You're probably going for things that are too large in terms of scope and/or scale and don't really solve any need for yourself. Think about bite-sized problem

You're probably going for things that are too large in terms of scope and/or scale and don't really solve any need for yourself. Think about bite-sized problems that will help you with your day to day work/hobbies and don't try to solve them in an 'enterprisey' or bullet-proof way... think the simplest thing that can possibly work. Then over time, start getting more ambitious. It doesn't matter if it only works for 2% of cases (as long as it's the 2% you care about), runs slowly, is a memory hog, crashes every 8th time you run it or any number of other things you'd never dare do for 'production' code. Assume it's a throwaway project and then be pleasantly surprised if it isn't.