Send us a text🌍 INAI • The Open AI HubThe Intelligence Atlas → the world’s most comprehensive, open hub of AI knowledge. 2 Million+ tools, models, agents, tutorials & daily news—free for all, updated every day.https://github.com/inai-sandy/inAI-wikiTop Highlights: Google expands Gemini AI Search to 200+ regions in 36 languages; Nvidia becomes first $4T company amid GPU scarcity; Google Quantum AI team wins Nobel Prize in Physics; open-source models (Qwen3-VL and GLM-4.6) lead benchmarks; OpenAI reports 800M+ users while Sora 2 launches globally.New Tools: Microsoft unifies agent frameworks with deep Azure integration; Anthropic launches safety auditing tools; Google DeepMind's CodeMender automates security fixes at scale; Granite Docling enables privacy-first document parsing; Hugging Face improves model customization; and multiple platforms expand access to frontier models.LLM Advances: Anthropic's Opus 4.1 outperforms GPT-5 Pro on difficult tasks; open-source models narrow the gap with proprietary systems; ultra-efficient models show impressive reasoning abilities; vision models advance with Tencent leading in China; and video generation capabilities expand with Sora 2.Research: Breakthroughs in distributed training efficiency; transparent safety evaluations from METR; AI advancements in oncology with human oversight still crucial; AI-designed bacteriophages combat drug-resistant bacteria; new benchmarks for embodied AI and reinforcement learning.Industry: Google's AI Search goes global as Gemini 2.5 enhances automation; Nvidia hits record valuation amid infrastructure constraints; Google Quantum AI's Nobel Prize boosts quantum computing legitimacy; Hugging Face sees explosive growth; competitive shifts among Cohere, Anthropic, and Perplexity; OpenAI focuses on apps while Oracle expands GPU offerings.Education: New AI courses from Andrew Ng and Oxford; practical RAG workshops; accessible guides for core AI mechanics; efficiency playbooks for model training; and infrastructure scaling guidance.Demos: Lightweight image understanding with Moondream; LlamaIndex agents automate complex workflows; VideoRAG enables better video comprehension; robotics advances including Tesla Optimus and open-source Reachy Mini; DeepMind's CodeMender demonstrates practical code maintenance.Discussions: Natural language interfaces vs. workflow automation; debates on visual UI limitations; infrastructure strategy considerations; challenges with AI-generated work quality; and broader reflections on AI consciousness and model behavior shaping.Support the show

Ver todos los episodios