Google AI Announcements

Google announced the launch of Genkit Agents, a new feature of the Genkit open-source framework that simplifies building full-stack, AI-powered applications with conversational capabilities. Genkit Agents provides a foundation for developers to create multi-agent applications with support for various programming languages.

Google released a new extension for VS Code, integrating Google Cloud Workbench Notebooks into the popular development environment. This allows data scientists and developers to access scalable cloud infrastructure directly from VS Code, streamlining the machine learning lifecycle.

Google released ADK 2.0, an updated version of its AI development kit, which adds a structured workflow runtime and task-collaboration model. This new release aims to bridge the gap between the flexibility of AI agents and the reliability of deterministic execution logic. ADK 2.0 allows developers to build AI applications with more control and predictability.

Google introduced a new skill for its coding agent that helps developers improve the quality of their agents by iterating through a flywheel of build, test, ship, monitor, and refine stages. The skill uses AutoRaters developed with Google DeepMind to evaluate and improve agent performance. This launch aims to address the challenges of ensuring agent quality in production environments.

Google has released ADK for Go 2.0, which allows developers to build reliable multi-agent applications with a graph-based workflow engine, human-in-the-loop, and dynamic orchestration. This new version aims to simplify the development of complex agent workflows.