I'm coming really late to this thread, but I'm going to comment anyway. What I'd like to propose is the idea of the "success loop". It seems to me that you are

I'm coming really late to this thread, but I'm going to comment anyway. What I'd like to propose is the idea of the "success loop". It seems to me that you are stuck in this place where you get no positive feedback from your work. The other way to work is to throw out a lot of small things, and then build on the things which work, and people find valuable. All of my large hacks have turned out to be the result of small hacks, positive feedback and collaboration. Consider not focusing on the whole vision, but choosing a kernel, just one novel contribution. The best ideas will attract other people. Once you have a bunch of minds working on things, it becomes like a relay where your analysis paralysis gets "routed around" by someone else. If you can create this environment for yourself and roll with it, I promise you will have the time of your life.