Why AI tools make some teams slower (articles.zimetic.com)

In a recent installment of the "Beyond the Coding Assistant" series, researchers explored the paradox of AI tools in software engineering, revealing that while individual developers may report productivity gains from AI adoption, teams collectively experience decreased delivery throughput and stability. The 2024 DORA Accelerate State of DevOps Report found a 1.5% decline in delivery throughput and a 7.2% drop in delivery stability correlating with a 25% increase in AI tool usage. This phenomenon, termed the "DORA anomaly," stems from AI accelerating certain coding tasks without corresponding adjustments in team workflows and processes. Significantly, this imbalance can lead to bottlenecks in code review and deployment processes as teams face increased workloads without improved coordination mechanisms. Anchor practices, such as architectural oversight, fast feedback loops, and clear communication protocols, are crucial for teams to leverage AI effectively. Without these foundational skills, teams may struggle, exacerbating stress and reducing overall productivity despite increased code output. The key takeaway for the AI/ML community is that to reap the full benefits of AI tools, organizations must adapt their development practices to integrate AI seamlessly into the entire workflow, ensuring that acceleration at one stage does not compromise overall efficiency.