Google Cloud Tech
Google Cloud TechVideo
- Format:
- YouTube channel
- New videos:
- Daily
- Typical length:
- about 15 minutes, not counting Shorts
- Latest video:
About this show
Welcome to the official YouTube channel for Google Cloud’s technical community. Explore our tutorials, see exciting demos, stay up to date with product news, and find new developer tools helping you build what’s next.
Videos come from the channel's public YouTube feed. The artwork is the newest video thumbnail. Last updated . Names, artwork and text belong to their publishers.
Cosmos
Similar by topic and creator
- The PrimeTime
Shares the topic Technology. - Just Have a Think
Shares the topic Technology. - Neso Academy
Shares the topic Technology.
Videos
More tokens spent lead to lower quality code - true or false?
· 45 sec
Does spending more tokens naturally lead to lower-quality code? True or false? YK Sugi, Senior Developer Relations Engineer at Google Cloud, explains why the answer is false, depending on how you spend those tokens. Using tokens only to generate more code with a smaller model can hurt code quality, but they can also improve quality across four workflows: writing tests, running interactive code reviews to understand the codebase, reviewing architecture, and simplifying existing code while keeping every feature intact. #Shorts #AIAgaveveents #AICoding #SoftwareEngineering #GoogleCloud Speaker: YK Sugi Products Mentioned: None
Not rated yetDeep dive into the Looker CLI
· 16 min
The Looker CLI really extends what you can do with Looker and Agents. The tool provides performance reports, continuous development tools, and the entire suite of API endpoints. This discussion has demos and examples of how you can leverage looker CLI to speed up your workflow. * Learn about it here → https://www.youtube.com/watch?v=e50CV4VbgqE 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech Speaker: Chrissie Goodrich Products Mentioned: Looker
Not rated yetWhy build your own AI agent harness?
· 48 sec
▶ Watch next: Harness Engineering Explained: Inside the Stack Behind Antigravity, Claude Code & Cursor https://www.youtube.com/watch?v=F8EZJAm9iO8
Not rated yetGemini 4 Argon demo: Remixing 3D apps & Firebase in Google AI Studio
· 5 min
Try Gemini 4 Argon in Google AI Studio → https://g.dev/ai/studio Explore Firebase Authentication & Cloud Firestore → https://g.dev/firebase/docs How fast can you take existing 3D web apps and remix them into custom, multiplayer experiences using natural language? In this standalone segment from The Google Cloud Show, Google builders Billy Jacobson, Luke Schlangen, and Mollie Pettit jump into Google AI Studio to put Gemini 4 Argon through a live, hands-on remix challenge—turning single-prompt ideas and voice-to-text instructions into working 3D game upgrades and full-stack backend integrations in seconds. Watch as Luke, Mollie, and Billy explore apps from the Gemini 4 Argon Builders Showcase and compete to customize them on the fly. First, they open a 3D chessboard app and use voice-to-text prompts in Google AI Studio to remix the experience side-by-side—swapping out abstract border symbols for traditional chessboard letters and numbers so they can call out moves like "Knight to E5," while transforming the pieces into bright neon colors. Next, they level up a 3D racing game by prompting Gemini 4 Argon to add user authentication and a real-time shared leaderboard. With a single click inside Google AI Studio to enable Firebase and Cloud Firestore, Billy spins up live login and database tracking in moments—a workflow that used to take weeks of manual setup—before racing head-to-head to claim the top spot on the board. Whether you're a full-stack developer, frontend engineer, or builder curious about rapid prototyping, you'll see how easy it is to