AI Changes Software Engineering

In an riveting conversation titled “How AI will Change Software Engineering,” hosted by The Pragmatic Engineer on November 19, 2025, the venerable Martin Fowler, Chief Scientist at Thoughtworks and renowned author, lends his profound insights on the evolving landscape of software engineering. As we step from the deterministic into the non-deterministic world of AI, the transformation echoes the monumental shift from assembly language to Fortran. Fowler believes this new era demands rigorous testing of large language models (LLMs) and emphasizes the ongoing relevance of refactoring. With the burgeoning capacity of AI models, the emphasis is on rapid prototyping and vibe coding, although caution dictates these remain in the realm of exploratory tools. The conversation segues into Fowler’s reminiscences about the Agile Manifesto’s creation, unveiling the timeless value of agile principles amid the shifting technological landscape. As Fowler reflects, the essence of great engineering—communication, continual learning, and adaptability—remains unchanged despite technological revolutions. He highlights that AI, while indispensable, must be integrated with traditional coding philosophies to optimize efficiency and maintain control. As I process these insights, Fowler’s conviction in blending AI’s capabilities with deterministic structures offers a balanced vision for the future of software development. This aligns with current discourse on AI’s immense potential and the caution needed in its unchecked deployment. Through Fowler’s lens, one sees a future that honors both innovation and the steadfast principles of engineering, offering a compelling roadmap amidst digital transformation.