From InfoQ · 40 stories
GitHub Code Quality, combining CodeQL and AI, is now generally available
GitHub Code Quality, a service integrating CodeQL analysis with AI-assisted detection of maintainability and reliability issues, is now generally available on GitHub Enterprise Cloud and GitHub Team. This release provides tools for engineering teams to manage code quality and maintainability, which is increasingly important with the rise of AI-generated code.
Google Releases Angular v22 with Stable Signal Forms, OnPush Default, and Experimental WebMCP
Google has released Angular v22, which stabilizes Signal Forms, Angular Aria, and asynchronous reactivity APIs, making OnPush the default change detection strategy. This update provides developers with new tools for reactive forms, data modeling, and improved performance, impacting how Angular applications are built and maintained.
Stripe Automates Database Remediation Using Graph Search and State Machines
Stripe's engineering team implemented a system that models its global infrastructure as a graph and uses graph search algorithms with state machines to automate database incident recovery. This new approach has reduced database-related pager alerts by 30%, eliminating 200 pages and 12 days of unhealthy shard states annually.
Instacart launches Blueberry, an AI assistant for incident response
Instacart introduced Blueberry, an AI-powered incident response system that helps on-call engineers investigate production issues by providing context and root cause hypotheses. This system aims to reduce the time engineers spend gathering information during incidents, improving diagnostic accuracy by leveraging historical data and operational knowledge.
Rootly Alters Code Review Process, Drops Small PR Rule Due to AI-Generated Code
Incident management platform Rootly has discontinued its small pull request (PR) rule, citing changes in code generation economics due to AI agents. The company now focuses on assessing "blast radius" and uses feature flags and rollback capabilities for safety, rather than limiting PR size. This shift reflects how AI agents produce complete feature implementations, making traditional small, stacked PRs inefficient for review.
Uno Platform 6.6 Adds Native AOT, Vulkan Rendering, and Accessibility Improvements
Uno Platform 6.6 introduces Native AOT publishing for five target platforms, an optional Vulkan rendering backend, and automatic registration for Model Context Protocol servers. These updates aim to improve application performance, simplify development, and expand cross-platform capabilities for WinUI-style applications.
npm Introduces Staged Publishing for Node.js Packages with Human Approval Step
npm has made staged publishing generally available, adding a mandatory human approval step with two-factor authentication before a package version goes live. This feature aims to enhance supply chain security by preventing direct, unreviewed publication of potentially malicious code.
yk Technology Automatically Adds JIT Compilers to Existing Language Implementations
A new technology called yk has been developed to automatically integrate just-in-time (JIT) compilers into existing language implementations, particularly for dynamically typed languages. This technology demonstrates significant speed improvements, with a Lua benchmark showing a 4x speedup and MicroPython showing a 2x speedup.
Ponytail AI Agent Skill Corrects Benchmark After Community Challenge
Ponytail, an open-source skill for AI coding agents, revised its performance benchmark after a contributor and Hacker News skeptics challenged its initial claims. The project, which instructs agents to prioritize minimal code solutions, rebuilt its benchmark against a more realistic agentic baseline.
SkiaSharp 4.0 Released, Aligning with Upstream Skia Milestones
Microsoft and Uno Platform released SkiaSharp 4.0, including versions 4.148.0 and 4.150.0, which updates the .NET bindings for Google's Skia 2D graphics engine. This release establishes a new cadence that aligns SkiaSharp package versions with upstream Skia milestones, modernizing the engine and improving performance in specific areas.
Microsoft Agent Framework Harness and Hosted Agents Reach General Availability
Microsoft has announced the general availability of the Agent Framework Harness and Foundry Hosted Agents, providing a production runtime for AI agents. This release offers platform teams tools to run and govern agents, moving beyond just building them, and addresses runtime execution and observability for AI agents.
HashiCorp Releases Public Beta of Vault Kubernetes Key Management for Encryption at Rest
HashiCorp has released the public beta of Vault Kubernetes key management, enabling Kubernetes clusters to use Vault Enterprise as their Key Management Service (KMS) provider for encryption at rest. This allows the Kubernetes API server to offload envelope encryption to Vault, protecting Kubernetes Secrets and other API resources stored in etcd by moving the encryption keys outside the cluster.
Embabel Agent Framework for Java AI Agents Reaches 1.0 General Availability
Embabel, a framework for building AI agents in Java and Kotlin, has released its 1.0 general availability version. This framework allows developers to define AI agents using typed domain objects, simplifying the creation of agents that can plan actions dynamically based on goals and changing conditions.
