AI news — 13 November 2026

Every morning we collect the most important AI news from dozens of sources and sum it up in a few lines, with the newest open-source projects beside it. Each card links to the original source.

RSS · Today's newsRSS · New repositories

News 7 stories

The Decoder
Business

Jensen Huang Resigns as Nvidia CEO Following Espionage Allegations

Jensen Huang has stepped down as CEO, president, and board member of Nvidia following reports of an extramarital affair with a suspected Chinese spy, an issue that erased roughly $1.4 trillion in the company's market value. CFO Colette Kress is serving as interim CEO while the board searches for a permanent successor. Huang's sudden departure poses a major leadership challenge for Nvidia, as he left no clear succession plan and had more than 60 employees reporting directly to him.

Original

Jensen Huang resigns as Nvidia CEO after espionage scandal wipes out $1.4 trillion in market value

Jensen Huang resigned as CEO, president, and board member of Nvidia effective immediately, according to a statement on November 12, 2026. The article Jensen Huang resigns as Nvidia CEO after espionage scandal wipes out $1.4 trillion in market value appeared first on The Decoder .

Read at the source Also covered by: TechCrunch، WIRED، فرانس 24، The Verge، سكاي نيوز عربية
The Decoder
Tools & products

Anthropic launches Claude Meetings to bring AI into Google Meet calls

Anthropic has released a beta feature called Claude Meetings for Team and Enterprise subscribers, allowing its AI to join Google Meet sessions directly. Claude can answer questions, take notes, and generate live documents and whiteboards during the call. Once the meeting concludes, participants can access all transcripts and generated materials via a dedicated page on claude.ai.

Original

Claude Meetings lets Anthropic's AI join your video calls

Anthropic is adding Claude directly to video calls. Claude Meetings joins Google Meet sessions and creates documents, whiteboards, and meeting notes in real time. The feature launches in beta for Team and Enterprise plans, with more platforms coming later. The article Claude Meetings lets Anthropic's AI join your video calls appeared first on The Decoder .

Ars Technica
Tools & products

Google and Samsung partner with eyewear brands on Gemini smartglasses

Google has collaborated with Samsung, Warby Parker, and Gentle Monster to release new smartglasses powered by Gemini. Preorders start at $395 for Warby Parker styles in the US and $449 for Gentle Monster designs in the US and South Korea.

Original

Google pushes privacy angle as Gemini-powered smartglasses debut starting at $395

The first Gemini smartglasses are a collab between Google, Samsung, Gentle Monster, and Warby Parker.

The Decoder
Research

Google DeepMind unveils AlphaGenome 2 to analyze DNA at single-cell resolution

Google DeepMind has introduced AlphaGenome 2, an AI model designed to predict how genetic variants affect specific cell types rather than whole tissues. Trained on more than 107 million cell profiles from 68 tissues, it identifies nearly twice as many disease-relevant gene changes as its predecessor. The model is available for noncommercial research through an API and will be integrated into Google's Antigravity platform.

Original

Google Deepmind shifts genomics AI from bulk tissue predictions to single cell type resolution with AlphaGenome 2

Google Deepmind has unveiled AlphaGenome 2, an AI model that analyzes genetic variants for individual cell types rather than across bulk tissue. Trained on more than 107 million cell profiles, it identifies nearly twice as many disease-relevant gene changes as its predecessor. The model is available for noncommercial research through an API and is set to be integrated into Google's agentic…

Hugging Face
Models & releases

IBM Granite Introduces LLMON Markup Language in Granite 4.1 3B and 8B Models

IBM Granite developed LLMON, a markup language designed specifically to provide structured and unambiguous context to large language models. The format uses special tags and unique identifiers to reduce model errors and help mitigate prompt injection attacks. IBM has released two experimental preview models derived from Granite to demonstrate this system.

