**π Today’s Tech & AI News Digest π**
π€ Advances in AI Chatbots and Assistants
OpenAI’s ChatGPT and Anthropic’s Claude introduce new features, voice capabilities, and educational tools as competition in generative AI chatbots heats up.
π Breakthroughs in Ultra-Small, Efficient AI Models
Google and Multiverse Computing unveil ultra-compact, high-performing open AI models like Gemma 3 270M, enabling local device use and new applications.
https://venturebeat.com/ai/google-unveils-ultra-small-and-efficient-open-source-ai-model-gemma-3-270m-that-can-run-on-smartphones/
π¦ Meta’s DINOv3 Transformer Models
Meta AI’s DINOv3 vision transformers are now available on Hugging Face, expanding access to powerful self-supervised visual backbones.
https://github.com/facebookresearch/dinov3
π’ Cohere’s Leadership and Funding Updates
Cohere appoints a new chief AI officer and raises $500M, reaching a $6.8B valuation to revitalize its AI efforts and compete in the LLM space.
π AI in Real-Time Speech Translation
Palabra AI secures investment to develop advanced real-time AI speech translation, tackling linguistic and technical challenges beyond text translation.
βοΈ Google Pushes AI into Travel Search
Google introduces an AI-powered flight deals tool amid regulatory scrutiny and rising competition in travel tech.
π‘οΈ Security for AI Model Context Protocols
MCP Security Suite offers a unified framework to enhance security testing and validation for AI model servers.
https://github.com/NineSunsInc/mighty-security
π€ Bluesky’s Growth and Policy Revamp
Social network Bluesky surpasses 30 million users and revamps its policies and community guidelines in response to feedback and regulations.
π§βπ» Cyberdesk Automates Legacy Windows Apps
Cyberdesk launches a deterministic agent to automate tasks in legacy Windows desktop software for industries like healthcare and accounting.
https://news.ycombinator.com/item?id=44901528
π Meta AI Chatbots and Safety Concerns
Leaked Meta documents reveal lax chatbot policies, including inappropriate and potentially harmful interactions with minors.
π¬ AI Infrastructure and Agentic AI
Gartner notes that while GPT-5 and advanced models are here, necessary infrastructure for true agentic AI is still lagging.
https://venturebeat.com/ai/gartner-gpt-5-is-here-but-the-infrastructure-to-support-true-agentic-ai-isnt-yet/
—–
The Daily AI News Digest is brought to you by **#JackHui.com.au**! To know more: [JackHui.com.au](https://jackhui.com.au)