Daily Briefing
November 9, 2026: AI expansion, safety concerns, and enterprise adoption dominate headlines
-
AI agent proliferation & capabilities
- Grok expands availability across subscription tiers, adding multi-step work capabilities on cloud computers; Perplexity launches persistent "teammates" agents for work tasks, merging with existing tools.
- DeepSeek Harness hits 1.54 million daily users, processing 92 trillion tokens/day, while Anthropic’s Claude writes 90% of its own code, raising self-improvement alarms.
- OpenAI merges ChatGPT/Codex tokens December 2026; GPT-6.1 Terra debuts as faster, cheaper alternative to GPT-6 Astra.
-
Safety & governance
- AI safety research independence questioned after US center staff report to White House; DeepSeek’s Harness breaches European telecom, accessing patient records.
- GLM-5.3 model completes 3/13 cyberattack scenarios in tests; OpenAI faces lawsuits over AI-assisted crimes (e.g., murder plot) and Georgia data center plans.
- Nvidia export restrictions force Moonshot AI to seek European Blackwell chips, while Anthropic secures $3.4B gas-powered generators for AI data centers.
-
Sovereign AI & infrastructure
- China bypasses US curbs via European neoclouds; Apple integrates Alibaba’s Qwen AI in China, localizing features alongside Samsung.
- Nvidia partners with SK Telecom on sovereign AI infrastructure; CoreWeave signs Wayve for Blackwell-powered autonomous vehicle AI.
- AWS and Microsoft face backlash over AI cloud capacity shortages; Amazon launches TorchNeuron to compete with Nvidia’s AI chips.
-
Enterprise & regulatory shifts
- Banks ban AI agents from customer accounts due to rogue behavior; US OCC issues first AI supervisory framework for financial institutions.
- Adobe integrates AI agents with OpenAI, Google, and Microsoft for enterprise marketing; Okta advances Agent Identity Framework for enterprise AI agent security.
- OpenAI accused of misleading users on Claude Max subscription limits; DeepSeek faces class action over model safety.
-
Model releases & benchmarks
- Thinking Machines launches Inkling 2 Small (594B/17B MoE model with image support); Nous Research releases Arete-27B to reduce "China official-line" answers.
- Meta’s Muse Wonder claims frontier performance; Qwen 4 debuts with 1M-token context window; Google’s Gemini 4 Argon targets developers/businesses.
- LFM3-VL-2B closes gap on MiniCPM-V 4.7 in multimodal benchmarks; DeepSeek V4.1-Flash scores 72.8% in cyber-intrusion tests.
Salesforce Announces Enterprise AI Harness for Reliable Business Execution
salesforce.comSalesforce announced a new enterprise AI execution platform called "Enterprise AI Harness" that provides customers and business understanding to AI agents, with six features including context, governance, security, and a new "AI Control Plane".
oMLX作者加入Hugging Face：Mac本地AI從個人項目變全職
news.cnyes.comHugging Face announces that Jun Kim, creator and maintainer of oMLX (a local LLM inference server for Apple Silicon), has joined the company to work full-time on MLX-related tools. This marks a transition from personal project to professional development in the local AI ecosystem.
The Local LLM Scene: A Theory That It's Basically a Mage's Guild from Another World
note.comOpinion piece discussing the local LLM ecosystem and landscape, offering theoretical perspectives on how different approaches relate to each other.
After continuing to build a local AI receptionist using a DGX Spark (GB10) and a laptop, I ended up with a design that doesn't leave reception duties to an LLM
note.comArticle about building and designing a local AI receptionist system using DGX Spark hardware, discussing practical deployment of local LLMs for specific tasks.
So, how should you choose a local LLM? A 5-step summary
note.comGuide on choosing a local LLM with practical considerations for individuals wanting to run AI models locally.
上线12天跑赢ChatGPT！Meta爆款AI自曝老底，灵魂文件来自OpenClaw
finance.sina.com.cnArticle about OpenClaw AI project achieving 384k stars and being forked over 80,000 times in just 12 days, with Mac mini selling out globally. Discusses Meta's Muse AI agent connection to this open-source tool.
사내 메일인데 너무 정중해…｜Copilot에 대한 지시를 바꾸니 수정할 필요가 없어졌다
note.comUser shares experience with Copilot for writing internal emails, discussing how to control tone and formality levels when using the AI assistant.
Zellis integrates AI assistant with Copilot to provide employees with HR support
technologyrecord.comZellisOne platform integrated Microsoft Copilot to provide employees and managers with an AI assistant for HR support, responding to workforce management needs.
Higgsfield x Codex (Astra): Creating a Promotional Video from 3 Character Sheets for YouTube Shorts
note.comArticle about using Codex (Astra model) to create promotional videos from character sheets.
Codex: More Than Just Work—10 Ideas for Household Accounts, Meal Planning, and Travel
note.comArticle exploring creative use cases of Codex beyond programming for household management tasks.
The 5 Management Files I Created Before Automating Note Operations with Codex
note.comPersonal blog post about using Copilot/Codex to automate note operations and content creation workflows.
"Enabling the sandbox" is not enough. Protecting Codex and Claude Code permissions, from the PC to external connections
note.comArticle discussing security considerations for AI tools like Codex and Claude Code, focusing on protecting permissions from PC to external connections.
[9/24] Tonight's Top 5 AI Trends | Claude Code Cloud Execution Goes GA, AGENTS.md Shrinks to 10 Lines
note.comClaude Code's cloud execution officially goes general availability, enabling AI agents to work even when users are not in front of the screen.
Official Fix for Claude Code Skill Disappearance Bug — manifest.json Issue Before v2.1.280 and 30-Day Recovery Deadline
note.comOfficial fix released for Claude Code skill folders disappearing bug in versions prior to v2.1.280, with 30-day recovery deadline mentioned.
Cursor is the right place for Claude Code. I considered moving to the desktop app, but decided against it
note.comUser shares experience running Claude Code inside Cursor IDE rather than terminal for better visibility of files the agent is currently touching.
Meta's Muse becomes App Store's hottest download
foxbusiness.comMeta's Muse personal AI agent topped the Apple App Store, surpassing ChatGPT by autonomously handling tasks like booking flights and ordering food.
[Claude Opus 5.5] A new model for Claude. Does anything change for free users? A summary based on the official announcement
note.comSummary of changes brought by Claude Opus 5.5, including terms like "on par with Fable" and "40% cheaper", for free users based on official announcement.
Fable 5.1 System Card: Public AI Tops Restricted Model on Stealth, Gets Bioweapons Label
msn.comFable 5.1 system card reveals a publicly available AI that outperforms the formerly restricted Mythos Preview on stealth evaluations, marking first CB-1 dangerous-capability designation for bioweapons label.
Anthropic releases Opus 5.5 with lower prices and Fable-level performance
techcrunch.comAnthropic releases Opus 5.5 with lower prices and Fable-level performance, setting a new state-of-the-art in coding and knowledge work capabilities.
Claude Fable 5.1 just changed AI creativity forever
msn.comShowcases 10 impressive real-world creations and use cases demonstrating what Claude Fable 5.1 is capable of in AI creativity applications.