Cloudflare Launches Meerkat, a Leaderless Global Coordination Service Based on QuePaxa
Cloudflare introduced Meerkat, an internal globally consistent control-plane service that uses the QuePaxa consensus algorithm. Meerkat allows leaderless writes, which improves availability across Cloudflare's global network by avoiding the downtime associated with leader failures in traditional consensus algorithms like Raft.
AWS Builder Center Launches Free Sandbox Environments for Workshops
AWS Builder Center now offers free, time-limited sandbox environments for workshops, eliminating the need for developers to use personal AWS accounts or worry about unexpected charges. This initiative addresses a common barrier to learning new AWS technologies by providing pre-provisioned, temporary environments for hands-on practice.
HashiCorp Introduces tfpolicy, an HCL-based Policy-as-Code Framework for Terraform
HashiCorp has launched tfpolicy, a new HCL-based policy-as-code framework for Terraform, now in public beta within HCP Terraform. This framework integrates policy creation and enforcement directly into Terraform workflows, aiming to simplify infrastructure governance by using the same language and workflow as infrastructure provisioning.
Dropbox Integrates MCP and Dash to Link Security Design with Code Review Workflows
Dropbox has implemented a new engineering approach that connects security design documentation directly with code review processes using its internal Model Context Protocol (MCP) and knowledge system, Dash. This integration aims to bridge the gap where security requirements defined during design often become disconnected from later code implementation and verification.
AWS Lambda Introduces Self-Managed Code Storage, Increasing Account-Level Quotas
AWS Lambda now allows self-managed code storage, enabling functions and layers to reference deployment packages directly from customer-owned S3 buckets. This change removes the per-Region code storage quota for large function fleets and increases the default Lambda-managed storage from 75 GB to 300 GB, but does not alter individual function size limits.
.NET 11 Preview 6 Updates MAUI CollectionView for Windows and Android Shell
Microsoft released .NET 11 Preview 6, introducing the next-generation CollectionView to Windows and re-implementing Android Shell with standard MAUI handlers. These changes modernize the MAUI framework by replacing older Xamarin.Forms infrastructure with platform handlers and source-generated code, improving architectural consistency and reliability.
Grafana Assistant Now Supports Over 30 Data Sources for AI-Powered Observability
Grafana Labs expanded Grafana Assistant to query and correlate data from more than 30 sources using natural language, including Snowflake, Oracle, and Jira. This update allows engineers to investigate incidents and troubleshoot distributed systems without switching between multiple monitoring tools, aiming to address fragmented operational data.
Remix 3 Beta Preview Rebuilds Framework with Web Standards, Drops React
Remix has released a beta preview of Remix 3, a complete rebuild that shifts the framework away from React and towards web platform primitives, expanding its scope to a full-stack offering. This change impacts web developers by introducing a new architectural approach for building web applications, potentially altering development workflows and toolchains.
Uber implements "Zero Growth Stack" for infrastructure and AI cost optimization
Uber introduced its "Zero Growth Stack" strategy to decouple infrastructure capacity from business demand, reducing physical hardware and optimizing AI costs. This initiative includes dynamic Go runtime tuning and new governance for generative AI tools, addressing a sixfold increase in AI-related expenses since 2024.
Netflix Details In-House LLM Serving Platform Using Triton and vLLM
Netflix has shared details about its internal platform for serving large language models (LLMs), which integrates Triton and vLLM to manage real-time and batch inference workloads across CPUs and GPUs. This platform allows Netflix to support diverse model sizes and hardware while maintaining a consistent production workflow for LLM inference.
TanStack Table V9 Beta Released with Tree-Shakable Features and Reworked State Management
TanStack Table V9 beta has been released, introducing tree-shakable features, state management built on TanStack Store, and reduced memory usage. This update reworks core aspects of the headless UI library to improve performance and extensibility across various JavaScript frameworks.
Zalando Implements In-Process Client-Side Load Balancer for High-Throughput API
Zalando's engineering team developed and deployed an in-process, client-side load balancer for its Product Read API, which handles approximately 1 million requests per second. This change improved latency predictability, reduced infrastructure costs, and enhanced visibility into failure origins by moving high fan-out internal traffic routing into the calling process.
