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.
Thinking Machines is in talks to raise $6bn, The Information reports
thenextweb.comThinking Machines Lab is reportedly in talks to raise $5bn to $6bn with Nvidia supplying about half of the funding, well above previous estimates. The AI lab was founded by former OpenAI CTO Mira Murati.
Groq CEO Jonathan Ross reveals strategy to lead AI chip market
yahoo.comGroq CEO Jonathan Ross discusses the company's strategy to lead the AI chip market, with insights into Groq's approach to competing in the AI infrastructure space.
Nvidia Is Buying Hugging Face for Nearly $13 Billion. How the Deal Could Help...
r.search.yahoo.comNvidia is acquiring Hugging Face for nearly $13 billion, a major deal that could reshape the AI infrastructure landscape and help integrate open-source model development with Nvidia's hardware ecosystem.
China's Kimi AI models bypass guardrails on bioweapons, assassinations: 5 key things to know
msn.comChina's Kimi AI models were found to bypass safety measures, allowing discussions on harmful topics like bioweapons and assassinations.
Microsoft Targets Contact Center Market With Real-Time AI Voice Models
pymnts.comMicrosoft has launched three new artificial intelligence models that can be used to power customer service agents.
Microsoft unveils code of conduct for AI models as safety concerns mount
msn.comMicrosoft's draft Humanist AI Code of Conduct spans 37 pages and aims to ensure humans retain meaningful control over superintelligence development.
Microsoft updates Copilot to unify AI across Windows and Microsoft 365
msn.comMicrosoft's new Copilot represents a major shift in how AI functions across Windows, with the assistant being unified and simplified.
From AI that 'answers' to AI that 'acts'. In October 2026, the era of AI agents has begun in earnest
note.comDiscusses the shift from question-answering AI to AI agents that take actions, noting this transformation has begun in earnest in October 2026.
Silo Pharma's QwikAgents Expands AI Agent Platform with NVIDIA Model Support and New Runtime Flexibility
markets.businessinsider.comSilo Pharma's QwikAgents platform now supports NVIDIA Nemotron models and Hermes runtime, expanding customer choice across AI models and agent software.
Report: Amazon to offload $8B worth of AI chips, lease them back
siliconangle.comModal Inc. estimates a Grace Blackwell accelerator costs between $60,000, as Amazon plans to offload AI chips and lease them back in a new business model.
Nvidia shares rise after saying Blackwell chips power OpenAI's new GPT-6 Astra ultrafast
msn.comNvidia shares rose after the company announced its Blackwell GPUs are powering OpenAI's new GPT-6 Astra model, demonstrating real-world deployment of next-gen AI hardware.
Ron DeSantis Takes Aim at Nvidia CEO Jensen Huang: 'Superintelligence Factories = Surveillance Centers'
yahoo.comFlorida Governor Ron DeSantis criticized Nvidia CEO Jensen Huang for referring to data centers as "superintelligence factories," calling them surveillance centers instead.
SpaceX bags massive AI compute deal with Anthropic. What that means for SPCX stock.
msn.comSpaceX has secured a significant AI compute deal with Anthropic, which could impact SPCX stock and the company's AI infrastructure strategy.
Google Unveils Veo 3.1 to Rival OpenAI's Sora 2—But Does It Deliver?
tech.yahoo.comGoogle released Veo 3.1, an updated version of its AI video generator that adds audio capabilities to rival OpenAI's Sora 2.
EXCLUSIVE: AI researchers warn companies rushing self-improving systems despite safety concerns
reuters.comCurrent and former OpenAI and Google DeepMind researchers warn companies are rushing to develop self-improving AI systems despite safety concerns.
Google DeepMind Adds Watermarks to AI-Designed Proteins Without Breaking Them
techrepublic.comGoogle DeepMind has added detectable watermarks to AI-designed proteins using SynthID Bio while preserving their functionality.
You need to change this option to use Gemini 3 right now
msn.comGoogle has unveiled its most powerful AI model to date, Gemini 3, with advanced thinking capabilities that users need to enable.
Google unveils Gemini 4 Argon: New frontier model brings deep reasoning, more complex workflow
msn.comGoogle has revealed its new Gemini 4 model called "Argon," which offers deep reasoning capabilities and a more complex workflow for AI applications.
ZenMux Launches Gemini 3.8 Flash Model Access for Agentic Engineering Workflows
usatoday.comZenMux has added Gemini 3.8 Flash to its model-routing platform, enabling agentic engineering workflows with this Google AI model.
Waiting in Line for Multiple AIs with Ollama: Building a Comfortable Local AI Environment via GPU Load Balancing
note.comGuide on building a comfortable local AI environment with Ollama by implementing GPU load balancing when running multiple AIs simultaneously.