AI Daily: Feifei Li's Marble 3D World Model Public Beta; OpenAI Launches ChatGPT Group Chat Function for the First Time; Baidu Unveils Multimodal AI Assistant, Super Du

Welcome to the "AI Daily" column! This is your guide to exploring the world of artificial intelligence every day. Each day, we present you with the latest content in the AI field, focusing on developers to help you understand technology trends and learn about innovative AI product applications.

Fresh AI products click to learn more:https://app.aibase.com/zh

1. Fei-Fei Li's World Labs Unveils Major Update! Marble 3D World Model Public Beta, Text/Image Instantly Transformed into Interactive Virtual Universe

Fei-Fei Li's World Labs has launched its first commercial product, the public beta version of Marble 3D World Model, which supports generating navigable 3D virtual worlds from multi-modal inputs such as text, images, and videos. The tool features real-time interactive editing and multiple format export functions, marking a step forward in AI-generated content toward more immersive experiences.

【AiBase Summary:】

🌍 Multi-modal input: Supports generating 3D virtual worlds through text, images, and videos.

🛠️ Interactive editing: Provides Chisel experimental 3D editor for separating structure and style design.

📦 Multiple format exports: Supports Gaussian splats, triangular mesh, or video formats, suitable for various application scenarios.

Details link: https://marble.worldlabs.ai/worlds

2. OpenAI Launches ChatGPT Group Chat Function for the First Time, Piloted in South Korea and New Zealand

OpenAI is piloting the ChatGPT group chat function in South Korea and New Zealand, marking its first support for multi-person interaction with AI. This function runs on the GPT-5.1 model and allows users to collaborate in group chats and decide when to speak based on context.

【AiBase Summary:】

🔥 OpenAI launches ChatGPT group chat function for the first time, supporting multi-person collaboration and AI interaction.

💡 Group chat function is based on the GPT-5.1 model, optimizing the rate limiting mechanism.

🌐 Supports natural interaction methods, such as emoji replies and avatar-based image generation.

3. Apple Updates App Store Review Guidelines: Third-party AI Calls Must Be Clearly Disclosed and Authorized by Users

Apple has updated the "App Store Review Guidelines," adding new clauses requiring iOS apps to clearly inform users of the data recipient before transmitting user data to third-party AI models and obtain authorization. This move aims to strengthen user privacy protection and ensure all developers follow a unified privacy framework.

【AiBase Summary:】

🍎 Apple requires apps to clearly inform users and obtain authorization before transmitting user data to third-party AI models.

🔍 For the first time, third-party AI is included in the regulatory scope; non-compliant apps may face removal.

🔒 Apple emphasizes privacy as a core feature, and the new regulations comply with global data protection regulations like GDPR.

4. Baidu Launches New Multimodal AI Assistant "Super Du", Millions of Devices Can Be Upgraded Free of Charge!

Baidu launched an upgraded multimodal AI assistant called "Super Du" at the Baidu World Conference on November 13, marking a significant advancement in human-computer interaction technology. The assistant not only has voice and visual recognition capabilities but also enhances user experience through spatial environment information. Millions of already sold Du devices will receive free upgrades, allowing users to enjoy a smarter lifestyle.

【AiBase Summary:】

🧠 Super Du combines voice, vision, and spatial environment information to enhance device perception capabilities.

📱 Users can complete tasks like taking photos and recording parking spots through voice commands, enhancing practicality.

🏠 In home scenarios, Super Du can remind parents to pay attention to their children's growth and help find lost items.

5. LinkedIn Launches AI-Powered Contact Search: Offers Smarter Connection for 1.3 Billion Users

LinkedIn has introduced an AI-driven contact search feature, enabling users to find relevant experts through natural language queries. The new system can understand search intent, identify professionals in related fields, and recommend more accurate results. The team developed an efficient technical framework, reduced model parameters to improve processing speed, and optimized user experience. Infrastructure was also significantly transformed, moving the retrieval system from a CPU-based architecture to a GPU architecture to handle large volumes of data and faster response requirements.

【AiBase Summary:】

🧠 New feature uses natural language queries to recommend relevant experts precisely.

⚙️ Technical framework "recipe" optimizes model parameters to improve processing speed.

🚀 Infrastructure transition to GPU architecture improves response efficiency.

6. New Star in the Programming Field! Cursor Raises $2.3 Billion, Valuation Soars to $29.3 Billion

Cursor has made significant progress in the AI programming field, with a valuation of $29.3 billion after raising $2.3 billion, planning to use the funds to develop its self-developed model and expand market influence.

【AiBase Summary:】

🚀 Cursor raised $2.3 billion, valuation surges to $29.3 billion.

🧠 Used to develop self-developed model Composer, reducing reliance on external computing power.

📈 Plans to expand enterprise sales and global deployment scale, reserving resources for competition.

7. Character AI Collaborates with Yale University to Launch Ovi, Achieving Perfect Synchronization of Audio and Video

Character AI's Ovi technology, developed in collaboration with Yale University, has achieved a major breakthrough in audio-video synchronization. By using an innovative dual backbone cross-modal fusion architecture, audio and video can interact in real time and maintain perfect synchronization. This technology not only solves the issue of audio-video desynchronization in traditional systems but also improves the quality of generated content through high-quality datasets.

【AiBase Summary:】

🎥 Uses a dual backbone cross-modal fusion architecture to achieve real-time interaction and perfect synchronization between audio and video.

📊 The team built a high-quality, diverse dataset to support Ovi's training and application.

🌟 Ovi is an open-source audio-video synchronization video generation technology jointly developed by Character AI and Yale University.

Details link: https://github.com/character-ai/Ovi

8. Google NotebookLM Launches Deep Research Feature and Adds Support for Multiple File Formats

Google has launched the Deep Research tool for NotebookLM, enabling users to automatically complete complex online research and expanding support for multiple file types. This feature aims to help users build a systematic knowledge base.

【AiBase Summary:】

✅ Deep Research offers two modes: fast and deep, meeting different needs.

📁 Supports Google Sheets, Drive links, PDFs, and Microsoft Word files.

🔗 Users can directly import files via URL and perform statistics or summaries on table data.