When you say you're working 50-60hr/wk and want to load more learning on top of that, I feel concerned for you. You've described a sure path to burnout.
If you're doing this for fun, well, make it fun. Take it easy on yourself. Just work on side-projects when you have time and energy. No one cares whether you're an expert in something that you just do for fun.
If the goal is more to make a career switch into some kind of IT or computer profession, I can say that I've never encountered anyone who expected me to know everything. We're all grownups and we all understand everyone has different strengths and weaknesses. When we're looking to recruit, we want someone who is helpful and can learn fast, not someone who is already an expert in our whole stack. You'll be more likely to get the job you want from networking than from already being an expert in techs a, b and c. When networking, it's nice to be able to talk about what skills or side-projects you've been working on, but think of these more as ice-breakers than as things which need to be perfect.
To focus on the question about analysis paralysis, try inventing side-projects which align with your goals. Setting a short-term goal will help you figure out what skills you need. If you want to get into game dev, either build a basic game from scratch or start modding existing games, it doesn't matter what. You'll figure out soon enough if you need to learn more theory.
Choose one stack and start small. It’s the only way. Just one rabbit hole can cost you months of time you don’t have.
My stack is node and Vue because it’s all JavaScript and Vue feels easier than react.
Why destroy yourself? Work on getting that 50-60 hr/wk down or at least re-point it at something you want to learn on the job.
Automate parts of your job. Programming is a tool, not a profession. Use that tool in ways that benefits you.
This doesn't have much to do with your being self-taught, IMO, it's probably a feature of your personality. I have such analysis paralysis sometimes, though not as much as you, I think.
The real answer is probably that you should pursue whichever subject of CS/programing/tech interests you. If you want a fast, reasonable answer, why not just go the curriculum route? It should serve as a foundation for anything else you may learn.
I took a class in fortran programming for geologist in 1968, then I built a career growing along with the industry as it grew from nothing to this huge affair today.
The way I did it was not to get ahead of the needs of the current project. New computer? Read the manual. New language? Read the manual. New paradigm? Read the manual. No immediate need? Ignore it.
In 65 years the only time I studied something before I needed it was in 1982 when I studied Pascal.
Neither you nor I can anticipate the future. Keep alert and quickly seize opportunities. Don't worry about what you don't have an actual need for.