AI can generate code, seniors can review it, and entry-level salaries aren’t trivial – the argument for cutting junior engineers sounds reasonable on a spreadsheet. Christine Miao’s counter-argument starts at a different level entirely.
We’re in a growing industry where companies need to be able to continue and carry forward. In order to do that, they need to hire juniors. Foundationally, it’s about survival.
Survival, AI adoption and, honestly, good vibes #
I had a chat with Miao, Founder and Researcher at Technical Accounting, after her talk at the CTO Craft Conference in Toronto. Her talk was followed by a presentation, one of the rare ones not (completely) generated with AI. It was fun and quirky, a bit like junior developers, who bring fresh perspective, followed by a bit of charming chaos.
They are essential for the companies’ and industry’s survival, Miao is certain, but that part is just the floor. Miao points to something harder to quantify: the organizational effect of having people experience a steep learning curve in public for the first time. Senior engineers who mentor a junior feel accountable in a way they simply don’t with an LLM.
I would certainly very much hope that no one feels responsible to an LLM the way they would feel responsible to a young person who is learning the ropes for the first time.
There is also an adoption argument. Juniors, she notes, have no prior workflows to protect. In organizations struggling to get engineers and other team members to actually use the AI tools they’ve licensed, hiring people without ingrained habits is a legitimate strategy.
The culprit for the current situation? #
I asked Miao who is responsible for the current contraction in entry-level hiring. She traces it up the chain further than most critics do. The culprit, in her view, is investors who conflate cost reduction with value creation:
They can’t tell the difference between cutting costs to maintain EBITDA, versus building a new product and creating a new growth area. In their minds, because the numbers are the same, it all adds up.
That mindset propagates downward into management, and juniors, whose output is hard to measure visibly, become an easy target.
How to mentor junior engineers who use AI? #
The mentorship question is where the conversation gets more complicated. Some senior engineers argue that** AI-assisted development is producing juniors who generate output without understanding it**, making them harder, not easier, to teach. Miao doesn’t dismiss the pattern, but she describes a bimodal distribution. On one end: juniors producing “really fast, really crappy slop” they can’t debug or explain. On the other: organizations where juniors leveled up in roughly three months, which brings down the cost of onboarding significantly.
The difference comes down to two factors. First, a genuine organizational commitment to mentorship. Not as policy language, but as actual practice. Second, engineering foundations that are already in reasonable shape.
If you have a good baseline, AI gets easier, junior engineers become easier. The inverse is also true: legacy codebases, poor documentation, and no mentorship infrastructure don’t just slow junior development, they actively compound each other. Adding AI to that environment and then blaming the junior for the output is a category error.
On the question of what makes a strong junior right now. Miao is direct: it’s not code, but a want to level up, having the aptitude to recognize when a field clicks, and the grit to work through the parts that don’t. The market tightening is, in her view, a reasonable filter. Previously, anyone who could code could find a job; now the bar requires genuine engagement with the craft.
Companies that don’t hire juniors will fall into their own trap #
The longer-term risk is structural. Organizations that stop refreshing their pipelines now will, Miao argues, find themselves trapped in five years: too brittle to adapt, carrying institutional knowledge that was never transferred, and too understaffed to train anyone new.
You hit a point where you are so busy keeping your head above water that you have no time to train anyone else. At that point, frankly, money won’t fix the problem.
For the organizations paying attention, the current environment is actually favorable: junior talent is abundant, under-hired, and, when developed properly, notably loyal. The companies that recognize this as a competitive window rather than a cost problem are likely to look very different from the ones that didn’t, and not in a recoverable way.