%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Welcome to the "AI Daily" section! This 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 to help you understand technology trends and innovative AI product applications.
Fresh AI products Click to learn more:https://app.aibase.com/zh
1、CapCut launches two major AI features: Canvas-style AI production studio Video Studio + embedded AI Video
CapCut launched two AI functions, Video Studio and AI Video, combined with the Dreamina Seedance2.0 model, improving video creation efficiency and quality.
AiBase Summary:
🎥 Video Studio: Uses an infinite canvas workflow to simplify the video creation process.
📹 AI Video: Embedded in traditional editors, providing instant material generation functionality.
🎬 Dreamina Seedance2.0: Multimodal generation technology, supporting high-quality video output.
2、Spending 4 billion in one quarter! Kuaishou's Q4 financial report is impressive: AIGC marketing materials dominate, AI completely reshapes the business chain
Kuaishou's Q4 financial report shows that AI technology has made a significant direct contribution to commercial revenue, with AIGC marketing materials dominating, and AI completely reshaping the business chain.
AiBase Summary:
🧠 Model advantage: Generative recommendation large model and intelligent bidding model optimization, increasing domestic online advertising revenue by about 5%.
🎨 Content productivity: AIGC marketing materials consumption reached 4 billion yuan in one quarter, showing advertisers' high recognition of AI content.
⚙️ Automation penetration: UAX automatic placement product penetration rate in non-e-commerce marketing services approaches 80%.
3、ByteDance opens up DeerFlow2.0: Building a "Chinese version" of the super agent orchestration framework
ByteDance has open-sourced its super agent orchestration framework DeerFlow2.0, which quickly gained popularity on GitHub and received a lot of attention. DeerFlow2.0 has core features such as deep integration of multidimensional capabilities, wide compatibility, and a secure sandbox file system, capable of meeting the needs of enterprise-level complex tasks and personal multi-step creation.
AiBase Summary:
🧠 Deep integration of multidimensional capabilities: DeerFlow2.0 is a SuperAgent scheduling center integrating various capabilities, supporting complex task decomposition and efficient collaboration.
🔧 Wide compatibility and ready-to-use: The framework supports mainstream multiple models and can seamlessly connect with MCP protocol and mainstream IM channels, suitable for enterprise-level complex tasks and personal multi-step creation.
🔒 Secure sandbox file system: Provides an isolated operating environment, supporting secure code generation, batch file reconstruction, and result retention.
More details: https://github.com/bytedance/deer-flow
4、Musician will be "unemployed"? Google DeepMind releases Lyria 3 Pro: AI can now independently compose complete long songs
The article introduces Google DeepMind's Lyria 3 Pro, which has achieved breakthroughs in the audio field from "short musical phrases" to "full song creation." It has structural awareness, capable of generating complete song structures, and supports high-fidelity output and multimodal interaction, marking that AI is evolving from an assistant tool to an independent producer.
AiBase Summary:
🎼 Lyria 3 Pro can arrange complete song structures, including intro, verse, chorus, and bridge.
🔊 Supports 24-bit high-fidelity output, meeting professional audio production needs.
🧠 Multimodal interaction allows users to quickly generate music that matches emotions and styles through text descriptions.
5、OpenAI tests new model "Spud": Will shut down Sora to integrate computing power, and transform into a desktop-level "super application"
OpenAI announced that the next-generation AI model, codenamed "Spud," has completed pre-training, with powerful performance, and is expected to be released within a few weeks. At the same time, the company is undergoing strategic contraction and organizational restructuring, shutting down Sora to integrate computing power, and plans to develop a desktop-level "super application" that integrates ChatGPT, Codex, and Atlas to cope with market challenges.
AiBase Summary:
🧠 The new generation AI model "Spud" has completed pre-training and has powerful performance, and will be released soon.
🔄 OpenAI is undergoing strategic contraction, shutting down Sora to integrate computing power, and transforming into a desktop-level "super application."
💼 To cope with the challenge from Anthropic, OpenAI plans to build a unified interactive center integrating ChatGPT, Codex, and Atlas.
6、Sweeping 11 list championships! Ant Group releases F2LLM-v2: "Hexagon" embedding model with full size and multilingual support
Ant Group and Shanghai Jiao Tong University jointly released the F2LLM-v2 series of Embedding models, which demonstrated outstanding performance in the MTEB evaluation, covering various languages and code fields, and provided developers with an efficient solution through full open source.
AiBase Summary:
🚀 F2LLM-v2 swept 11 list championships in the MTEB evaluation, demonstrating strong performance.
🌐 Supports 282 natural languages and more than 40 programming languages, achieving global coverage.
🔧 Provides a full-size model family ranging from 80M to 14B, suitable for various scenario needs.
7、DingTalk Wukong AI officially released: Enterprise-level "digital employee" with double-click usage
The release of DingTalk Wukong AI marks a new era of more convenient and secure enterprise-level AI applications, redefining the standard of office AI with its simple and easy-to-use characteristics.
AiBase Summary:
🧠 Wukong AI realizes enterprise-level AI deployment through simple operations, lowering the technical barrier.
🔒 Emphasizes data security and privacy protection, addressing enterprises' concerns about AI.
💡 Provides an intuitive "computational grain" counter, allowing users to better understand AI resource consumption.
8、Performance beats opponents ten times its size: Apple releases RubiCap image description framework
Apple Company jointly with the University of Wisconsin-Madison released the RubiCap AI training framework, aiming to achieve more accurate image descriptions, solving the hallucination problem in traditional image annotation, and outperforming large models in performance.
AiBase Summary:
🍎 RubiCap framework is designed for "dense image description," accurately capturing image details.
🧠 Through a reinforcement learning mechanism, Qwen2.5 acts as a referee to improve model accuracy.
🚀 The RubiCap model outperforms billion-scale models despite its smaller parameter scale.