Tech news, September 22, 2026
Tuesday: 297 pieces the paper read from primary sources, by section, newest first. Each links to its source; the line under it is the source's own opening.
News
106 pieces, the newest 20 here.
- Claude Opus 5.5 on Snowflake Cortex AI
Claude Opus 5.5 is in public preview on Snowflake Cortex AI, bringing improvements in coding, financial analysis, and visual understanding.
- Faster C++ code intelligence with whole codebase indexing
C++ code intelligence in GitHub Copilot CLI is now faster with support for whole codebase indexing.
- Now on Amazon CloudWatch Omni: collaborative AI-powered observability for your applications
Amazon CloudWatch Omni is the next evolution of CloudWatch — unified observability that brings your applications and AI agents into one reimagined experience, with auto-discovered…
- Introducing Amazon CloudWatch Omni: AI-powered observability for generative AI and agentic workloads
Learn how Amazon CloudWatch Omni delivers AI-powered observability purpose-built for generative AI and agentic workloads.
- Better prompt caching for GPT-6
Learn how GPT-6 improves prompt caching with higher cache hit rates, new diagnostics, explicit breakpoints, and controls that reduce latency and costs.
- What if your app could fix itself?
With EAS Workflows and EAS Simulator, a bug report becomes a pull request with a fix and recordings showing it works.
- The operations gap between deploying an application and running it forever
This post features the new Cluster Mode deployment model in AWS Elastic Beanstalk that extends the operational promise of the service to containerized application portfolios.
- ClickHouse appoints Mike Scarpelli, former Snowflake and ServiceNow CFO, to Board of Directors
ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced the appointment of Michael Scarpelli to its Board of Directors
- AI Agent Observability coming to Observe by Snowflake
Observe by Snowflake's Agent Observability (coming soon) helps teams monitor, debug and evaluate AI agents — tracking quality, cost and performance.
- Retirement: Support for Node.js 22 ends on April 30, 2027
Support for Node.js 22 will end on April 30, 2027, please transition to Node.js 24 by that date.
- [In preview] Public Preview: Flex Nodes for AKS
Flex Nodes for AKS is a deployment option for application and platform teams to use hybrid and edge infrastructure as connected worker nodes to an AKS control plane in Azure.
- Anthropic unifica Claude Chat, Cowork, Artifacts e Design
Anthropic unifica Claude Chat e Cowork em interface única, integrando Artifacts, Design, Docs e Slides.
- GPT-6 Astra, Sol, and Luna: For production agents in Microsoft Foundry
Explore GPT-6 Astra, Sol, and GPT-6 Luna in Microsoft Foundry, with scalable model options for production AI agents, complex workflows, and high-volume tasks.
- Bring more intelligence to everyday work with GPT-6 Sol and GPT-6 Luna on Amazon Bedrock
GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock, giving you more options to match intelligence and efficiency to each workload.
- OpenAI GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock
Today, AWS announces the general availability of GPT-6 Sol and GPT-6 Luna from OpenAI on Amazon Bedrock.
- AI Agent Observability: Making Silent Production Failures Visible and Actionable
By Archana Kumari, Yushu Yao, and Priya Lal.
- Introducing GPT-6 Sol and Luna
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
- Spotlight on SIG Apps
As Kubernetes adoption has grown, the conversation has shifted beyond running containers to managing increasingly complex application lifecycles.
- Next.js Security Update for a Critical Upstream Issue
The September 22, 2026 out-of-band security update for Next.js is now available
- Bringing Enterprise Identity and Key Control to AI on CoreWeave
Enterprise AI security without a rebuild.
New Models
9 pieces.
- New model: Cohere: Command A+
Command A+ is Cohere's flagship model for enterprise agentic workflows.
- New model: OpenAI: GPT-6 Luna Pro
GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on com…
- New model: OpenAI: GPT-6 Luna
GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol.
- New model: OpenAI: GPT-6 Sol Pro
GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on comple…
- New model: OpenAI: GPT-6 Sol
GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier.
