Tech news, September 28, 2026

If you write Java, Python, Node.js, or.NET, you have been able to add OpenTelemetry to an application without editing its code for years: attach an agent at startup and telemetry…

Amazon EventBridge relaunched the Custom event bus so a platform team can share one governed bus across the organization while application teams publish and subscribe from their o…

xAI's Grok 4.7 is now available on Amazon Bedrock: a frontier model for coding, long-running agents, and knowledge work.

Vector search and vector indexes are now generally available in Azure SQL Database.

Azure SQL Dev Hub gives you a central starting point for building applications with Azure SQL, whether you’re developing directly or using AI agents.

AWS Lambda now supports scalable network bandwidth for functions with 2,048 MB of memory or more running outside a VPC.

Perimeters draw a policy boundary around every part of an AI system, so a project moves from experimentation to production without touching the code.

Since launching a year ago, the Microsoft Research Asia — Singapore lab has established a strong foundation, deepened collaboration across government, academia, and industry, and…

Arize AX now integrates directly with TypeSafe AI, bringing native Jev-as-a-Judge evaluations into the Arize evaluation workflow.

Software factories change what engineers do.

Amazon Rekognition Face Liveness now returns Feedback Codes that help customers understand why a liveness check received a low score and how users could retry successfully.

Traditional OLTP systems weren't built for the search demands of AI agents.

EDB and Red Hat have closed a critical gap with a validated, supported way to run VM-based Postgres on OpenShift Virtualization without re-architecting it first.

Highlights Rév O’Conner added a settings popup to control the layout of the Inspector.

The enforcement date for GitHub Actions minimum version requirements for self-hosted runners on GitHub Enterprise Cloud has changed.

Claude Sonnet 5.5 is coming to Cortex Inference in public preview, with Cortex AI Functions, Snowflake CoWork, Snowflake CoCo, and Cortex Agents coming soon.

OpenAI apologises for incidents involving Australian government websites and outlines stronger safeguards and support to strengthen Australia’s cyber defences.

Watch the winning trailer from the Future Vision XPRIZE, The Gifted.

Claude Sonnet 5.5 is now available on Amazon Bedrock and Claude Platform on AWS.

By Sohini Arya and Manish Kumar Jha.

Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade.

Solar Decide is Upstage's structured decision model, served as a System One endpoint on Solar Mini 4.

Trending on the Hugging Face Hub.

Apple today released patches for all of its operating systems.

The edge isn't the AI model, but how you assemble the system around it.

USN-8487-1 fixed vulnerabilities in curl.

Michał Majchrowicz and Marcin Wyczechowski discovered that libevent incorrectly handled certain DNS responses.

Huntress researchers reveal how attackers are exploiting ChatGPT Custom GPTs to spread ClickFix lures and DLL-sideloaded malware.

TL;DR: Frontier AI has collapsed the vulnerability exploit window from weeks to mere hours.

Overview Authlib (versions up to and including 1.7.2) contain a signature‑verification bypass in the JSON Web Signature (JWS) general JSON serialization handling.

A look at the targeted AI taskflows behind these findings, the critical Android bugs they uncovered, and how to run the same open-source agent on your own app.

Posted by Robert Rothenberg on Sep 28 ======================================================================== CVE-2026-88815 CPAN Security Group =================================…

On Saturday, October 24, 2026, we will conduct an exercise demonstrating that the AWS European Sovereign Cloud can operate without depending on any infrastructure outside of the E…

TLDR: An agent, by definition, can take actions and is more than just a language model.

TLDR: An agent, by definition, can take actions and is more than just a language model.

A manufacturing defect rarely belongs to one system.

Property Finder built end-to-end autonomous incident management with AWS DevOps Agent.

Our early guidelines for safety cases in frontier AI training cover technical safeguards, operational practices, and investigating misalignment incidents

An MLOps workflow for evaluating deploying and rolling back AI applications

After a major or minor version upgrade on Amazon Aurora MySQL or Amazon RDS for MySQL, some queries regress because the optimizer's cost models, defaults, and execution strategies…

Moving from AI pilot to production in government requires more than technology.

Quantum computers promise to solve problems beyond the reach of classical machines, but they share a stubborn adversary: noise.

Deploy a text-to-speech model on Amazon SageMaker AI with the AWS vLLM-Omni Deep Learning Container and stream generated speech over a persistent bidirectional connection.

Surface Mount Technology (SMT) polarity programming defines how a polarity-sensitive component feeds from tape reel or tray to the pick-and-place machine, designed to help maintai…

Retail companies managing finished goods face mounting pressure to maintain competitive pricing while protecting profit margins in today’s dynamic marketplace.

Speech At the Oakland Tech Week Opening Keynote, cohosted by the Kapor Center, Oakland, California

In the wake of NVIDIA's announcement of a new agent safety platform, AI Now's Aya Ibrahim joins "Balance of Power" to warn against "one-size-fits-all" security steps.

The AI Research Resource (AIRR) is a suite of advanced supercomputers that provides AI-specialised compute capacity to researchers, academia, and industry.

How to nominate a UK organisation for the Pro-Worker AI Adoption Prize, recognising organisations using artificial intelligence (AI) that benefits workers.

GLM-5.3, KV Cache Offloading, HiSparse, DeepSeek Sparse Attention, IndexShare, Single-rollout Asynchronous Optimization, Cybersecurity

Headlines and History The headlines keep coming: such and such an AI system has solved such and such a math problem.

You know we don’t have to do this anymore, right?

Only on first execution.

Welcome to Import AI, a newsletter about AI research.

ArsTechnica is reporting on a “new” attack against RSA, one that bypasses factoring.

Realistic and editable animal fur reconstruction from multi-view images is challenging due to fine-scale detail, self-occlusion and obfuscation, and, unlike human hair, the lack o…

One deployed language model must often serve many compute budgets, yet serving each budget still means a separate training or compression run per point.

Modern video diffusion models require tens of denoising evaluations over long spatiotemporal token sequences.

Unified multimodal models can both look at and render images, so in principle they can repair their own generations: diagnose what an image gets wrong, revise it, observe the resu…

Identifying intertextual references is central to literary scholarship, but computationally difficult when source material is transformed through paraphrase, allusion, historical…

\emph{Distributional training} provides collective supervision for one-step visual generation by matching real and generated features in frozen representation spaces.

LLMs have demonstrated strong capabilities in creative writing.

When a large language model (LLM) agent executes the same task, token consumption can vary by over an order of magnitude across runs.

We present a representation-learning framework for composite adaptive tracking control under dynamically coupled disturbances.

Sparse inertial pose estimation promises camera-free motion capture from consumer devices, but consumer sensors are unreliable: firmware-fused orientations are biased, mounting va…