Engineers use AI to write more code. Where's the AI to fix it? (www.xurrent.com)

A recent discussion highlighted a critical gap in the AI tools available to engineers, specifically in addressing the burgeoning number of software bugs produced by rapid code generation facilitated by AI. While tools have accelerated code writing—enabling engineers to work three times faster, according to Google's 2025 DORA report—this surge has led to increased delivery instability, with more code shipping but also breaking more frequently. As a result, site reliability engineers (SREs) face greater challenges in managing incidents, often spending more time on "toil" rather than seeing a reduction as expected. To effectively bridge this gap, industry experts emphasize the need for AI tools designed specifically for fixing production issues. Such tools must understand the unique architecture of a company's systems, display confidence levels in their analyses, and facilitate decision-making by providing relevant context without attempting to replace human judgment. The introduction of AI tools like Sera, which collaborates with SREs by investigating incident causes and managing stakeholder communications, aims to empower engineers rather than replace them. This approach not only enhances productivity but also allows engineers to focus on designing more robust systems while minimizing the chaotic response to incidents.