Tech news, October 12, 2026
Monday: 391 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 13:30 UTC, in the sources' own words.
- LeadThe EU KIDS Act Should Make Services Safer, Not Create an Internet Badge
Age assurance proposals raise questions on how these systems would work in practice, and if they would protect minors without creating new risks to privacy, security, and access.
- NewsWhich LLM should review your code?
New frontier models ship almost every month.
- NewsEU inference: control where AI processes your requests
Enterprise teams can route supported AI model requests to European endpoints, both while building in Lovable and in the AI features of their published apps.
- OutageVoice Call Failures from a Subset of Twilio Phone Numbers to Germany
Twilio customers may be experiencing voice call failures from a subset of Twilio Phone Numbers to network subscribers in Germany.
- SecurityChromium: CVE-2025-12429 Inappropriate implementation in V8
Information published.
- Deadlinegemini-2.5-flash shuts down in 8 days
Replacement: gemini-3.8-flash or gemini-3.5-flash-lite or gemini-3.1-flash-lite.
News
87 pieces, the newest 20 here.
- Bring data quality to your Amazon EMR workloads with Amazon EMR Data Quality built on AWS Glue Data Quality
AWS Glue Data Quality is now available natively on Amazon EMR, bundled in AWS runtime for Apache Spark (emr-spark-8.1.0).
- Amazon CloudFront now supports AWS Lambda Web Functions as an origin
Starting today, you can deliver web applications built on AWS Lambda Web Functions through CloudFront's global edge network.
- Client telemetry user attribution and configurable capture
Enterprise administrators can now use telemetry in managed settings to attribute GitHub Copilot activity to a user without collecting their prompts, responses, or tool content.
- Announcing AWS Lambda Web Functions
AWS Lambda Web Functions is a serverless compute primitive that runs your HTTP framework natively, delivering fast, consistent responses through multi-concurrent execution and aut…
- Amazon EC2 G7 instances now available in Europe (Frankfurt) and Europe (Stockholm) Regions
Amazon Elastic Compute Cloud (Amazon EC2) G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs are now available in Europe (Frankfurt) and Europe (Stockholm)…
- Amazon Connect Customer now prefills step-by-step guides for after contact work
Amazon Connect Customer can now fill in the step-by-step guides that agents complete after a contact, using information extracted from customer conversations, such as booking iden…
- OCP Summit 2026: Powering AI with Open Innovations in Scale-Up, Scale-Out, and Scale-Across Networks
At the 2026 Open Compute Project Global Summit, Meta is sharing the next generation of open networking infrastructure, including: Our first liquid-cooled network rack (IcePack), a…
- Workday and Snowflake Zero-Copy Integration Now Available
Workday Data Cloud and Snowflake's zero-copy, bidirectional integration is generally available.
- .NET and .NET Framework October 2026 servicing releases updates
A recap of the latest servicing updates for.NET and.NET Framework for October 2026.
- AWS Lambda announces Lambda Web Functions
AWS Lambda announces Lambda Web Functions, a new serverless compute primitive where your code handles HTTP requests directly.
- Cloudflare One Client - Cloudflare One Client for Windows (version 2026.8.2110.0)
A new GA release for the Windows Cloudflare One Client is now available on the stable releases downloads page.
- Find your routine next steps on the post-meeting screen in Google Meet
We are introducing next steps in Google Meet.
- From alerts to answers: The future of PostgreSQL operations is intelligent
PostgreSQL earned its place in the modern application stack through openness; the next evolution of PostgreSQL operations is moving from reactive to intelligent.
- Configure Access Management and Access Approvals policies using classification labels with rules in Assured Controls
We’re introducing Assured Controls rules in the Google Workspace Admin Console, giving administrators a more granular, resource-centric way to enforce Access Management and Access…
- Together AI integration: Evaluate open models and optimize evaluation costs in Arize AX
Evaluate open models with Together AI in Arize AX.
