InfoQ Homepage News
-
Enhancing A/B Testing at DoorDash with Multi-Armed Bandits
While experimentation is essential, traditional A/B testing can be excessively slow and expensive, according to DoorDash engineers Caixia Huang and Alex Weinstein. To address these limitations, they adopted a "multi-armed bandits" (MAB) approach to optimize their experiments.
-
Two Missing Characters: How a Regex Flaw Exposed AWS GitHub Repos to Supply-Chain Risk
AWS recently published a security bulletin acknowledging a configuration issue affecting some popular AWS-managed open-source GitHub repositories. Dubbed CodeBreach, the critical vulnerability could have resulted in the introduction of malicious code and hijacking of the repositories leveraging AWS CodeBuild.
-
Google’s Universal Commerce Protocol (UCP) Powers Agentic Shopping
Google has launched the Universal Commerce Protocol (UCP), an open-source standard designed to enhance commerce on AI-powered platforms. UCP creates a common language for agentic shopping, enabling seamless interactions among consumers, businesses, and payment providers.
-
Expired Oracle Patent Opens Fast Sorting Algorithm to Open Source Databases
A recent article reports that an Oracle patent on a fast sorting method has expired, allowing open source databases to use it freely. Mark Callaghan, the inventor behind the sorting algorithm, shows how this 20-year-old approach can speed up sorting similar data and could make database systems faster and more efficient.
-
Ramp Builds Internal Coding Agent That Powers 30% of Engineering Pull Requests
Ramp has shared the architecture of Inspect. This internal coding agent has quickly reached about 30% adoption for merged pull requests in the company’s frontend and backend repositories. The fintech company shared a detailed technical specification. It explains how they created a system that gives AI agents the same access to the development environment as human engineers.
-
VoidZero Announces Oxfmt Alpha with Rust-Powered Performance and Prettier Compatibility
VoidZero has unveiled Oxfmt, a cutting-edge Rust-based code formatter that offers over 30x faster performance than Prettier for JavaScript and TypeScript projects. Compatible with existing Prettier configurations, Oxfmt addresses developer needs for efficiency and style consistency. Enjoy seamless migration, enhanced capabilities, and a commitment to community-driven improvements.
-
DoorDash Applies AI to Safety across Chat and Calls, Cutting Incidents by 50%
DoorDash deploys SafeChat, an AI-driven safety system for moderating chat, images, and voice calls between Dashers and customers. Using a layered text moderation architecture, machine learning models, and human review, SafeChat detects unsafe content in real time, enabling immediate actions and reducing low- and medium-severity safety incidents by roughly 50 percent.
-
OpenCost Looks Back on 2025 Milestones and Charts a Roadmap for 2026
The OpenCost project, an open-source cost and resource management tool hosted by the Cloud Native Computing Foundation (CNCF), has published a year-in-review reflecting on its progress in 2025 and outlining priorities for 2026.
-
Open Payment Standard x402 Expands Capabilities in Major Upgrade
After six months of real-world usage, the open payment standard x402 has received a major update, extending the protocol beyond single-request, exact-amount payments. The release adds support for wallet-based identity, automatic API discovery, dynamic payment recipients, expanded multi-chain and fiat support via CAIP standards, and a fully modular SDK for custom networks and payment schemes.
-
European Initiative for Data Sovereignty Released a Trust Framework
The Danube release of the Gaia-X trust framework provides mechanisms for the automation of compliance and supports interoperability across sectors and geographies to ensure trusted data transactions and service interactions. The Gaia-X Summit 2025 hosted facilitated discussions on AI and data sovereignty, and presented data space solutions that support innovation across Europe and beyond.
-
AWS Launches European Sovereign Cloud amid Questions about U.S. Legal Jurisdiction
AWS has launched its European Sovereign Cloud with a €7.8 billion investment, designed to meet EU regulatory demands and address data privacy concerns amid geopolitical tensions. Despite its operational separation from global regions, questions linger about legal protections against U.S. data access. Competitors like Microsoft and local providers may present stronger sovereignty options.
-
How CyberArk Protects AI Agents with Instruction Detectors and History-Aware Validation
To prevent agents from obeying malicious instructions hidden in external data, all text entering an agent's context must be treated as untrusted, says Niv Rabin, principal software architect at AI-security firm CyberArk. His team developed an approach based on instruction detection and history-aware validation to protect against both malicious input data and context-history poisoning.
-
Anthropic Announces Claude CoWork
Introducing Claude Cowork: Anthropic's groundbreaking AI agent revolutionizing file management on macOS. With advanced automation capabilities, it enhances document processing, organizes files, and executes multi-step workflows. Users must be cautious of backup needs due to recent issues. Explore its potential for efficient office solutions while ensuring data integrity.
-
Tracking and Controlling Data Flows at Scale in GenAI: Meta’s Privacy-Aware Infrastructure
Meta has revealed how it scales its Privacy-Aware Infrastructure (PAI) to support generative AI development while enforcing privacy across complex data flows. Using large-scale lineage tracking, PrivacyLib instrumentation, and runtime policy controls, the system enables consistent privacy enforcement for AI workloads like Meta AI glasses without introducing manual bottlenecks.
-
MIT's Recursive Language Models Improve Performance on Long-Context Tasks
Researchers at MIT's CSAIL published a design for Recursive Language Models (RLM), a technique for improving LLM performance on long-context tasks. RLMs use a programming environment to recursively decompose and process inputs, and can handle prompts up to 100x longer than base LLMs.