Robots Are Coming — But They Haven’t Met My Plumber Yet
Where Automation Meets the Real World
A few weeks ago, I had a plumbing issue.
Nothing dramatic. Just one of those problems that starts small and then slowly reveals itself to be something else entirely.
At first glance, it looked straightforward. A leak. Maybe a fitting. Possibly a simple replacement.
That lasted about five minutes.
What the plumber actually found was a piping system that had been installed throughout our neighborhood when the homes were originally built. In our case, we had made only minor changes over the years—a split pipe at the water heater and a failed bathtub drain. Otherwise, the system had been left alone.
Which sounds fine… until you realize what it means.
We are essentially sitting on a potential time bomb.
The original plumbing system—Kytec—has a known history of failure. Sometimes it fails in a few years. In our case, it has lasted more than twenty. That doesn’t make it reliable. It just means we’ve been lucky.
And now, any repair brings a new complication.
Different materials.
Different standards.
Different decades of thinking about how plumbing should be done.
In other words, a perfectly normal house.
The Problem Wasn’t the Problem
The issue wasn’t just fixing a leak.
The issue was figuring out:
- what had been done before
- what still worked, what didn’t, and why
- how to connect things that were never designed to work together
There was no manual.
No clean diagram.
No step-by-step instructions.
Just a situation that required someone to:
- look at it
- think about it
- try something
- adjust
- and make it work
What made it more interesting was the material involved.
The house has Kytec piping—something that, at one point, seemed like a good idea. It’s now been out of business for more than a decade, and for good reason.
At this point, more than a third of my neighbors have replumbed their entire homes to get rid of it. That was the preferred solution from my plumber as well—a clean, complete fix.
The estimate was somewhere north of $16,000.
That gets your attention.
The alternative was to work with what was already there. And that’s where things got complicated.
Kytec doesn’t connect easily to modern copper systems. In fact, there aren’t standard fittings for it anymore. So the problem wasn’t just fixing the issue—it was figuring out how to connect incompatible systems safely and reliably.
There was no clean answer.
Just a series of decisions.
This Is Where the Conversation About Robots Gets Interesting
We hear a lot about robots and AI replacing work. And in many cases, that’s probably true.
But it depends very much on the type of work.
There’s a difference between:
- structured tasks
- and real-world problem solving
Where Robots Will Do Very Well
Robots are going to be excellent in environments that behave.
- manufacturing
- logistics
- warehouses
- new construction
- repetitive industrial work
These are places where:
- the environment is known
- the variables are limited
- the process can be defined
That’s exactly what the recent RoboForce article was getting at.
👉 Build systems designed for specific tasks in controlled environments.
That works.
It seems likely that plumbers, electricians, auto repair technicians, and others who fix things will be around for quite a while—without an immediate threat of being displaced by robots.
Where Things Get Complicated
Now take that same robot and put it in a house that has been:
- partially remodeled
- patched as needed
- and adjusted over time
Now ask it to:
👉 figure out what’s going on
That’s a different problem.
The plumber didn’t just fix something.
He interpreted a situation.
He made judgment calls.
He adapted in real time.
In our case, he gave us a choice:
- install a workaround using temporary fittings to keep the recycle system functioning
- or remove part of the system and simplify the plumbing so standard fittings could be used
The first option came with no warranty. If something failed, it was entirely on us.
The second option was cleaner and supported—but required giving up part of the existing system.
Some choice.
The Gap No One Talks About
There is a gap between:
👉 what can be defined
👉 and what has to be figured out
Robots handle the first very well.
The second is harder.
Not impossible—but harder than most of the headlines suggest.
Experience Matters More Than We Think
Watching the process, it was clear that the solution didn’t come from:
- a set of instructions
- or a predefined method
It came from experience.
From having seen enough variations of similar problems to recognize patterns and possibilities.
That kind of knowledge doesn’t always show up in a clean, structured way.
It’s built over time.
One situation at a time.
This Connects to Something Bigger
There’s a broader theme here.
We’re building robotics systems that are very good at:
- executing defined tasks
- optimizing processes
- working within known constraints
But much of the real world operates outside of that.
Especially:
- repair work
- older infrastructure
- anything that has evolved over time
Which is… most things.
A Branch We’ll Come Back To
Could robots eventually do this kind of work?
Probably.
But it will require:
- better sensing
- better adaptation
- something that looks a lot like judgment
- and the ability to evaluate alternatives and assess risk in real time
That’s a higher bar.
In the meantime, the first wave of robotics is likely to stay where the problems are cleaner.
So What Does This Mean?
It means two things can be true at the same time.
Robots and AI will:
- transform many industries
- handle large amounts of structured work
And at the same time:
- struggle with messy, real-world problem solving
At least for a while.
Final Thought
There’s a tendency to talk about automation as if it moves evenly across everything.
It doesn’t.
It moves where the environment allows it.
And where it doesn’t, people still matter.
Especially the ones who can walk into a situation, look at it for a few minutes, and say:
👉 “Well… that’s not how that’s supposed to work.”
And then fix it anyway.