- Community Spotlight: Slava Rozhnev and SQLtest.online
Open source runs on people.
- Introducing the Googlebook Linux Terminal Preview: A VM-Backed Environment for AI and Development
Posted by Cyrus Stoller, Group Product Manager, Laptops & Tablets Today, getting the most out of AI workflows requires access to a terminal with a command line, even if someone do…
- Can we change an AI model’s architecture without retraining it from scratch?
“Priming” reuses pretrained Transformer weights to build models with more efficient architectures at less than 1% of the original training cost and with throughputs up to twice as…
- Our commitment to open source: show the code, run the numbers
SurrealKV and SurrealMX are open source under Apache 2.0, and every benchmark we publish comes with the command to run it.
- Google Ads API Release and Sunset Schedules for 2026-2027
Last year, we shifted to a monthly release schedule for the Google Ads API to provide you faster API access to Ads features available in the Google Ads UI.
New Models
5 pieces.
- New model: Perplexity: Decider 27B Latest
This model always redirects to the latest model in the Decider 27B family.
- New model: Nace.AI: Drex v2.0
Drex v2.0 is a 35B mixture-of-experts decision model (3B active) from Nace.AI, the larger successor to [Drex v1.5](https://openrouter.ai/nace-ai/drex-v1.5).
- Open model: allenai/HiRO-ACEv2
New open weights from Ai2 on the Hugging Face Hub.
- Open model: nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
New open weights from NVIDIA on the Hugging Face Hub.
- New model: Meta: Muse Wonder
Muse Wonder is a frontier reasoning model from Meta, designed for complex personal and professional agentic tasks.
Security Desk
51 pieces, the newest 10 here.
- USN-8915-1: FFmpeg vulnerabilities
Momoko Shiraishi discovered that FFmpeg did not correctly handle certain memory operations.
- Swann Communications SWIFI-SE2KIC
Swann Communications SWIFI-SE2KIC khardwood Oct 12, 2026 Release Date October 13, 2026 Description View CSAF Summary Successful exploitation of these vulnerabilities could allow a…
- [vim-security] Command Execution via Rmd Syntax Highlighting in Vim < 9.2.1178
Posted by Christian Brabandt on Oct 12 Command Execution via Rmd Syntax Highlighting in Vim < 9.2.1178 =============================================================== Date: 12.10.…
- [vim-security] Arbitrary Code Execution via Ruby Omni-Completion in Vim < 9.2.1177
Posted by Christian Brabandt on Oct 12 Arbitrary Code Execution via Ruby Omni-Completion in Vim < 9.2.1177 =================================================================== Date…
- Playing It SFAPE: Code Execution in Isolated User Mode
It is tempting to think of Windows in terms of a simple model: user mode and kernel mode.
- MyReport Updater: From gRPC to SYSTEM
MyReport is a report generation software often used as a component of web-based applications.
- GuardDuty Extended Threat Detection has added five new indicators to surface attack sequences
Amazon GuardDuty Extended Threat Detection correlates weak GuardDuty signals into a single, critical-severity attack sequence finding when their composition crosses a threshold of…
- Inductive Automation Ignition Software
Inductive Automation Ignition Software khardwood Oct 12, 2026 Release Date October 13, 2026 Description View CSAF Summary Successful exploitation of this vulnerability could allow…
- Quantum-resistant signing on AWS, Part 2: Enforce your ML-DSA algorithm policy with SCPs and IAM
You want to set a cryptographic algorithm policy and stop non-compliant choices from being made in your Amazon Web Services (AWS) environment.
- Quantum-resistant signing on AWS, Part 1: Build a software supply chain with ML-DSA
Software supply chain attacks have grown in both frequency and sophistication.
Technical Guides
39 pieces, the newest 10 here.
- How Halliburton Landmark enhanced seismic interpretation performance and data management using Amazon FSx for NetApp ONTAP
Halliburton Landmark provides exploration and production software solutions through their iEnergy® cloud platform.
