AI-for-software-engineering survey maps a decade of progress

A new arXiv paper looks back at a decade of AI for software engineering and software engineering for AI. The survey-style framing captures a major shift: AI is now changing how software is built, while software engineering practices are becoming essential to building reliable AI systems.

That convergence is visible in coding agents, evaluation pipelines, model governance, and AI infrastructure. The field is no longer just about applying models to code completion; it now includes lifecycle management for AI systems themselves.

For practitioners, the value of this kind of review is orientation. It helps separate durable engineering problems from short-lived tooling hype.