Tech news, September 21, 2026
Monday: 190 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
60 pieces, the newest 20 here.
- Simplifying Model Serving Across Multiple GPUs with NVIDIA TensorRT Multi-Device Integration in NVIDIA Dynamo-Triton
The compute and memory demands of generative AI increasingly exceed what a single GPU can provide.
- TinyTorch: Don’t Just Import PyTorch. Build It.
A framework you write yourself, tensors through transformers TL;DR Every mature systems project eventually needs a teaching version.
- Amazon launches research initiative with Stanford University to advance AI and science
The collaboration aims to advance research while broadening participation and translating discovery into real-world solutions.
- [Launched] Generally Available: Azure Sphere OS version 26.09 is now available
Azure Sphere OS version 26.09 is now available in the Retail feed.
- xAI’s Grok 4.6 is now available in Amazon Bedrock
xAI's Grok 4.6 is now available in Amazon Bedrock: a frontier model for long-running agents, coding, and knowledge work, with a 500K token context window and four reasoning effort…
- Kubernetes v1.37: Tracking When a PersistentVolumeClaim Was Last Used (Beta)
Kubernetes v1.37 promotes the PersistentVolumeClaimUnusedSinceTime feature gate to Beta (enabled by default).
- NVIDIA Launches DSX Ready to Qualify Power and Cooling Products for AI Factories
Every AI factory needs power and cooling that fit its computing architecture.
- Advancing AI for biology: Teaching models to design and characterize antibodies
Three new papers from Amazon Bio Discovery address bottlenecks in AI-driven antibody engineering, from benchmarking binding predictors to experimentally validating de novo design.
- Source material: Graphic designer Aitana Picó wants us to appreciate bad trends
The multidisciplinary designer and artist shares what’s in her “mental database”: an off-putting tarot deck, a Jockstrap CD, and a recurring dream about aerial silks.
- Vercel Connect now supports Microsoft Teams
Vercel Connect now includes a managed connector for Microsoft Teams.
- Announcing the 2026 OpenTelemetry Governance Committee Election
The OpenTelemetry project is excited to announce the 2026 OpenTelemetry Governance Committee (GC) election.
- [In preview] Public Preview: Introducing a Guided Copilot Experience for Building Azure Apps in VS Code
Today we're previewing a new way to build cloud apps with GitHub Copilot in VS Code: a guided Copilot experience that takes you from idea to a deployed Azure app through a structu…
- [Launched] Generally Available: Azure Functions support for PowerShell 7.6
Azure Functions support for PowerShell 7.6 is now generally available.
- Attach an Ubuntu Pro subscription to your AWS Image Builder image
We are pleased to announce that you can now attach an Ubuntu Pro subscription directly to any base image in AWS Image Builder, allowing you to create enterprise-grade custom image…
- Reuse builds across multiple services
Build a commit once and deploy the same immutable artifact across multiple Render services and environments.
- AWS Weekly Roundup: AWS Builder Center mobile apps, Amazon Connect Talent GA, Amazon Corretto 27, and more (September 21, 2026)
Living in the Netherlands, I spend a fair amount of time on trains, and that is usually where I catch up on what the builder community is writing.
- Bring your Android game to the car screen today
Posted by Jan Kleinert, Developer Relations Engineer, Android for Cars Today, the games category for Android Auto and cars powered by Android Automotive OS with Google built-in is…
- Open-Sourcing Rebalancer: A Generic, High-Performance Library for Solving Assignment Problems
We’re open-sourcing Rebalancer, the assignment-problem solver that has been used to solve resource allocation problems throughout Meta for over nine years.
- From Enablement to Execution, Egypt’s AI Ecosystem Reaches Production Scale
Today, Egypt’s AI builders gathered in the Grand Egyptian Museum for a reception that highlighted the nation’s rapidly growing AI ecosystem — spanning AI natives, developers, rese…
- Scale your AI workloads faster and more efficiently with GKE Pod snapshots
When running modern AI workloads, there’s often a conflict between performance and cost.