fork, customize, and connect cloud backends to interactive apps in Google AI Studio without starting from a blank file. Chapters: 0:00 - Exploring the Gemini 4 Argon Builders Showcase 0:20 - Spotting a Missing Feature on the 3D Chessboard 0:48 - Remixing the 3D Chess App with Voice-to-Text Prompts in AI Studio 1:16 - Testing Coordinate Labels (Knight to E5) & Neon Chess Pieces 2:27 - Adding Firebase Auth & a Live Leaderboard to a 3D Racing Game 2:45 - Enabling Firebase & Cloud Firestore in One Click Inside AI Studio 3:13 - Live Multiplayer Race & Testing the Real-Time Leaderboard 4:15 - Why Automated Auth & Database Setup Saves Weeks of Dev Time 4:39 - Recap: How Anyone Can Remix & Share Interactive Apps in AI Studio More resources: Read the Gemini 4 Argon Announcement → https://g.dev/ai/intro-gemini-4-argon Read the Gemini 4 Carbon Announcement → https://g.dev/ai/intro-gemini-4-carbon Explore Google Antigravity Custom Subagents Documentation → https://g.dev/ai/custom-subagents Explore Firebase Authentication & Cloud Firestore → https://g.dev/firebase/docs 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech Speakers: Billy Jacobson, Luke Schlangen, Mollie Pettit Products Mentioned:, Products Mentioned: Gemini 4, Gemini 4 Argon, Google AI Studio, Firebase, Firebase Authentication, Cloud Firestore
Not rated yetWatch me build a full AI app (no code)
· 55 min
You have the specs and the implementation plan. Now it is time to actually build. In this episode of Build It, we hand our phase-gated implementation plan to Google Antigravity and watch it build our smart personal trainer app milestone by milestone, from project setup and the backend server to the workout tracker, AI coach chat, and living memory. This is the real build, bugs and all. Along the way, things break. Instead of hitting enter and hoping Gemini fixes everything, you will see how to troubleshoot like a builder: reading the browser console, testing whether an API key is actually loading, and asking "should this be working?" before asking for a fix. Then we tackle the biggest bug of the episode, where editing one workout rewrote the whole week, by changing how the system works instead of requesting a one-off fix. Chapters: 0:00 - Intro: The Full Build, Bugs and All 0:31 - Recap: From Prototype to Specs to Implementation Plan 1:18 - Milestone 1: Project Setup and Automated Checks 2:55 - Milestone 2: Backend Server and Validation Tests 3:52 - Milestone 3: App Shell and Workout Tracker 5:21 - Moment of Truth: Fixing the First Error With the Browser Console 7:44 - Fixing a Browser Caching Bug 9:41 - Milestone 4: AI Coach Chat and Workout Proposals 14:30 - Troubleshooting the Gemini API Key 18:25 - Testing the Live AI Coach: Modify, Questions, and Voice 25:26 - Adding a Single-Day Workout Card 29:32 - Investigating Inconsistent Workout Edits 35:15 - Finding the Root Cause: Where the Coach Gets Its Context 39:40 - Changing the System: Passing Context From the App 42:10 - Testing Surgical Workout Edits 44:36 - Milestone 5: Progress, Settings, and AI Living Memory 48:02 - Ask "Should It Be Working?" Before Asking for a Fix 53:31 - Wrap-Up and What's Next More resources: Get the Prompts & Artifacts → https://g.dev/ai/build-it-resources Builder Central → https://g.dev/ai/builder-central Google Antigravity → https://g.dev/ai/build-it-with-agy Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleCloud #Gemini #Antigravity #NoCode #AIAgents #BuildIt Speakers: Devin Chasanoff, AI Builder Advocate Products Mentioned: Gemini, Google Antigravity, Firebase, Google Cloud
Not rated yetWhat Is Jev? Fast discriminative models vs. LLMs
· 1 min
📖 Codelab next: Getting started with Discriminative (Jev/DiffusionGemma) models https://g.dev/ai/discriminative-models.
Not rated yetDecide which Looker MCP integration is right for you?
