Here's the thing though: you don't have to settle down and there's no reason why you should if you don't want to. If you feel better running your own show and make it work, continue doing that.
I'm completely unsold on working on someone else's ideas rather than my own, and I'm not really a fan of being a little worker bee in someone else's business either -- it feels deeply unsatisfying for me. Perhaps this is just how I'm wired up (and, without meaning to mangle Ben Franklin's words), I value freedom over job security. I just can't get excited about the idea of "settling down" with a company.
Try to do what makes you happy if at all possible. If you end up in a situation that's no longer making you happy, work to change it. Bonus points if you can choose the option that makes for the better story.
If what your goal is is to be able to pick up whenever you want, move where ever you want, change what your doing on the drop of a hat then ya joining a company doesn't make sense. But I would argue then you're using programming as a means to a type of lifestyle. I would much rather commit to an idea and team that I believe in and make something special happen.
If your on the same page as the company you are working for then having more man hours, capital, resources and connections made available by the company can only help you do what you want to do better and faster.
Thinking up apps and launching prototypes is fun, sure -- but it's also the easy part if you're a little creative. The hard -- and valuable -- part is growing it into a business, successfully scaling it, building a company, and all that goes with those things.
I think if you looked on average at people who "just start startups and bolt" vs. people who help build startups beyond their first phases, the latter tend to participate more in the equity upside. I can't imagine the appeal of being the former.
To the OP -- you don't actually say why you don't like the idea of being tied down. Can you elaborate?
For example, Amazon's suggestion feature is a brilliant implementation of K-means (or something like it). The small group of developers that implemented it probably got to face unique and intellectually difficult challenges that not many other people have had to face. Sure, I can imagine how they did it, but I didn't do it.
One person's small company is another's large.
Enjoy where you are right now and every moment as it happens.
Individual insurance in the U.S. is a recipe for disaster. (It may not befall you, but the odds are far greater than miniscule that it may. And as you get older, individual insurance -- or no insurance -- becomes increasingly risky, along with the likelihood that you will have a family including dependents.)