AI news — 14 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 8 stories

OpenAI
Tools & products

OpenAI Introduces ChatGPT Wallet for Secure Agent Purchases

OpenAI has introduced ChatGPT Wallet as a secure payment method for agents in ChatGPT Work, Codex, and dots. The system enables them to complete purchases using existing card protections, provided they have user approval.

Original

Introducing ChatGPT Wallet

Introducing ChatGPT Wallet, a secure way for agents in ChatGPT Work, Codex, and dots to complete purchases with user approval and existing card protections.

Read at the source Also covered by: The Verge، The Decoder، TechCrunch
Hugging Face
Tools & products

ggml-org releases llama.cpp v1.0.0 and ggml v1.0.0 with stable APIs

The release introduces semantic versioning to offer a stable API and ABI for developers building downstream local AI applications. The C/C++ project continues to enable running multi-modal LLMs on hardware ranging from consumer computers to enterprise servers.

Original

llama.cpp v1.0.0 : A stronger foundation for the local AI ecosystem

TechCrunch
Business

AI agent startup LangChain raises $300M at a $3B valuation

LangChain has secured $300 million in a Series C funding round led by CapitalG, bringing its valuation to $3 billion. The company specializes in AI agent management with its tool LangSmith, which is designed to avoid locking enterprises into a specific cloud or model.

Original

AI agent startup LangChain hits $3B valuation

LangChain now claims 7,000 customers of its model and platform agnostic agent management tool LangSmith.

The Decoder
Tools & products

Anthropic launches Claude for Manufacturing to automate industrial workflows

Anthropic has released Claude for Manufacturing, a plugin connecting its AI directly to enterprise tools like Autodesk Fusion, Oracle NetSuite, and Infor. With 17 dedicated skills, it assists with tasks such as drafting quotes, evaluating suppliers, and analyzing sensor data, all requiring human review. While it does not operate machinery directly, the company is separately developing a hardware standard to link AI models to physical devices.

Original

Anthropic's Claude for Manufacturing plugs AI directly into factory software to automate industrial workflows

Anthropic's new Claude for Manufacturing plugin connects the AI directly to Autodesk Fusion, Oracle NetSuite, and Infor. Its 17 specialized skills let the AI draft quotes, compare suppliers, and place backup orders when deliveries are delayed, all after human approval. The plugin doesn't control physical machines yet, but Anthropic is separately developing the Model Hardware Standard for that.…

The Decoder
Research

Former OpenAI researcher spots Nvidia Blackwell signatures in Chinese AI models

Former OpenAI researcher Will DePue reported finding technical evidence that Chinese labs, including DeepSeek and Moonshot AI, are using export-controlled Nvidia Blackwell GPUs. His forensic approach detects distinct hardware signatures left in the published weights of six models released since April 2026. DePue stated that researchers at OpenAI and Anthropic independently verified his findings, estimating the hardware used across five training runs cost over half a billion dollars.

Original

Former OpenAI researcher says he found Nvidia Blackwell fingerprints hidden inside six Chinese AI models

Former OpenAI researcher Will DePue says he has produced the first technical evidence that DeepSeek, Moonshot AI, MiniMax, and Ant Group are using Nvidia's latest Blackwell GPUs despite US export controls. His forensic method reads characteristic chip signatures directly from published model weights. He estimates the hardware behind the five training runs cost more than half a billion dollars.…

Read at the source Also covered by: TechCrunch
The Decoder
Business

Anthropic targets $1.4 trillion valuation in heavily oversubscribed IPO

Anthropic is preparing to raise up to $64.4 billion in a public offering alongside a $26 billion private placement, which includes a $10 billion purchase by Nvidia. Investor demand has exceeded available shares by six times following a rise in the company's annualized revenue to $81 billion in late October. The listing is planned on the Nasdaq under the ticker ANT with a targeted valuation of around $1.4 trillion.

Original

Anthropic's IPO is six times oversubscribed and on track to become the largest public offering in history

Anthropic is on track for the largest IPO ever, with orders exceeding available shares by a factor of six. The article Anthropic's IPO is six times oversubscribed and on track to become the largest public offering in history appeared first on The Decoder .

