AI Sawed Off the First Rung of the Career Ladder

I'm so tired of the demos. Aren't you?

The slick videos, the breathless LinkedIn posts from tech bros, the keynote speeches promising a glorious new era of "human-AI collaboration." They show an AI writing marketing copy, generating code, or summarizing a mountain of legal documents in seconds. And they always frame it the same way: "Look how much faster you can be! Look how much more you can do!"

They call it an assistant. A co-pilot. A tool.

I call it what it is: a replacement. Specifically, a replacement for the person who used to do that work. And that person was almost always young, hungry, and learning. That person was the entry-level employee.

The Great Unlearning

For decades, the path was clear. You got your degree, you landed a junior role. You were a junior copywriter, a junior developer, a paralegal, a design assistant. Your job was to do the grunt work. The stuff the seniors didn’t have time for. You wrote the fifth, sixth, and seventh versions of that ad copy after the client rejected the first four. You hunted for that missing semicolon in the codebase for three hours. You spent a week in a dusty room reading depositions and highlighting names.

It wasn't glamorous. But it wasn't supposed to be. It was the training ground. It's where you learned by doing. It’s where you made mistakes that were low-stakes enough for the company to absorb, but painful enough for you to never make again. It's where you watched the seniors, figured out how they thought, and learned the thousand tiny, un-teachable things that separate a novice from a professional.

Generative AI eats this world for breakfast. All that "grunt work"? That's its entire menu.

Why pay a junior copywriter to brainstorm social media posts when ChatGPT can spit out 50 in a minute? Why have a budding programmer write unit tests when a Github Co-pilot can autocomplete them? Why pay a paralegal to summarize a document when a language model can do it instantly?

Tech companies, in their infinite wisdom, have built a tool that automates apprenticeship.

They’ve created a system that is excellent at doing the tasks of a beginner, but completely incapable of teaching the wisdom of an expert.

We are systematically eliminating the very roles that create the next generation of skilled, experienced, and valuable senior employees. We’re so obsessed with short-term productivity gains that we’re strip-mining our own future talent pool. Where do the senior designers and lead engineers of 2035 come from if the junior roles they would have occupied in 2025 are filled by an algorithm?

"Prompt Engineering" Is a Lie

"Ah," the tech evangelist will tell you, "but new jobs will be created! People just need to learn to be 'prompt engineers'!"

This is the biggest, most condescending lie of the whole revolution. Telling a generation of aspiring graphic designers, writers, and lawyers that their new career is to get good at typing sentences into a text box is an insult of the highest order.

It’s a niche skill, not an industry. It’s the art of politely asking the robot to do the job you used to do. It's like telling a generation of carpenters that their future isn't building houses, but getting really good at describing houses to a magical hammer. It completely misunderstands that the value wasn't just in the final product, but in the process of creation—the thinking, the iteration, the mistakes, the learning.

This isn't a new career ladder. It’s a footnote. a trapdoor. It’s a way to placate the people whose jobs are being automated by pretending they have a new, even more important role in the assembly line. They don't.

What are we doing to an entire generation of young people? We're sending them into a world of work where the entry-level is a locked door. We’re telling them to acquire skills, only to then build machines that make those skills obsolete on arrival. The anxiety is palpable. College students I talk to aren’t excited; they’re terrified. They see the ladder, but they also see the smug, glowing robot hovering right where the first rung is supposed to be.

This isn’t just a bad business decision, it’s a societal dead end. We’re trading the messy, inefficient, beautiful process of human learning and mentorship for the sterile, instantaneous, unthinking output of a machine. And we will pay the price for it, one missing rung at a time.