Emacs
Regex
Latex
Bash/Terminal/Powershell
Unicode
APL/J/K
Linear Algebra
Category Theory * LearningList
** TODO Emacs
** TODO Regex
** TODO Latex
** TODO Bash/Terminal/Powershell
** TODO Unicode
** TODO APL/J/K
** TODO Linear Algebra
** TODO Category Theory
For LaTeX: I very strongly recommend picking up a copy of the Lion Book (Leslie Lamport) and running through the examples. I actually picked up a great deal by taking documents I'd found online in bastardised for and tagging them up manually in LaTeX for output into flexible formats -- HTML, PDF, ePub, and more. It turns out that the basics of LaTeX are, well, pretty basics. Paragraphs are double returns, which buys you a lot. Italic and bold text, and the ``quoting'' style, are the main adaptations I had to make.
Digging into the equations and images features takes some doing, but you can produce excellent docs well before that point.
Stack Exchange has a wonderful LaTeX community.
Of the rest of your list:
1. Emacs: I actually got pretty good at it at one point, then signed on to a gig (in the days just before running one's own Unix was readily possible) in which it wasn't available and the boss (idiot) didn't believe in installing anything which wasn't vendor-provisioned. That said: commit to using it. I've found that I prefer vim's editing modes, but there's viper-mode in emacs. The power of having all text interactions under your editor is ... immense.
2. Regex: Start with the basics. The Sed & Awk book (O'Reilly) is a good start. Most of it is simple substiutions, and you can get tremendous mileage from that, though as with many tools, you can abuse the concept. Perl remains probably the canonical home for advanced regex use, though most languages have some facility for it now.
3. Bash: I'd recommend this, if only because Microsoft has a poor track record of integrating with others and/or supporting other tools over a long period of time. Bash has existed on Windows since the 1990s under various toolkits, most notably Cygwin, and is now pretty much native.
4. Unicode: mostly a matter of figuring out what bits you need within a given project. I'd deprecate that on your list and consider it something to pursue on an as-needed basis.
I can't much speak to the remainder of your list, though I'd suggest treating them as tools adapted to uses and goals.
Deep dives are fine, but sometimes you just wanto to get a few toes in the water.
I'd argue that if you've got performance issues with regexes, you're (arguably) abusing the tool.
https://github.com/melling/LanguageLearning
https://github.com/melling/ios_topics/blob/master/README.org
It doesn't take long, and I found the explanation clear enough that it's stuck. It's also a good reference site if you want to quickly remind yourself what X means in Y dialect.
I also have a long todo list that more things go on than get done. But, I think it's probably unreasonable to expect to learn more than say one or two programming-language level things a year if you've got a job. Certainly if you're going to do the learning thoroughly and not just forget it all by the next year.
https://www.youtube.com/playlist?list=PLZHQObOWTQDPD3MizzM2x...