AI News Top 10 — Sunday 21 June 2026
- Raspberry Pi AI HAT+ 2 — Generative AI on Pi 5
Raspberry Pi has released the AI HAT+ 2, specifically designed for generative AI on the Raspberry Pi 5. Equipped with a Hailo AI accelerator, it enables hobbyists and homelabbers to run LLMs and image generation models locally on an affordable platform.
Raspberry Pi
🏷️ Relevant for: homelab, Hermes - US Export Ban on Anthropic’s Mythos and Fable AI Models
The US government has forced Anthropic to disable Fable 5 and Mythos 5 for foreign users due to national security concerns. These were Anthropic’s most powerful models ever, now subject to White House negotiations.
Forbes | Al Jazeera
🏷️ Relevant for: AI Industry - Apple Unveils ‘Siri AI’ — Brand New AI Assistant at WWDC 2026
Apple has revealed a fundamentally redesigned Siri at WWDC 2026, powered by Apple’s own LLM. The Wall Street Journal calls it ‘genuinely good’ after a week of testing. Siri AI is coming to iPhone, iPad, Mac, and Apple Watch with watchOS 27.
Apple | CNN
🏷️ Relevant for: AI Assistants - MyHeritage Launches Scribe AI — Genealogy AI Tool
MyHeritage has launched Scribe AI, an AI tool that transcribes, interprets and provides insights on historical documents and photos for family history research. FamilySearch and RootsTech are also presenting guidelines for responsible AI use in genealogy.
MyHeritage Blog | New York Post
🏷️ Relevant for: Genealogy - Trump Signs AI Executive Order for AI Innovation and Security
A new AI policy that both stimulates innovation and establishes cybersecurity guidelines for frontier AI models, including deployment on classified networks through deals with major tech companies.
Politico | Hogan Lovells
🏷️ Relevant for: AI Policy - Norway Imposes Near-Total Ban on AI in Primary Education
Norway has become one of the first countries to introduce strict rules that virtually ban AI in primary schools, citing concerns about privacy, learning development, and replacement of human contact.
Reuters
🏷️ Relevant for: AI & Society - Tech Workers Consciously Reducing AI Use After ‘Maxed Out’ Experiment
A notable trend: tech workers who have used AI to its fullest are now consciously choosing to minimize their usage. They report that excessive AI reliance leads to sloppiness and reduced critical thinking.
The New York Times
🏷️ Relevant for: Hermes, AI Workflows - Nvidia’s Jensen Huang: ‘Society Needs New Social Norms for the AI Era’
The Nvidia CEO advocates in an AP interview for new societal agreements around AI usage, including rethinking work, education, and privacy.
AP News
🏷️ Relevant for: AI Future - ‘World First’: Vaccine Fully Designed by AI
An AI system identified a unique protein structure that addresses a broad range of virus variants — something that would have taken human researchers years to accomplish.
BBC
🏷️ Relevant for: AI Breakthroughs - Building a Self-Hosted AI Gateway with LiteLLM and Open WebUI
A practical guide to setting up a fully self-hosted AI gateway with LiteLLM as the model router and Open WebUI as the interface — perfect for homelabbers managing multiple LLM providers.
HackerNoon
🏷️ Relevant for: homelab, Hermes
Other Notable Items
- Bernie Sanders Unveils Plan for Public Ownership of AI Companies (AP News)
- DeepSeek V4 Preview Released (CNBC) — Chinese AI arms race intensifies
- OpenClaw — Viral AI Agent Software with Critical RCE Vulnerabilities (The Hacker News)
- Meta’s ‘AI Mode’ on Facebook (TechCrunch)
- Google TurboQuant: 6x Less LLM Memory via AI Compression (Ars Technica)
- G7 Summit: World Leaders + OpenAI, Anthropic, Google Together on AI (CNBC)
- AI Breakthrough: 100x Energy Savings with Higher Accuracy (ScienceDaily)
This is an automated daily roundup. Generated 21 June 2026.
Wat vond je van dit bericht?
