Tech news, November 5, 2026
Thursday: 627 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.
The morning briefing
As the paper printed it on /brief/ at 08:46 UTC, in the sources' own words.
- LeadVivaldi 8.3 for Desktop: make bad ads sad
Vivaldi 8.3 is bad news for bad ads.
- NewsVivaldi 8.3 for Mobile: A faster, cleaner web in your pocket
Vivaldi 8.3 for Mobile brings the redesigned Ad and Tracker Blocker to Android and iOS, smarter media controls for Android, and a massive performance upgrade on iOS.
- NewsIntroducing Teel - Turbo-Efficient Execution Library for Agents (but not only)
Teel is an open-source Python runtime that overlaps independent LLM and tool calls in plain sequential code.
- OutageProblem with Cloud SIEM APIs in North America 2 (US2)
We have implemented a fix for the issue affecting Cloud SIEM APIs.
- SecurityWant to Rent an Internet Address at the White House?
Several staffers at the White House are using computers or mobile devices equipped with residential proxy services that render their internet address available for others to rent,…
- Deadlinegemini-3.6-flash shuts down in 14 days
Replacement: gemini-3.8-flash.
News
202 pieces, the newest 20 here.
- Real-time fraud detection with Feature Views
Real-time fraud detection depends on three kinds of features: what a customer did...
- [Launched] Generally Available: Azure Front Door TLS key exchange policy with post-quantum support
Azure Front Door now supports configurable TLS key exchange policies, including hybrid post-quantum key exchange, to help you prepare for evolving cryptographic threats while meet…
- Bringing conversational analytics to Gemini with Agentic Data Cloud connectors
Data Cloud connectors in Gemini for business, now in preview, give business users and analysts a direct way to query live enterprise data in natural language.
- Chapter 11 Update and Transaction Notice
We wanted to provide an update on the status of the Character’s Chapter 11 case.
- The Kotlin Multiplatform Contest Is Open: Win a Trip to KotlinConf
The Kotlin Multiplatform Contest is back!
- Accelerate data access for inference and training with the open-source distributed accelerator for cloud storage
The goal is simple: reduce repeated reads without replacing the durable storage for all AI workloads.
- How to assess your organization’s AI readiness
AI adoption moves fast.
- AWS Interconnect - last mile announces new metro in Northern Virginia, USA
Today, AWS announced a new AWS Interconnect - last mile metro in Northern Virginia, USA.
- How to Trace and Evaluate a Google ADK Agent with Opik
Trace a multi-agent Google ADK app with Opik, fix handoffs that split into two traces, and evaluate it on 23 hard support cases to compare two Gemini models on cost, speed and ref…
- PHP 8.6.0 RC5 available for testing
The PHP team is pleased to announce the fourth release candidate of PHP 8.6.0, RC5.
- Snowflake CoCo for Excel and Google Sheets: Governed AI
Snowflake CoCo brings governed AI to Microsoft Excel and Google Sheets, with live Snowflake data, agentic workbook building and role-based security.
- How I built a video sharing site with Claude Code and Mux after years away from coding
See how Eric built ScubaTube with Claude Code, using Video.js v10 for a dive data overlay and TV mode, Mux Robots for chapters, and Mux Data for engagement.
- Announcing Web Answer on Amazon Bedrock AgentCore
Today, we are announcing the general availability of Web Answer on Amazon Bedrock AgentCore.
- Microsoft Foundry extension for Azure Developer CLI is generally available
Set up, test, evaluate, and deploy hosted agents with the generally available Microsoft Foundry extension for Azure Developer CLI, keeping agent and deployment configuration along…
- Gemini Omni 1.2 Flash now in Vids with 15-second generations and improved quality
Users now have access to Gemini Omni 1.2 Flash directly within Google Vids.
- Inside GitHub’s new Git storage architecture
GitHub's new Git storage architecture is built so thousands of agents can push to a single repository at once.
- Rust Program Manager Updates
RustConf 2026 Shortly after joining the Foundation full time, I attended two conferences: RustConf and Kangrejos.
- Interactive Visual Gym: Evaluating Visual Interaction Models Through Simulation
AI assistants are rapidly evolving beyond text and image chat boxes into real-time, vision enabled copilots for the physical world.
- Vera Delivers the Best of Both Worlds: Highest Throughput and Fastest Agents
AI agents do more than generate answers.
- Cult.fit builds operational muscle with Codex
Cult.fit is using Codex to help teams across functions build tools, move faster, and change how work gets done.
New Models
8 pieces.
- Open model: ConwayResearch/Underdog-Woof
Trending on the Hugging Face Hub.
- Open model: Cactus-Compute/mirage
Trending on the Hugging Face Hub.
- New model: Google: Gemini Omni 1.2 Flash
Gemini Omni 1.2 Flash is Google's native multimodal video generation and editing model, the successor to Gemini Omni 1.1 Flash.
- New model: KT: Mi:dm K 3.0 Pro
Mi:dm K 3.0 Pro is a Mixture-of-Experts reasoning model from KT, with 120B total and 12B active parameters.
- New model: Auto Router
Auto is OpenRouter’s flagship model router, built for general-purpose usage.
