AI helps strong dev teams and hurts weak ones, according to Google's 2025 DORA report (www.zdnet.com)

Google Cloud’s 2025 DORA report—based on a 5,000-developer survey and 100+ hours of interviews—finds AI now mainstream in software development (90–95% usage, median two hours/day). Most devs (80%) report higher productivity, but only 59% see better code quality and 70% trust AI outputs. Crucially, DORA frames AI as an amplifier: it magnifies the strengths of well-run teams and the dysfunctions of weak ones. Top-performing teams (roughly the top 30%) combine speed and stability, producing fast, high-quality output, while disorganized teams can suffer more instability when they adopt AI. For practitioners and the AI/ML community the technical takeaway is organizational, not purely tool-focused. Seven practices drive positive AI impact—clear AI policy, robust data ecosystems, accessible internal data, version control, small-batch changes, user focus, and shared quality platforms. Two decisive enablers are strong internal platform engineering (now adopted by ~90% of orgs) and value stream management (VSM), which maps flow from idea to delivery and "dramatically" amplifies AI benefits. The report’s implication: treat AI adoption as systems engineering—invest in platforms, data, VSM and engineering fundamentals, enforce QA/version control, and AI will scale real productivity rather than merely shifting risk.