A good first step is to imagine something fun you would like to create for yourself (or whatever idea you already have), and then try building that thing as a side project. Limit the experiment to no more than 3 months. Learn how to heavily reduce the scope and build almost every day. See what you learn about yourself and then regroup.
Why do I suggest this? Many reasons. Here are a few:
First, you'll learn that the biggest hurdles to success are your own mind and habits. To even have a chance, you need to learn how to motivate yourself and work without a lot of structure surrounding a project. This is really useful.
Also, regardless of what kind of product, and regardless of whether you want to start a VC backed or bootstrapped company, the first step is getting started and showing some sort of traction. Again, keep the time limited, or you'll waste a lot of time building something no one actually wants. This part is when you'll start to learn about the all important product-market fit.
Finally, you'll learn about what happens when you have something besides your normal work to care about. This is an interesting dynamic. It might change how you feel about work, or it might not.
As far as the concern about stagnation, I've found that rarely goes away if you still have ambition about employment. Once you have a family, this only intensifies. Your personal tolerance for risk might change this dynamic. If you do enough interesting tech stuff on your side project and/or startup, then you can satisfy some of that concern.
