When you are young, you like new things. New things are kool, it does not fit into main stream, and it won't be easy to blend into the lifestyle that most people practices.
But you like it, it might be built into your blood, so if I ask you not to like new things, you rather die. That's right, instead of isolating yourself out of the main stream, sometimes you could hide or vomit or do whatever minimizes it.
This is not good.
Because ten years later, you might find out that most people likes what you like back then, or maybe most people has moved to something so revolutionary than anything you thought of. It's not worth vomitting everyday for that ten years ago, isn't it now ?
Vision of new form
Building a new way or system is risky, time-consuming, with possibility of no gain in return. People say it's worth a shot, only with managed risk, or it could lead to more chaotic situation.
Starting new is always temptating and painful. Soon you finish building something, you realize that you are missing some components and you are running out of money or time. It seems you are still so far away to catch up with any existing system in the market.
It's not smart to compete for foundation when things are in its infant stage. Instead, if we could skip the foundation and focus on the vision, we might be able to save tremendous resources while failing fast and winning quickly. Is that how all kool things actually wins out in the end?
Playing with the mainstream product to a point that you can not stand them is always the good start. The most important is that you do not have to spend any dime till that point. What you have to overcome is to understand the limitation of the current product and take it as a fact. Keep in mind what you aim at is the vision of new form, not to judge/criticize the legacy form. If appreciating the legacy form is one of the effective way in saving money, you should start to practice it.
Sometimes you'll be lucky to find a startup product which is doing very close to what you want. You should be thrilled that you can have a even more powered jump-start to focus on your vision, since once again you have some finished work to test your visions, which is really valuable economically.
Foundation of new system
At some point you are struggled by building your own system for accommodating the experimented features or ideas.
Let me tell you, your own system is very important piece, but it's not must-have. If you focus on it too much too early on, it'll only cost you quite a bit. Because you are pushing things to a form that is not necessarily taking place yet. Very much like an asian parent want their kids to be making a living on something even when their kids are still kids. The vision is actually Correct, but the system is not There yet.
I have seen some people not care about money that much, but if part of your lifestyle is building new things, you probably should care about putting the right money to the right place in the long run. If you are really smart, what you should do is wait. Skip building new system and focus on the vision with your limited resources.
Ways to your Rome
At one point, you'll be so ready to build your own system.
One mistake that we tend to make is shifting from one framework to another framework only realizing none of them fits us. Don't over-committing to any framework, pick a light weight or no framework to start with. Keep it in mind, your goal is to construct a new framework for your own benefit. Starting lightly gives you benefits of flexibility.
Now remember you had some mainstream or startup projects that you were having fun. I assume you already tested your idea on those systems, now it's time to migrate it to your light weighted framework by only building the infrastructure that will power your idea.
It's the process of spotting the new features, migrating into your system and refining the foundation to power these features. This process is slow since it doesn't build on top of the giants.
As much as "Don't reinvent the wheels" spirit, you might end with stand-alone components, some of them might be borrowed from legacy systems. You then somehow wire them together through a build process, and then they just work together to power your idea beautifully.
Try not to reproduce any existing framework, focus on the vision and the execution for it. Along this baseline, everything else should unfold towards a structure, therefore a new system.