Podcast: How Will We Train Developers If AI Does the Routine Work: A Conversation with Scott Hanselman

In this podcast, Michael Stiefel interviewed Scott Hanselman about the future of training developers in light of AI taking over routine work. Hanselman proposes a preceptorship model akin to nursing, where a dedicated group of preceptors would solely train new engineers. These preceptors would be judged based on their ability to train individuals, rather than the volume of code they produce.

AI agents can generate software features but struggle with software architecture, which requires a broader perspective and human oversight. Engineers must continue to supervise AI tools and maintain a human-centric approach to development. The conversation also delved into the appropriate use of agentic AI, emphasizing the importance of long-term investments in human capital and mentorship to foster future senior talent.

Furthermore, Hanselman highlighted the sociological implications of remote work, social media, and AI, stressing that these factors contribute to developers' isolation. To counteract this isolation, he stresses the need for intentional human connection and face-to-face interactions to promote healthy growth among the next generation of developers.