It's Time to Be a Plumber | AI Security Now

Every time someone expresses concern about AI taking their job, there's a familiar response waiting in the wings: "People said the same thing about agriculture mechanization, about factory automation, about computers. And yet here we are, with more jobs than ever. Humans adapt. New opportunities always emerge."

It's a comforting story. It's also backed by centuries of evidence. But I recently encountered an argument—originally articulated by Forrest Maready in a podcast—that pinpoints exactly why this reassurance might not hold this time around.

The Historical Pattern That Gave Us Confidence

The optimistic narrative goes like this:

New technology automates work in Domain X
Workers are displaced but eventually retrain
Entirely new Domain Y emerges (often something we couldn't have predicted)
Humans fill those new roles because the technology that automated X can't do Y yet
Repeat

Farmers became factory workers. Factory workers became office workers. Typists became software developers. The thought-line was always the same: yes, your current job might disappear, but you'll find something new to do. Something the machines can't do. Yet.

That last word—yet—was always the escape hatch. The technology was specialized. It could do X, but when Y emerged, we had a head start.

Why AI Breaks the Cycle

Here's what's different now, and why Maready's observation cuts so deep:

AI's generality means there is no "yet" anymore.

When Domain W emerges tomorrow—some new field we can't even imagine today—why would we expect humans to have any particular advantage there? If AI can already handle knowledge work across domains X, Y, and Z, the company confronting this new opportunity isn't going to think, "We need to hire people for this." They're going to think, "Let's point our AI at this."

The cycle breaks because the gap between technological capability and new opportunity has collapsed.

In previous revolutions, there was breathing room. The steam engine couldn't do accounting. Factory automation couldn't do programming. Computers couldn't do creative strategy. That gap—measured in years or decades—was where humans found their next foothold.

But a sufficiently advanced AI doesn't need to wait for a specialized version of itself to be developed for Domain W. It's already general-purpose. It already handles language, reasoning, analysis, synthesis, planning. Whatever W turns out to be, if it's knowledge work, AI is already 80% of the way there on day one.

The Efficiency Imperative

Even if we grant that some new opportunities will emerge that humans could do, there's a second problem: economic pressure.

Companies facing competitive markets don't optimize for "employing humans." They optimize for efficiency. If AI can deliver 90% of the value at 10% of the cost, the decision isn't even difficult. And because AI can be replicated and deployed at scale almost instantly, the company that hesitates is the company that loses market share.

This isn't about corporate callousness. It's about survival in competitive markets. The firm that says "we'll hire humans for this new domain out of principle" will be outcompeted by the firm that doesn't.

So... Plumber?

Which brings us to the title.

I'm not literally saying everyone should go to trade school tomorrow (though honestly, it's not a bad hedge). The point is this: the work that remains resistant to AI replacement is work that exists in the physical world, requires hands-on problem-solving in unpredictable environments, and involves real-time adaptation to messy, non-standardized situations.

Plumbing. Electrical work. HVAC repair. Carpentry. Equipment maintenance.

These jobs require:

Physical presence and manipulation
Dealing with unique, non-standardized environments (every house is different)
Real-time problem-solving with incomplete information
Adaptation to unexpected complications

Could robots eventually do these things? Maybe. But that's a much harder technical problem than making AI that can write code, analyze data, create marketing campaigns, or even design buildings.

The bitter irony is that we've spent decades pushing people toward knowledge work as the "safe" career path. Get a degree. Work with your mind, not your hands. That's where the good jobs are.

Turns out we might have had it backwards.

The Uncomfortable Question

I don't have a policy prescription here. I'm not sure anyone does yet. But I think we need to be more honest about the question:

If the pattern that saved us before was "new domains emerge where humans have an advantage," and if AI's generality means humans won't have that advantage in new domains, then what exactly is the plan?

"Humans will adapt" isn't an answer anymore. It's an assumption. And it might be an assumption that's about to be tested in a way it never has been before.

Maybe it really is time to be a plumber.

Credit to Forrest Maready for the core insight explored in this piece, which I first encountered in a podcast discussion.