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.
Tesla integrates Grok Bot into its vehicles for the ultimate personal assistant
teslarati.comTesla has expanded Grok from an in-car chatbot into a hands-free work assistant, officially launching Grok Bot as a personal AI tool.
ByteDance, Runway and Luma Now Split the AI Video Market OpenAI Left Behind
startupfortune.comArticle about Sora's API shutting down on September 24 and how competitors like ByteDance, Runway, and Luma are now splitting the AI video market.
Генерация видео в Gemini с помощью Omni и Veo, лимиты на использование, оплата из России, как создать видео, стоимость подписки, доступ к функциям.
vc.ruRussian article discussing Google's Veo video generation model for Gemini, covering usage limits, payment options from Russia, and how to create videos with the AI tool.
The Turning Point Was 200 Yen per Task: A Thorough Comparison of Claude Opus 5.5 vs. GPT-6 Sol
note.comComparison between Anthropic's Claude Opus 5.5 and OpenAI's GPT-6 Sol models, analyzing pricing at around 200 yen per task and their respective capabilities in the AI model race.
ChatGPT Voice rolls out global update with plugin support and GPT-6 models
cryptobriefing.comOpenAI's ChatGPT Voice updates globally, bringing new GPT-6 Astra, Sol, and Luna models with 50% cheaper API pricing, plugin support, and worldwide availability.
Google's new Gemini TTS models can clone a voice from 30 seconds of audio
thenextweb.comGoogle introduced new Gemini TTS models that design voices from text prompts, offer 2,000+ voices and clone a voice from 30 seconds of audio after consent check.
Nvidia acquires AI platform Hugging Face for $13 billion
eastbaytimes.comEast Bay Times reports that Nvidia Corp. has agreed to acquire AI startup Hugging Face in a transaction valued at about $13 billion, including an equity-based retention program for key employees.
Workload-Driven HBF Substrate For Capacity-Scalable LLM Inference (Huawei, ETH Zurich, HUST)
semiengineering.comSemiconductor Engineering reports on research from Huawei, ETH Zürich, and HUST publishing a technical paper titled FLINT about efficiently leveraging high bandwidth flash for LLM inference acceleration.
MiniMax 开源编程 Agent，实测更强更省钱！
sohu.comMiniMax announced the official open-source release of MiniMax Code CLI under MIT license, with version v0.4.12 released simultaneously. The programming agent is tested to be stronger and more cost-effective for developers.
China's AI firms DeepSeek, Moonshot invited to attend UN Security Council meeting on AI-related risks: report
globaltimes.cnRepresentatives from Chinese artificial intelligence firms DeepSeek and Moonshot are expected to brief the United Nations Security Council on AI-related risks, according to a report. The meeting focuses on international cooperation around emerging AI safety challenges.
Nvidia Just Qualified Tesla and Vertiv Products for Its New AI-Factory Standard
finance.yahoo.comNvidia is extending its influence beyond GPUs into electrical and cooling equipment, qualifying Tesla and Vertiv products for its new AI factory standard.
My Experience Running Chinese Open-Weight Models on Ollama
note.comUser experience running high-performance open-weight models from Chinese research teams and companies using the Ollama local LLM runner.
The monitoring job I set up with pgrep -f detected 147 instances in 11 days. Only 1 was the real thing, and 146 were commands where I was searching for ollama
note.comTechnical analysis of Ollama running in triplicate with llama-server consuming 9.7GB physical memory and 10.4GB swap, investigating runaway instances.
OpenAI makes development moves to counter SpaceX and Meta
thestreet.comOpenAI is developing new agentic capabilities as part of its strategy to compete with other AI companies.
iX-Workshop: KI für eigene Daten: RAG und Embeddings in der Praxis anwenden
heise.deGerman workshop article covering practical implementation of Retrieval Augmented Generation (RAG) and embeddings for developing customized AI solutions with personal data.
Better than DeepSeek: Xiaomi's MiMo-V2.6-Pro debuts as the top open weights model in the world alongside cheaper V2.6 Flash
venturebeat.comVentureBeat report on Xiaomi's MiMo-V2.6-Pro outperforming DeepSeek and other open weights models, with a cheaper V2.6 Flash variant also released in September 2026.
What Happened in September When We Were Told 'Welcome to the AGI Era': A Full Explanation of GPT-6 Astra, ChatGPT Images 2.5, and EU Regulations [September 2026 Update]
note.comAnalysis of OpenAI CEO Sam Altman declaring the AGI era in September 2026, covering GPT-6 Astra launch with advanced computer-use capabilities and EU regulatory context.
My Experience Running Chinese Open-Weight Models on Ollama
note.comUser experience running high-performance open-weight models from Chinese research teams locally via Ollama, discussing practical deployment and performance.
AI Kill Switch: The Race to Control Rogue Frontier AI & AGI Models (Veterans First for America)
pr.cullmantimes.comArticle discussing zero-knowledge privacy shields and efforts to control rogue frontier AI and AGI models, featuring perspectives from veterans' advocacy.
NASC Colloquium: Exploring AI's Influence on Higher-Order Reasoning
colgate.eduColgate University hosts colloquium exploring AI's influence on higher-order reasoning, featuring research from Lupina Foundation Postdoctoral Scholar.