If you are concerned about learning too much, you are doing it WRONG.
The more you learn, the easier it will be for you to solve common problems.
We are getting questions DAILY here which would be solved trivially by having some (very little) programming experience - which is very obviously missing. And in some cases, even willingness to learn is missing. Possibly not in your case, so I decided to write an answer.
Common "rule of thumb" says it takes 10K hours to master something.
But also Pareto principle says 80% of anything is irrelevant.
To be competent programmer, you need to learn few very different languages. Blub Paradox is term describing how less-skilled programmer looks at more advanced languages, and why.
If learning Lisp is too much for you, try Python AND JavaScript. Hint: JavaScript is Lisp with C syntax.
So if you apply both 10K hours rule and Pareto principle: You can learn new language much faster, if you have the experience to tell which parts of the new language are same or can be ignored. But to gain this experience, you need to learn few languages before that, so...
I have no experience with CodedUI, but IMHO all the attempts to replace programmers and creating some magical UI allowing users to generate code are doomed for failure. I have personal experience with one such system/framework: Fit/FitNesse. Idea was to provide users with table-like language so they can write tests. It never happened, so programmers ended up writing code is bad limited DSL (domain-specific language) instead of the one they already mastered. Why Fit failed - and yes, we still have to maintain many test written in bad DSL.
Avoid that, use real language, they are free and make you more productive beyond first learning curve. At minimum, allow you to use Page Object design pattern, increasing QA test engineer productivity immensely.
See also Greenspun's tenth rule:
Any sufficiently complicated C or Fortran program contains an ad-hoc, informally-specified, bug-ridden, slow implementation of half of Common Lisp
So instead of such half-implemented DSL (CodedUI, FitNesse), programmers prefer to have full (Turing-complete) language, which they can enhance as requirements evolve.