InfoQ - Architecture & Design - News

Android Bench 2 Adds Support for Long-Horizon Tasks, Agentic Evaluation, and Continuous Scoring

Sergio De Simone — Fri, 09 Oct 2026 17:00:00 GMT

Google has released Android Bench 2.0, a major update to its benchmark framework for evaluating AI models and agents on Android development tasks. The update introduces long-horizon tasks (LHTs), agent-based evaluation, and continuous scoring to better assess performance on complex, multi-step development tasks.

By Sergio De Simone

Github Migrates Copilot Runtime to Rust with AI-Assisted Rewrite

Leela Kumili — Fri, 09 Oct 2026 14:29:00 GMT

GitHub migrated more than 800,000 lines of Copilot runtime code from TypeScript and Node.js to Rust in about 14.5 weeks using AI-assisted development. The incremental migration used N-API interoperability, automated testing, human review, and 128 pull requests while continuing to ship releases.

By Leela Kumili

Cloudflare K2 Builds Event Streams on R2 Object Storage, at a Second of Produce Latency

Steef-Jan Wiggers — Fri, 09 Oct 2026 05:06:00 GMT

Cloudflare has launched K2 in public beta, a serverless event streaming service implementing a durable log on R2 object storage. Cloudflare states produce latency of about one second at p99; a tester reported end-to-end p99 of 7.5 seconds, which a commenter answering for the team said was higher than expected. Consume charges match produce charges per GB.

By Steef-Jan Wiggers

Technological Sovereignty Requires Choice, Skills, and Support, Panelists Say During OSS EU

Olimpiu Pop — Thu, 08 Oct 2026 08:08:00 GMT

European organisations can enhance their technological sovereignty by developing credible alternatives and acquiring necessary skills, rather than completely avoiding global providers. A panel discussion at the OSS EU highlighted the importance of understanding dependencies in cloud and hardware, and emphasized the need for practical choices in software for effective operational sovereignty.

By Olimpiu Pop

Cloudflare Open Sources Decision Models for AI Agents

Renato Losio — Thu, 08 Oct 2026 06:54:00 GMT

During its recent “Birthday Week”, Cloudflare announced Clef, a set of open-weight AI models designed to choose between predefined options rather than generate text. Cloudflare released 9B- and 27B-parameter models, along with a platform for adapting them to specific decision-making tasks.

By Renato Losio

Grab Redesigns Counter Service Storage for 50% Lower P99 Latency

Leela Kumili — Wed, 07 Oct 2026 14:30:00 GMT

Grab migrated its high-volume Counter Service from a wide column database to Aerospike using storage abstraction, shadow traffic, data parity validation, and gradual traffic migration. The redesigned data model consolidated time buckets into map-based records. Grab reports about 50% lower production p99 read latency, 1 TB versus 3 TB of disk usage, and 45% to 50% lower cost per node.

By Leela Kumili

Cloudflare Uses an AI Harness to Probe and Harden Its WAF

Matt Foster — Wed, 07 Oct 2026 07:21:00 GMT

Cloudflare placed frontier AI models inside a controlled testing harness to probe its Web Application Firewall (WAF), using blocked attacks as starting points for models to generate and refine new variations.

By Matt Foster

Spanner Omni Reaches GA, Replacing Google's Atomic Clocks and File System with Software

Steef-Jan Wiggers — Wed, 07 Oct 2026 04:43:00 GMT

Google has made Spanner Omni generally available, running its distributed SQL database on-premises, across clouds or on a laptop. Doing so meant replacing Colossus with a Colossus-like abstraction layer and TrueTime with software-based time synchronization. There is no availability SLA, and practitioners are pricing the move in tail latency and operational toil.

By Steef-Jan Wiggers

QCon London 2027 Announces 15 Tracks on Production AI, Architecture, and Engineering at Scale

Artenisa Chatziou — Tue, 06 Oct 2026 10:00:00 GMT

A preview of the 15-track QCon London 2027 program, covering agent evaluation and guardrails, AI-era architecture, distributed-system debugging, modern data platforms, high-performance engineering, and Staff+ leadership.

By Artenisa Chatziou

Java News Roundup: JobRunr 9, OpenXava 8, Quarkus, LangChain4j, JNoSQL, Introducing Lathe

Michael Redlich — Mon, 05 Oct 2026 17:00:00 GMT

This week's Java roundup for September 28th, 2026, features news highlighting: the GA releases of JobRunr 9.0 and OpenXava 8.0; point releases for Quarkus, Micronaut, LangChain4j, Eclipse JNoSQL; ADK for Java and ADK for Kotlin; and introducing Lathe, a new Java language server.

By Michael Redlich

Akka Tests Spec-Driven AI Delivery across 65 Open Source Projects

Leela Kumili — Mon, 05 Oct 2026 13:58:00 GMT

Akka used 65 open-source projects to examine how specification structure, context, model selection, automated validation, and delivery guardrails affect AI assisted software porting. The experiment measured time, token use, code size, test parity, and performance, finding substantial variation across models, effort levels, and project types.

By Leela Kumili

Cloudflare Fixes Cross-Tenant Data Exposure in Containers

Steef-Jan Wiggers — Mon, 05 Oct 2026 08:48:00 GMT

Cloudflare has disclosed a cross-tenant data exposure vulnerability in Containers and Sandboxes, caused by thin-provisioned storage pools configured to skip zeroing reused blocks. Researchers recovered directory structures, database pages and complete SQLite databases across four continents. Cloudflare remediated it and found no evidence of exploitation.

By Steef-Jan Wiggers

Istio 1.31 Adds Agentgateway Waypoints and Moves Release Artifacts off Google Cloud

Mark Silvester — Sat, 03 Oct 2026 08:30:00 GMT

Istio 1.31 adds agentgateway waypoints in ambient mode, with a canary configuration fix included in 1.31.1. It also ends the publication of images and Helm charts to Google Cloud, requiring repository migration ahead of the 13 October outage test and signing-key updates for teams verifying images.

By Mark Silvester

AI Agents Are Disrupting Open Source Security Disclosure

Renato Losio — Sat, 03 Oct 2026 06:46:00 GMT

A recent article by Anil Madhavapeddy argues that AI agents can turn publicly available clues about software vulnerabilities into working exploits, reducing the effectiveness of traditional disclosure embargoes in open source projects. The author highlights the need for faster patching and release processes as the time between vulnerability disclosure and exploitation shrinks.

By Renato Losio

Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50%

Leela Kumili — Fri, 02 Oct 2026 14:22:00 GMT

Uber has rebuilt major parts of the Uber Eats search pipeline, reporting a 50% reduction in end-to-end latency. Changes include Above-the-Fold measurement, reduced retrieval work, parallel hydration, advertising data redesign, infrastructure optimizations, and an agentic coding workflow. Uber is also exploring microbatching, product-based retrieval, and HTTP multipart streaming.

By Leela Kumili