· 4 min
Choosing the right integration path for your Looker AI agents can be tricky. In this video, we break down the key differences between the native Looker-Managed MCP Server and the open-source MCP Toolbox for databases to help you pick the best approach for your workload. Resources: * Use the Looker Managed MCP Server → https://g.dev/cloud/looker-managed-mcp * Get Started with MCP toolbox for databases → https://g.dev/cloud/mcp-toolbox-for-databases-looker Speaker: Chrissie Goodrich Products Mentioned: Looker
Not rated yetComputer use vs. WebMCP: How AI agents use the web
· 1 min
▶ Watch next: Make your website agent ready with WebMCP https://www.youtube.com/watch?v=FARxSG_EY98
Not rated yetHow to compare documents 🍿 Just a moment #02
· 1 min
🔗 Document understanding → https://g.dev/cloud/document-understanding 🔗 Gemini models → https://g.dev/cloud/gemini-models ⚡ How to compare documents 🍿 Just a moment #02 Have you ever had to compare two versions of a complex document? What if one is a digital PDF and the other is a scan, spanning hundreds of pages, with different layouts and pagination? In this episode, see how Gemini's native multimodality spots subtle differences, from tiny typos to $100M contract updates, in a single prompt. 🍵 Grab your tea, coffee, or matcha… It only takes a moment! Prompt template used in the video example: [DOCUMENT_A] … [/DOCUMENT_A] [DOCUMENT_B] … [/DOCUMENT_B] TASK: Perform an exhaustive, character-by-character textual comparison between DOCUMENT_A and DOCUMENT_B. CRITICAL REQUIREMENTS: 1. You must systematically verify every numbered paragraph sequentially from start to finish. 2. You must catch every single difference, including single-character changes, minor typos, spelling corrections (e.g., missing letters in common words), and punctuation changes, in addition to substantive changes (numbers, dates, names). OUTPUT FORMAT: ### Difference #[1, 2, 3…] - Location: [location] - Evidence A: "[short quote from Document A]" with difference in **bold** - Evidence B: "[short quote from Document B]" with difference in **bold** - Explanation: [brief description] 📌 Follow Laurent online: 🔗 LinkedIn → https://goo.gle/laurent-picard-linkedin 🔗 X → https://goo.gle/laurent-picard-x 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleGemini #GenerativeAI #GoogleCloud #DocumentUnderstanding #NativeMultimodality #JustAMoment #TechIn2Minutes Speaker: Laurent Picard Products mentioned: Gemini
Not rated yetWhy 5 tools beat 25 tools for AI agents
· 1 min
📖 Read next: Developer's Guide to Building ADK Agents with Skills (Google Developers Blog) https://developers.googleblog.com/developers-guide-to-building-adk-agents-with-skills/
Not rated yetBuilding a custom evaluation for your agent
· 13 min
Build your custom evaluation with ADK → https://g.dev/cloud/cpm-vid-adk Codelab → https://g.dev/ai/self-evolving Learn how to build, test, and automate custom evaluations for your AI agents using Google's Agent Development Kit (ADK). In this step-by-step tutorial, Annie Wang teaches how to define what "success" means for complex agents, compare using requirement-based checkers, and integrate agent tests directly into CI/CD deployment pipelines using pytest. Chapters: 00:00 - Intro & the 4 pillars of agent evaluation 00:28 - What counts as success? 01:56 - Evaluation approaches: Comparators 03:45 - How to check: Deterministic code vs. LLM as a judge 05:50 - Building a custom metric in Python 08:00 - Configuring ADK eval & validating the evaluator 11:00 - Automating agent testing in CI/CD with Pytest Speaker: Annie Wang Products Mentioned: Agent Development Kit
Not rated yetWhat is native multimodality? 🍿 Just a moment #01
· 1 min
🔗 Gemini models → https://g.dev/cloud/gemini-models ⚡ What is native multimodality? 🍿 Just a moment #01 Did you know there are three tokenizers involved when Gemini processes multimodal inputs? In this episode, we peek under the hood to see how text, vision, and audio tokenizers work together. 🍵 Grab your tea, coffee, or matcha… It only takes a moment! 📌 Follow Laurent online: 🔗 LinkedIn → https://goo.gle/laurent-picard-linkedin 🔗 X → https://goo.gle/laurent-picard-x 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleGemini #GenerativeAI #GoogleCloud #NativeMultimodality #JustAMoment #TechIn2Minutes Speaker: Laurent Picard Products mentioned: Gemini
Not rated yetMultimodal AI agent architecture: Voice, vision, tools & memory
· 6 min
