I call this the "Hello-world gulf" - the HUGE gap between "hello world" and what are basically samples and actual code.
In Python creating an auxiliary library in C++ that does the computationally intensive things is absolutely brilliant. The ability to link languages like this is also something you should consider when choosing what language to do stuff in, so great work there.
However this doesn't apply to flash - you should basically drop Flash.
The best technologies here are "open" and I don't mean this in the hippy-ish sense (although I really do fall under that banner....). Look at PDFs, these are great, look at Adobe's ability to do PDFs - it's crap. But they have specific extensions (things you can only do in Adobe's reader) that "locks you in". Flash is a bit like that in the sense it is a closed aboimination (there are open source implementations but whether or not they'd work with your library, or your mentioned compiler is another issue)
I know of no "serious" game (that is made by someone who's aware of what MinGW is) that is compiled using Microsoft's C++ compiler for example. This is often because open stuff either plays nice with other things, or can be made to play nice with other things. Before someone dismisses this as off topic, please read the next paragraph.
If it's the browser based bit you like, this is the way to go.
The Hello-World gulf
Many years ago I started with this thing called "DarkBasic" - it held me back years by:
- Not working like any engine should (so I had this warped idea of APIs engines should present to users)
- Selling dreams and false-hope (buy this ad-on for AI! Because implementing such stuff in it was not feasible, this was simple money-making from "dreams" as I said)
- Letting you make "samples" quickly but being the most awful thing ever for "actual software" (this is the "hello-world gulf")
You're at the point now where you're looking to cross it (from the sounds of it) and this is the toughest part of "growing up" programmer wise. I see this post as a nudge in the right direction because with game design especially there are HUGE markets for false hope.
How to cross the gulf
With systems programming I say "do actual projects" like things that let you put in data and analyse it with graphs and such, this shows you user input validation and the importance of planning ahead and over time you'll learn.
I've always said: "one cannot be taught how to model object orientated things" unlike the Matrix (which you can explain without showing) you actually do need to see it to understand it. After a while everything will just "click" and finally those diagrams will be something you actually want.
The same largely goes for games, only they're more complicated. Once you can plan ahead you'll see the huge massive gaping chasms that crap like DarkBasic doesn't show you and be like "I shouldn't use that for the job, because of this"
I know this seems like a rant but I am totally serious, from the sounds of it you're teetering at the edge you now need to cross into the "real" world.