Program it yourself. RPG makers are dumbed-down, limited in features and you won't improve your programming skill with them.
Frank Lloyd Wright didn't build houses with his bare hands.
The amount of overhead to make a game from scratch is tremendous. I do it, but I've been doing this for years, and I still work with others to lighten the load. I've seen numerous projects crash and burn because they ran into technical difficulties and couldn't fulfill their vision -- or because the teams got bogged down on boring fiddly implementation bits, reinventing the same wheel already used hundreds of times by hundreds of other teams. If someone wants to make a game, and is drawn to using a pre-made toolset, but gets scared off because somebody on a forum told them they should be doing everything from scratch instead, they're almost certain to meet the same fate. On the other hand, use tools, and the feedback loop from work you put into the game is shortened to a fraction of what it would be. Effort turns into results quickly and powerfully.
I used to think using game maker programs was weak, a crutch for the less capable and less ambitious, one I myself only leaned on because I couldn't do any better. Then I went to college, majored in software programming for games, became a design intern at a studio developing an MMO, and eventually graduated and started my own business as an independent developer. Now that I've worked on more than half a dozen finished games myself, I no longer have such elitist preconceptions.
I've personally seen programmers come up with amazing technology, only to desperately search for someone with the creativity to put it good use. I've personally seen designers who hate even basic scripting, but can perform absolute wizardry when given a good editor. DIY game development may sound badass, but in the real world, companies spend hundreds of thousands of dollars paying programmers to do nothing but write tools exactly like this for their own internal use. Not only does it save money by accelerating development times, it enables their often very talented designers to forget about programming everything from scratch and focus on what's really important -- making a fun game.
My initial reaction was also the essential essay
Make games not engines. However, there's a reason that Yume Nikki is the only RPGMaker game to receive an overall positive reaction (Ao Oni, The Way, and Exit Fate could also be argued, but they are at best heavily flawed and extremely subject to user reaction). Most of these games required their developers to work around RPGMaker's natural flaws to make their game, perhaps taking longer than it would to replicate the functionality in code. My personal opinion is that RPGMaker is valued more for the inclusion of default sprites and music rather than the engine.
I've seen great things done with GameMaker, but I don't know if it's actually capable of reducing the time needed to become good at game development. We've seen a crop of people who've developed the ability to use it extremely well (Dan Remar comes to mind), but it's worth noting that they've been using it for 5+ years.
In the end, I think the best option is the proper integration of a high-level language of your choice with pre-existing libraries (or even an engine, if you can find one that fits your idea). This gives you the ability to treat components as "black boxes" like with a game maker, but doesn't limit you or require you to use workarounds if you want it to do something different.
tl;dr Learning to do anything well takes time, and learning to use a language won't take that much longer and will leave you much more flexible
Some languages you might consider:
C/C++: I wouldn't use this for gamedev unless you have previous experience with it or work with it professionally. On the other hand, it's fast, solid, and has an extensive and ever-growing number of libraries. You'll probably want a bit of fluency in it at some point to look at the large number of examples written in it (reading the Nethack source is pretty much a right of passage)
Java: It's clean, portable, and has excellent networking. There aren't as many game focused libraries for it though. It isn't as slow as people often claim, but I wouldn't program something with huge memory requirements like Minecraft in it though.
Python: It's a great place for non-programmers to start, but if you have any experience with other languages you'll hate it
C#+XNA: If you aren't concerned with portability, this is what to use. It's fast, clean and hassle free. Visual Studio is the best IDE I've used.
Flash: I haven't used it, but I suspect it's one of the best things for aspiring devs, especially given the development of Flixel and Flashpunk
That said, there's nothing really wrong with GameMaker. In the end the biggest challenge with RPGs (the RPGMaker-styled JRPG ones, at least) is that unlike other types of games that the bulk of the work is in the art assets rather than the programming.