All news with #google tag

⚠️ Google’s AI-assisted coding tool Antigravity, launched in early November, has a critical vulnerability discovered by researchers at Mindgard within 24 hours that can install a persistent backdoor and execute malicious code each time the application starts. The flaw arises because the assistant follows custom user rules unconditionally and gives excessive weight to rules embedded in project source, while a global configuration directory can hold files specifying arbitrary commands that are read and acted on at startup. Mindgard also identified two additional vulnerabilities that could expose user data, and no patch is yet available.

⚠️ Google’s newly released Antigravity IDE has drawn security warnings after researchers reported vulnerabilities that can allow malicious repositories to compromise developer workspaces and install persistent backdoors. Mindgard, Adam Swanda, and others disclosed indirect prompt injection and trusted-input handling flaws that could enable data exfiltration and remote command execution. Google says it is aware, has updated its Known Issues page, and is working with product teams to address the reports.

🚧 Gemini 3’s release on 18 November 2025 signals a structural shift: beyond headline performance gains, it accelerates embedding large multimodal assistants directly into enterprise workflows and infrastructure. That continuation of a trend already visible with Microsoft Copilot effectively makes AI assistants a new enterprise perimeter — changing where corporate data, identities, and controls must be enforced. Security, compliance, and IT teams need to update policies, telemetry, and incident response to this expanded boundary.

🚀 Anthropic's Claude Opus 4.5 is now generally available on Vertex AI, delivering frontier performance for coding, agents, vision, and office automation at roughly one-third the cost of Opus 4.1. The model introduces advanced agentic tool use—programmatic tool calling (including direct Python execution) and dynamic tool search—plus expanded memory and a 1M-token context window to support long, multi-step tasks. On Vertex AI, Opus 4.5 is offered as a Model-as-a-Service on Google's high-performance infrastructure with prompt caching, efficient batch predictions, provisioned throughput, and enterprise-grade controls for deployment. Organizations can leverage the Agent Builder stack (ADK, A2A, and Agent Engine) and Google Cloud security controls, including Model Armor and Security Command Center protections, to accelerate production agents while managing cost and risk.

⚠️ This week’s recap spotlights multiple actively exploited vulnerabilities, supply‑chain compromises, and a record cloud DDoS that forced rapid vendor responses. Fortinet disclosed a FortiWeb OS command injection (CVE-2025-58034) that was observed chained with a recent critical fix, raising concerns about silent patching and disclosure timing. Google patched an actively exploited Chrome V8 0‑day (CVE-2025-13223), and attackers continued to abuse browser notifications, malicious updates, and SaaS integrations to phish and persist. The incidents underscore urgent priorities: patch quickly, scrutinize integrations, and strengthen monitoring and response.

🌐 Today Google is announcing TalayLink, a new subsea cable that will extend the previously announced interlink cable from the Australia Connect initiative to establish a diverse path between Australia and Thailand via the Indian Ocean. The project includes planned connectivity hubs in Mandurah (Western Australia) and South Thailand, the latter in partnership with AIS, plus local landing support from IGC. These investments are designed to integrate Google Cloud’s upcoming Thailand region and data center into its global network, improving resilience, routing diversity, and onward connectivity across the Indian Ocean.

📱 Google has made Quick Share interoperable with Apple's AirDrop, enabling two-way file transfers between Pixel devices and iPhones starting with the Pixel 10 family. The implementation uses AirDrop's "Everyone for 10 minutes" direct, device-to-device mode with no server intermediaries. Google says it applied threat modeling, internal security and privacy reviews, Rust parsing to reduce memory risks, and independent NetSPI testing. Users must manually confirm recipients before sharing.

🚀 Vertex AI Studio now centers developer workflows around Gemini and introduces agents-as-tools to streamline prompt engineering and app creation. The Studio adds three core agent commands — /Prompt, /Evaluate, and /Build — to refine prompts, assess outputs with custom autoraters, and generate working code. Team features include cross-account prompt sharing, version history, and notes. Onboarding is simplified with one-click API keys, an /Ask helper, express mode, and loginless model trials.

🚀 The Gemini CLI now includes Looker and Looker Conversational Analytics extensions, enabling direct terminal access to Looker data and dashboards. These additions let users ask complex questions, generate reports, and create dashboards without leaving the command line. Installation requires the Gemini CLI (npm), the two extensions, and configuration of Looker API credentials and optional Google Cloud settings. The update aims to streamline workflows and make data exploration more accessible from everyday development environments.

🚀 Google Cloud engineers demonstrated an experimental GKE cluster running 130,000 nodes to validate extreme scalability for AI/ML workloads. The test sustained control-plane throughput near 1,000 operations per second, supported over one million datastore objects, and achieved a baseline of 130,000 Pods launching in 3 minutes 40 seconds. The project combined API-server caching KEPs, a Spanner-backed key-value storage backend, and job-level orchestration via Kueue to enable predictable admission, rapid preemption, and efficient utilization at massive scale.