Airbus Selects Scaleway as Sovereign Cloud Partner, Prioritizing Protection from Extraterritorial Laws
Airbus selected French provider Scaleway as its sovereign cloud partner, making protection against non-European extraterritorial legislation a formal evaluation criterion in its tender process. This decision highlights a growing concern among major industrial buyers regarding data sovereignty and legal jurisdiction in cloud services.
Ink & Switch Releases Bijou64, a Canonical Variable-Length Integer Encoding
Ink & Switch introduced bijou64, a new variable-length integer encoding designed to ensure each number has only one byte representation, eliminating a class of security vulnerabilities. This format also decodes significantly faster than LEB128, addressing issues in protocols like their Subduction sync protocol.
Meta Ports React Compiler to Rust for Faster Builds and Tighter Toolchain Integration
Meta merged a Rust port of the React Compiler into its main repository, replacing the previous TypeScript version. This change aims to deliver significantly faster build times and tighter integration with Rust-based JavaScript toolchains, improving compilation performance for React applications.
GitHub Enhances Issues Navigation, Boosts Instant Access from 4% to 22%
GitHub redesigned the navigation architecture for GitHub Issues, enhancing client-side efficiency to increase instant navigation from 4% to 22%. The shift reduces latency by leveraging client-side caching, predictive prefetching, and a local-first approach, improving user experience significantly for developers.
Walmart Discusses Building Agents Using English as Programming Language
Walmart's Jake Mannix discussed the development of LLM-based agents at a conference, emphasizing the use of English as a programming language for building these agents. This shift represents a return to early programming styles, reminiscent of BASIC, and highlights the industry's ongoing evolution in AI agent technology.
Rspack 2.0 launched with performance improvements and leaner dependencies
Rspack 2.0, ByteDance's Rust-based JavaScript bundler, has been released, featuring performance enhancements and reduced dependencies. This version offers a pure ESM core, support for React Server Components, and requires Node.js 20.19+.
GitLab Introduces Carbon Emission Metrics for CI/CD Pipelines
GitLab has launched a feature to measure carbon emissions from CI/CD pipelines, integrating environmental metrics into DevOps workflows. This development aims to help engineering teams optimize sustainability alongside traditional performance measures.
Engineering AI for Creativity and Mobile Visual Search Presented by Bhavuk Jain
Bhavuk Jain outlined two projects utilizing AI for mobile applications: AI-generated wallpapers and a visual search tool. These innovations aim to enhance user experience by reducing friction in finding products and images.
Yelp Launches Training Orchestrator for Unified Machine Learning Model Training
Yelp has introduced Training Orchestrator, a new internal framework designed to streamline machine learning (ML) model training. This framework replaces fragmented team-specific scripts with a standardized, configuration-driven orchestration model, addressing issues of code duplication and inconsistency across ML training processes.
GitLab 19.2 Introduces AI Automation for Security Backlog Management
GitLab has launched version 19.2 of its DevSecOps platform, which includes new AI-driven features aimed at alleviating security review backlog caused by rapid AI-assisted coding. Key new functionalities include Dependency Scanning Auto-Remediation and Security Review Flow, addressing vulnerabilities and logic flaws in code efficiently.
Java Roundup: Value Objects Preview, WildFly 41, Open Liberty 26.0.0.7 Released
Key updates in Java include the reintroduction of Value Objects (Preview) and the GA release of WildFly 41. Additional notable releases include Open Liberty 26.0.0.7 and other updates for TornadoVM and LangChain4j, among others.
DoorDash Launches Entity Cache for 1.5M RPS with 99.99999% Availability
DoorDash has introduced Entity Cache, a proxy caching platform to efficiently manage service requests across its microservices. This system streamlines service communication, achieving over 1.5 million requests per second with 99.99999% availability, reducing backend load and improving response times.
InfoQ Launches Three Certification Cohorts Starting August 2023
InfoQ has launched three five-week online certification cohorts in August 2023 for experienced professionals. These cohorts focus on Architecture, Engineering Leadership, and AI Security & Privacy Engineering, helping senior engineers enhance their skills through real-world applications.
Max Korbacher Discusses Failures in Internal Platform Engineering
Max Korbacher emphasizes that success in platform engineering depends on understanding user needs, not just coding. He outlines common pitfalls in platform development, such as unmet management expectations and lack of user adoption.
Clare Liguori Discusses Strands Agents SDK Growth in InfoQ Podcast
Clare Liguori explains the evolution of the Strands Agents SDK from a Python project to a production-level agent harness. The conversation highlights the SDK's model-driven architecture and the importance of a modular approach in constructing AI agents, offering insights for developers in the space.