Over the last few years, countless predictions have emerged about the supposed end of the junior developer in tech. The logic seemed straightforward: if an AI can write code, explain concepts, and solve problems in seconds, why hire professionals who are just getting started?
The reality we're seeing in the industry looks quite different. AI is transforming how engineering teams work, but that doesn't mean junior talent has lost its relevance. It means the skills a junior developer builds and the way they learn are evolving.
And that opens a much more interesting question: how do we train the next generation of engineers in a world where answers are just one prompt away?
What research says about AI and technical learning
A recent study published on arXiv by researchers from Carnegie Mellon University (CMU), UCLA, the University of Oxford, and MIT analyzed more than 1,200 interactions between people and AI assistants to understand how these tools impact learning and problem-solving.
The results showed that people who used AI performed better in the short term, but when the assistance was removed, they tended to perform worse and abandon tasks more quickly.
At first glance, this might read as a critique of AI use. But a closer look at the results reveals an important nuance.
The negative effects were concentrated mainly among those who used AI to get direct answers. In contrast, those who used it to ask for explanations, clarifications, or hints showed considerably better outcomes.
The conclusion doesn't seem to be that AI hurts learning. It seems that how we use AI determines what we actually learn from it.
AI as an accelerator for junior developers, not a replacement
This view aligns with what we believe at Howdy and with what our partners believe too. As Arik Yelovitch, CTO and co-founder of Adaptive Insurance, put it, AI isn't replacing the junior developer; it's acting as an accelerator for their capabilities.
A professional who once needed days to research a technology can now grasp its fundamentals in hours. A developer who used to get stuck on a complex bug can now get context, alternatives, and explanations instantly.
More than accelerating learning on its own, AI accelerates access to knowledge and significantly lowers the barriers to experimenting, building, and validating ideas.
But accelerating isn't the same as replacing. Because there are still deeply human capabilities that no tool can develop for us:
- Technical judgment
- Decision-making
- Architectural thinking
- Communication
- Collaboration
- Business understanding
- Leadership
And those are precisely the skills that end up distinguishing the engineers who grow within an organization.
At the same time, it would be naive to claim nothing is changing. Many tasks traditionally associated with junior profiles, like generating boilerplate, writing basic documentation, or handling repetitive implementations, are already being accelerated by AI tools.
What's disappearing isn't the need to develop junior talent. What's changing is how that talent creates value and builds experience.
How to develop senior engineers from junior talent in the AI era
Maybe the most important question isn't what AI can do today. The question is: who is developing the seniors of tomorrow?
Every generation of professionals is built on thousands of decisions, mistakes, learnings, and conversations with more experienced people.
If AI solves every problem automatically, there's a real risk of accelerating execution without developing the judgment needed to make good decisions when there's no obvious answer.
That's why interesting initiatives are starting to emerge, focused not just on AI-generated code, but on the reasoning behind technical decisions. One of them is Seniorify.dev, a platform that challenges developers to clarify their intent and reasoning before AI writes a single line of code.
The idea is simple but powerful: push developers to justify architectural decisions, evaluate trade-offs, and defend their technical choices. Because the goal isn't to stop AI from writing code. It's to make sure humans keep developing judgment.
What it means to be an AI-Enabled Engineer
At Howdy, an AI-Enabled Engineer isn't someone who delegates their thinking to a tool. It's someone who uses AI to accelerate research, explore alternatives, validate hypotheses, and learn faster, while maintaining full ownership of technical decisions.
The difference is subtle, but it matters.
Passive AI use looks for answers. Active AI use looks for understanding.
For example, instead of simply asking for an implementation, a junior developer can use AI to explore different solutions, understand their trade-offs, identify edge cases, or challenge their own assumptions.
In other words: AI can write code. The engineer is still responsible for understanding why that code exists.
Why the human factor matters more than ever
There's a line from the study that perfectly captures this challenge: "A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results."
A mentor doesn't just answer questions. They help someone learn, accompany their progress, and prioritize the other person's growth over the immediate result.
That's exactly what organizations need to preserve.
AI can offer quick answers, generate alternatives, and speed up execution. But it isn't designed to ask whether the person on the other side of the screen is developing judgment, confidence, or autonomy. Its goal is to solve the current problem, not necessarily to prepare someone to solve the next one.
So as AI becomes more capable, the value of human mentorship doesn't decrease. It increases.
The speed AI brings is incredibly valuable, but turning that speed into sustainable professional growth still requires human accompaniment.
How we live this at Howdy
At Howdy, we believe the future doesn't belong to engineers who work without AI. Nor to those who fully delegate their work to it. We believe in AI-Enabled Engineers: professionals who use AI intentionally to amplify their capabilities while continuing to develop judgment, autonomy, and critical thinking.
That's why, alongside actively adopting AI tools within our teams, we maintain something that remains irreplaceable: human accompaniment.
Our Engineering Mentors work alongside professionals to guide their technical and career growth, helping them develop skills no tool can build on its own. Their role isn't simply to answer technical questions. It's to help each professional understand the context behind decisions, strengthen their judgment, and build the autonomy to take on increasingly complex challenges.
And this isn't just the responsibility of those starting out in their careers. It's also the responsibility of seniors, tech leads, and engineering managers to build environments where AI accelerates learning rather than replacing it.
Because the goal isn't simply to ship faster. It's to develop professionals who can make better decisions. And we believe that combination, the speed of AI and human accompaniment, is what will allow the next generation of great engineers to grow.
More AI, more mentorship
The conversation about AI and tech talent is often framed as a false dichotomy: humans or artificial intelligence. The reality is probably far more interesting.
AI is enabling junior developers to be productive, but turning that productivity into sustainable growth still requires context, experience, judgment, and mentorship.
The organizations that make the most of this new era won't be the ones that replace people with AI. They'll be the ones who combine both to develop the next generation of engineers.
Because if AI can accelerate learning, the role of mentors isn't less important. It's more important than ever.
