The future of AI software is here – it's just not evenly distributed (thelastsoftwareengineer.substack.com)

AI is revolutionizing software development, enabling faster coding than ever before. Despite this advancement, the growth in new features for popular software tools remains disappointingly stagnant. The disparity between code production and feature deployment raises questions about how AI's contributions are being realized in large software projects. Anecdotes suggest that while AI-generated software is being released (such as new operating systems and tools), established products have not seen a corresponding boom in innovative features, which is supported by research highlighting only marginal increases in feature deployment metrics. This phenomenon reveals the complexity of software engineering beyond mere coding. As articulated in classic literature on software productivity, such as Fred Brooks' "No Silver Bullet," the inherent complexity of determining what software should do—known as essential complexity—poses greater challenges than the tools and practices that can streamline coding (accidental complexity). As AI accelerates the coding process, project coordination and feature decision-making still require thoughtful collaboration among teams, resembling the intricate communication needed in construction projects. Thus, while AI enhances efficiency, it does not eliminate the crucial discourse necessary for navigating complex software development, highlighting the uneven distribution of AI advancements across the industry.