A new guide shows how to develop AI agents with the Agent Development Kit (ADK) that run reliably over weeks, using structured persistent storage instead of stateless chatbots to work continuously on multi-step processes like hiring without losing context.
Google’s ADK enables AI agents to dynamically load specialized domain expertise and independently generate new expertise areas without overwhelming system prompts.
Google demonstrates how to build long-lived AI agents with the Agent Development Kit that work for weeks, pause, and retain their memory – without the context loss of traditional chatbots.
ADK’s SkillToolset enables AI agents to dynamically load domain-specific knowledge at runtime; progressive disclosure saves tokens and integrates context information strategically. A developer guide presents four practical application patterns.
Gemini CLI now supports specialized subagents that handle complex tasks in parallel, each with its own tools and context window, making the main session faster and more efficient.
At the Google Cloud AI Agent Bake-Off, developers learned that autonomous AI agents in production require systematic engineering discipline, not just good prompt engineering; prototypes and production are different worlds.
At the Google Cloud AI Agent Bake-Off, developers learned under extreme pressure that production-ready AI agents require precise system architecture rather than just good prompts—disciplined agentic engineering is essential to move from prototype to robust application.
Google is developing an AI-powered mouse pointer designed to understand what the user intuitively wants, replacing complex commands with simple pointing and speaking through four interaction principles that automatically capture context.
Google DeepMind introduces Gemma 4 – high-performance open-source models in four sizes (2B to 31B parameters), with the 31B model ranking 3rd globally, outperforming models 20x its size and running efficiently on edge devices to professional GPUs.
Gemini Robotics-ER 1.6 enhances spatial understanding and enables robots to read complex instruments for the first time through specialized embodied reasoning training.
Google DeepMind announces strategic partnership with South Korea: A new AI Campus in Seoul will connect leading research institutions with Google’s AI experts to accelerate breakthroughs in life sciences, climate research and other areas.