- Cross-Region Amazon WorkSpaces Migration at Scale – How Wolters Kluwer Migrated 3,000 Desktops Without Disruption
A customer case study on migrating Amazon WorkSpaces from Asia Pacific (Singapore) to Asia Pacific (Mumbai) Introduction When organizations first adopt Amazon WorkSpaces, they pro…
- Apache Iceberg Table Statistics: NDV, Puffin and Query Plans
Iceberg keeps per-file bounds automatically but table-level NDV statistics only when you compute them.
- Disaster Recovery for Apache Polaris: Backup and Restore
Recovering Polaris means recovering a metastore and object storage separately.
- Apache Iceberg Concurrency: Commit Conflicts, Isolation and Retries
What causes an Iceberg commit conflict, the retry properties that govern it, the two isolation levels, and how to design conflicts out of a busy table.
- Parquet vs Avro vs ORC: Choosing a File Format in 2026
Parquet and ORC are columnar and built for reading.
- Why Apache Arrow and Apache Parquet Are Not the Same Thing
Parquet is how columnar data sits on disk and Arrow is how it sits in memory.
- Apache Iceberg vs Apache Parquet: Table Format or File Format?
Iceberg is a table format and Parquet is a file format.
- Accessing OpenAI GPT models on Amazon Bedrock with global cross-Region inference in Thailand, Malaysia, Singapore, Indonesia, and Taiwan
Learn how to run OpenAI GPT-6 Sol, Luna, and Astra on Amazon Bedrock through global cross-Region inference from Singapore, Jakarta, Thailand, Malaysia, and Taipei.
- Agentic AI in Manufacturing: Scale From Pilot to Production
See how manufacturers scale agentic AI from pilot to production with a four-stage roadmap for governed data, autonomous operations and C-suite approval.
AI in the Field
5 pieces.
- Building responsible AI in EdTech: A learner protection framework
In this post, we look at why these three make up the AWS CAF-AI perspectives EdTech leaders should mature first, and what building each one looks like in practice.
- Finding the perfect moment: How Formula 1® uses AWS AI to search archives in seconds
F1 worked with AWS to migrate petabytes of footage from ageing tape robots, in multiple formats and proxy files, into Amazon Simple Storage Service (Amazon S3) alongside the live…
- From authoring to editing: a competency framework for trainee use of ambient AI clinical documentation
npj Digital Medicine, Published online: 12 October 2026; doi:10.1038/s41746-026-03351-5 Ambient AI has spread through clinical settings faster than curricula can adapt, shifting t…
- Transformative AI and demand bottlenecks
Transformative AI promises to accelerate growth through productivity enhancement and automation.
- AI-based hourly prediction of emergency department demand integrating social determinants of health
npj Digital Medicine, Published online: 12 October 2026; doi:10.1038/s41746-026-03378-8 AI-based hourly prediction of emergency department demand integrating social determinants o…
Columns
9 pieces, the newest 6 here.
- OpenAI Fires Three Safety Researchers
Three OpenAI safety researchers, Tomek Korbak, Mikita Balesni and Jasmine Wang, were fired, ostensibly for ‘mishandling sensitive company information.’ Those employees have publis…
- How Anthropic’s Dynamic Workflows Coordinate Claude Agents
At AI Engineer New York, Anthropic drew on lessons from Unix and Linux to explain how agents can tackle bigger projects without overwhelming developers.
- Hello Python 3.15, goodbye 3.10, and now I can use 3.11
Python 3.15 is here!
- FBI Arrests Suspect in FBI-Hacking Case
In September, hackers from a group called ShinyHunters announced that they stole a trove of FBI data about its own agents.
- C++ futures part 1: The mental model
Managing a value that is produced later.
- The 2026 curl team
The curl security team happens to also double as the curl core team.
Research
9 pieces.
- AI models may be capable of self-replicating without access to a datacenter
Small AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.
