How to be less frustrated with computational reproducibility?

After a semester of seeing students struggling with basics of working with R, I started reflecting on my own frustrations with trying to achieve greater reproducibility of my research workflows over the last year. I ended up writing a post about it which took a very different direction from where I was planning to go (I hope it's not inappropriate to share it here):

There's a ton of practical resources and tutorials online about reproducibility, but I'm interested in how do you all deal with the complexity and abstraction creep that seems to inevitably come with most reproducibility-oriented tools? Am I just getting old and struggling to learn new tools or do others who otherwise consider themselves fairly technically capable have similar issues?