People who code and automate things (whether it’s for a big tech shop, or a smaller company’s line-of-business pipeline) like to think we’re all that. We’re really not.

We’re digital carpenters who are paid well only because demand outstrips supply…and because a lot of people think what we do is super-complicated wizardry. It’s not - it’s advanced sudoku at best.

Sure you still need to be smart and experienced to get good at sudoku (coding automation pipelines, too). But while coding skills are not easily learned quickly, it also isn’t the Rain Man + Gandalf undertaking that many would have you believe. But the practice of creatively making…something useful–seemingly out of thin air? That will always be nothing short of magic to a lot of people. Myself included. That said, the tech layoffs of the past 20 months have taught me that coding for automation, site-resilience, and overall operations is slowly going to become… less rare in terms of supply. This will be due to AI and its rise, and people who get really, really good at creating output pipelines using AI: they will be earliest discruptors in DevOps.

I have always said: it’s not AI that will take your job, it’s people leveraging AI that will take your job. And I see a lot of junior engineers and non-coders wielding it, for DevOps stuff these days.

The “wielding it well " – that part isn’t here yet. But it’s coming, and yes it will “scale”. Tech workers using AI with precision and proper risk-management, with a discerning ability to QC generated outputs, are already here and steadily making gains at their jobs using AI to iterate and improve on what they already do. And these tech workers are gaining ground to be miles ahead of those who don’t use AI. Sorry, it’s just facts.

As a recent example, I began learning Terraform which uses HCL syntax to essentially map and manage JSON value types and expressions (an over-simplified description, it does so much more). It’s heady stuff, very new to me, yet I was able to leverage a simple GPT-4o OpenAI Playground model to work me through the basics. In no time at all, I had my first Terraform AzureStack deployment rolling complete with a virtual network and compute resources. For leading-edge cloud enginering and DevOps Infra as Code deployment – learning (and then quickly using) something like Terraform could take months of iteration and labbing. It only took me a week to get from zero-to-deployed lab, and another week to learn ins and outs using AI and the docs as a starter-guide. I even pushed live HCL through GPT-4o, generating JSON outputs outside of my Terraform lab, just expand my fundamental understanding of how to tweak its operation in my test-dev scenario. AI literally shaved months from what would have been an extended learning cycle for me.

I believe code may eventually become, for your average non-tech company, a commodity. So unless you’re an engineer at OpenAI or other big tech shop, or you are working with pipelines that millions of users depend on (basically, a relative few in the future), what you do in dev right now could become less bespoke in the next few years. That’s not at all to say coding is going away, it isn’t. But the practice of developing software and creating solutions is changing, and leveraging AI may become as important as which framework or platform you work with. And if you’re avoiding AI altogether in Enterprise IT and DevOps, “not ever leveraging AI” could eventually translate to being “less in-demand”. Less bespoke.

What we do in software development and operations is still quite valuable and, for the moment, still a rewarding craft to take pride in doing well. But do mark my words: more and more people will develop published commercial and open source projects coming not from dev teams and sprints, but from small groups creating AI-driven deliverables. So if you’ve been ignoring AI at all, you aren’t doing your tech career any favors over the next 5-10 years by continuing to ignore it.

That is not to say that AI does everything better than teams of humans in anything right now. But when it comes to software engineering – coders and technical teams are already seeing disruption. AI hasn’t quite had its “Why Software Is Eating the World” moment, but trust me it’s only a matter of time. The potential gains in software engineering, operations, and management are just too great for orgs to ignore. Yes it’s early days…yes there remain standards and safety discussions, debates, and the usual human rule-making (read: legislation) cycles; however: nothing stops this train. OpenAI, Anthropic, and anybody else with signifiant Compute are iterating smarter and more capable training models each month. The progression cycles are too heavy, and moving too fast, and nobody is slowing down.

So what can good software and DevOps engineers focus on, besides learning to leverage AI, to protect and grow their careers in these coming disruptions? I think the most important thing for you to remember as a DevOps Engineer, Software Developer, or anyone driving automation pipelines right now: People, not code, are the most important part of your job.

If you don’t stray from the concept that Software is ultimately a people undertaking, you’ll secure some stability during the disruptions in the years ahead. AI usage is going to shake things up, and DevOps project-work is already being affected, but remember: ultimately Projects arePeople. That’s doubly-true for any software dev or enginering project. Keep that understanding as your high-order bit, and software engineering in the age of AI need not be a dead-end job in the decade ahead.