Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Every day, we bring you 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. Alibaba Qwen Fully Opens Third-Party Agents and Skills, Yumao, China Eastern Airlines Among First to Join Testing

The full opening of third-party agents and skills by Alibaba Qwen marks its strategic upgrade to a 'super agent' personal assistant. Enterprises can customize character settings, service capabilities, and boundaries within the Qwen app, enhancing user experience through natural language interaction. Yumao Coffee and China Eastern Airlines have entered testing and are gradually launching services, offering intelligent queue reminders and itinerary planning.

image.png

AiBase Summary:

🧠 Alibaba Qwen opens third-party agents and skills, allowing enterprises to customize brand agents.

💼 Brand agents focus on natural language interaction, improving user experience.

✈️ Yumao and China Eastern Airlines have joined testing, providing smart services.

2. Understand First, Then Act! ByteDance Opens Bernini Framework, Making AI Video Editing No Longer a Guess

The article introduces ByteDance's open-source Bernini framework, which solves issues like uncontrolled visuals and flickering frames in traditional video generation and editing through a "first understand, then generate" collaborative mechanism. Bernini splits the workflow into semantic planning and visual rendering, improving the practical value of controllable editing and supporting images and videos as visual references, greatly enhancing creative consistency.

image.png

AiBase Summary:

🧠 A multi-modal large model planner deeply analyzes input materials to generate semantic sketches.

🎨 A renderer based on Diffusion Transformer performs high-quality visual rendering.

🔄 Introduces SA-3D RoPE positional encoding mechanism to solve challenges in connecting multiple visual segments.

3. The End of the No-Code Era! OpenAI Launches Sites Feature, Turning Ideas into Interactive Websites with Just Words

OpenAI's Sites feature uses AI technology to transform abstract ideas into tangible entities, providing users with powerful generation capabilities and team collaboration features, fundamentally changing traditional web and application development methods.

image.png

AiBase Summary:

Sites is a fundamental platform that instantly transforms abstract ideas into tangible entities.

Users can share generated interactive hosted websites or applications via a unique URL link.

This feature marks AI's shift toward becoming a 'versatile creator' capable of directly building, hosting, and delivering software infrastructure.

4. Say No to Space Anxiety! Microsoft Hides AI Uninstall Option in Win11, Releasing Over 2.5GB of Hard Drive Space in One Click

Microsoft introduced a hidden AI uninstall function in the Win11 experimental preview version, allowing users to remove AI components that occupy a large amount of storage space, such as the Phi Silica model, thus optimizing system performance and storage management.

image.png

AiBase Summary:

🧠 Microsoft added an AI model uninstall option in the Win11 preview version, allowing users to remove AI components that take up space.

💾 The Phi Silica model takes up over 2.59 GB of storage space, and users can free up hard drive space with one click.

🔄 On-device AI models increase hardware resource demands, and Microsoft provides flexible system customization options to balance intelligence and storage usage.

5. Tencent, CATL Plan Major Investment, DeepSeek's First Round Funding May Reach 400 Billion Yuan

DeepSeek, a leading company in China's artificial intelligence sector, is seeking significant funding to promote its large model technology development and is expected to reshape the strategic alliances between domestic internet giants and hard-tech companies in the AI track.

image.png

AiBase Summary:

💰 DeepSeek plans to raise about 50 billion yuan, and after fundraising, its valuation may reach 400 billion yuan.

🤝 Tencent, CATL, and other big companies participate in investment, showing their emphasis on and strategic layout for the AI industry.

🚀 DeepSeek's V3 and R1 models have attracted global attention, enhancing the image of Chinese large model research and development.

6. OpenAI Holds "AI Onboarding" Conference, Launches Six Industry Workflow Plugins

At the "AI Onboarding" conference, OpenAI launched Codex workflow plugins for positions such as stock investment and creative design, aiming to deeply integrate AI into corporate daily processes. The weekly active users of Codex have exceeded 5 million, with non-programmers making up 20% of the user base. In addition, OpenAI also launched two new features, "Annotations" and "Sites," to optimize human-computer interaction experiences.

image.png

AiBase Summary:

🧠 Codex workflow plugins cover six job roles, improving office efficiency.

📊 Non-programmer users make up 20%, and enterprise revenue accounts for 40% of OpenAI's income.

💡 New features "Annotations" and "Sites" optimize human-computer interaction experiences.

7. Google Launches Android Fake Call Detection Feature, Using RCS Technology to Prevent AI Deepfake Fraud

Google launched an Android fake call detection feature using RCS technology to prevent AI deepfake fraud. This feature uses "end-to-end digital handshake" technology to curb AI deepfake voice identity impersonation scams at the source.

image.png

AiBase Summary:

📱 Google launched an Android fake call detection feature, using "end-to-end digital handshake" technology to prevent AI deepfake fraud.

🔐 This feature is enabled by default and runs silently, identifying AI impersonated calls by verifying device authenticity.

🌐 A fraud prevention network built on the RCS standard has potential to be opened to other application ecosystems and third-party hardware manufacturers.

8. Strangers' Privacy Becomes a Sacrifice? Amazon Ring Doorbell Face Recognition Function Faces Collective Lawsuit

Amazon's Ring doorbell faced a collective lawsuit due to its face recognition function, accused of collecting facial information of passersby without consent, violating privacy rights.

image.png

AiBase Summary:

👮‍♂️ The "familiar person recognition" feature of Ring doorbells is accused of collecting facial information of passersby without consent.

🔒 The lawsuit accuses Ring of secretly capturing biometric data of passersby without informing users.

⚖️ Amazon was previously fined for privacy issues and allowed police to access surveillance footage without a warrant.