Original

LLMON: An LLM-Native Markup Language, now in Granite 4.1 3B and 8B

MIT Technology Review
Policy & safety

Nvidia Privately Confirmed AI Chips Can Be Tracked to Prevent Smuggling

Nvidia representatives told US lawmaker Bill Foster that the company could technically implement location tracking on advanced AI chips to prevent them from being smuggled to China. Although the Semiconductor Industry Association publicly called such measures potentially infeasible, Nvidia privately acknowledged they were doable. The proposal is part of the Chip Security Act, which is gaining renewed attention following the resignation of Nvidia CEO Jensen Huang.

Original

Nvidia privately confirmed it could track chips to fight smuggling, even as industry publicly questioned feasibility

Representatives from Nvidia told Illinois representative Bill Foster that it would be technically possible to implement the anti-smuggling measures called for in the Chip Security Act, Foster told MIT Technology Review in an interview on Thursday, even while an industry group claimed they were “potentially infeasible.” Foster, a Democrat who serves on the Committee on…

Open-source repositories 8 repositories

Trending on GitHub★ 12,679
Cybersecurity

openai/codex-security

Codex Security is a CLI and TypeScript SDK that scans code repositories to detect, investigate, and fix security vulnerabilities. It can automatically generate patches, verify fixes, and integrate with CI workflows or platforms like Jira and GitHub.

  • TypeScript
  • Apache-2.0
  • #security
  • #code-analysis
  • #cli
Trending on GitHub★ 11,638
Coding

shadcn/improve

This agent skill audits codebases using capable models to identify issues and generate detailed implementation plans. It helps optimize costs by using expensive models solely for analysis and specifications, leaving the actual execution to cheaper models.

  • MIT
  • #agents
  • #coding
  • #code-review
New on GitHub★ 3,536
Coding

expo/exact

Exact is a cross-platform application runtime that enables developers to write apps once and run them natively on the web, macOS, iOS, and Linux. It is designed so AI coding agents can build, run, inspect, and test applications across all supported platforms without relying on webviews.

  • Rust
  • MIT
  • #cross-platform
  • #agents
  • #rust
Open repository about +314 stars a day
Hugging Face model♥ 2,166
Models

facebook/dinov3-vitl16-pretrain-lvd1689m

This is a pre-trained Vision Transformer model designed for image feature extraction using the DINOv3 architecture. It is built for computer vision tasks and works seamlessly with the Hugging Face Transformers library.

  • other
  • #vision
  • #transformers
  • #feature-extraction
Trending on GitHub★ 1,774
Audio & video

sgl-project/sglang-omni

SGLang-Omni is a high-performance serving runtime designed for interactive multimodal and speech AI systems. It supports full-duplex interaction, text-to-speech, and audio generation, while also powering tools like Voxt for fast local voice input on Mac.

  • python
  • Apache-2.0
  • #speech
  • #audio
  • #serving
New on GitHub★ 1,572
Tools

RunanywhereAI/trnrt

TrnRT is an open-source inference engine designed to serve large language models on AWS Trainium with an OpenAI-compatible API. It features custom NKI megakernels for decode performance and supports context windows of up to one million tokens.

  • Python
  • Apache-2.0
  • #inference
  • #llm
  • #aws
Open repository about +116 stars a day
New on GitHub★ 1,037
Tools

wici-ai/inferssd

InferSSD builds on llama.cpp to run MoE models larger than your RAM or VRAM by streaming experts directly from an SSD. It works with standard GGUF files and includes a browser chat interface alongside an OpenAI-compatible API.

  • C++
  • Apache-2.0
  • #inference
  • #moe
  • #llm
Open repository about +47 stars a day
Hugging Face model♥ 502
Models

NousResearch/Arete-27B

It cuts propaganda-containing answers from 22% to 0.9% while maintaining general-capability and agentic benchmark performance.

  • apache-2.0
  • #llm
  • #fine-tuning
  • #alignment

Summaries are written automatically from the sources' headlines and excerpts and may contain mistakes. The full story is always at the source. To have a link or source removed, write to us from the contact section.

Let's start

Have an idea or a system that needs transforming?

Write us a few lines. Your message reaches us instantly, and we usually reply within one business day.

scouthub@lab: ~/contact

scouthub@lab:~$ contact --new --lang=en