Welcome to the "AI Daily" column! Here is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers, helping you understand technical trends and innovative AI product applications.
Fresh AI products click to learn more:https://top.aibase.com/
1. Report: Bilibili is about to launch "Project H" AI creation tool, promoting video podcast business
Bilibili is actively expanding its video podcast business and plans to launch a series of support policies, including an AI creation tool called "Project H." This tool aims to help podcast creators save time searching for video materials and editing, improving their creative efficiency. At the same time, Bilibili is expected to attract podcast creators to join during the summer this year, promoting video podcasts as an important growth point.
AiBase Summary:
🎙️ Bilibili will launch the "Project H" AI creation tool, helping improve the efficiency of video podcast creation.
📊 The consumption time of video podcasts reached 2.59 billion minutes in Q1 2025, with user scale exceeding 40 million.
🏙️ Bilibili provides traffic support and free recording venues as part of its support policies to promote content creators' transformation.
2. Zhiyuan Launches "Naozha Robot Lixi X2-N": Dual Form Switching
Zhiyuan's Naozha Robot Lixi X2-N, with its unique dual-form design, demonstrates strong adaptability and flexibility, performing excellently in different scenarios.
AiBase Summary:
🤖 Dual form design, switching between wheel and leg modes, suitable for various scenarios and complex terrains.
⛰️ In leg mode, it has excellent obstacle-crossing capabilities, can climb stairs blindly and carry heavy objects steadily.
🛞 In wheel mode, it achieves efficient movement, with the ability to "move while sliding," easily handling complex terrains like single-bridge and slopes.
3. Yushu Technology Rushes for Sci-Tech Innovation Board IPO, Valued at Billions with Support from Alibaba and Tencent
Yushu Technology is accelerating its rush for the Sci-Tech Innovation Board IPO and has completed a C-round financing of approximately 700 million yuan, with a post-money valuation of 12 billion yuan. This round of financing was led by several industry giants, indicating that its listing process has entered a critical stage.
AiBase Summary:
🚀 Yushu Technology plans to go public on the Sci-Tech Innovation Board (IPO).
💰 Completed a C-round financing of about 700 million yuan, with a post-money valuation of 12 billion yuan.
🤝 The financing team is impressive, including well-known institutions such as China Mobile, Tencent, and Alibaba.
4. Open Source Multimodal Large Model EarthMind: A Revolutionary Tool for Analyzing Earth Observation Data
EarthMind is an open-source multimodal large model designed to efficiently analyze and understand complex earth observation data. It introduces a spatial attention prompt (SAP) module to enhance the accuracy of pixel-level understanding, and through cross-modal fusion and multi-granularity understanding, it effectively integrates and analyzes data from different sensors.
AiBase Summary:
🧠 Introduces the spatial attention prompt (SAP) module to enhance the accuracy of pixel-level understanding.
🔄 Through cross-modal fusion and multi-granularity understanding, EarthMind realizes effective integration and analysis of data from different sensors.
🌍 EarthMind is an open-source multimodal large model specifically designed for processing complex earth observation data.
5. Gemini CLI Major Update! Audio-Video Processing + Privacy Features, Developers' Good News!
The latest version of Gemini CLI brings multiple functional improvements and optimizations, including audio-video processing, enhanced Markdown, upgraded privacy protection, compatibility optimization, and improved stability. These updates further expand its application scenarios, providing developers with a more efficient and flexible working experience.
AiBase Summary:
🎥 New audio-video processing capabilities, expanding tool application scenarios
🔒 Enhanced privacy protection features, making user data control more transparent
⚙️ Compatibility optimization, supporting more editors and cross-platform use
Details link: https://github.com/google-gemini/gemini-cli
6. Invisible AI Desktop Assistant Glass: Open Source Hits Big, Smartly Records Life Moments
Glass is an open-source AI desktop assistant developed by the Pickle team, aiming to become the 'digital brain extension' for users. It is specifically designed for macOS, runs in the background, captures screen activity and audio in real-time, intelligently analyzes and transforms information into structured knowledge, improving work and life efficiency.
AiBase Summary:
✨ Glass is a lightweight and fast desktop tool, specifically designed for macOS, capturing screen activity and audio in real-time.
🧠 Has strong contextual understanding capabilities, transforming scattered information into a practical knowledge base.
🔒 Uses an 'invisible design,' not interfering with user privacy or operational smoothness.
Details link: https://github.com/pickle-com/glass
7. Claude Will Launch Claude Neptune v3 Model, Strong Mathematical Capabilities
Anthropic is testing a new AI model called "Claude Neptune v3," which may be a predecessor to Claude4.5 or a new breakthrough. Currently in the internal red team testing phase, it focuses on testing the robustness of its constitutional AI system and shows excellent performance in mathematical reasoning capabilities.
AiBase Summary:
🔍 Claude Neptune v3 is currently in the internal red team testing phase, focusing on testing the robustness of the constitutional AI system.
🧠 The model shows outstanding performance in mathematical reasoning, possibly comparable to OpenAI's o3Pro and Google's Kingfall models.
🚀 Anthropic plans to optimize the model's context window and tool usage capabilities through Neptune v3 to meet complex task requirements.
8. OpenAI Announces GPT-5 Will Integrate Multiple Models, Achieving a New Breakthrough
OpenAI announced that GPT-5 will integrate multiple models, achieving a new breakthrough. The model is planned to be released in the summer, combining the reasoning capabilities of the O series with the multimodal functions of the GPT series, enhancing overall performance and reducing the need for users to switch between different models.
AiBase Summary:
🧠 GPT-5 will integrate reasoning capabilities and multimodal functions
📅 GPT-5 is expected to be released in the summer
🔄 The new model aims to reduce the need for users to switch between different models