- New model: Auto Code
Auto Code is OpenRouter’s router built specifically for coding agents and harnesses.
- Open model: nvidia/MetricTrail
New open weights from NVIDIA on the Hugging Face Hub.
- Open model: allenai/maniflow-tb
New open weights from Ai2 on the Hugging Face Hub.
Security Desk
50 pieces, the newest 10 here.
- Threat actors use agentic AI tooling to find and abuse exposed secrets
Microsoft Threat Intelligence has observed threat actors using the agentic offensive security framework ARTEX to search GitHub for OAuth application secrets, which can be used to…
- Phone return scams: where a real delivery and a fake return end up in a bill for you
Phone return scams are part of a very convincing social engineering method where the criminals get a phone and you get billed.
- NIKKE PC launcher compromise put players’ passwords and personal files at risk
Players of Goddess of Victory: NIKKE are being urged to check their PCs after attackers tampered with files used by the game’s Windows launcher.
- Chrome for Android Update
Hi, everyone!
- Scammers use AI-generated images of missing pets to extort owners
Scammers are using AI images to enhance their "I've found your missing pet and want money to return it to you" schemes.
- Barracuda Networks is your new best friend (Barracuda Email Security Gateway Pre-Auth Remote Root RCE CVE-2026-116606 / CVE-2026-111560)
Yes, that's right!
- Even gem list can run someone else's code
Ruby gems can run code on install and on every gem command, even gem list.
- Introducing AWS account access manager: IAM role flexibility with centralized workforce access
AWS Identity and Access Management (IAM) is introducing account access manager, a flexible new way to manage workforce access to your Amazon Web Services (AWS) accounts.
- Case Study: Kubescape Completes All Six Security Slam Badges with OpenSSF
Kubescape completed all six project-level badges in the Fall 2026 Security Slam, bringing maintainers and community contributors together to strengthen security across its open so…
- Bring a tent, I've got the coffee: Lessons from it-sa
Thorsten may have had to camp out in Nuremberg to attend it-sa, but the real story is the initiative to add a formal IT Security apprenticeship program in Germany for the thousand…
Technical Guides
12 pieces, the newest 10 here.
- Getting started with Web Answer on Amazon Bedrock AgentCore
Web Answer on Amazon Bedrock AgentCore returns a synthesized, source-cited answer to a natural-language question in a single tool call, with no search, retrieve, synthesize, and c…
- How Amazon Marketplace evaluates content quality with Amazon SageMaker AI
Learn how Amazon Marketplace evaluates A+ Content quality at scale on Amazon SageMaker AI by rendering each product page as a single snapshot, scoring it with a vision-language mo…
- Configure, Certify, and Operate Kubernetes GPU Clusters with NVIDIA DSX OS
One underperforming GPU can slow an entire distributed training job.
- Building an agent skill optimizer with AlphaEvolve, Harbor and Gemini 4 Argon
Every day, more and more people are using coding agents like Google Antigravity and OpenCode to complete software engineering tasks.
- Five Ways NVIDIA Co-Design Maximizes the Value of Every Byte in the AI Factory
In an AI factory, high bandwidth memory (HBM)is the fastest, most expensive, and most contested memory tier.
- Introducing presigned URL downloads in the AWS SDK for Java 2.x
Until now, the AWS SDK for Java 2.x could generate presigned URLs through S3Presigner but could not consume one to download an object.
- Learn With Mentor AI: Getting More Out of Video Courses
Use Mentor AI while you watch Real Python video courses.
- The True Cost of LLM Inference: GPUs vs. Token Pricing
What does LLM inference really cost?
- Quiz: Beautiful Soup: Build a Web Scraper With Python
In this quiz, you'll test your understanding of web scraping using Python.
- Quiz: Learn With Mentor AI: Getting More Out of Video Courses
Check how you use Mentor AI with Real Python video courses, from asking what the instructor said to quizzing yourself after a lesson.
AI in the Field
10 pieces, the newest 6 here.
- Reinventing software delivery with AI-SDLC on Amazon Bedrock AgentCore
Verisk partnered with AWS to build AI-SDLC: specialized AI agents spanning requirements, grooming, development, code review, and QA on Amazon Bedrock and Amazon Bedrock AgentCore,…
- Into the Omniverse: How Developers Turn Ideas Into Simulations With Frontier AI Agents
Turning a simulation idea into a working application means assembling assets, connecting physics and rendering, and checking that the scene behaves as intended.
- Clearing up a robot’s fuzzy memory
A new system identifies and corrects unreliable object descriptions stored in a robot’s memory, improving the robot’s ability to interact with its environment.
- Proposed Priorities and Definitions-Secretary's Supplemental Priorities and Definitions
The Secretary proposes to amend the Notice of Final Priorities and Definitions--Secretary's Supplemental Priorities and Definitions on Evidence-Based Literacy, Education Choice, a…
- Medical Devices; Hematology and Pathology Devices; Classification of the Digital Cervical Cytology Slide Imaging System With Artificial Intelligence Algorithm
The Food and Drug Administration (FDA) is classifying the digital cervical cytology slide imaging system with artificial intelligence algorithm into class II (special controls).
