Ask HN: How do your find your itch to scratch?

After being a user of open source software for many years (with a sporadic contibution here and there), I have the urge to start my own project. I have always read about "finding an itch to scratch", which I am having trouble with. For the most part, I'm pretty happy with the software I use and I don't frequently have complaints that need solving. Its very possible that I have too much tunnel vision and don't even see the things that could be better anymore.

How do I break through this, or are there any general approaches to finding a need that many users have and aren't solved a dozen times over?