The Terraform Google Provider v6.0.0 is now generally available, featuring enhancements and bug fixes. The key changes include an opt-out default label 'goog-terraform-provisioned', deletion protection fields added to multiple resources, and the ability to reduce the suffix length in 'name_prefix' for multiple resources. This major release does not change support for earlier versions of Terraform.

Source: https://cloud.google.com/blog/

A new era of search is emerging that allows customers to search for an image or video, or information within it, using natural language processing (NLP) and multimodal embeddings. This technology enables cross-modal semantic search scenarios, such as searching images using text or finding text in images.

Source: https://cloud.google.com/blog/

Businesses across all industries are seeking data to better understand their customers and drive sales. A major consumer packaged goods brand can gain valuable insights by understanding customer actions on a retailer's website before making a purchase. However, retailers are hesitant to share sensitive customer data, making collaboration difficult. As a result, CMOs are being asked to find a solution to get detailed level insights, safely analyze data from different sources without compromising privacy, and use those insights for smarter decisions.

Source: https://cloud.google.com/blog/

Building or modernizing applications with generative AI foundation models like large language models (LLMs) requires a clear understanding of business goals and use cases. The choice between a managed LLM solution like Vertex AI and a self-hosted option on a platform such as Google Kubernetes Engine (GKE) is crucial. This decision hinges on a thorough grasp of project requirements.

Source: https://cloud.google.com/blog/

Google Kubernetes Engine (GKE) can help organizations serve large language models (LLMs) at scale while minimizing the cost per token. To achieve this, GKE features like workload and infrastructure autoscaling and load balancing are used. A performance benchmarking tool for GKE was created to automate the end-to-end setup of serving LLMs.

Source: https://cloud.google.com/blog/

AI21 Labs has launched the Jamba 1.5 Model Family, a family of open models available in public preview on Vertex AI Model Garden. The model family includes two models: Jamba 1.5 Mini, designed for speed and efficiency in tasks like customer support and document summarization, and Jamba 1.5 Large, capable of handling advanced reasoning tasks like financial analysis. Both models feature a 256K context window, Mamba-Transformer architecture, and support advanced developer features.

Source: https://cloud.google.com/blog/

ChromeOS was designed with security at its core, featuring verified boot, sandboxing, and on-device encryption. This has earned it recognition as the most secure operating system out of the box, surpassing MacOS and Windows. ChromeOS achieves this high level of security by restricting third-party access to its core operating system, including the kernel.

Source: https://cloud.google.com/blog/

Box.Inc, a global leader in Cloud Content Management, implemented Google Cloud Dataplex to streamline its Data Platform operations, transforming its approach to managing complex analytics use cases and driving product innovation and growth. The company's Data Platform is built on a multi-tenant model grounded in data mesh architecture, decentralizing data ownership to teams with the greatest contextual understanding.

Source: https://cloud.google.com/blog/

Google Cloud Storage offers three types of bucket location types: single region, dual-region, and multi-region. The company is adding new dual-region pairings to help businesses navigate European data compliance rules. The new pairings include Frankfurt/Zurich, Frankfurt/London, and Belgium/London. This enables data replication across geographically distinct regions, providing data reliability, availability, and performance.

Source: https://cloud.google.com/blog/

Mandiant observed a malware called PEAKLIGHT, which contained configurations related to the SHADOWLADDER malware and dropped a malicious DLL called WebView2Loader.dll. The malware also utilized the comp.exe utility to drop two additional files, an AutoIt3 binary and the LummaC.V2 payload. The files were found in the AppDataLocalTemp directory.

Source: https://cloud.google.com/blog/