Tech news, October 9, 2026
Friday: 551 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 10:09 UTC, in the sources' own words.
- LeadAdded Ultrafast mode for GPT-6.1 Sol in the Responses API.
Feature · gpt-6.1-sol · v1/responses.
- NewsA green exit code is not evidence that the work happened
Agents don't build trust for another reason, structurally worse than the first.
- NewsRevamping Skills in Deep Agents
Deep Agents now lets you bind tools to skills, pin skills at runtime, and reload skills mid-thread, so agents with expansive skill repositories stay context-efficient and effectiv…
- OutageItaly SMS - Delivery degradation towards Telekom Italia Network
Please be advised that we are observing SMS delivery degradation towards Telekom Italia Network We are working towards resolution of the issue and will issue an update as soon as…
- SecurityFwd: Go 1.27.2 and Go 1.26.9 are released
Posted by Alan Coopersmith on Oct 08 -------- Forwarded Message -------- Subject: [security] Go 1.27.2 and Go 1.26.9 are released Date: Thu, 8 Oct 2026 18:08:50 +0000 From: announ…
- AI in the Field · HealthArtificial intelligence for opportunistic osteoporosis screening on chest radiographs
npj Digital Medicine, Published online: 09 October 2026; doi:10.1038/s41746-026-03333-7 Artificial intelligence for opportunistic osteoporosis screening on chest radiographs
News
68 pieces, the newest 20 here.
- GitHub Copilot weekly releases — October 5
This week’s updates make Copilot easier to use across accounts and environments, with more control over what agents can access and how you manage their work.
- dbt now supports Amazon Redshift data sharing — and your pipelines just got a lot faster
dbt now supports Amazon Redshift data sharing: write across databases, isolate dev and CI, and run faster without copying data.
- CodeQL 2.27.2 improves C++, Go, Rust, and JavaScript analysis
CodeQL 2.27.2 is now available, adding a C++ regular-expression parser and analysis improvements across several languages.
- Retirement: Azure Key Vault Secrets Provider Extension for Azure Arc enabled Kubernetes clusters
The Azure Key Vault Secrets Provider Extension for Azure Arc-enabled Kubernetes will be retired on October 9, 2027.Customers currently using the extension should plan to migrate t…
- Agents can now buy domains with the Vercel CLI
Your agent can now use the Vercel CLI to buy a domain.
- DigitalOcean MicroVMs: Fast, isolated compute for your AI agent infrastructure
Coding-agent platforms, sandbox products, and code-execution services all need the same thing: isolated machines that start up fast, retain their state between bursts of work, and…
- Reducing cross-AZ latency with the AWS Lambda metadata endpoint
AWS Lambda now exposes Availability Zone metadata through an endpoint in the execution environment, so your function can discover its AZ ID with an HTTP request.
- How to build great out-of-the-box user experiences with Managed Deep Agents
Managed Deep Agents includes a new API for managing reactions for your distributed agents, and a system to dynamically assign emoji responses with your instrument of choice.
- Google Workspace Weekly Recap - October 9, 2026
Secondary calendars will now follow the lifecycle of their owner As announced earlier this year, we are now rolling out the secondary calendar lifecycle changes for non-personal W…
- Introducing Clef-omni with full multimodality, plus a faster Clef and a cheaper Clef-flash
We are expanding the Clef decision model family with Clef-omni, natively processing audio, video, images, and text in a single pipeline.
- New Pro teams now default to 30-day deployment retention
New Pro teams now default to a 30-day deployment retention period.
- Deployment Storage billing begins for existing Pro teams
Deployment Storage and Functions Storage billing will begin for all Pro teams at $0.10 per GB-month.
- New logic and search steps in Workspace Studio help expand automation capabilities
We are introducing several new feature enhancements to Google Workspace Studio as part of the October 2026 feature release.
- Amazon Bedrock now supports reasoning summaries for OpenAI models
Amazon Bedrock now supports the reasoning.summary parameter for OpenAI models through the Responses API.