Full Season 1 playlist → https://www.youtube.com/playlist?list=PLc2r9FdYHRlg What actually separates a real-time multimodal AI agent from a chatbot with voice and vision bolted on? In the Season 1 finale of The Omni App, we connect all six layers of a bidirectional live multimodal agent, voice, Google ADK, tool calling, browser control, hybrid memory, and real-time vision, and distill the entire architecture into two core design questions that keep an open live session feeling like one natural conversation. 📺 Watch all of season 1 here: * Trailer → https://youtu.be/YGgErBnx6po * Ep 1 · Voice → https://youtu.be/pFc-HcUgFgY * Ep 2 · ADK → https://youtu.be/yQEKMsCtsmE * Ep 3 · Tools → https://youtu.be/Sdrpso7IKtU * Ep 4 · Browser → https://youtu.be/xDL5DFMo1-Y * Ep 5 · Memory → https://youtu.be/8C4bh59FtsI * Ep 6 · Vision → https://youtu.be/dNGchf4InU0 🧠 Watch along and learn: * Omni App (Perceive → Reason → Express) vs. turn-based request-response chatbots * How all 6 layers work together in one open live loop * Question 1 (Input): Continuous streams (mic, live camera) vs. complete inputs (screenshots, photos, text) * Question 2 (Execution): Blocking vs. non-blocking work—acknowledging fast while moving slow work off the live hot path * Architecture boundaries: Live path vs. background, SQL facts vs. semantic vectors, snapshots vs. frame streams ⏱️ Chapters: 00:00 - Intro: What Is an Omni App & Bidirectional Live Loop? 00:55 - Chatbot vs. Omni App: Request-Response vs. Live Workflow 01:32 - Ep 1 Recap: Real-Time Bidirectional Voice & Barge-In 01:47 - Ep 2 Recap: Structuring the Live Loop with Google ADK 02:05 - Ep 3 Recap: Live Tool Calling Without Breaking the Flow 02:24 - Ep 4 Recap: Voice-Driven Browser Control via Screenshots 02:41 - Ep 5 Recap: Structured Facts vs. Semantic Vector Memory 03:04 - Ep 6 Recap: Real-Time Vision — Snapshots vs. Frame Streams 03:21 - Architecture Question 1: Continuous Stream vs. Complete Input 04:05 - Architecture Question 2: Blocking vs. Non-Blocking Work 04:42 - Core Takeaway: Designing Multimodal Agent Boundaries 05:05 - Season 2 Preview: Deeper Perception & Creative Expression 🔔 Subscribe and stay tuned for Season 2, where our agents go deeper on multimodal understanding and creative generation! More resources: Gemini Live API docs → https://g.dev/cloud/gemini-live Agent Development Kit (ADK) docs → https://g.dev/cloud/adk-docs Google Cloud SQL docs → https://g.dev/cloud/docs-sql AlloyDB docs → https://g.dev/cloud/docs-alloydb Vertex AI Agent Engine Memory Bank → https://g.dev/ai/memory-bank-overview Follow Annie online: LinkedIn → https://goo.gle/annie-linkedin X → https://goo.gle/annie-x Watch more The Omni App → https://g.dev/cloud/modern-ai-app 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #AIAgents #MultimodalAI #SoftwareArchitecture #GeminiLiveAPI #GoogleADK #VoiceAI #GoogleCloud #AgenticAI #LLM #VertexAI Speaker: Annie Wang Products Mentioned: Gemini Live API, Agent Development Kit, Google Gen AI SDK, Google Cloud SQL, AlloyDB, Vertex AI Agent Engine Memory Bank
Not rated yetThe first prompt engineers run when a new LLM drops
· 47 sec
▶ Watch next: Gemini 4 Argon: 10 incredible AI demos https://www.youtube.com/watch?v=gRhM1h10T3U
Not rated yetExplain the A2A protocol without saying 'AI' or 'agent'
· 54 sec
▶ Watch next: Introduction to Agent2Agent (A2A) Protocol https://www.youtube.com/watch?v=Fbr_Solax1w
Not rated yetWhy code is breaking production (Google Engineers Explain)
· 1 hr 2 min
Order the Site Reliability Engineering book here → goo.gle/site-reliability-engineering-book Writing code just got 10x faster with AI which means breaking production just got 10x easier. As AI accelerates the software development lifecycle, how do you keep systems from collapsing under the weight of constant code churn? In this episode, the authors and editors behind Google’s Site Reliability Engineering (SRE) book reunite to launch the Site Reliability Engineering (2nd Edition) and break down how system reliability, incident management, and developer careers have fundamentally changed in the AI era. From why 100% uptime is actually the wrong target, to why Google intentionally triggers outages to test dependencies, to the "irony of automation" leaving junior engineers unprepared for catastrophic failures—this is the ultimate blueprint for Intentional Reliability in modern software engineering. What We Cover in This Episode: -Why 100% reliability leads to developer burnout and runaway cloud costs -How to use Service Level Objectives (SLOs) and Error Budgets to push back on feature velocity -Rolling back vs. Rolling forward when AI generates 10x more Pull Requests -Using AI agents safely in production and incident response -Why a 4,000-person incident bridge call is a disaster (and how to actually run incident comms) -The Junior Engineer dilemma: How do new devs build "muscle memory" when AI handles the easy bugs? Speakers: Betsy Beyer, Chris Jones, Christof Leng, David Huska, Jennifer Petoff: Niall Richard Murphy,
Not rated yetGiant LLM vs. 100 small models: which wins?