New Models
9 pieces.
- New model: Xiaomi: MiMo-V2.6-Pro-UltraSpeed
MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro.
- New model: Xiaomi: MiMo-V2.6-Flash
MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi.
- New model: Xiaomi: MiMo-V2.6-Pro
MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi.
- New model: SpaceXAI: Grok 4.7
Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6.
- Open model: XiaomiMiMo/MiMo-V2.6-Flash-RL
Trending on the Hugging Face Hub.
- Open model: XiaomiMiMo/MiMo-V2.6-Pro-RL
Trending on the Hugging Face Hub.
- Open model: Edge0/Audio8-ASR-Infinite
Trending on the Hugging Face Hub.
- Open model: SupraLabs/Supra2-IMG
Trending on the Hugging Face Hub.
- New model: Qwen: Qwen3.8 Omni Flash
Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding.
Security Desk
15 pieces, the newest 10 here.
- 21st September – Threat Intelligence Report
For the latest discoveries in cyber research for the week of 21st September, please download our Threat Intelligence Bulletin.
- OpenPLC Runtime v3
OpenPLC Runtime v3 khardwood Sep 21, 2026 Release Date September 22, 2026 Description View CSAF Summary Successful exploitation of this vulnerability could allow an attacker to hi…
- Mind the (Patch) Gap, Part 2: Fake Websites Used to Deploy Chrome & Windows 0-Day Exploits
On September 9, 2026, Volexity published a blog post detailing the simultaneous use of multiple chained zero-day exploits in Google Chrome (CVE-2026-85046, CVE-2026-87491) and Mic…
- Transforming Bedrock Guardrails events into OCSF with CloudWatch
Security teams investigating possible AI-related security events need guardrail intervention data alongside their existing security telemetry.
- CVE-2026-7273: Zyxel GS1900 Series Switches Stack-Based Buffer Overflow Vulnerability
Zyxel GS1900 series switches contain a stack-based buffer overflow vulnerability in the CGI program which could allow a LAN-based, unauthenticated attacker to exploit the flaw and…
- One does not simply defend agentically
Defenders can’t use AI in the same way attackers can, but there’s much they can do to unlock the potential of agentic cyber defence.
- SAML: A fractal of bad design
Born out of academia and raised in corporate IT departments, the Security Assertion Markup Language (SAML) authentication protocol continues to be a staple in these organizations.
- From Exposure to Lockdown: How AWS Neutralizes Compromised IAM Credentials through Managed Policies
We explore how AWS neutralizes exposed IAM credentials using managed policies, detailing GitHub secret scanning and CloudTrail monitoring strategies.
- Group Policy hijacked: PAYLOAD ransomware weaponizes Active Directory GPO
Kaspersky GERT experts dive into the technical incident analysis of PAYLOAD ransomware: an encryptionless, binary-less operation that abused Active Directory mechanisms for managi…
- The SMB cybersecurity squeeze: AI agents at work, old attacks in overdrive
As AI opens new paths to company data while making familiar attacks faster and cheaper, SMBs need protection designed around the time and expertise available to operate it
Technical Guides
32 pieces, the newest 10 here.
- Helix: The internal tool powering our Shopify app's native migration
We built Helix to help LLMs rebuild the Shopify app in Swift and Kotlin, using small checkpoints and strict quality gates to keep the code shippable.
- How to Evaluate AI Agents From Tool Calls to Task Completion
When you ship an AI agent, the key question is whether it can execute a chain of work across dozens of sequential tool calls against a live environment, and...
- MLOps Solutions for Production Machine Learning
Learn how MLOps solutions support experiment tracking, model serving, monitoring, feature management, governance, and production rollouts.
- How We Think About AI Red Teaming: A Field Framework for Enterprise Deployments
A solutions engineer's field framework for AI red teaming: attack surface, adversarial testing, and what actually gets fixed.
