Architect

Contributing to U.K. financial sector resilience as a critical third party

Google Cloud has been designated as a critical third party (CTP) to the U.K. financial sector by the U.K. Treasury, reflecting its significant role for U.K. firms. This designation will enable oversight by the Bank of England and other financial regulators to enhance operational resilience in the sector. Google Cloud is committed to supporting customers in meeting regulatory requirements and fostering closer collaboration with regulators. The company aims to maintain its focus on being a secure and resilient platform for digital transformation in finance.

Frontier and Center: Who evaluates the evaluations?

The article explores the complexities of evaluating AI agents, particularly in data retrieval and question-answering contexts. It criticizes traditional pass/fail benchmarks that provide limited insights into an agent’s capabilities, advocating for a more nuanced approach through the use of a meta-benchmark called Discovery Bench. This framework uses information theory to modulate query difficulty by adjusting terms based on their informative power. The aim is to generate detailed performance maps, highlighting not only successes but also specific failure modes, thereby enabling targeted improvements in AI systems. The piece emphasizes the need to rigorously evaluate both the performance of AI agents and the quality of evaluation benchmarks themselves.

What’s new with Google Cloud

Google Cloud provides regular updates, announcements, and resources on its services. Upcoming webinars include one on July 16 introducing Google Cloud NGFW with advanced malware protection powered by Palo Alto Networks, and another on optimizing infrastructure for agentic AI. Public previews for Cloud Run sandboxes and new capabilities in Apigee are also highlighted. Reports indicate many organizations need infrastructure upgrades for agentic AI. Various tech talks will cover AI governance, security, and optimizing API deployments.

Slack Introduces Agent Driven End-to-End Testing to Improve Resilience in UI Test Automation

Agentic testing is an AI-driven approach to end-to-end test automation introduced by Slack engineering. It uses AI agents that execute workflows based on intent rather than fixed scripts, adapting to UI and system changes at runtime. The approach aims to reduce brittle tests in distributed systems while complementing deterministic unit, integration, and E2E testing strategies. By Leela Kumili

Improving Smart Tiered Cache for Public Cloud Regions

Smart Tiered Cache allows for precise upper tier selection for origins hosted on AWS, GCP, Azure, and Oracle Cloud with customer-provided cloud region hints.

How Datadog Used Claude and Cursor for Test-Driven Production Migration

In a recent article, Datadog engineer Arnold Wakim shared what worked, what didn’t, and the lessons they learned while evolving a critical production system using AI to overcome hard limits in its storage backend and significantly improve performance. By Sergio De Simone

Specification-driven composition for flexible data workflows

Specification-driven composition offers a solution to data pipeline scalability issues. As pipelines evolve from simple scripts, duplicated transformation logic and cascading changes create management challenges. This approach aims to avoid code duplication and improve maintainability by centralizing transformation definitions.

Navigating the ingress-NGINX retirement

1. The Post-March 2026 landscape ⚠ The CatalystAcknowledge the March 2026 retirement of the Kubernetes SIG Network ingress-nginx controller. Staying on this controller introduces severe operational risks, including unpatched CVEs and a complete halt of feature…

Frontier models and production agents: Advancing Microsoft Foundry for the agentic era

Introducing OpenAI’s latest frontier model series, the Asia Pacific Data Zone, and product agent capabilities, all generally available in Microsoft Foundry. The post Frontier models and production agents: Advancing Microsoft Foundry for the agentic era appeared first on Microsoft Azure Blog.

Safely run AI-generated code in Cloud Run sandboxes

Google Cloud has introduced Cloud Run sandboxes in public preview, allowing developers to run AI-generated code and untrusted binaries without risking host applications or sensitive data. This feature provides lightweight, isolated execution environments that start in milliseconds, ideal for tasks like executing dynamic Python scripts or running headless browsers. Security measures include credential and environment isolation, denied outbound network access by default, and a read-only filesystem. Cloud Run sandboxes leverage existing resources without incurring extra costs, making them a cost-effective solution for executing untrusted code securely.

Migrate Amazon EC2 to EKS Auto Mode using Kiro CLI and MCP servers

This post provides a detailed guide on migrating a Node.js web application from EC2 instances to a scalable, containerized service using EKS Auto Mode. It covers the configuration and utilization of AWS and Amazon EKS MCP Servers with the Kiro CLI. The tutorial includes steps for automating essential migration tasks, such as creating Dockerfiles, optimizing images, generating Kubernetes manifests, and deploying the application in a production environment on EKS Auto Mode.

Why we cannot wait for better post-quantum signature algorithms

NIST is advancing nine new post-quantum signature algorithms as potential candidates for future standardization. We take a closer look at all of them, and argue that while they are in the works and show great potential, we should use ML-DSA for now — the best one currently available.