- AI-coupled HPC Workflow Applications, Middleware and Performance
AI integration is revolutionizing the landscape of HPC simulations, enhancing the importance, use, and performance of AI-coupled HPC workflows.
- A Survey on Tabular Data: From Tree-based Methods to Tabular Deep Learning
Tabular data remains one of the most challenging modalities for deep learning due to its heterogeneity, lack of spatial or sequential inductive bias, and small-sample regimes.
- Beyond Efficiency: A Systematic Survey of Resource-Efficient Large Language Models
The burgeoning field of Large Language Models (LLMs), exemplified by sophisticated models like OpenAI’s ChatGPT, represents a significant advancement in artificial intelligence.
- Sentiment Analysis in Digital Spaces: An Overview of Reviews
Digital data generated via social media have become a prosperous entity for sentiment analysis researchers seeking to understand individuals’ feelings, attitudes, and emotions.
- Exploring Backdoor Vulnerabilities and Fairness Challenges in Generative Models: A Survey
Global connectivity has increased societal diversity worldwide, fostering growth and dispelling negative stereotypes.
- Less can be more for predicting properties with large language models
Nature Computational Science, Published online: 12 October 2026; doi:10.1038/s43588-026-01045-2 Researchers show that large language models cannot learn from spatial coordinates r…
- Designing dielectrics for microchip capacitors
Nature Electronics, Published online: 12 October 2026; doi:10.1038/s41928-026-01726-w The miniaturization of capacitors for advanced electronics requires dielectrics that balance…
- Unsupervised domain adaptation for dialogue summarization with augmented heuristic intermediate domains
Releases
46 pieces, the newest 12 here.
- DSPy 3.4.1: DSPy 3.4.1 makes GEPA's structured InstructionProposer the default instruction proposer and lets dspy.RLM…
- langgraph-checkpoint-postgres==3.1.3
- b11577
- Browser Use + Qwen-Image-2.1-Turbo
- langgraph-checkpoint-sqlite==3.1.2
- Envoy 1.39.4: Summary of changes:
- b11576
- Podman 6.1.4: Fixed an issue where Podman would not function when run under WSL 3 (#29749)
- b11575
- Envoy 1.38.7: Summary of changes:
- Podman 5.8.9: Fixed an issue where Podman would not function when run under WSL 3 (#29749)
- langchain==1.4.5
New Repos & Packages
53 pieces, the newest 8 here.
Watch
22 pieces, the newest 6 here.
- How to Build AI-Scale Caching with Memorystore for Valkey
- Google Cloud Live: Inside the Gemini at Work 2026 keynote
- Open Source Inference at Production Scale
- AI Code Needs Specs: Formal Verification & Safer AI Agents
- Fuzz Testing a Discrete Event Simulator — by Apoorv Kothari — Seattle Rust User Group, Sept 2026
- BAML’s Rust-Powered Telemetry: 1000× Faster & Lossless—by Vaibhav Gupta—Seattle Rust UG, Sept 2026
Package releases
65 from the most used packages' own repositories, the newest 15 here.
- @tanstack/history release-2026-10-12-2358 · every release →
- @asamuzakjp/dom-selector v10.0.2 · every release →
- @tanstack/history release-2026-10-12-2321 · every release →
- checkov 3.3.28 · every release →
- react-resizable-panels 4.14.5 · every release →
- google/cloud-storage v2.6.0 · every release →
- google/cloud-core v1.74.0 · every release →
- @azure/abort-controller @azure/arm-commvaultcontentstore_1.0.0 · every release →
- langchain-aws==1.8.2 · every release →
- pendulum 3.3.0 · every release →
- schemars v1.2.3 · every release →
- postcss 8.5.30 · every release →
- smart-open v8.0.3 · every release →
- postcss-discard-duplicates cssnano@9.8.0 · every release →
- strands-agents python/v1.60.0 · every release →