· 55 sec
▶ Watch next: Intro to multi agent systems with ADK https://www.youtube.com/watch?v=0Z0GUDakR_A
Not rated yet6 Front-End Concepts Every AI Builder Needs to Know
· 23 min
Get the Prompts & Artifacts → https://g.dev/ai/build-it-resources Builder Central → https://g.dev/ai/builder-central Google Antigravity → https://g.dev/ai/build-it-with-agy Debugging a vibe coded app can feel like an endless game of bug whack-a-mole. In this episode of Build It, we bridge the gap between no-code prompting and systems thinking by using a custom Front-End Concepts Tutor agent in Google Antigravity to build a live interactive RPG Quest and Inventory Tracker that teaches us the exact mental models behind the app as we build it. Using the app's interactive X-Ray Mode and side-by-side comparisons of vague prompts vs. system-level architect prompts, we walk through the six core pillars of front-end architecture without writing code: Components and Boundaries, Props vs. Local State, Calculated Values (Computed State), Dynamic Rendering and Conditional Styling, Dynamic Actions and Event Bubbling, and Local vs. Global State. You can grab the Front-End Concepts Tutor prompt from the resources below, pick your own theme (golf, coffee, gaming, or fitness), and follow along in Google Antigravity. Chapters: 0:00 - Intro: Bridging the Gap From Bug Chasing to Systems Thinking 0:53 - Customizing a 1-on-1 Front-End Concepts Tutor Agent in Google Antigravity 3:13 - Pillar 1: Components, Boundaries, and X-Ray Mode 5:01 - Pillar 2: Props vs. Local State (Where Data Lives) 10:25 - Pillar 3: Calculated Values and Computed State 14:53 - Pillar 4: Dynamic Rendering, Conditional Styling, and Boss Phases 17:48 - Pillar 5: Dynamic Actions, Button States, and Events Bubbling Up 19:33 - Pillar 6: Local State vs. Global State (App-Wide Theme and Currency) 22:45 - Recap: The 6 Front-End Pillars for AI Builders Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleCloud #Gemini #Antigravity #NoCode #AIAgents #BuildIt Speakers: Devin Chasanoff, AI Builder Advocate Products Mentioned: Gemini, Google Antigravity, Google Cloud
Not rated yetGemini for regulated industries: Cited & auditable AI
· 34 sec
In banking and capital markets, you can't act on an AI recommendation if you can't verify its sources or methodology. In this Google Cloud Live clip from Gemini at Work 2026, Archana Kannan (Senior Director of Product Management, Applied AI at Google Cloud) explains how Gemini for Financial Services pairs built-in governance and identity management with workflow-level traceability—giving analysts clear citations and transparent methodology on every day-to-day output. In the full Google Cloud Live episode, we cover: * *Zero trust governance & traceability:* Cryptographic identities, isolated sandboxes, and end-to-end auditability for regulated workflows * *One persistent cloud agent:* 24/7 continuity across desktop, mobile, Workspace, and CLI * *Zero copy enterprise data:* Grounding agents in live business context via Knowledge Catalog & Borderless Lakehouse #Shorts #GoogleCloud #GeminiAtWork #FinancialServices #AIAgents Speaker: Archana Kannan Products Mentioned: Gemini
Not rated yetHow Gemini 4 Argon & Carbon were built (ft. DeepMind’s Head of Gemini Product)
· 41 min
