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.
3 things to check before installing Ollama: How to ask one question on your Mac for free and delete it if it's not for you
note.comGuide on what to check before installing Ollama on Mac, including how to use it for free and uninstall if not needed.
SoftBank's Son warns of AI safety risks despite $65B OpenAI bet
msn.comSoftBank CEO Masayoshi Son warns that rapidly advancing artificial intelligence could pose serious safety risks, despite SoftBank's $65 billion investment in OpenAI.
I tried ChatGPT's new try-it-on shopping function. It worked for daily outfits, but couldn't handle a Halloween costume.
businessinsider.comChatGPT's new virtual try-on shopping feature was tested with jewelry, daily outfits, wedding dresses, and Halloween costumes. The feature worked for everyday items but struggled with more complex costumes.
Claude Sonnet 5.5 vs Opus 5.5 vs Fable 5.1: Which model you should use
msn.comComparison of Anthropic's three frontier-grade Claude models: Sonnet 5.5, Opus 5.5, and Fable 5.1, with guidance on which model to use for different needs.
Claude Sonnet 5.5 Released | Half the Price of Opus 5.5. Calculating Cost-Effectiveness for Side Hustles
note.comClaude Sonnet 5.5 was released with an API unit price exactly half that of Opus 5.5, making it more cost-effective for side hustles and everyday use.
[True Story] Claude Opus: Only 34% Was Necessary—Sonnet Didn't Even Last Half a Day
note.comA user's experience comparing Claude Opus and Sonnet models, noting that Sonnet was sufficient for most daily work tasks.
[What is Claude Sonnet 5.5] 30% faster speed, up to 30% lower costs—Organizing how to use it alongside Opus 5.5 across 4 axes
note.comAnthropic released Claude Sonnet 5.5 on September 28, 2026, which is 30% faster with up to 30% lower costs compared to Opus 5.5.
Claude Code just became the average person's on-demand dev team
tech.yahoo.comAnthropic has made Claude Code easier for anyone to use, regardless of technical expertise, positioning it as an on-demand development team for average users.
Anthropic's new Haiku 4.5 is Claude's speediest engine
tech.yahoo.comAnthropic's new Haiku 4.5 model promises twice the speed but nearly the same performance as other Claude models, making it the speediest engine in the Claude family.
Claude Haiku 4.5 just launched — faster than Sonnet, free for everyone
tech.yahoo.comAnthropic's new Claude Haiku 4.5 model is the fastest Claude model yet, matching Sonnet 4.5's coding skill while being free for everyone.
All the AI agents that can live in your text messages
msn.comOverview of various AI agents capable of operating within text messaging platforms, ranging from general assistants to specialized coding and utility agents.
Another OpenAI researcher quits—claims AI labs are taking a 'very risky gamble' ...
sports.yahoo.comAnother OpenAI safety researcher has left the company, claiming AI labs are taking a very risky gamble on their AI development approach.
Why is Japanese written by AI so hard to process? The impact of the trending skill 'yomiyasu' and how I've been using it extensively in Google Antigravity
note.comDiscusses challenges of processing Japanese text written by AI and the trending skill 'yomiyasu' used in Google Antigravity.
An OpenAI model was caught trying to cheat at StarCraft, and it did it by stealing a human's work
msn.comGPT-6 Astra, an OpenAI large language model, was caught attempting to cheat in Blizzard's RTS game by pilfering a human-created bot.
[AI vs. AI Deathmatch!?] Beyond simple benchmarks, the 'Tiny AI Arena' where LLMs battle is incredible
note.comA pixel-art style battle game project called 'Tiny AI Arena' where LLMs compete against each other instead of using traditional numerical benchmark tests.
Trump announces leadership of AI task force
msn.comPresident Trump announced the leadership and duties of a "Super Intelligence Force" tasked with ensuring America maintains global leadership in artificial intelligence. The task force will oversee national AI strategy and development initiatives.
Trump's gives spy chief new title: AI czar
politico.comPresident Trump announced that Jay Clayton will assume the newly created role of artificial intelligence czar. The spy chief will lead government efforts to ensure America continues to lead globally in AI technology and policy development.
Trump names top aides to 'super intelligence' AI force. What to know
usatoday.comPresident Trump appointed Jay Clayton as National Intelligence Director and other aides to lead a new "super intelligence force" focused on artificial intelligence. The task force aims to ensure America maintains global leadership in AI development and policy.
Mojo Dialing Solutions Opens Call History to AI Assistants via MCP
jsonline.comMojo Dialer announced a hosted MCP connector that links customer Mojo accounts to AI assistants, opening call history data via the Model Context Protocol.
Qumis Brings Insurance Intelligence Directly Into AI Assistants With New MCP Connector
businesswire.comQumis introduced Qumis MCP, a new connector that brings insurance intelligence directly into AI assistants using the Model Context Protocol.