4th July - AI News Daily - Hugging Face Revolutionizes Research with FineWeb2: The Multilingual Dataset Changing Global AI

Send us a textNews:https://s.server489.com/AI-2025-07-04Tweet Summarieshttps://s.server489.com/TwAI-2025-07-04Industry Expansion & Open Source: Baidu's ERNIE 4.5, Hunyuan A13B, and DeepMind's AlphaGenome are now open source, boosting global innovation. Cohere expands in Montréal while China launches an Open Source AI Heatmap. LangChain joins top cloud transformers as AI startups achieve record revenue growth. Waymo extends robotaxi service to Boston, and Chris Manning joins AIX Ventures as General Partner. New Tools: Hugging Face released FineWeb2, a multilingual research dataset. Agentic's platform enables monetization of LLM services. Other launches include Higgsfield's Soul Inpaint for image editing, CoreWeave's NVIDIA GB300 NVL72 for inference, AllTracker for video point tracking, VibeKit for secure coding agents, and Mirage for AI-powered game development. LLM Advancements: New models include Baidu's ERNIE 4.5, Arcee's AFM-4.5B-Preview, and Hugging Face's Open-R1, smolagents, and SmolVLM2. Osmosis-Apply-1.7B demonstrates smaller specialized models outperforming larger ones. Polaris 4B now offers MLX compatibility, while Cursor supports local testing of Hugging Face models. Feature Enhancements: Anthropic's API now includes RAG citations, while LlamaIndex integrates Anthropic's tool calling. Cursor added productivity features and local LLM support. The Gemma 3n vision model faces challenges with a MobileNet V5 encoder bug. Learning Resources: The AI Evals FAQ is now available as an enhanced PDF, and Jeremy Howard and John Whitaker are showcasing the SolveIt tool in upcoming courses. Demonstrations: Weaviate's Verba app showcases HIPAA-compliant healthcare AI. Hugging Face provides a testing space for vision-language models, while AllTracker and Mirage demonstrate advances in video tracking and game creation. Industry Discussions: Calls for critique tracks in ML conferences aim to improve scientific rigor. Microsoft's Mustafa Suleyman predicts chatbot-dominated interfaces, while experts debate AI's impact on programming careers and drug discovery. Major Deals & Infrastructure: OpenAI and Oracle finalized a $30B cloud deal with 4.5 gigawatts of computing power. OpenAI is diversifying with Google TPUs alongside Nvidia GPUs and Azure. The company announced a July shutdown to address employee burnout amid fierce talent competition. Enterprise Adoption & Regulation: Organizations like BBVA are deploying AI tools, saving employees hours weekly. Google is integrating Gemini into Workspace and launching Veo 3 for video generation. Legal challenges include Disney's lawsuit against Midjourney, while regulatory frameworks evolve differently in Europe and the US. Research & Security: Studies warn about AI's impact on critical thinking, while research reveals the importance of "super weights" in LLMs. AI-generated student essays and phishing recommendations highlight security concerns, with companies like Cloudflare developing AI-driven defenses. Support the show🌍 INAI • The Open AI Hub The 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-wiki

There are no comments yet.

Please log in to write the first comment.