Try Gemini 4 Argon & Carbon in Google AI Studio → https://g.dev/ai/studio Build with AI Agents in Google Antigravity → https://g.dev/ai/use-antigravity How do you build a frontier model that can work autonomously on complex codebases for days while bringing indirect prompt injection and cheating rates down to 0%? In this episode of The Google Cloud Show, we talk to Tulsee Doshi, Senior Director and Head of Product for the Gemini model family at Google DeepMind to unpack Google's back-to-back model releases of Gemini 4 Argon and Gemini 4 Carbon through an exclusive Google DeepMind interview and a hands-on developer demo showing how persistent, long-horizon models are powering the next generation of AI agents and software engineering workflows. In our main segment, Smitha sits down with Tulsee to take you inside the lab to share how Google dogfoods early checkpoints through internal experiments. We will also discover why Gemini 4 Argon (built for deep reasoning and multi-step workflows with a 1M token output capacity) and Gemini 4 Carbon (#1 on the Artificial Analysis Intelligence Index) rolled out first to trusted cyber defenders via the Fairwind program. You'll also hear how Carbon improves web development aesthetics, reduces fabrication, and achieves state-of-the-art defensive cybersecurity performance, plus why agentic code submissions inside Google have jumped 30% since Argon launched, how the Chrome team saw Carbon perform over 10x better at generating high-quality code patches, and how Tulsee uses audio models in Antigravity alongside Argon to turn voice rambles into structured documents and vibe-code a custom nanny app. Stick around after the interview for a hands-on builder segments that put Gemini 4 to work in Google AI Studio by using voice-to-text prompts to remix a 3D chessboard with coordinate labels and bright neon pieces, and adding Firebase Authentication and a live Cloud Firestore leaderboard to a 3D racing game in minutes instead of weeks. Finally, Rody Davis from the Antigravity DevRel team demonstrates how to build an automated "Software Factory" inside Google Antigravity powered by Gemini 4 Argon by using Markdown-defined custom subagents with isolated context windows and read-only permissions to audit Web (Node.js), Android (Jetpack Compose), and iOS (SwiftUI) code in parallel, strip out legacy polyfills, and steer subagents directly using /learn, /boost, and Teamwork Preview. Chapters: 0:00 - Intro 0:46 - Gemini 4 Argon (1M token output) & carbon (#1 intelligence index) overview 2:24 - Interview: Tulsee Doshi, Head of Product for Gemini at Google DeepMind 4:07 - Relentless shipping: How Gemini 4 iteration compares to Gemini 2 & 3 6:23 - Gemini 4 vs. Gemini 3.1 Pro: Model persistence on long-horizon tasks 7:46 - Argon vs. Carbon: Less fabrication, web dev aesthetics & SOTA Cyber 13:36 - Why Argon & Carbon rolled out to defensive cyber teams first 14:43 - DeepMind PM’s workflow: Voice rambling to structured docs in Antigravity 18:01 - How a DeepMind PM tests every new model checkpoint 21:21 - Why you should treat Gemini 4 as a thought partner, not the thinker 22:24 - Why Gemini 4 models are named after periodic table elements 25:27 - Remix 3D apps in Google AI Studio 28:28 - Adding Firebase Auth & a Live Firestore leaderboard to a 3D racing game 30:26 - How to build a multi-agent software factory in Antigravity 31:09 - Defining custom subagents in markdown & platform version policies 34:20 - Using the /learn command & parallel subagent context windows 35:31 - Enforcing read-only subagent permissions & direct subagent messaging 38:00 - Scheduling monthly codebase audits & using /boost and teamwork preview 40:07 - Wrap up & where to start building with Gemini 4 & Antigravity 🔗 More resources: * Read the Gemini 4 Argon announcement → https://g.dev/ai/intro-gemini-4-argon * Read the Gemini 4 Carbon announcement → https://g.dev/ai/intro-gemini-4-carbon * Explore Google Antigravity custom subagents…
Not rated yet