- Anthropic Claude Sonnet 5.5 and Claude Opus 5.5 are now available on Kiro in AWS GovCloud (US)
Two new Anthropic models are now available in the Kiro IDE and CLI for the AWS GovCloud (US) Regions.
- MongoDB 9.0: More 8.4 than 9.0?
MongoDB 9.0 is here.
- Project Beacon: Bringing safety signals to open models at scale
Baseten partners with Goodfire AI to bring safety controls to open model inference.
- Amazon S3 Vectors metadata pre-filtering is now available in AWS GovCloud (US) Regions
Amazon S3 Vectors metadata pre-filtering is now available in the AWS GovCloud (US-East) and AWS GovCloud (US-West) Regions.
- Hack the World: Why hackathons are still the best place to learn to build
The barriers to building software have collapsed.
- Modernizing Unstructured Data Workflows: Alteryx Live Query meets Google Cloud BigQuery
Alteryx One Live Query and Google Cloud BigQuery redefine how enterprises handle complex, unstructured data at scale by reducing reliance on disconnected tools, minimizing data mo…
New Models
9 pieces.
- New model: Microsoft: Microsoft-Decision-1
Microsoft-Decision-1 is a small model built for fast decision-making.
- New model: Nace.AI: Drex v1.5
Drex v1.5 is a 9B decision model from Nace.AI.
- Open model: LiquidAI/LFM2.5-24B-A2B-Base
New open weights from Liquid AI on the Hugging Face Hub.
- New model: Cloudflare: Clef Omni
Clef Omni is the mixture-of-experts member of Cloudflare's open-source Clef decision model family, a fine-tune of Qwen3-Omni-30B-A3B (30B total, 3B active parameters) served on Wo…
- Open model: nvidia/agile_one_s_place_ssd_n17_24050
New open weights from NVIDIA on the Hugging Face Hub.
- Open model: nvidia/agile_one_s_walk_n2_1680
New open weights from NVIDIA on the Hugging Face Hub.
- Open model: nvidia/agile_one_s_pick_ssd_n17_58000
New open weights from NVIDIA on the Hugging Face Hub.
- Open model: nvidia/agile_one_s_pick_ssd_n17_40000
New open weights from NVIDIA on the Hugging Face Hub.
- Open model: tencent/Youtu-Parsing-Omni
Trending on the Hugging Face Hub.
Security Desk
58 pieces, the newest 10 here.
- Stable Channel Update for ChromeOS / ChromeOS Flex
The Stable channel is being updated to OS version 16805.33.0 (Browser version 154.0.8037.151) for most ChromeOS devices. If you find new issues, please let us know on…
- Critical: TinaCMS admin preview iframe loads an attacker-controlled origin from the URL fragment
Affects tinacms, @tinacms/app.
- Critical: Contao: Cross-site scripting in the comments bundle
Affects contao/comments-bundle.
- Critical: Nginx UI: Authenticated Remote Code Execution via Backup Restore App Config Overwrite
Affects github.com/0xJacky/Nginx-UI.
- Critical: Vikunja: Link-share token reads any tenant's kanban buckets and enumerates usernames/IDs instance-wide (BOLA)
Affects code.vikunja.io/api.
- Unfixed vulnerabilities in Cyrus SASL
Posted by Alan Coopersmith on Oct 09 https://github.com/cyrusimap/cyrus-sasl/issues/891 points out there are a number of publicly reported security issues in the Cyrus SASL projec…
- SmarterTools security advisory (AV26-1026)
- SonicWall security advisory (AV26-1017) – Update 1
- CVE-2026-13762 and CVE-2026-13763 - Issue with HTTP/2 multi-frame request body inspection in AWS WAF
- Issue with containerd CRI Plugin - CVE-2026-50195, CVE-2026-53488, CVE-2026-53492, CVE-2026-53489, CVE-2026-47262
Technical Guides
42 pieces, the newest 10 here.
- How California built its Behavioral Health Public County Profile on Databricks
The California Department of Health Care Services (DHCS) built the Behavioral Health...
