I'm asking the HN community because you all seem smart and I believe programmers are very natural high-focus people.
You don't.
People tend to overestimate how much work can they do in the short run, but vastly underestimate how much can they in longer timespan. This means you should simply start doing whatever you have at hand, and do it again and again. Start accumulating knowledge and skills and finished side projects.
You say yourself you're interested in many things; take notes when you're working, so it takes you less time to resume your work later. It will still take a lot to resume work, especially if the field is completely different from what you were doing lately. Maybe at some point you'll decide to narrow your focus, so switching becomes easier; maybe you'll get better at switching context; maybe you'll just get used to this.
You can't gain expert level in everything, so you must accept that you'll leave some fields out. But don't let it prevent you from learning any of them.
After two or three years you'll see real progress, and after five you'll be amazed how much have you done.
Instead of sticking with it and finishing them completely.
I think it's about prioritising what could have the biggest impact to your life and being honest with yourself as to what you want to get out of it.
I have now allocated time for "intellectual exercises" but I don't kid myself that I'm doing these for my career. And if I find a justifiable angle for it to benefit my professional career I make sure I put in place the steps to promote it on my résumé, blog etc.
There's nothing wrong with having lots of interests and playing around with lots of pet projects. But the more you do that, the less likely you'll ever finish anything. There's certainly a lot of sacrifice in order to see something to completion.