I started out doing what basically amounts to working in an agency. A business guy I graduated with wanted to start a marketing company and he got the jobs and I built the websites. It was still a contract setup and I made $50 an hour because Hacker News told me that $50 was a good starting rate so I asked for it and he gave it to me.
Then I got into doing Django web development and charged around $90/hr because I got a lead and I didn't really want the work so I said, "My rate is $90/hr" and he said "okay" and from then on my rate was $90/hr. (Ha!) I just recently got $125/hr in a recent project.
Starting when I was still sort of in college helped a lot. Living in small town Indiana where the college was and this company started out helped a lot lot. Rent at the time was $150/month with a roommate. Having very low expenses meant I could experiment more and not have to worry as much about paying rent. Not having any dependents helped too.
The biggest advice I have to anyone who wants to go out on their own: learn business skills. If you already know how to develop, you may think that's all you need. It's a meme at this point that "If you build (a SaaS) they will come." No, they really wont. You need marketing, marketing, marketing, and coding to build a webapp.
This is why I recommend people start out freelancing first. Then go into building consulting services. And _then_ build products. Freelancing is so much easier to make money with. Any competent developer can get $100/hr+ with some networking. That will teach you how to talk to clients and how to sell yourself. All things you need if you are going to sell products, but it's way easier and you don't also have the burden of designing a good product.
Consulting services are very lucrative, but require even more business sense. If you can business and code at the same time, you are in a verrrrry small percentage of people, and have superpowers.
Build your audience somehow. You'll need one if you are going to sell a product. Make the product first as a service that "doesn't scale". That is, manually implement the business outcome your software would give your customer manually yourself.
Then if you feel the need build a product based on your consulting business. Create rungs on a "product ladder" - free content you put out, paid info products, higher price consulting services, and then at the very top automate all of it in software. It's a model I've seen a number of people do that will be a lot "less risky" than quitting your job and building a webapp blind without anyone watching.
I tried this once. I quit all my freelance work for a year and lived off of savings. I then proceeded to not make anything anyone wanted and just ended up spending all of my money. Then I got back into freelancing and have money again.
In general, this strategy is: "Have safe money and risky money". Your employment is safe money. It's boring, but safe. If you want to get out on your own with less risk, figure out a way to get more time but keep that safe money (maybe try to reduce hours, maybe try to reduce your commute, maybe try to work 4 days a week), then pick up freelancing part part time. Then once freelancing becomes "safe money", quit your job. Work on freelancing part time and build a business that is "risky money".
This is what I'm now doing and I have managed to make a reasonable living while having a lot more time to pursue ventures I'm passionate about. I still spend about 20 hours of my day doing "boring" web development work, but it pays for the rest of my life. I'll probably always keep at least a little of that work because it pays so well for the time put in.
Happy to talk about this more, this subject is of great interest to me and is something I've thought about a lot. My email's in my profile if you want. I think your mindset is more reasonable than many people seem to think, and I wish you the best in your quest to find work you love. It's a long term process but in 10 years of doing this I wouldn't have it any other way.