The future does not need fewer junior developers. It needs better systems for helping both people and agents become useful, safe, and trusted participants in the work.

AI is changing how software gets written, but it is also revealing something software teams have always struggled with: context is the real curriculum.

Junior developers are early in career. AI agents are early in context. Both can be surprisingly capable, both can be confidently wrong, and both need structured guidance before they can be trusted with production systems.

In this keynote, Scott Hanselman takes a practical, demo-driven look at what AI changes in software development and what it does not. Through live demos, real talk, and lessons from recent work with Mark Russinovich in Communications of the ACM, Scott will explore the gap between producing code and understanding systems. He’ll connect the future of AI-assisted development with the future of human talent pipelines, arguing that preceptorship is not nostalgia for a slower era. It is how complex engineering cultures transfer judgment, context, taste, debugging instincts, and responsibility.