Google says 75 percent of its new code is now written by AI

Editor's take

Google now utilizes AI to generate three-quarters of its newly written code, with human engineers performing the subsequent review. This significant adoption by a foundational tech giant underscores the rapidly maturing capabilities of AI in software development, moving beyond simple code completion to substantial generation. The impact is felt across the developer workforce, potentially shifting roles towards oversight and complex problem-solving rather than routine coding, and highlights Google's commitment to leveraging its internal AI advancements, like Gemini, to boost productivity.

This development prompts scrutiny on the efficiency and security of AI-generated code at scale. The effectiveness of human review processes in catching subtle bugs or security vulnerabilities introduced by AI will be crucial. Future attention should focus on the long-term trends in developer productivity metrics, the evolution of AI coding tools beyond generation to more sophisticated design and architecture, and whether other major tech firms, such as Microsoft with its GitHub Copilot integration, will follow suit with similar adoption rates.