That sounds like an issue a close friend of mine has and he refers to it as a blank page problem. He can modify existing projects but has issues starting a new one. And the approach you suggested works for him - he starts with a very rough and broken poc and then fixes it bit by bit.
I believe my problem is different. Once everything is figured out I lose interest. So if there problem requires a lot of thinking and investigation where you end up changing few lines of code, I love it. But if I see a ticket, the solutions is obvious to me and the rest of the time has to be spent coding, it just feels boring.
> But if I see a ticket, the solutions is obvious to me and the rest of the time has to be spent coding, it just feels boring.
I only have ~3 years of experience and I feel the same as well. Certain colleagues and I meet, discuss some possible small side projects, and after selecting one I usually participate in the planning and outlining or the algorithmic issues, boiler plate part is given off to a junior if interested enough to work on same.
I believe my problem is different. Once everything is figured out I lose interest. So if there problem requires a lot of thinking and investigation where you end up changing few lines of code, I love it. But if I see a ticket, the solutions is obvious to me and the rest of the time has to be spent coding, it just feels boring.