- New model: inclusionAI: Ming Image 0.1 Design
Ming Image 0.1 Design is a text-to-image model from inclusionAI aimed at graphic-design output, with an emphasis on legible text rendering inside the generated image.
- New model: Anthropic: Claude Opus 5.5
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5.
- New model: AssemblyAI: Universal-3.5 Pro
Universal-3.5 Pro is AssemblyAI's speech-to-text model served through its Sync API, returning a complete transcript with word-level timestamps in a single synchronous response for…
- Open model: black-forest-labs/flux-3-action-base
Trending on the Hugging Face Hub.
Security Desk
34 pieces, the newest 10 here.
- Critical: Home Assistant: XSS in Statistics Graph Card
Affects homeassistant.
- Critical: lightrag-hku: No Rate Limiting on /login Endpoint Allows Brute-Force Attacks
Affects lightrag-hku.
- Critical: OpenBao's Recovery Mode Vulnerable To Token Leakage via Timing Attack
Affects github.com/openbao/openbao.
- Critical: [mcp-atlassian] Authentication bypass in HTTP transport: AtlassianOpaqueTokenVerifier accepts any non-empty token
Affects mcp-atlassian.
- CVE-2026-94450 - Potential denial of service when configured to send Retry packets in s2n-quic
Bulletin ID: 2026-116-AWS Scope: AWS Content Type: Important (requires attention) Publication Date: 09/22/2026 13:00 PM PDT Description: s2n-quic is a Rust implementation of the Q…
- CISA Adds Four Known Exploited Vulnerabilities to Catalog
CISA Adds Four Known Exploited Vulnerabilities to Catalog ggaylor Sep 22, 2026 Release Date September 22, 2026 Description CISA has added four new vulnerabilities to its Known Exp…
- Chrome for Android Update
Hi, everyone!
- Stable Channel Update for Desktop
The Chrome team is delighted to announce the promotion of Chrome 154 to the stable channel for Windows, Mac and Linux.
- VU#738147: Vendor-signed UEFI Shell applications allow Secure Boot bypass
Overview Vendor-signed UEFI Shell applications may allow an attacker to bypass Secure Boot protections by abusing commands such as mm (Memory Modify).
- 2026-013: Critical Vulnerability in F5 BIG-IP APM
On 22 September 2026, F5 published an advisory addressing a critical vulnerability affecting its BIG-IP APM product.
Technical Guides
44 pieces, the newest 10 here.
- The Reliability Layer for Healthcare AI: Common LangSmith Use Cases
See how LangSmith helps healthcare AI teams turn clinical review into reusable evaluators, datasets, and release gates for safer AI in production.
- AI Change Management
Learn what AI change management involves and how enterprises can adapt roles, workflows, and governance as adoption scales.
- Enabling Private High-Performance Production AI Inference with NVIDIA Confidential Computing
As large language model (LLM) inference increasingly processes sensitive information and proprietary model context across personal, enterprise, and regulated...
- Evaluate skill-equipped agents with Strands Evals and Amazon Bedrock AgentCore
Skills let you encode domain-specific procedures as reusable, portable instructions for agents, but a fluent answer doesn't prove the agent picked the right skill or followed it.
- Topology-Aware Workload Scheduling with NVIDIA Topograph
AI factories are power-limited systems that deliver maximum value when fully optimized.
- Creating a memory dump in C#
Learn how to create a memory dump in C# to capture the state of your application for debugging purposes.
- Today I will… debug a production crash
We’ve all experienced this, the desktop application is hanging, the window is greyed out, and we’re just waiting for it to start to respond.
- Implement per-pod image pull permissions with ECR repository policies on Amazon EKS
Learn how to scope Amazon ECR image pull permissions to individual Kubernetes pods on a multi-tenant Amazon EKS cluster using KEP 4412 credential providers and ECR repository deny…
- Intuitive risk investigation with Amazon Neptune and Linkurious Enterprise
Learn how to configure Linkurious Enterprise to connect to Amazon Neptune and give business users intuitive, code-free graph exploration for risk investigation, from setup and ful…
- Adding custom domains to AWS Lambda MicroVMs with Application Load Balancer
Many teams want to expose their AWS Lambda MicroVMs under a custom domain they own, and satisfy CORS for browser clients, without changing the application.
