Friday 16 October 2026

The Difference Between Using and Relying on AI

I recently read JP’s article, “Things Are Still Worth Doing Yourself”, and while I agree with some of what he’s saying, I think he’s placing AI usage into one rather narrow category.

I agree that there’s value in creating something yourself. I also agree that relying on AI to do everything for you, without understanding the process or learning anything along the way, can be detrimental. Particularly when we’re talking about creative endeavors like writing, art, music, or game development.

But I disagree with the idea that depending on AI in any meaningful capacity automatically means we’re abandoning critical thinking or opportunities to learn.

Over the past several months, I’ve used AI extensively as a technical troubleshooting assistant. I’m a longtime Linux user, and between my desktop systems, servers, virtual machines, networking equipment, and self-hosted applications, there are plenty of opportunities for something to go wrong.

I’ve worked through problems involving Linux kernel modules, conflicting virtualization software, DNS server failures, GPU driver upgrades, systemd services, firewall configurations, remote connectivity, and server migrations, among other things.

In these situations, AI didn’t simply do the work for me.

I described the problem, provided command output and error messages, evaluated possible solutions, executed commands, and verified the results. When something didn’t work, we went back and investigated further.

That’s troubleshooting. That’s learning. And that’s critical thinking.

Before AI, I would have approached these problems by searching Google, reading documentation, digging through forum discussions, and trying various solutions until I found one that worked. Depending on the problem, that could take anywhere from a few minutes to several hours, sometimes longer.

AI hasn’t eliminated that investigative process. It has made it considerably more efficient by helping narrow down the possibilities and explaining why particular solutions might work.

And I’ve learned things along the way.

I’m not suggesting AI is infallible. It isn’t. It can provide bad advice, misunderstand a problem, or recommend something completely inappropriate. You still need enough understanding to recognize when something doesn’t make sense, particularly when you’re making changes to an operating system or server.

But that is precisely my point. Used properly, AI doesn’t have to replace your ability to think. It can supplement it.

There’s a substantial difference between asking AI to create something because you don’t want to put in the effort and using AI as a tool to help you accomplish something you are actively working on.

I think the former deserves scrutiny. I think the latter can be enormously beneficial.

Ultimately, I agree that things are still worth doing yourself. I just don’t believe doing something yourself means you can’t use modern tools to help you along the way.