'If organizations focus only on short-term efficiency... they risk hollowing out the next generation of technical leaders': Microsoft execs say senior workers must mentor juniors to fix AI mistakes

Companies risk future skills shortages if they stop hiring junior developers today, Microsoft execs sayAI promises productivity boosts, but we need humans to manage agentsHuman-AI collaboration is more important than volume of code

Microsoft Azure CTO Mark Russinovich and Developer Community VP Scott Hanselman have argued senior engineers must actively mentor junior workers to avoid future skills shortages, suggesting AI coding agents are affecting younger and newer workers disproportionately.

In a research paper, the two executives outline how AI coding assistants can boost senior engineer productivity.

However for early-in-career workers, AI actually slows them down, causing them to guide, check and carefully integrate AI-generated code with their own work.

AI helps coding now, but it could wipe away future skills

In the paper, the two authors present some common issues with AI coding assistants, including the introduction of bugs, duplicating code or writing code which passes certain tests but fails more generally.

Though these are totally legitimate issues that are mirrored across multiple studies and in practice, it's the effects on human workers (and particularly younger generations) that the Microsoft execs worry most about.

At the moment, companies are hiring fewer developers in response to a rise in AI usage. But this means future generations will not be so well-equipped with coding and AI management skills.

"f organizations focus only on short-term efficiency – hiring those who can already direct AI – they risk hollowing out the next generation of technical leaders," the conclude.

Although smaller companies with limited resources might struggle not to fall into the pitfalls of AI's short-term promises, the two researchers and Microsoft execs urge larger organizations to continue hiring early-in-career developers.

"The future of software engineering will be defined not by the volume of code AI can generate but by how effectively humans learn, reason, and mature alongside these systems," they add, indicating that while AI isn't going anywhere, neither are human workers.