What happens when you stop hiring juniors

I studied web development. If I were starting that course now, I’m not sure what I’d be walking into.

The entry-level work that used to train people — the ticket that teaches you how a codebase actually fits together, the first-draft report nobody reads but everyone learns from writing, the data formatting task that makes you understand what clean data looks like — most of that is handled by AI tools now. In seconds, not days.

So companies are making what feels like a rational call: stop hiring juniors. The work they’d do is already covered. One less salary. One less person to manage. Clean saving.

I get it. I’ve had the same thought myself. But I’ve also spent long enough inside teams to know that junior roles were never just about the output. They were infrastructure. And when you pull infrastructure out because you’ve stopped noticing what it does, things start to break in ways that are hard to trace back.

What entry-level work actually was

The tasks juniors did weren’t just cheap labour dressed up as development. They were the mechanism through which people learned how a business actually works.

Writing that first status report taught you what your team was building and why. Formatting data taught you what “good” looked like before anyone explained it. Fixing a small bug in a large codebase taught you how systems connect — not from a diagram, but from the experience of breaking something and having to understand why.

There’s another side to this that’s easy to miss. When a junior asks a question, the senior has to answer it. That forces the experienced person to articulate something they’ve been doing on autopilot for years. Sometimes they realise their reasoning doesn’t hold up anymore. Sometimes they catch an assumption nobody’s tested since the team was half its current size.

That friction was productive. Not comfortable — productive.

Michael Polanyi wrote about tacit knowledge in the 1960s — the idea that we know more than we can tell. The way most organisations passed that knowledge along was through proximity. Juniors sitting near seniors. Asking things. Watching how decisions got made. Absorbing context that never made it into documentation.

Remove the junior and you don’t just lose the output. You lose the transmission mechanism.

What AI has done to the entry point

Code generation tools handle the kind of tasks that used to fill a junior developer’s first six months. First-draft content gets produced in minutes. Basic research, data cleaning, formatting — faster by machine now.

Dropbox cut roughly 16% of its workforce in 2023, citing AI’s ability to handle work previously done by people. Large tech companies ran significant layoffs through 2023 and 2024, and multiple analyses noted that junior and mid-level roles were hit harder than senior ones. By early 2025, the World Economic Forum was reporting a contraction in roles requiring less than two years’ experience, with 40% of employers expecting to reduce headcount where AI can automate tasks.

None of this is doom. It’s just the reality of what’s happened to the bottom rung of the ladder.

The problem is that the bottom rung wasn’t just where you stood. It’s where you learned to climb.

The mentoring collapse

Here’s what happens when you remove the people being mentored: the people doing the mentoring stop mentoring.

That sounds circular. It isn’t. Mentoring is a habit. It’s maintained by practice, not by policy. When there’s a junior on the team, seniors explain things as a matter of course. They narrate decisions. They pair on problems. They review work and leave comments that teach, not just correct.

Take the junior away and none of that happens. Not because the senior person has become selfish — but because there’s no trigger for it anymore. The knowledge stays locked in their head. The craft they’ve built over fifteen years becomes a private asset instead of a shared one.

Remove the entry point, and the pipeline doesn’t just slow. It breaks. And it breaks slowly enough that nobody notices until the experienced people start to leave and there’s no one behind them.

The culture cost

This is the argument that gets dismissed as soft. It isn’t.

Amy Edmondson’s work on psychological safety — later confirmed by Google’s Project Aristotle — shows that the best-performing teams aren’t the ones with the most talented individuals. They’re the ones where people feel safe to ask questions, admit mistakes, and challenge assumptions.

Juniors contribute to this whether anyone plans it or not. They ask questions that senior people have stopped asking. They haven’t yet learned which topics are politically sensitive or which processes are “just how we do things.” They bring a kind of constructive naivety that forces teams to re-examine habits.

When a team is composed entirely of experienced people, something shifts. The shared assumptions get thicker. The willingness to question norms gets thinner. Everyone knows the same things, so nobody challenges them. The team becomes more capable on paper but less adaptable in practice.

I worked with a delivery team a while back — six people, all experienced, all good at their jobs. They’d been together long enough to develop real shorthand. Decisions moved fast. But when I joined, it took me a week to work out that nobody was documenting handoffs anymore because “everyone just knows.” When I asked why a particular step worked the way it did, three different people gave me three different answers — and none of them matched what was actually happening. They’d drifted so far inside their own norms that they’d lost the ability to see them. There was no clarity, just a shared assumption that there was.

One new person asking questions would have caught that years earlier.

The pipeline problem — industry scale

Everything above is about individual companies. The bigger picture is worse.

If enough companies stop hiring juniors at the same time — and that’s what’s happening — the next generation of experienced professionals doesn’t develop. Not at any single company. Across the whole sector.

Think about the maths. If a significant share of tech and digital businesses reduce or eliminate entry-level hiring between 2024 and 2027, and it takes roughly five to seven years for a junior to develop into a confident mid-level professional, then by the early 2030s there’s a generation-sized hole in the talent market.

The experienced people who are carrying things now will retire, burn out, or move on. And the people who should’ve been ready to step up won’t exist — because nobody invested in developing them.

The companies making these decisions now aren’t thinking about 2032. They’re thinking about this quarter’s headcount. That’s rational in the short term. But the collective effect of everyone being rational in the short term is an irrational outcome for the industry.

What to actually think about

I’m not going to argue that every company should keep hiring juniors regardless of context. That’s not honest. If the entry-level work genuinely doesn’t exist anymore, hiring someone to do it is a waste of their time and yours.

But “the work doesn’t exist” is different from “we haven’t thought about what we lose.”

The first question is where your next seniors come from. If you’re not developing them internally, you’re betting that someone else will — and that bet only works if enough people take the other side. Right now, everyone’s making the same call. Alongside that sits the knowledge transfer problem: seniors who aren’t teaching anyone aren’t articulating what they know. That knowledge becomes fragile. When they leave, it walks out with them. No handover document captures fifteen years of context about how your clients think and what breaks under pressure.

The more interesting question is whether the role can be redesigned rather than eliminated. Some companies are already doing this. The entry-level work has changed, so the role changes too. Instead of writing boilerplate code, a junior developer might spend their time reviewing AI-generated output, learning system architecture, or building the judgment that tools can’t provide. The tasks are different. The development is still happening.

There’s a version of this where AI tools actually make junior development faster — if you design for it deliberately. A junior paired with good tooling and a structured mentoring programme could build in two years what used to take four. That’s genuinely interesting. But it requires intent. It requires someone to think about what they’re trying to develop, not just hand over a Copilot licence and leave them to it. The good boring work of building a pipeline doesn’t happen by accident.

The trap is treating this as a binary: hire juniors the old way, or don’t hire them at all. The honest answer is somewhere else. It’s about being deliberate — knowing what you’re giving up and having a plan for how you replace it.

The bill arrives later

Every decision to cut junior hiring makes sense on its own. The salary saving is real. The productivity argument is sound. The AI tools genuinely do handle the work.

But the costs are real too. They’re just deferred. Mentoring atrophies. Knowledge stops transferring. Teams become less adaptable. And the pipeline that produces your future experienced people gets thinner every year nobody feeds it.

The companies that will be in the strongest position in five to ten years aren’t the ones that saved the most on headcount in 2025. They’re the ones that asked: what are we actually losing here — and who’s going to be in the room when the people who know everything decide to leave?

That’s not sentiment. That’s a risk most teams aren’t tracking. And by the time it becomes obvious, the people who could’ve fixed it are long gone.