- Grafana Alerting: Scale alert routing without scaling complexity using multiple notification policies
Alert routing often starts simple.
- How Ramp runs GPU AI workloads at scale with ECS Managed Instances
Ramp runs GPU-powered AI inference continuously on Amazon ECS.
- From structures to dynamics: scaling AI for molecular dynamics in drug discovery
Why molecular dynamics matters for drug discovery Molecular dynamics (MD) is an important tool in modern drug discovery.
- Stay focused wherever you work with Firefox mobile browsers
When it’s Monday morning and you’re on the subway, in a coffee shop, or squeezing in a few email replies after your morning workout, your phone becomes your digital office.
- Migrate SQL Server multi-result-set procedures to PostgreSQL
SQL Server stored procedures can return multiple result sets from one call, but PostgreSQL cannot.
- From Two Codebases to One: How DataVue Brought Phonetic Match Keys into SingleStore with Python UDFs
See how DataVue used SingleStore Python UDFs to unify phonetic matching in one codebase, halve processing time and keep regulated data on-platform.
AI in the Field
1 piece.
- Why Deploying Physical AI at Scale Demands Safety at Every Layer
Physical AI is moving rapidly from research to large-scale deployment.
Columns
9 pieces, the newest 6 here.
- Jev introduces a new shape of LLM - System One, aka Decision Models
Last week TypeSafe AI unveiled Jev, their first example of a new category of model that they are calling "System One models" (I'm with Maggie Appleton, I think "decision models" i…
- Where do AI norms come from?
A dialogue with Dr Cat Hicks on ethics, learning, and what we owe each other.
- In Search of a Compositional Theory of Self-Stabilization
My quest for a principled solution for metastable failures has taken me back to my roots on self-stabilization, as this recent paper related the problem to composition of self-sta…
- Writing Rust code that's faster than state-of-the-art libraries by asking agents to make the code faster
c’mon, try doing a breakthrough
- dlab Open Source Week: Frontier AI on Your Own Hardware
In one of my classes I asked the question I was afraid to ask but I just needed the answer to: “Who is afraid of not getting a job after graduating?” About eighty percent of the 1…
- Import AI 473: The US’s superintelligence strategy; human brain in a mouse skull; and machine hermeneutics
Welcome to Import AI, a newsletter about AI research.
Research
2 pieces.
- Global Conference in Pittsburgh Puts Carnegie Mellon’s Robotics Leadership in the Spotlight
Pittsburgh will host the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Sept.
- Algorithmic systems, human agency and the future of platform research
Nature Computational Science, Published online: 21 September 2026; doi:10.1038/s43588-026-01038-1 While there are widespread concerns about the potential negative societal impacts…
Releases
12 pieces.
New Repos & Packages
5 pieces.
Watch
11 pieces, the newest 6 here.
- Understand the Gemma 4 model family
- Figma Gave GPT-6 Astra a Moonshot. Here's what happened.
- One Prompt. A Feature Built and Tested. | Ramp × GPT-6 Astra
- Build mission-critical geospatial applications using ArcGIS Maps SDK for Flutter | Martino Yovo
- AWS Summit Amsterdam 2026 | AWS Events
- How to move AI from code completion to agentic workflows
Package releases
34 from the most used packages' own repositories, the newest 15 here.
- type-is v3.0.0 · every release →
- redshift-connector v2.1.17 · every release →
- boxen v9.0.0 · every release →
- lazy-loader v0.6 · every release →
- markitdown v0.1.8 · every release →
- firebase/php-jwt v7.2.0 · every release →
- google/cloud-storage v2.5.4 · every release →
- google/gax v1.50.0 · every release →
- google/cloud-core v1.73.4 · every release →
- google/auth v1.55.0 · every release →
- holidays v0.105 · every release →
- livewire/livewire v4.4.6 · every release →
- notebook v7.6.3 · every release →
- eventsource v5.1.2 · every release →
- @typescript-eslint/types v8.70.1 · every release →