All news with #google tag

Google fixes new Chrome zero-day exploited in attacks

🔒 Google released an emergency update to address a newly discovered Chrome zero-day, CVE-2025-13223, which is being actively exploited. The high-severity flaw stems from a type confusion vulnerability in Chrome's V8 JavaScript engine and was reported by Clement Lecigne of Google's Threat Analysis Group. Patches are available in versions 142.0.7444.175/.176 for Windows, 142.0.7444.176 for macOS, and 142.0.7444.175 for Linux; users should check About Google Chrome and relaunch to apply the update.

Google Chrome fixes actively exploited V8 type bug

🛡️ Google has released emergency Chrome updates addressing two V8 engine type confusion flaws, including an actively exploited vulnerability tracked as CVE-2025-13223 (CVSS 8.8) that can lead to arbitrary code execution or crashes. The patch also fixes CVE-2025-13224 flagged by Google's AI agent Big Sleep and completes a set of seven zero-days addressed this year. Users should update Chrome to 142.0.7444.175/.176 (Windows/macOS/Linux) and apply fixes for other Chromium-based browsers when available.

Google Gemini 3 Appears on AI Studio Ahead of Release

🤖 Google’s Gemini 3 has been spotted in AI Studio, suggesting an imminent rollout that could begin within hours or days. The AI Studio entry references how temperature influences reasoning — noting "For Gemini 3, best results at default 1.0. Lower values may impact reasoning" — and highlights controls such as context size and temperature. Earlier sightings on Vertex AI show a preview build named gemini-3-pro-preview-11-2025, while Google is also testing an image model codenamed GEMPIX2 (Nano Banana 2).

Google announces Dhivaru subsea cable and regional hubs

📡 Google announced Dhivaru, a new Trans-Indian Ocean subsea cable linking the Maldives, Christmas Island, and Oman, and will build two new connectivity hubs in the Maldives and Christmas Island. The investment builds on the Australia Connect initiative and aims to improve reach, reliability, and resilience across the Indian Ocean, supporting AI services such as Vertex AI. The hubs will provide cable switching, content caching, and colocation to reduce latency, improve availability, and support local ecosystems.

2025 Google Cloud Partner All-Stars: Recognizing AI Leaders

✨The 2025 Google Cloud Partner All‑stars honor individual partners advancing innovation and delivering measurable customer outcomes. This year introduces a new AI Innovation category that recognizes strategic leaders guiding enterprise-wide AI adoption and applying agentic AI to transform core processes and create new revenue streams. Recipients also demonstrate excellence across delivery, marketing, sales, and solutions engineering, reflecting technical rigor and long-term customer focus.

Hands-on with Gemma 3: Deploying Open Models on GCP

🚀 Google Cloud introduces hands-on labs for Gemma 3, a family of lightweight open models offering multimodal (text and image) capabilities and efficient performance on smaller hardware footprints. The labs present two deployment paths: a serverless approach using Cloud Run with GPU support, and a platform approach using GKE for scalable production environments. Choose Cloud Run for simplicity and cost-efficiency or GKE Autopilot for control and robust orchestration to move models from local testing to production.

Android Memory Bugs Drop as Google Expands Rust Use

🛡️ Google reports that adopting Rust across Android has reduced memory-safety vulnerabilities to under 20% for the first time and claims a 1000x lower vulnerability density versus legacy C and C++ code. The company says Rust changes have a 4x lower rollback rate, require about 20% fewer revisions, and cut code review time by roughly 25%, improving overall delivery speed. Google plans to extend Rust to kernel, firmware and critical first-party apps while maintaining layered defenses.

Google to Flag Android Apps for Excessive Battery Use

🔋 Google will begin flagging Android apps on Google Play that show high background activity and cause excessive battery drain. The change centers on a new Android Vitals metric called excessive partial wake locks, and apps that cross the bad-behavior threshold may be labeled as battery drainers and lose prominence in discovery surfaces. Developers will receive alerts in their Android Vitals dashboard and have until March 1, 2026 to remediate issues.

Using BigQuery ML to Solve Lookalike Audiences at Zeotap

🔍 Zeotap and Google Cloud describe a SQL-first approach to building scalable lookalike audiences entirely within BigQuery. They convert low-cardinality categorical features into one-hot and multi-hot vectors, use Jaccard similarity reframed via dot-product and Manhattan norms, and index vectors with BigQuery’s VECTOR_SEARCH. By combining pre-filtering on discriminative features and batching queries, the workflow reduces compute, latency, and cost while avoiding a separate vector database.

Advancing Text-to-SQL: Gemini's BIRD Benchmark Breakthrough

🚀 Google Cloud reports a new state-of-the-art Single Trained Model Track score on the BIRD benchmark, achieving 76.13 with a fine-tuned Gemini 2.5-pro. The team credits rigorous data filtering, multitask supervised fine-tuning, and test-time self-consistency selection for the gains. These improvements bolster NL2SQL features in AlloyDB AI and BigQuery, and enhance developer tooling such as Gemini Code Assist for reliable SQL generation.

