[sixtofour]

I can think of a couple ways to approach this.

1. Pick a very small project, something useful to you but borderline under-ambitious. Anything that bugs you. Nothing that you would necessarily point proudly to on github; don't be thinking about github. I'm talking small, maybe even just a simple command line utility, or not much more. Implement it. Don't give up, learn it. Improve it. Pick another small project. Another. A slightly bigger one. What you're trying to do here is to get a giant heavy flywheel moving, but it's going to take some time before it's really spinning. Once it's spinning fast it will almost sustain itself and you won't have enough time in the day to implement all your ideas.

2. If you can't think of any small thing to build (hard to believe, but OK), then poke around for some technology that you'd like to learn, that has good documentation, and start going through its tutorials and documentation. Learn as much as you can about it. At some point you'll start to get ideas about things you can build. Start building them. It doesn't have to be an OSS project, it can just be your own if you want. At this stage it's more important that you get that flywheel spinning any way you can.

N. Don't give up.