- Deployment Isn't the Bottleneck: Inference Economics Is
Deploying an AI model takes an afternoon.
- How to Build a Model Router in the Harness
How we built a model router into Open SWE's harness that cut median cost per coding task by 64% with no measurable drop in quality, and how to build your own.
- Cloud Performance Testing with Postman: From Laptop to CI
Cloud performance testing gives you what a laptop can’t: load spread across many machines, a clean baseline, and results your whole team...
- Shared vs. Dedicated AI Inference: Embed & Rerank
Shared or dedicated AI inference?
- How iSAM Funds built an options research platform for a 10,000x data problem with ClickHouse Cloud
iSAM Funds uses ClickHouse Cloud to research full options history, achieving 35–50x compression and 100x faster single-core ingestion while simplifying its data pipelines.
- Alignment is a human-AI workflow: how Custom Views make agent review possible
Automated evals can score every agent session, but they do not automatically reflect what your team considers a good outcome.
- From post-training to production: How TinyFish runs its web-native model on Databricks AI Runtime
TinyFish builds infrastructure for enterprises that need agents to work across the live web.
- OTel-Native by Design - Building Products That Export to Any Observability Stack
With contributions from Dan Gomez Blanco (New Relic).
- ICYMI: What landed for AI builders in September 2026
A monthly recap of the latest Amazon Bedrock, Amazon Bedrock AgentCore, and Strands updates from September 2026: broader model choice, faster serverless agents with built-in evalu…
AI in the Field
9 pieces, the newest 6 here.
- RareDx: Building a multi-agent AI system to transform rare disease diagnosis
In this post, we demonstrate RareDx, an agentic AI-powered application prototype, built on Amazon Bedrock AgentCore, Amazon Web Services (AWS) purpose-built AI services and open s…
- Commission holds special meeting of Scientific panel on frontier AI safety and risks
Commission holds special meeting of Scientific panel on frontier AI safety and risks Anonymous (not verified) Fri, 10/09/2026 - 09:16 The Commission holds a special meeting of the…
- Artificial intelligence for opportunistic osteoporosis screening on chest radiographs
npj Digital Medicine, Published online: 09 October 2026; doi:10.1038/s41746-026-03333-7 Artificial intelligence for opportunistic osteoporosis screening on chest radiographs
- Integrating generative artificial intelligence facilitates faculty interaction network and role transition in a biostatistics curriculum
npj Digital Medicine, Published online: 09 October 2026; doi:10.1038/s41746-026-03376-w Integrating generative artificial intelligence facilitates faculty interaction network and…
- Implementing and Scaling Artificial Intelligence in Low-Resourced Radiation Oncology: A Systematic Review of Deployments
npj Digital Medicine, Published online: 09 October 2026; doi:10.1038/s41746-026-03391-x Implementing and Scaling Artificial Intelligence in Low-Resourced Radiation Oncology: A Sys…
- Early prediction of prolonged ICU stay in sepsis patients using an explainable hybrid deep learning model
npj Digital Medicine, Published online: 09 October 2026; doi:10.1038/s41746-026-03387-7 Early prediction of prolonged ICU stay in sepsis patients using an explainable hybrid deep…
Columns
7 pieces, the newest 6 here.
- I expect rapid progress but not towards general superintelligence
I’ve often been surprised when I hear from top researchers in industry that they think AI will be better than them at their job in a few years, and I didn’t really know why I doub…
- Friday Squid Blogging: I Caught a Squid
On Wednesday I spent a day fishing, on a small boat out of Gloucester, MA.
- 2026.41: It’s Not You, It’s Me
The best Stratechery content from the week of October 5, 2026, including drifting apart from Apple, Facebook complications, and the delightful absurdity of U.S.-China dynamics.
- Lobbying
Have you tried to explain lobbying to a non American unfamiliar with the concept?
- One way of thinking about access specifiers in C++/WinRT implementation types
Mapping them to concepts from C# and the now-defunct C++/CX.
- New Math from OpenAI
It is kind of a huge deal.
Research
157 pieces, the newest 10 here.
