The article argues that software engineering cannot be fully resolved through code generation with AI, citing challenges in translating natural demands into formal code. It discusses the unique position of software engineering between practical real-world problems and abstract mathematics, highlighting human intention as a key factor in development processes. The author reflects on the inefficiencies of AI in understanding and solving ill-defined software requirements, emphasizing the need for human oversight and iteration in software design.