Daily AI briefing

AI Briefing: MANGOS Take the Stage as Gemini Shatters Benchmarks

AI Briefing June 13, 2026: The rise of MANGOS stocks, Google's Gemini-SQL2 record, Meta's internal AI revolt, and the G7's AI summit.

  • Briefing date: 2026-06-13
  • Editorial AI analysis
  • The AI Wrap

Stories covered on 2026-06-13

  1. How to Make AI Worthy of Clinician Trust: A Framework That Actually Works

    Towards AI · 2026-06-13

    A new framework proposes a structured approach to building trustworthy AI for healthcare, emphasizing transparency in model development and performance validation.

  2. LLM Observability with LangSmith — Part 2: Eval Gates, Prompt Versioning & Choosing Your Stack

    Towards AI · 2026-06-13

    LangChain's LangSmith platform has introduced features for LLM observability, specifically Eval Gates, prompt versioning, and stack selection tools.

  3. Amazon security research reportedly led to the White House’s Anthropic Fable ban

    The Verge · 2026-06-13

    According to the Wall Street Journal, the export control directive that led to Anthropic cutting off access to Fable 5 and Mythos 5 was triggered in part by cybersecurity research

  4. Police officer investigated for using AI to 'create evidence' in multiple cases

    Hacker News · 2026-06-13

    A police officer is under investigation for allegedly fabricating evidence by using generative AI tools in multiple criminal cases.

  5. Your LLM Needs a Map!

    Towards AI · 2026-06-13

    A recent analysis highlights the critical need for Large Language Models (LLMs) to possess internal, dynamic "maps" or knowledge graphs to improve reasoning and accuracy

  6. KPMG pulls report on AI usage due to apparent hallucinations

    TechCrunch · 2026-06-13

    Once again, AI proves to be an unreliable source of information about AI.

  7. LLM Observability with LangSmith -Part 1: Tracing Everything & Building Audit-Grade Callbacks

    Towards AI · 2026-06-13

    LangChain's LangSmith platform now offers detailed tracing and audit-grade callback capabilities for Large Language Model (LLM) applications.

  8. Terraform MCP Server Enables AI Assistants to Interact with Terraform Infrastructure

    InfoQ · 2026-06-13

    HashiCorp has announced the general availability of the Terraform MCP Server, an open-source MCP server that enables ag

  9. Building a Gemini Live voice app with React, FastAPI and your own WebSocket protocol

    Towards AI · 2026-06-13

    The quickest way to try Gemini Live from a browser is to open a WebSocket straight to Google. For a weekend experiment, that is fine.

  10. Amazon CEO reportedly raised Anthropic model concerns before government crackdown

    TechCrunch · 2026-06-13

    Amazon CEO Andy Jassy may have been the source of security concerns that led Anthropic to cut off worldwide access to two models on Friday.

  11. Apple Wrote a $1 Billion Check to Google Four Days After Settling a $250 Million AI Lawsuit

    Towards AI · 2026-06-13

    the $1 billion Google Gemini deal just shifted Apple’s AI strategy from building to renting. No, that doesn’t mean they surrendered. Here…

  12. How to Build a QwenPaw Agent Workspace with Custom Skills, Model Providers, Console Access, and Streaming API Testing

    MarkTechPost · 2026-06-13

    In this tutorial, we implement a QwenPaw workflow that provides a practical environment for building and testing an agent-powered assistant.

  13. Larger Context Windows Don’t Fix RAG — So I Built a System That Does

    Towards Data Science · 2026-06-13

    Increasing context size in RAG systems doesn’t improve accuracy for aggregation tasks—it makes errors harder to detect.

  14. New AI model called "Count Anything" does exactly what it says, and that's harder than it sounds

    The Decoder · 2026-06-13

    "Count Anything" is intended to be the first AI model capable of counting objects in any type of image, from crowds to cell samples under a microscope

  15. OpenAI faces investigation from state attorneys general

    TechCrunch · 2026-06-13

    It's not clear which states are involved, but they're asking about everything from OpenAI's ad policies to its handling of health data.

  16. Microsoft Now Ships Four MCP Servers for Your Data Stack. Here’s Which One Actually Fits Your Job.

    Towards AI · 2026-06-13

    Power BI Modeling MCP. The remote Power BI MCP. The Fabric MCP Server. The Data Agent MCP. Four official ways to connect AI to your…

  17. Parse PDFs for RAG Locally with Docling: Rich Tables, No Cloud Upload

    Towards Data Science · 2026-06-13

    Enterprise Document Intelligence [Vol.1 #5ter] - Table cells, OCR, captions, headings: cloud-grade structure, running on your own machine.

  18. Solving the 3Blue1Brown String Probability Problem (Without AI)

    Towards Data Science · 2026-06-13

    Let's practice data science thinking through a probability problem

  19. Microsoft CEO Satya Nadella admits he's a token-maxer, too: "It's addictive"

    The Decoder · 2026-06-13

    Microsoft CEO Satya Nadella is warning against "token-maxing," throwing the most powerful AI models at every problem.

  20. Anthropic cuts off Fable 5 and Mythos 5 access following government order

    The Verge · 2026-06-13

    On Friday evening, the government ordered Anthropic to block access to Fable 5 and Mythos 5 for all foreign nations, both inside and outside the US

  21. My yard is dying, so I made an app for that

    The Verge · 2026-06-13

    When I returned to my computer five minutes after giving Gemini a lengthy prompt, I had two things: a functional app in a preview window, and a message about a bug.

  22. Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin

    The Decoder · 2026-06-13

    Google Research's Gemini-SQL2 turns natural language into executable SQL queries. Built on Gemini 3.1 Pro, it tops the BIRD benchmark at 80.04 percent accuracy

  23. A report on the benefits of AI was reportedly full of AI hallucinations

    Engadget · 2026-06-13

    KPMG published a paper about the benefits of AI last year. An investigation found that it was full of AI hallucinations.

  24. Anthropic to disable its most advanced AI models after US order limiting foreign access

    The Guardian AI · 2026-06-13

    Company said US government believes safeguards can be bypassed and product used to identify software vulnerabilities Anthropic said it will “abruptly disable” its most advanced AI

  25. Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file

    The Decoder · 2026-06-13

    Microsoft and three Chinese universities have developed SkillOpt, a method that optimizes instruction documents for AI agents using principles from traditional model training.

  26. Momfluencers are co-parenting with AI. Is it better than a man? | Arwa Mahdawi

    The Guardian AI · 2026-06-13

    Women in heterosexual marriages continue to do most of the caregiving. Now some are offering guides to AI-fying parenting In honour of Pride I’d like to share some important news:

  27. Apple’s new AI photo editing tools mostly work, for better and worse

    The Verge · 2026-06-13

    The most popular camera in the world just got its first set of serious AI photo editing features, and I don't think any of us are ready.

  28. Building a Custom AI Agent with SAP Joule Studio: The Complete Guide Nobody Wrote

    Towards AI · 2026-06-13

    SAP has released a comprehensive guide detailing the construction of custom AI agents leveraging its Joule Studio platform.

  29. UK sets out AI infrastructure push at London Tech Week – how does it stack up?

    The Guardian AI · 2026-06-13

    Government announces plans to invest billions, but questions linger over how its proposals on chips, social media and more will work Ownership of the commanding heights of the AI

  30. Pioneering UK Nerve Lab harnesses AI to map effect of children’s screen time

    The Guardian AI · 2026-06-13

    Other projects include developing tools to help visually impaired people navigate video games Parents are constantly being told to limit their children’s screen time.