- Collaborative generalist-specialist AI for safe worklist deprioritization of non-contrast brain computed tomography
npj Digital Medicine, Published online: 05 November 2026; doi:10.1038/s41746-026-03447-y Collaborative generalist-specialist AI for safe worklist deprioritization of non-contrast…
Columns
9 pieces, the newest 6 here.
- The benchmark where AIs have to lie to win
AIs are now good enough to run a conspiracy
- The Iron(wood) Throne: Google's TPU Stakes Its Claim on Inference
Viable Alternative of CUDA Moat, NVIDIA scared and working nights & weekends, Jensen’s Counter Submission, Kimi K3, Muse Spark Open, Qwen, GLM5.3, TorchTPU, vLLM, SGLang, llm-d, R…
- AI #193: Progress May Dramatically Accelerate
Anthropic is warning us.
- Scaling Slack and GitHub with Maude Lemaire
Maude Lemaire shares lessons from scaling Slack and GitHub, improving performance at Cursor, and using AI while keeping engineering judgment in human hands.
- This robotics concept helps explain the jaggedness of frontier capabilities
Sim-to-real gaps are everywhere.
- You shouldn’t change the style of a Win32 static control after it has been created
It doesn't really handle that case.
Research
156 pieces, the newest 10 here.
- User Embeddings are Superpositions of Interpretable Behavioral Modes
User embeddings are central components of modern recommendation, personalization, and behavioral modeling systems.
- Do You Miss the Nights We Spent Crafting Figures? Rethinking Teaser and Framework Figures in the Age of Generative AI
Scientific figures give research ideas a visual form through choices of abstraction, composition, and emphasis.
- Where and Why is Grounded Reasoning Useful?
Grounded reasoning, where a model outputs text coordinates of relevant locations in chain-of-thought reasoning, is a common strategy for solving complex visual reasoning problems.
- SwarmEvolve: Latent Agent Swarms for Evolutionary Discovery and Optimization
LLM-guided evolutionary search uses AI agents to improve programs for mathematical and computational problems by iteratively refining solutions and combining high-scoring methods.
- Scientist World: Grow a Research World for AI-Scientists
AI-Scientists arrived, but a world for them to grow in did not.
- EgoGeneralist: Unifying Egocentric Action-Video Tasks through Video-Native Modeling
Modern video models possess strong visual priors, yet do not natively represent human action in metric 3D space.
- Monus vs. Interpretations in Semiring Semantics
Interpretations have been widely used in the study of provenance of queries on semiring-annotated databases.
- XStack: A Benchmark for ML Repo Migration
ML code should run where researchers choose, not where it was first written.
- Same Behavior, Different Mechanism: What Finite Experiments Can Certify About Integrated Information
Interpretability, AI auditing and assessments of machine consciousness increasingly rest on mechanism-level quantities: functionals of a system's causal transition law that descri…
- Towards interpreting any residual stream direction by scaling activation inversion
To interpret a direction in a language model's residual stream, we can look for text that strongly activates it.
Releases
63 pieces, the newest 12 here.
- LiteLLM 1.108.1: Verify Docker Image Signature
- langchain-typesafe==0.0.2
- langchain 1.6.2
- @astrojs/vercel@11.0.20
- @astrojs/node@11.1.11
- @astrojs/mdx@8.1.3
- OpenTofu 1.13.9: Fix local backend state persistence bug when the first apply crashes. (#4811)
- OpenTofu 1.12.12: Fix local backend state persistence bug when the first apply crashes. (#4811)
- b12158
- @anthropic-ai/sdk 1.6.0
- sdk-v1.6.0
- Anthropic Python SDK 2.6.0: api, add error code for ramp limit errors
New Repos & Packages
24 pieces, the newest 8 here.
Watch
36 pieces, the newest 6 here.
- Client side functions for Flutter GenUI
- Securing Intelligent Media Workflows with Microsoft Entra ID | Avid
- KYC Becomes KYA: Sentinel Bio's Joshua Monrad on DNA Screening, Uplift Studies & Agent Monitoring
- CoreWeave Registry: The System of Record for Enterprise AI Assets
- Unlocking Media Archives with Agentic Intelligence | Support Partners
- SIG Scalability Meeting for 2026-11-05
Package releases
57 from the most used packages' own repositories, the newest 15 here.
- uncalled-for 0.6.0 · every release →
- cfn-lint v1.60.0 · every release →
- tox 4.67.12 · every release →
- zircote/swagger-php 7.3.0 · every release →
- phpstan/phpstan-doctrine 2.1.1 · every release →
- @opentelemetry/instrumentation-mongodb instrumentation-browser-navigation-v0.18.0 · every release →
- cuda-bindings cuda-pathfinder-v1.8.6 · every release →
- jsonpointer v3.9 · every release →
- cookie v2.0.2 · every release →
- markdown 3.11.3 · every release →
- streamlit 1.67.0 · every release →
- @aws-sdk/types v3.1163.0 · every release →
- aws/aws-sdk-php 3.403.0 · every release →
- semgrep v1.184.0 · every release →
- @opentelemetry/core v3.1.0 · every release →