A conversation about EMPs and the end of civilization made me look hard at my own skillset. What I found was more durable than I expected — and what I found missing from the industry was more concerning.
"AI and technology will be gone. Taken out by an EMP in both of our lifetimes."
Paraphrased slightly, but that's the gist of a conversation I had recently. The kind of conversation that sounds extreme until you sit with it for a while.
Also shared on LinkedIn.
I'm not here to argue probability. Whether an EMP event at scale is likely or not is a different post, and honestly, a different domain.
But the question it raised is worth taking seriously: in a world where companies are racing to go "100% AI," laying off entire engineering teams to run leaner on automated tooling, and the average person can't sustain two minutes away from a screen — what actually happens if the technology disappears?
This is territory that "preppers" have been thinking about for years, each with their own theory of what the breaking point looks like. An author I've been reading, Liz Koss, explores the EMP scenario through fiction in her novel The Rebuild — available on Amazon — and it's a compelling read if you're the kind of person who processes big hypotheticals better through story than through survival manuals.
My first instinct when that conversation ended was to look at my own skillset on paper.
I write code. I test code. I work with tools that exist entirely within technology. If the infrastructure disappears, what exactly do I have left?
More than it looks like, actually.
The way I process problems doesn't require a computer. It requires a mindset: question something, test it, observe results, adjust, repeat. That's the scientific method. It's how I approach a legacy codebase I've never seen before. It's also how I'd approach herbal medicine, carpentry, gardening, or any other domain that has nothing to do with software.
The tools change. The process doesn't.
The more specific question is what disappears when AI specifically goes away.
For me, AI functions as a sounding board. A junior developer I can ask questions at midnight without burning anyone out. A starting point for exploring ideas in domains I don't have formal training in. It's made the learning process faster — not because it's doing my thinking, but because it doesn't get bored or annoyed when I ask seventeen follow-up questions about the same topic.
Without it, I'd do what I did before: find people who know things, exhaust their knowledge base, find books, find more people, document what I learned, and share it forward. The process is slower. The process is otherwise identical.
That's actually a reassuring realization. The knowledge compounds in me, not in the tool. When the tool goes away, the knowledge stays.
Here's where I want to be direct, because I think it's getting glossed over in a lot of conversations about AI adoption.
Companies have been cutting experienced engineers, testers, and QA professionals to run leaner on AI tooling. This is a real trend. Teams that spent years building domain knowledge — understanding how users actually behave, where the edge cases live, what the code is really doing underneath the feature — have been downsized in favor of models that generate faster and cost less.
But that institutional knowledge didn't transfer to the model. It walked out the door with the people.
AI doesn't have context. It doesn't know your specific users, your specific codebase, or the specific ways your system has evolved over time. It can generate something that looks right. It can't always tell you whether it is right for your situation. The people who could tell you that — the ones who'd spent years learning to "act dumb" and think like an unexpected user, to catch the thing the model missed — a lot of them are gone.
If the technology underdelivers, or disappears entirely, what's left?
The companies that kept people who knew things will adapt. The ones that replaced knowing with generating are in a more precarious position than the efficiency numbers suggest.
My skills are more portable than my resume makes them look. The mindset I've built — curious, methodical, willing to sit with a problem until I understand it — applies everywhere.
Without technology, I'd build, grow, and learn with what I have available. I'd document it. I'd share it. I'd find the next problem worth understanding.
That part doesn't change. It never really depended on the tools to begin with.
The day-to-day survival question — what preppers are actually preparing for — is a genuinely different conversation, and one worth having another time.