Who will train the next generation of developers?
Candidates are pasting AI-generated answers into interviews. They sound perfect until you ask them to explain their reasoning. Then it falls apart.
In response to AI coding tools infiltrating engineering teams, companies have cut hiring to prioritise seniors and staff engineers, the ones who already know what good looks like and can evaluate AI outputs.
Great news if you're already senior. Catastrophic if you're trying to break in.
Junior developers don't know what "good" looks like yet. They can't evaluate AI outputs because they haven't built the mental models that come from years of actual work. And now they won't get the chance to learn. Entry-level positions are vanishing because companies ask: "Why invest in training when we can hire experienced engineers who are productive from day one?"
It's the capitalist calculus: optimise for now, worry about later when later arrives.
But here's the problem, every senior engineer today was once a junior who needed mentorship and made mistakes. The industry has always depended on companies investing in the next generation, even when the payoff wasn't immediate. That implicit contract is breaking.
In five to ten years, who will be the senior engineers if no one is training juniors today?
I get it. Not every company has the resources for extensive training programs right now. Economic pressures are real. But we're collectively creating a crisis while solving for the quarter.
So what happens next?
Juniors need to find other paths: open source contributions, side projects, mentorship outside traditional employment. Build in public. Make your learning visible. Some have done this already. It might become more necessary now.
Companies with runway should see this as competitive advantage. Invest in juniors while everyone else fights over the scarce pool of seniors.
Senior engineers: mentor where you can. The industry gave you a chance to learn. Pay it forward.
And as an industry, we need new models. Apprenticeships. Shared training initiatives. Ways to spread the cost of developing talent. AI hasn't made junior developers obsolete. It's made them harder to develop and easier to overlook. Those aren't the same thing.
We're at a crossroads. We could keep optimising for the short term and watch the talent pipeline hollow out. Or we could recognise that investing in the next generation, even when it's expensive, is investing in our collective future.
The market will do what markets do. But markets are made of human decisions. The question is: what will we decide?