The year I was sure AI would end my job

Back from lunch. I'd just worked out and taken my hour, the one slice of the day that's actually mine. I walk into the office and someone waves me over to a screen. A few people are already gathered around it. "Come look at this. Watch, this is about to make your job disappear."

The person saying it was smiling. It was a joke. It just wasn't only a joke.

On the screen, an app was building itself. A real one, right there in front of everyone. And I stood in my post-workout calm, watching a machine do, in minutes, a rough version of the thing I'd spent years learning to do by hand.

That was a couple of years ago. The first time I watched a tool like that work, and the first morning of a year I spent quietly certain my work was ending.

I went looking for higher ground

It moved fast after that. I'd open LinkedIn and everyone was on it. Lovable, then Bolt, growth curves that didn't look real. Every week the tools got better and the distance between "I have an idea" and "here's the app" got shorter.

I was a builder. No-code, Bubble, product. My whole edge was taking something from nothing to shipped, and here was software doing the nothing-to-shipped part on its own.

So I did what you do when you're scared. I looked for higher ground. Product management, I told myself. That will survive. Strategy, judgment, talking to users, harder to automate. I started steering back toward the PM in me, toward the part of the job that felt safe.

Around me, people were making bigger moves. People in tech, in growth, in sales, talking about leaving the whole thing behind for something they could touch. Something physical, real. You heard it out of San Francisco too, people walking out of the industry that had defined them.

The dread came at the wake-up

I want to be honest about what that year felt like, because the honest part is the point.

It didn't keep me up at night, and that surprised me. I've been anxious my whole life. Fifteen years of insomnia, and I don't mean falling asleep at 4 instead of 10, I mean not sleeping at all, sometimes two days straight, every week or two. I'd finally beaten that. So the dread didn't come at night anymore.

It came at the wake-up. In the metro. Packed into a train that kept breaking down, on my way to work, and this strange disconnection would settle over me. Look at all of us, crammed in here, rushing to jobs that in three months might not exist.

I wasn't afraid of going from dev to software engineer, or from lead to product owner. I thought the entire field was about to be gone. All of it.

I talked to my girlfriend about it. My parents. But you know how it is. At the bottom of it you're alone with the thing, and I thought about it every single day.

The snake eating its tail

Here's what made it worse: the companies.

They wanted AI, badly, but they didn't want you coding with AI. And they wanted you to ship in hours, as if you were. The snake eating its own tail. Deliver like a machine, but by hand, and don't ask questions.

By the end of that year I hit a small burnout. A contract that ended badly. And I remember thinking, I want to work in tech, but at what cost.

The boat was empty

There's an old Taoist story I wish I'd known that year.

A man is crossing a river in his boat. Another boat is coming straight at him, about to hit. He shouts, he gets angry, he braces to blame whoever's steering. Then he sees the boat is empty. No one's in it. Just a current, carrying it toward him. And his anger vanishes on the spot, because there was never anyone to be angry at.

That was the whole year in one image. I'd been shouting at an empty boat.

I'd decided AI was coming for me, that there was intent, a threat with my name on it. There wasn't. It was a current. And the moment you see the boat is empty, you stop drowning in it and you start learning to row.

I turned toward the thing I was scared of

Then something shifted. I stopped running toward higher ground and turned to face the water. Claude Code was everywhere. Instead of protecting the PM in me, I put my hands in the code, the exact thing I thought would kill me.

I'd always loved no-code. Bubble was my comfort zone, my happy place. So when I started Nudje, my own product, I made the uncomfortable call and built it straight in code. Not the "hey, build me an app" kind of vibe coding. Weeks, months of scoping. Trying to talk to potential users, one percent reply rate, but one percent is still something. Real product work, with AI as the engine underneath.

And the thing I was sure would erase me did the opposite. It multiplied me. Today I run four projects at once, not all the same weight, without my brain melting, and the work is good. That still amazes me.

Velocity was never the job

Here's what I finally understood.

No-code's whole promise was velocity. We shipped features in days instead of weeks. AI ships them in hours. So yes, on pure speed, no-code lost its reason to exist.

But speed was never the actual job. Shipping the feature fast is the easy part now. The work, the real work, is everything around it. Scoping it right. Testing what comes out, because AI writes fast and wrong just as easily as fast and right. Reading the code well enough to make it better. Convincing a client that QA isn't a cost, it's the whole point.

That's where product and technical people live now. Not in typing the feature. In deciding what to build, and proving it holds. AI didn't take that. It handed me more of it.

These days, I've never seen so many roles open. AI PM, product builder, a few people reaching out every week. The year I was sure the field would vanish, I got the luxury of choosing what came next.

If you're the one in the metro

So if you're where I was a year ago, waking up in that metro doing the math on how long you've got left, here's the only thing I'd tell you.

The boat is empty. There's no one in it coming for you. It's a current, and currents you can learn to read.

Turn toward the thing you're scared of. That's usually where the work went.