The Disappearing Junior
The grunt work AI just took off your plate was the thing that quietly turned juniors into seniors.
Walk into almost any knowledge-work team right now and you will find a small, quiet change that nobody planned and almost nobody is worried about yet. The grunt work has gone. The first-draft memo, the boilerplate code, the initial research pass, the tedious cleanup, all the small unglamorous tasks that used to land on the newest person's desk, are increasingly done by AI before a junior ever touches them. Everyone agrees this is a good thing. The junior is freed from drudgery. The work moves faster. What is there to worry about. Here is what there is to worry about. That drudgery was not a tax on the junior. It was their education.
We mistook the training for waste
Think about how anyone actually became good at this kind of work. Not through a course, and not in a week. They became good by doing a thousand small, boring, low-stakes things. The junior who wrote the dull first draft learned, draft by draft, what made writing good. The one who fixed the trivial bugs built, bug by bug, an instinct for where bugs hide. The tedium was the gym.
The grunt work was never the cost of training juniors. It was the training.
Boring. Beneath them. A tax on the new hire. Obvious drudgery, ripe for automating away.
A thousand small reps. Building the instinct. The apprenticeship in disguise. The thing that quietly turned a junior into a senior.
The paradox nobody mentions
So we have arranged something genuinely strange. AI makes a junior more productive and less trained at the same time. The output looks better than any beginner could produce alone, which makes everyone feel the arrangement is working. But the thousand small reps that used to build judgment are now performed by the machine, and judgment, as anyone senior will tell you, is the whole job.
The junior is shipping senior-looking work while missing the experience that made it senior. The output went up. The learning went down. Only one of those is visible this quarter.
The ladder with the bottom rungs removed
Every senior person you respect was once a junior who did the boring things. That is not sentiment, it is the mechanism. Expertise is accumulated, rep by rep, mostly through work that felt beneath the person doing it at the time. Now picture an organisation that has automated away exactly those reps.
Remove the bottom rungs and you do not get a taller ladder. You get a generation that can start, but never climbed.
The bill does not arrive now. It arrives in about ten years, when an organisation looks around for the seniors it forgot to grow, and finds a generation fluent in operating the tool and strangely thin on the judgment the tool cannot supply.
What to actually do about it
The answer is not to make juniors do pointless busywork out of nostalgia. It is to be honest that the busywork was doing something, and to replace that something on purpose.
Have juniors check and critique the AI's work, not just accept it. Evaluating a thousand drafts builds judgment almost as well as writing them did. Make them reconstruct why the machine's answer is right or wrong. The reps still exist. They just moved to the gap between what the tool produces and what good requires.
"But maybe we just need fewer juniors now"
Here is the objection that sounds like clear-eyed economics. If the machine does the entry-level work, then the entry-level role was overstaffed to begin with, and the market is simply correcting. Fewer juniors, more output, lower cost. Every efficiency in history has thinned some rung of some ladder, and the world kept turning. Maybe seniority finds a new on-ramp the way it always has.
Maybe. But notice the shape of the gamble. Look at how law firms minted partners: armies of first-year associates ground through document review and due diligence, dull work that quietly taught them how deals actually fit together. AI now eats most of that review. Cut the associates and the firm's costs drop this year, while its supply of partners collapses a decade from now, long after the executive who made the call has moved on. That is the trap. Cutting juniors is a real saving today and an invisible debt tomorrow, and the person who books the saving is never the one who pays the debt. Fewer juniors is a defensible call. Pretending it has no cost is not.
The whole point
The work was never the point. The learning hidden inside the work was. Lose sight of that, and you automate away your own future seniors without ever deciding to.
I write about AI, data, and learning regularly at pinaldave.com, and I have been teaching this hands-on in my AI workshops.