I was on a flight a while back and got to talking with the person next to me. It turned out they ran a high-end general contracting firm, and before long they were showing me photos of their work. This was really good stuff, beautiful work. He even had some videos he had shot with drones of waterfront work.

Then they told me a story I haven’t been able to stop thinking about. I’ll warn you in advance, this one’s uncomfortable. Somebody lost their job in it, and I don’t want to be tone deaf about that. But the business math underneath is one of the clearest examples I’ve seen of where AI and human labor are about to collide, so let’s look at this closely.

For years, this company kept a CAD creative on payroll. Their job was to produce the drawings and visual artifacts that went into bids and plans. It was important work and apparently the person was good at it.

But over time, the owner got frustrated. Not because the work was bad, but because it never got faster. The owner literally said, if only one small thing is different from a previous drawing, shouldn’t the next one be quicker to make? (Sound familiar, software devs?) But instead it still took days, sometimes weeks. From the owner’s perspective it felt like there were no efficiency gains, no skill improvements, and no reuse.

Then suddenly one day, the owner discovered something (you can probably guess that part). They figured out they could buy some GenAI tokens and vibe their way to the marketing drawings themselves. Faster, and far cheaper.

Oof.

So what do you think happened next? Yup. The owner laid off the CAD person.

Look, I get it. The CAD person, and many others, are going to say, “AI stole that person’s job”. And probably also insist that the work coming out of the GenAI tools is “not as good” as the CAD drawings. I mean, I’ve met people by now that if they knew would literally reject the contractor’s bid ON THE SOLE BASIS of the drawings were made by AI.

I suspect this story is not unique.

But let me present an alternative take on the story. Here’s a piece of business math that’s hard to accept.

The role stopped justifying its cost.

Why was this?

Again, let’s look through the lens of a software developer. What was that CAD person doing for years? It felt like there should have been a goldmine of past performance here. But instead, no reusable library of standard components. No templated starting points. No streamlined pipeline that turned the tenth similar drawing into a twenty-minute job instead of taking three days.

It is very easy to argue both sides here. I’ve been a software developer for three decades, I get it. “Nine women can’t make a baby in one month.” But let’s be honest with ourselves here…if after years you are not increasing your velocity in your development, you’re doing something wrong.

THAT’S the part I want you to hear, because it’s the part you can actually do something about.

In almost any knowledge worker role, there are two ways to spend your time. You can do the same unit of work over and over at roughly the same cost each time. Or you can turn your own repetition into a system, and become the person who makes the whole operation faster. The first kind of work is a commodity, and commodities get cheaper. The second is leverage, and leverage is exactly what makes you hard to replace.

(For more on this topic, try The E-Myth Revisited, by Michael Gerber.)

What sucks is that the CAD person could have been a hero. What if they’d walked into the owner’s office a year earlier and shown the owner the portfolio of all the previous work, and all the ready-to-go assets and pre-made drawings. That person doesn’t get replaced by tokens. That person helps the business level up, running circles around the competition.

I want to be careful here because again, me being tone-deaf means this comes across as a boomer telling you to “just pull yourself up by your bootstraps”. Obviously its not always that easy, and obviously not all business owners are going to see it this way. I’m sure someone will tell me that if the CAD person had given the business owner the assets, then the owner would have still just laid them off anyway.

Maybe. Or maybe the owner would see that this person brings value to the business. That the person’s REAL job isn’t “create a CAD drawing” but “help the business figure out what customers want and then show it to them”. And the CAD drawing was just a tool and a deliverable.

The reality is, we are in the midst of a disruption on the order of the steam engine, printing press, the Internet. And if you want to emerge successful on the other side, the answer can’t be “ban AI” or “just do the same thing you’ve been doing”. That didn’t work with any other revolutionary technology, and it won’t work here.