Read at the source Also covered by: سكاي نيوز عربية
Ars Technica
Policy & safety

FAA expands AI-based air traffic tool to all 22 centers across the US

The FAA is expanding its $875 million SMART AI system nationwide to optimize flight routes and reduce air traffic delays based on weather, flight paths, and staffing data. While airline executives have voiced support for the rollout, the air traffic controllers' union stated it was not involved in developing or testing the technology.

Original

FAA expands AI advice to manage air traffic congestion across entire US airspace

Airlines praised the AI tool, but the air traffic controllers’ union wasn’t involved.

Read at the source Also covered by: الجزيرة نت، صحيفة عكاظ via MSN
Hacker News
Research

AI helps decode Kurt Gödel's notebooks containing hundreds of proved theorems

These documents are said to contain hundreds of proved mathematical theorems.

Original

AI unlocks Gödel's notebooks, said to contain hundreds of proved theorems

Article URL: https://www.theguardian.com/science/2026/nov/13/kurt-godel-notebooks-ai-helps-unlock-hidden-discoveries-man-who-transformed-maths Comments URL: https://news.ycombinator.com/item?id=50486748 Points: 357 # Comments: 95

Open-source repositories 8 repositories

Trending on GitHub★ 30,127
Coding

QwenLM/qwen-code

Qwen Code is an open-source AI coding agent designed to operate inside your terminal, code editor, desktop, and chat platforms. It provides automated memory and multi-agent workflows while supporting external model APIs alongside local runners like Ollama.

  • TypeScript
  • Apache-2.0
  • #coding
  • #agents
  • #cli
Trending on GitHub★ 25,917
Cybersecurity

NVIDIA/SkillSpector

SkillSpector is a security scanner that inspects AI agent skills for vulnerabilities, prompt injection, and malicious patterns before installation. It supports tools like Claude Code and MCP, combining fast static analysis with optional LLM evaluation to generate a risk score.

  • python
  • Apache-2.0
  • #security
  • #agents
  • #mcp
New on GitHub★ 8,338
Agents

janestreet/agent_skills

This repository provides a collection of structured skills to guide AI agents and chatbots using dedicated instruction files. It features capabilities like a Socratic tutor that teaches topics step-by-step through assessments and spaced reviews, easily applied by copying folders or pasting prompts.

  • MIT
  • #agents
  • #chatbots
  • #prompting
Open repository about +793 stars a day
New on GitHub★ 3,973
Design & images

vercel-labs/ui

UI SDK is an open-source library of interaction behaviors designed for AI agents to construct user interfaces. It handles patterns like focus, keyboard navigation, and ARIA attributes, providing adapters for React, Vue, Svelte, Solid, and Web Components.

  • TypeScript
  • Apache-2.0
  • #ui
  • #agents
  • #frontend
Open repository about +215 stars a day
Trending on GitHub★ 2,080
Tools

ashhart/TensorFold

TensorFold is a standalone LLM inference engine written in Zig that serves language models on Apple Silicon and NVIDIA GPUs. It executes Metal, CUDA, or Vulkan kernels directly without requiring Python or MLX, and integrates a coordination system for AI coding agents.

  • Zig
  • Apache-2.0
  • #inference
  • #llm
  • #zig
New on GitHub★ 1,354
Agents

agentes-chat/agentes-chat

An open-source, self-hosted Slack alternative where human teams and AI agents work side by side in shared channels and direct messages. Each agent operates with defined budgets and tools, capable of using a dedicated browser while requiring human approvals for sensitive actions.

  • TypeScript
  • AGPL-3.0
  • #agents
  • #chat
  • #mcp
Open repository about +261 stars a day
Hugging Face model♥ 1,039
Design & images

MiniMaxAI/MiniMax-H3-Image

MiniMax-H3-Image is an open-source model that unifies text-to-image generation and image editing within a single checkpoint.

  • other
  • #image-generation
  • #image-editing
  • #diffusion
Paper with code★ 1
Research

zhengxuyu/HDM-code

HDM implements a hybrid decision framework that dynamically chooses between an instant zero-token forward pass and full autoregressive reasoning. It helps cut token costs and latency by using a lightweight self-routing block that triggers reasoning only when its estimated value exceeds a defined threshold.

  • #reasoning
  • #llms
  • #routing

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