Use simple methods to make complex problems either go away or become routine. Take the same tools I learned and used for weapons targeting, systems vulnerability assessments and damage expectancy calcs and use it to make machines - real, virtual and imaginary - do what I want.
I started out in tech as a system administrator. These days, I'm a Op - my tools still include multidrivers and wire crimpers, but also include cloud APIs, Python, Bash and git.
I tell my teams that "good Ops are lazy." If they grok the correct meaning of this, they're Ops. You've got to always be thinking about how to not have to do what you're doing again. Have the computer do it; that's what they were built for.
I want to build my rackfarm once. I want to manually spin up my cloud once, so that I know exactly how it was built and how it behaves. Then I want my tools to do it faster and more consistently than I can. I'm valuable because I can think, and that's what the tools can't do.