🤖 BigQuery AI consolidates BigQuery’s built-in ML, generative AI functions, vector search, and agent tools into a unified platform. It enables users to apply generative models and embeddings directly via SQL, perform semantic vector search, and run end-to-end ML workflows without moving data. Role-specific data agents and assistive features like a data canvas and code completion accelerate work for engineers, data scientists, and business users.

🔬 Google Cloud outlines an agentic AI framework to accelerate life sciences R&D by orchestrating specialized, fine-tunable models into modular workflows. It describes four agents—MedGemma for deep literature and data synthesis, TxGemma for in-silico preclinical prediction, Gemini 2.5 Pro as the cognitive orchestrator, and AlphaFold-2 plus docking tools for molecular design. The architecture maps data flows, tooling, and cloud services (Vertex AI, HPC, search) to move from target discovery through iterative Design→Dock→Predict→Refine cycles toward lab-ready lead nomination while preserving version control and compliance.

📈 Google announces it was named a Leader in the 2025 Gartner Magic Quadrant for Cloud Database Management Systems for the sixth consecutive year and positioned furthest in vision. The post presents the company's AI-native Data Cloud—a unified stack integrating BigQuery, Spanner, AlloyDB, Looker, and Dataplex—to support agentic AI. Google highlights embedded specialized agents, developer tooling (Data Agents API, ADK, Gemini CLI) and Agent Analytics in BigQuery to accelerate AI-driven applications while asserting cost and governance benefits on a single, open platform.

🤖Google has begun showing ads in its AI mode, the company's answer-engine experience rather than a traditional search engine. AI mode has been available for about a year and is free to all, with Google One subscribers able to toggle advanced models such as Gemini 3 Pro. Until now Google avoided ads to keep the conversational experience compelling; the new placements are labeled “sponsored” and typically appear at the bottom of AI-generated answers rather than in the right-side citation area. This looks like an experiment or optimization to improve click-through rates while complying with ad disclosure rules.

📡 Google updated Quick Share to interoperate with Apple's AirDrop, enabling direct file transfers between Pixel 10 devices and iPhone, iPad, and macOS. Transfers require the Apple device to be discoverable to Everyone for 10 minutes, while Android users must set Quick Share visibility to Everyone or use Receive mode. Google said the implementation is built in memory-safe Rust, avoids routing data through servers, and was independently assessed and hardened after a low-severity information-disclosure issue was fixed.

🔒 The Google Ads account for Cologne-based retailer Music Store was reportedly taken over by attackers on 19 October 2025. Criminals have linked more than 2,500 foreign advertising accounts to the company’s payment profile and are running persistent campaigns promoting online casinos and crypto exchanges that administrators cannot remove. The assigned Google account manager has reportedly been unable to stop the activity, and formal attempts to get intervention via official channels have so far failed. Police cybercrime investigators and consumer protection authorities have been notified, and reported losses exceed €4 million.

🔐 Google Threat Intelligence Group (GTIG) disclosed a previously undocumented loader, BadAudio, used by China-linked APT24 in a multi-year espionage campaign that employed spearphishing, watering-hole infections, and supply-chain compromises. The loader is heavily obfuscated, leverages DLL search-order hijacking and control-flow flattening, and exfiltrates encrypted system data to hard-coded C2 servers. In at least one observed case it delivered an Cobalt Strike Beacon, and many samples remained undetected by most antivirus engines.

🔒 Google announced cross-platform file sharing between Android and iOS by making Quick Share interoperable with AirDrop, beginning with the Pixel 10 Family. The company emphasizes a "secure by design" approach that included threat modeling, internal security and privacy reviews, and in-house penetration testing. The interoperability layer is implemented in Rust to reduce memory-safety risks in parsing wireless data, and transfers are direct peer‑to‑peer without routing content through servers. Google also engaged third‑party testers and experts who validated the implementation and found no information leakage.

🎨 Google is unveiling Nano Banana Pro (Gemini 3 Pro Image), a high-fidelity image generation and editing model available today in Vertex AI and Google Workspace, with a rollout to Gemini Enterprise coming soon. The model supports multi-language text rendering and on-image translation, connects to Google Search for context-aware outputs, and accepts up to 14 reference images and 4K inputs for production-grade assets. Built-in SynthID watermarking and planned copyright indemnification address commercial use and responsible deployment.

🎮 Google has released Gemini 3 Pro, a multimodal model that posts strong benchmark results and produces notable real-world demos. Early tests show top-tier scores (LMArena 1501 Elo, high marks on MMMU-Pro and Video-MMMU) and PhD-level reasoning in targeted exams. Designers reported one-shot generation of a 3D LEGO editor and a full recreation of Ridiculous Fishing. Adherence remains imperfect, so the author suggests Claude Sonnet 4.5 for routine tasks and Gemini 3 Pro for more complex queries.