AI in the Field
5 pieces.
- Building AI Agents for Telecom Network Operations
This post shows how to close that gap with AI agents that reason about telecom alarms the way a senior engineer does using a portable, framework-agnostic skill that packages 3GPP…
- Poitras Center to fuel early careers of 50 young scientists dedicated to psychiatric disorders research
Patricia and James Poitras ’63 provide fellowships for graduate students and postdocs who will shape the future of mental health research.
- NVIDIA Isaac ROS 5.0 Advances Agentic, Open Source Robotics Development
To build and deploy sophisticated robotics applications that can perceive, reason and act in dynamic environments, developers need new physical AI models and tools.
- Beyond ADMET: AI for medicines safety - survey
The MHRA and the Regulatory Innovation Office (RIO) invite organisations to share their views on the use of artificial intelligence (AI) in medicines development.
- Introducing transit rewards
Today, Waymo is excited to announce our transit rewards program, which we will initially offer in the San Francisco Bay Area, with future cities to follow.
Columns
5 pieces.
- Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Yesterday was Grok 4.7 ( pelicans ) and MiMo v2.6 Flash/Pro ( more pelicans ).
- An Oscar, Two Asteroids, and the Algorithm in Your sklearn: John Platt on AI for Science
We talked to Google’s Oscar winning “Giganerd” about automating science, solving climate change, and how future generations can contribute to science in the age of superintelligen…
- How will AI change operating systems? Part 2: Windows
Deepdive into the Windows team’s efforts to make the OS “AI agent-friendly” and win back developers by going all-in on Linux on Windows, local models, GPUs, & more
- MiMo-V2.6 Pro Architecture and Training Notes
Notes on MiMo-V2.6 Pro's GQA and sliding-window attention, agent training tasks, reward signals, and large RL batches.
- Debating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI
Podcast #19
Research
4 pieces.
- Can AI Be Slowed Down? Stanford HAI Experts Weigh the Risks, Rules and Race Ahead
In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and…
- Intuitive enzyme design with LLM agents
Nature Computational Science, Published online: 22 September 2026; doi:10.1038/s43588-026-01052-3 A collaborative system of large language model agents brings protein engineering…
- Computational active materials and embodied intelligence in extreme conditions
Nature Computational Science, Published online: 22 September 2026; doi:10.1038/s43588-026-01051-4 Embodied intelligence has enabled machines to act autonomously, yet intelligence…
- Reproducibility in the era of large language models
Nature Computational Science, Published online: 22 September 2026; doi:10.1038/s43588-026-01061-2 Reproducibility has long been a concern in machine learning research.
Releases
33 pieces, the newest 12 here.
New Repos & Packages
11 pieces, the newest 8 here.
Watch
9 pieces, the newest 6 here.
- Gemma 4 12B: The encoder-free model explained
- Ship a GitHub PR From a Slack Message with Managed Deep Agents
- Sound processing tricks in Flutter that you MUST TRY | Andrii Khrystian
- LGND Wins Snowflake Startup Challenge 2026
- Don’t Let AI Go Rogue: Human Governance for Agentic Systems
- If you had a superpower to make a single change to JavaScript – what would it be?
Package releases
37 from the most used packages' own repositories, the newest 15 here.
- meson 1.12.1 · every release →
- url-normalize v3.0.1 · every release →
- mako rel_1_4_3 · every release →
- datadog v0.54.0 · every release →
- oracledb v26.0.1 · every release →
- timm v1.0.30 · every release →
- hydra-core v1.3.7 · every release →
- json-repair v0.63.5 · every release →
- sass 1.105.0 · every release →
- nest-asyncio2 v1.7.3 · every release →
- @napi-rs/wasm-runtime napi-async-runtime-v0.2.3 · every release →
- python-slugify v9.1.1 · every release →
- @eslint/core mcp-v0.3.13 · every release →
- clickhouse-connect v1.9.0 · every release →
- laravel/telescope v5.25.0 · every release →