Build long-running agents with Google’s Agentic Stack | The Agent Factory

Google Cloud Tech · Beginner ·🤖 AI Agents & Automation ·1mo ago

Key Takeaways

Builds long-running agents with Google's Agentic Stack, including AI systems that persist, sleep, self-correct, and execute complex workflows

Original Description

Google’s Agent Studio → https://goo.gle/43BUUjA ADK 2.0 → https://goo.gle/3RR3VTm Read more about Long-running agents → https://goo.gle/49xvNC1 In this episode, we explore "long-running agents": AI systems that persist, sleep, self-correct, and execute complex workflows over hours, days, or even weeks. Join host Smitha Kolan and Addy Osmani, Director of Google Cloud AI, as they break down the three non-negotiable rules for production-grade agents: durable state checkpoints, event-driven dormancy (true sleep), and separated evaluation. You'll see three live demos of long-running agents: a multi-day HR onboarding coordinator that delegates tasks and waits for human signatures; a custom operating system built autonomously; and a complex, highly optimized 3D video store generated entirely in Blender by an AI agent over several days. Whether you're a cloud architect, an AI engineer, or a technical founder, you'll walk away with a blueprint for building long-running agents that actually finish the job. You'll learn how to implement the three-agent setup (planner, generator, evaluator). Chapters: 0:00 - Introduction to Long-Running Agents & ADK 2.0 1:19 - Why AI Agents Need to Sleep (State & Dormancy) 4:49 - The 3-Agent Setup for Multi-Day Workflows 5:16 - Demo: Long-Running HR Onboarding Agent 11:56 - Demo: Building an OS 16:08 - Demo: Generating a 3D Video Store in Blender 22:47 - "Beyond Vibe Coding" & Developer Pushback 26:07 - Avoiding Cognitive Debt and Cognitive Surrender 28:17 - What is Agent Engine Optimization (AEO)? 31:06 - Addy Osmani’s Favourite Developer Tool More resources: 👔 Long Running AI Agent Team for New Hire Onboarding → https://goo.gle/4fgJmJS Agent’s CLI → https://goo.gle/4fKemC8 Agentic Releases at I/O 2025 → https://goo.gle/4dWoUf2 Speakers: Smitha Kolan X → https://goo.gle/4uEGz1P LinkedIn → https://goo.gle/4353lDX Addy Osmani X → https://goo.gle/3PMeLcO LinkedIn → https://goo.gle/4wZaNyc Watch more of The Agent Factory →
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related Reads

📰
A Regression-Test Workflow for Consistent AI Characters Across Chat, Images, and Video
Learn to create a regression-test workflow for consistent AI characters across chat, images, and video to ensure quality and coherence
Dev.to · Joe Wu
📰
issues are prompts, PRs are context, github is the queue
GitHub can be viewed as a queue where issues are prompts and PRs are context, enabling automation with agents
Dev.to AI
📰
Automating Quarterly Data Aggregation: Connecting Portfolios, Performance, and Benchmarks
Automate quarterly data aggregation for independent financial advisors using AI, connecting portfolios, performance, and benchmarks
Dev.to AI
📰
Re: ACP already exists — you're right, that's why we built ATC instead
Learn how to differentiate between existing protocols and your own when building new AI agent technologies, and why renaming can be a crucial step in product development
Dev.to AI

Chapters (10)

Introduction to Long-Running Agents & ADK 2.0
1:19 Why AI Agents Need to Sleep (State & Dormancy)
4:49 The 3-Agent Setup for Multi-Day Workflows
5:16 Demo: Long-Running HR Onboarding Agent
11:56 Demo: Building an OS
16:08 Demo: Generating a 3D Video Store in Blender
22:47 "Beyond Vibe Coding" & Developer Pushback
26:07 Avoiding Cognitive Debt and Cognitive Surrender
28:17 What is Agent Engine Optimization (AEO)?
31:06 Addy Osmani’s Favourite Developer Tool
Up next
I built a custom Hermes plugin. #HermesAgent #Claudecode #openaicodex #openclaw #nousresearch
Tech Friend AJ
Watch →