You should scale your expectations down, come up with the most simple, minimum viable product. If you still don't feel the scale is small enough, try to break off a chunk of that MVP you think you /can/ complete.
Then try.
You're probably going to make an initial version as you're learning, and your design is either going to radically change as you learn more about the problem you're actually trying to solve and the tools you're solving it with, or you'll end up throwing it away and refactoring the entire thing when you do have an understanding.
Then, when you've got something that works, you'll start tweaking it. Making X better, or adding Y; or even removing Z because you realize it isn't something that should be there.
That's the programming equivalent of the 50 pots thing. Each iteration you run being like a small test-firing, and each deployment being like a pot you actually feel like putting through use tests.