What Software Engineers Can Learn from the Aviation Industry (mwalterskirchen.dev)

A recent exploration into "Agentic Engineering" highlights crucial lessons for software engineers from the aviation industry's experience with automation. As AI coding tools become more integrated into development workflows, similar to how autopilots were adopted in aviation, engineers may risk "coding atrophy," where crucial skills diminish due to over-reliance on AI. Research shows a troubling trend: code churn has doubled, refactoring has dropped significantly, and AI-generated code is introducing more security flaws compared to human-written code. The aviation analogy warns that while safety metrics might improve superficially, deeper errors may lurk beneath the surface, calling for awareness among developers regarding automation dependency. The author suggests that to counteract these risks, the software engineering community should adopt training practices from aviation, emphasizing “automation management.” Proposed strategies include scheduling regular manual coding sessions to keep skills agile, confronting complex challenges without AI assistance, and developing realistic training environments akin to flight simulators for engineers. While AI tools promise productivity gains, the narrative stresses the importance of maintaining sharp critical thinking and coding skills, as the long-term health of software engineering will depend on engineers who can adeptly navigate both automated tools and intricate problem-solving scenarios.