The future of IT

My thoughts on what the future holds for systems administration.

Tags: sysadmin aiPosted on: 2026-08-31

Recently I saw a Reddit post arguing that becoming a Systems Administrator wasn't worth it anymore. The main argument was around the fact that companies aren't hiring junior administrators anymore. Then there is also a lot of talk about how AI is affecting the tech industry more than most, replacing most of the work that IT workers used to do.

It's true that you can see a lot of layoffs in the news related to AI. However, I disagree with both arguments. and I personally don't see IT going anywhere. I want to briefly explain why I don't think tech is now a bad career to go into.

IT has always been an evolving field

I started playing with technology in the 1990s, when being in IT basically meant setting up Windows 95 PCs by hand, or keeping a simple web site up on an Apache and PHP platform. There was no automation in the 90s. No DevOps, no Jenkins, no deployment pipeline. At the most, you would have some Bash or Perl scripts that you composed yourself. Since then, the field has evolved tremendously, and workers have been expected to evolve with it. We're expected to do everything as code, use automation tools, DevOps, GitOps, DevSecOps, and of course being cloud native. I see AI as just the next evolution, and not as some kind of revolution like many tech CEOs tend to believe.

First of all, AI is non-deterministic. That's okay when you're talking about processing human language, but having a process that gives you different answers every time you run it is the exact opposite of how you want your technology to work. This is how we get the spectacular AI failures publicized in the news, things like hallucinations, and frustrations when it comes to relying on AI agents. While AI can be a very useful tool for some tasks, I think if you look closely at companies that push AI where it shouldn't be, you'll see humans are the ones that spend their days picking up the slack where the AI models failed.

Big Tech companies love to push simplicity, promising that their latest innovation will make it much simpler for everybody. But if anything, I think the IT world of tomorrow is going to be even more complex than the world of today. Every time a new technology is introduced, it means more skills are required to integrate it with current systems, ensure security, data governance, and so on. Some people think all we'll be doing in the future is prompting AI, but I think that's highly unlikely. To get good results, those models need massive amounts of scaffolding. You need people to pick the right models, put safeguards in place, build the harnesses, and while you can build a complete automation pipeline by asking OpenClaw to build it for you, you'll always get a much better result if you build a deterministic pipeline, and then insert AI only where it makes sense, like parsing human text.

And that isn't even talking about the cost of tokens. We're basically 3 years into the AI world, and we're already seeing lots of headlines about companies pulling back on token quotas. That means we're going to see more needs for creative uses of AI, including smaller models, local models, etc. All of that will require experts in the field, and that's more work for system administrators.

What are the risks going forward?

It's true that executives expect IT to do more with less, and the push to replace workers with AI will keep causing short term disruptions. But I actually think these next few years will be crucial for anyone wanting to start or continue a career as a systems administrator. Looking through a lot of social media threads, it seems like a lot of modern IT workers think the job is simply pressing buttons on a cloud dashboard. Emails are failing? Better open a ticket with Microsoft. The VMs are slow? Let's just open a support ticket with AWS. The network is failing? Better find the right vendor to contact. If that's your day to day, then you may be right that your future in the industry is bleak.

Instead, I feel like system administrators will need to do what we've done for decades: evolve and adapt. Learn fundamentals, learn about scripting, security, networking, etc. Make yourself more than a help desk worker who can only follow a checklist and open a ticket with a cloud vendor. This is why I built and run my own home lab. I built an MCP server, plugged it into my wiki, my private Git repo, and tested various AI integrations with it. That's why I learn about supply chain cyber attacks, and how to protect a network against click jacking. I constantly improve my automation pipelines. Not because I was told to, but because I find it genuinely fascinating, and the type of skill I should probably know about. I think the next 5-10 years will indeed see the IT landscape evolve, but it's not going anywhere. After all, when the AI model fails and the cloud isn't the solution, you're the last line of defense, assuming you actually have the knowledge needed to solve the problem.