Google Enhances Gemini Code Assist with New Agentic Capabilities

Google has introduced new 'agentic' capabilities to its Gemini Code Assist, allowing AI agents to perform complex programming tasks across various coding environments. During its Cloud Next conference, Google announced that these agents can now create applications from product specifications in Google Docs, perform code transformations, and more.

The upgraded Gemini Code Assist is now available in Android Studio and other integrated development environments (IDEs). These agents can generate work plans, report progress, and handle tasks such as implementing new app features, executing code reviews, and generating unit tests and documentation. The enhancements are seen as a response to competitive pressures from other AI coding assistants like GitHub Copilot and Cognition Labs' Devin.

Despite these advancements, Google advises users to review the code generated by Gemini Code Assist, as AI-generated code can sometimes introduce security vulnerabilities and bugs. The tool is available in various editions, including a no-cost version for individuals and more comprehensive options for businesses, as detailed on the Google Code Assist website.