Google reverses Android developer verification plan

🔁 Google has softened its planned Developer Verification requirements after widespread backlash, saying it will create a dedicated account type for limited app distribution and an advanced sideloading flow for experienced users. The original rule would have blocked installation of apps from unverified developers on certified devices beginning in 2026. Google says these changes respond to concerns from students, hobbyists, and power users who need accessible or higher-risk pathways to install apps.

Agent Factory Recap: Building Open Agentic Models End-to-End

🤖 This recap of The Agent Factory episode summarizes a conversation between Amit Maraj and Ravin Kumar (DeepMind) about building open-source agentic models. It highlights how agent training differs from standard ML, emphasizing trajectory-based data, a two-stage approach of supervised fine-tuning followed by reinforcement learning, and the paramount role of evaluation. Practical guidance includes defining a 50-example final exam up front and considering hybrid setups that use a powerful API like Gemini as a router alongside specialized open models.

Google Sues to Dismantle 'Lighthouse' Smishing Kit

🛡️ Google has filed a civil lawsuit in the Southern District of New York to dismantle Lighthouse, a phishing-as-a-service kit used to power large-scale SMS phishing (smishing) campaigns. The company says the kit — likely run from China and marketed on Telegram — offered more than 600 templates mimicking over 400 organizations and targeted more than one million people across 121 countries. Google is pursuing legal remedies and supporting new legislation while deploying technical protections such as AI-powered scam flagging and expanded account recovery options.

Fortinet Named Google Unified Security Recommended Partner

🔒 Fortinet has been named the inaugural Google Unified Security Recommended partner for network protection, integrating FortiSASE and the FortiGate NGFW to run natively on Google Cloud. The integration delivers unified policy and shared telemetry with Google Security Operations, combining FortiGuard Labs and Google threat intelligence for AI-driven detection and response. Customers gain consolidated management, improved performance via Google’s backbone, and simplified procurement through the Google Cloud Marketplace.

Four Steps for Startups to Build Multi-Agent Systems

🤖 This post outlines a concise four-step framework for startups to design and deploy multi-agent systems, illustrated through a Sales Intelligence Agent example. It recommends choosing between pre-built, partner, or custom agents and describes using Google's Agent Development Kit (ADK) for code-first control. The guide covers hybrid architectures, tool-based state isolation, secure data access, and a three-step deployment blueprint to run agents on Vertex AI Agent Engine and Cloud Run.

Google Announces Unified Security Recommended Program

🔒 Google Cloud is launching the Google Unified Security Recommended program to validate deep integrations between its security portfolio and third-party vendors. Inaugural partners CrowdStrike, Fortinet, and Wiz bring endpoint, network, and multicloud CNAPP capabilities into Google Security Operations. Partners commit to cross-product technical integration, a collaborative support model, and investment in AI initiatives such as the model context protocol (MCP). Qualified solutions will be available via Google Cloud Marketplace for simplified procurement and consolidated billing.

Looker Conversational Analytics Reaches General Availability

💬 Google Cloud has made Looker Conversational Analytics generally available, bringing natural-language data queries to all Looker users. Built on the Looker semantic layer and powered by Gemini and Google’s agentic frameworks, the feature provides instant, explainable answers and supports multi-turn exploration across up to five connected Explores. Analysts can build and share agents, use LookML for fine tuning, and rely on a governed foundation that surfaces “How was this calculated?” explanations. Admins can enable the capability now to accelerate data discovery and improve self-service across teams.

Fortinet Named Google Unified Security Recommended Partner

🔒 Fortinet has been named the inaugural Google Unified Security Recommended partner for network protection, recognizing FortiSASE and FortiGate NGFW running natively on Google Cloud. The collaboration delivers a cloud-native SASE that unifies networking and security with global PoPs on Google’s private backbone, centralized policy and telemetry via FortiManager, and AI-enhanced threat protection from FortiGuard Labs. Customers can deploy through Google Cloud Marketplace and expect lower TCO through a consolidated architecture and simplified operations.

Google Sues to Disrupt China-Based SMS Phishing Operation

📱 Google has filed suit in the Southern District of New York to unmask and disrupt 25 unnamed operators tied to Lighthouse, a China-based phishing kit that has victimized over one million people across 120 countries. The complaint alleges Lighthouse powers a “Smishing Triad” that spoofs trusted brands, blasts mass text lures, and automates enrollment of stolen cards into mobile wallets using one-time verification codes. Google asserts trademark infringement and RICO claims and seeks to dismantle the coordinated groups behind the service.

Google Cloud expands Hugging Face support for AI developers

🤝 Google Cloud and Hugging Face are deepening their partnership to speed developer workflows and strengthen enterprise model deployments. A new gateway will cache Hugging Face models and datasets on Google Cloud so downloads take minutes, not hours, across Vertex AI and Google Kubernetes Engine. The collaboration adds native TPU support for open models and integrates Google Cloud’s threat intelligence and Mandiant scanning for models served through Vertex AI.