My first real job was teaching first grade, and a huge amount of it was, honestly, repetitive. Sitting on a tiny hard chair with one kid for the tenth time on the same letter sound. Long repetitive stretches before suddenly seeing it click. The material was the same, but how each kid learned it was different. In that way, the repetition made room for me to notice the subtleties. The way it clicked for one kid. The way the exact same approach fell flat with another kid. How sometimes no matter how you explain it, today won’t be the day it lands. And how on other days it can feel like a giant leap forward. It wasn’t glamorous, but it was also exactly where I learned to teach. You don’t get the judgment without first doing the small, repetitive, humbling stuff. That’s how almost all of us learned to do anything.
I’ve been thinking about that a lot, because AI is remarkably good at precisely that kind of repetitive work. The first drafts, the simple tickets, the basic analysis, the repetitive tasks that used to make up a new person’s entire first year. AI does them in seconds now. Which sounds wonderful, right up until you ask the question sitting underneath it: if no one does the simple work anymore, how does anyone ever learn to do the hard work?
This isn’t a thought experiment, and it’s moving fast. Postings for entry-level roles in the US are down about 35% in the last 18 months, a lot of it tied to AI. A Harvard working paper found that at companies adopting generative AI, entry-level hiring has fallen by roughly 80%. Recent college grads are now facing higher unemployment than the national average, which almost never happens. And the roles still labeled “entry-level” increasingly aren’t: in software, more than 60% of them now ask for three or more years of experience. Researchers even have a name for it. They call it seniorization, the strange new world where entry-level jobs suddenly demand senior skills like judgment and strategic thinking, the exact things you used to spend years earning your way toward.
So here’s the crisis. We’re automating away the bottom rungs of the ladder while still expecting people to somehow arrive at the top of it. CNBC called it the end of the career ladder as we know it, and I think that’s about right. This isn’t only a young-person problem, either. Every senior expert you depend on was once a junior who got to make cheap mistakes on small things. If we stop building those rungs, we are deciding, whether we mean to or not, to stop growing our next generation of experts. That bill comes due in about a decade, and it is a brutal one.
But I don’t think the ladder is gone. I think it has to be rebuilt on purpose, and the companies that do it will have an almost unfair advantage ten years from now. A few things I’ve watched work:
Treat your juniors as apprentices, not cheap labor. AI does the simple tasks now, so a new person’s real job is to build judgment fast. Give them the interesting problems sooner, with more support around them.
Let them work alongside AI, not get replaced by it. Use the tools as training wheels while a human mentor pushes them toward the calls only people can make.
Capture what your experts know before it walks out the door. That knowledge used to pass down by osmosis, over years of shared repetitive work. That osmosis is exactly what’s disappearing, so you have to make it deliberate now.
Rebuilding that ladder inside a company, keeping people growing while the tools keep changing under them, is a big part of the work I do with teams. If it’s on your mind, I’d love to talk.
I still think about that tiny hard chair in my first classroom, and how much I learned in the least glamorous hours of that job. Our kids, and our newest colleagues, deserve a version of that too. This time, it’s on us to build it for them, because it won’t build itself.