- Can Language Models Play the Wikipedia Game?
Language models increasingly support or automate information retrieval from document collections too large to index.
- Towards Physically Grounded Articulated 3D Assets: A Multi-Agent Pipeline and Benchmark
In this paper, we study physically grounded articulated 3D asset generation---moving the goal from reproducing how objects look to capturing how they move and function.
- Splat2Act: From Captured Scenes to Personalized VLA Manipulation
A user who brings a generalist manipulation policy into their home or office wants it to pick up their objects from their shelves, counters, and tables.
- InstantFusion: A Shared Latent Interface for Heterogeneous Image Generators
Diffusion/flow models have emerged as the de facto paradigms for image generation, giving rise to a diverse family of modern generators with increasingly heterogeneous architectur…
- Images Talk Back: Guiding MMDiTs by Suppressing Image-to-Text Feedback
Multimodal diffusion transformers (MMDiTs) allow the evolving image to update internal text representations through joint attention.
- Hide and Seek: HIDE Concepts to SEEK explanations for black box models
Interpreting machine learning model predictions is a central challenge, particularly for black-box vision models.
- Constructive Bounds for the Topological Expressivity of Neural Networks
The expressivity of a neural network $Φ$ describes the class of functions that a network can express.
- Chronos Policy: Learning Robot Policies on Physical Time
Robot policies often convert sensing and control into fixed-rate observation-action sequences, leaving physical time implicit in sequence position.
- AI Models Can Provably Hide Arbitrary Capabilities
Capability evaluations aim to surface dangerous model behaviors before deployment, but their reliability depends on the assumption that hidden capabilities can be elicited.
- Can machines reason non-verbally?
Modern machines can solve equations, follow logic and reason through language, yet they still struggle with problems that require visual non-verbal reasoning.
Releases
69 pieces, the newest 12 here.
- v2026.10.7: Background tool updates at the shell prompt, a verified download cache, and `mise upgrade --global`
- Nuxt 4.6.1: v4.6.1 is the next patch release
- b11539
- CrewAI 1.15.27: Add XPU to OpenCLIP provider device options
- vfox-v2026.10.7
- openai 7.32.0
- OpenAI Python SDK 3.28.0: api, add agent environment suspension and expiration (#4046) (6aa25a9)
- OpenAI Node SDK 7.32.0: api, add agent environment suspension and expiration (#2904) (6e4ed63)
- b11538
- Ruff 0.17.0: Released on 2026-10-09
- uv 0.13.0: Released on 2026-10-09
- OpenAI Codex 0.162.1: Fixed a TUI crash when asynchronous questions contain multiple lines, preserving line breaks and complete…
New Repos & Packages
40 pieces, the newest 8 here.
Watch
36 pieces, the newest 6 here.
- Desktop for Windows, Projects from Chat & TikTok Ads | This Week in Replit (Oct 9, 2026)
- What is a Skill?
- Accelerating Multi-Cloud Workloads with Google Cloud and CoreWeave
- I Gave OpenAI’s Dots Full Control of My Laptop (Here’s What Happened)
- High Performance Serving Framework for Offline and Real-time Diffusion Models - SGLang-Diffusion
- Embed a service catalog for agents directly into your app with the Provisioning API
Package releases
56 from the most used packages' own repositories, the newest 15 here.
- @tanstack/store @tanstack/svelte-store@0.12.3 · every release →
- postcss-discard-duplicates cssnano@9.5.0 · every release →
- ty 0.0.86 · every release →
- redshift-connector v2.2.0 · every release →
- @aws-sdk/types v3.1149.0 · every release →
- llama-cloud v2.18.0 · every release →
- opentelemetry-0.33.1 · every release →
- tox 4.65.0 · every release →
- faker v40.43.0 · every release →
- filelock 4.1.0 · every release →
- p-map v7.1.0 · every release →
- httptools v0.9.0 · every release →
- bare-fs v4.8.4 · every release →
- pydantic-ai-slim v2.55.0 · every release →
- geventhttpclient 2.6.1 · every release →