It is remarkably tempting to take on big projects. But the more I try, the more I fail.
I had one success creating Rocket Spelling, which was—at least by my standards—a big project, which I completed in a single summer. I've been trying to replicate that success for a while, but it isn't working. And I'm becoming convinced that it was a bit of an outlier.
Most of the time, trying to tackle something big all at once just ends in failure. Instead, my most successful work tends to be when I create lots of little things. (Such as when I used to create Scratch projects prolifically.)
Specifically, I think I should focus on ideas that I can complete, from idea to execution to publication, all in one sitting (ideally under an hour).
Some of these mini projects will suck, but at least I'll be getting something done, rather than finishing nothing.
(For example, this blog post is very short and not particularly exceptional. But I was able to complete and publish it in one sitting, which is a better output than if I had tried and failed to create something bigger and better.)