Welcome to the 【AI Daily】 section! Here is your guide to exploring the world of artificial intelligence every day. We present the hottest topics in the AI field, focusing on developers to help you gain insights into technology trends and understand innovative AI product applications.

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

1. 360 Launches Multimodal Content Creation Engine "Nano Search"

At a press conference, 360 introduced "Nano Search," an innovative multimodal content creation engine designed to provide a simple and user-friendly search and creation experience for users of all ages. This engine can not only analyze user needs but also generate and rewrite content, redefining the functionality of search engines. With various input methods, users can easily access information and transform it into creative resources, pushing search technology into the 3.0 era.

image.png

【AiBase Summary:】

🧒 Suitable for users of all ages, easy to use, supports multiple input methods like voice and photo.

💡 By analyzing user needs, Nano Search can generate and rewrite search results, promoting content creation.

📹 Users can customize video voices, using various voices to generate creative videos, enhancing the interactive experience.

Details link: https://www.n.cn/

2. Alibaba Releases QwQ-32B-Preview: Challenging OpenAI's Inference Models with Self-Fact-Checking Capabilities

The Qwen team at Alibaba has launched the QwQ-32B-Preview inference AI model, marking a significant breakthrough in the field of artificial intelligence. This model has 32.5 billion parameters and can process text up to 32,000 words long, outperforming OpenAI's o1 series in several benchmark tests. Its unique self-fact-checking mechanism effectively avoids the inference traps of traditional AI models, although it still has some limitations, such as language switching and insufficient common sense reasoning.

image.png

【AiBase Summary:】

🚀 The QwQ-32B-Preview model has 32.5 billion parameters, supports long text processing, and demonstrates exceptional reasoning capabilities.

🧠 This model performs well in key tests such as AIME and MATH, especially in complex logic and mathematical problems.

🔍 Utilizing a self-fact-checking mechanism, it avoids traditional AI's inference traps and has been released on the Hugging Face platform.

Details link: https://huggingface.co/Qwen/QwQ-32B-Preview

3. iFLYTEK's Spark Low-Code Intelligent Agent Platform Upgraded

iFLYTEK Co., Ltd. has recently conducted a comprehensive upgrade of its Spark low-code intelligent agent platform, aiming to significantly shorten development cycles and expand application scenarios through rapid building and flexible customization. The new platform allows users to create AI agents for free and reduces development barriers through fine-tuning and template functions, meeting the needs of various users. Additionally, the platform integrates rich functional modules, supports complex task handling, improves decision-making accuracy, and further broadens its application boundaries.

image.png

【AiBase Summary:】

✨ After the platform upgrade, users can quickly create AI agents, achieve customized optimization, and lower development barriers.

🛠️ The new template feature allows users with no programming background to easily get started, meeting personalized needs.

📡 Provides AI podcast functionality, converting public account articles into radio programs, enriching users' auditory experience.

4. Meitu's WHEE Launches "IP Image Customization" Feature: Cartoon Characters Easily Transformed from 2D to 3D

The recently launched "IP Image Customization" feature on the WHEE platform uses advanced AI technology to allow users to easily convert 2D images into 3D models, creating personalized IP images. This innovative tool not only simplifies the conversion process from 2D to 3D but also provides users with a vast creative space suitable for various application scenarios. This marks a significant advancement for WHEE in the field of AI visual creation, providing new inspiration and tools for designers and creative workers.

image.png

【AiBase Summary:】

🖼️ Through the "IP Image Customization" feature, users can easily convert 2D images into 3D models.

💡 This feature provides users with limitless creative space, suitable for various application scenarios.

🚀 WHEE has taken a significant step in AI visual creation, providing new tools for designers and creative workers.

5. ByteDance Sues Former Intern for 8 Million Due to Code Tampering and Model Training Attack

ByteDance has recently filed a lawsuit against former intern Tian, accusing him of tampering with code and attacking the company's internal model training, resulting in severe losses. ByteDance demands 8 million yuan in compensation and a public apology from Tian. The incident stemmed from Tian's dissatisfaction with the allocation of team resources, leading to his malicious attack on the company's research project. ByteDance emphasized that this action is to maintain company security and prevent similar incidents from happening again.

image.png

【AiBase Summary:】

⚠️ ByteDance is suing former intern Tian for 8 million yuan and reasonable expenses of 20,000 yuan.

💻 Tian tampered with code and attacked the company's model training due to dissatisfaction with resource allocation, causing resource loss.

📉 ByteDance emphasized that the impact of this incident is limited, has terminated Tian's internship agreement, and notified relevant authorities.

6. Competing with NotebookLM! ElevenLabs Launches New Feature GenFM for AI-Generated Multichannel Podcasts

ElevenLabs has recently launched the GenFM feature, allowing users to upload videos or texts to generate multichannel podcasts, supporting up to 32 languages. This feature not only automatically selects two voices but also incorporates human-like filler words to enhance the naturalness of conversations. In the future, ElevenLabs plans to add more customization options and expand its business in Poland and India, demonstrating its ambition and innovation in the AI podcasting field.

image.png

【AiBase Summary:】

🌐 The GenFM feature allows users to upload videos or texts to generate multichannel podcasts, supporting 32 languages.

🎙️ This feature automatically selects two voices and enhances the natural experience of conversations by adding human-like filler words.

🚀 In the future, ElevenLabs will support more customization options and expand its business in Poland and India.

7. ComfyUI Major Update: Desktop Version Officially Released, Making AI Drawing Easy for Beginners!

The desktop version of ComfyUI has been officially released, providing users with a more straightforward AI drawing experience. This version supports both Windows and Mac systems and comes with detailed operation guides, lowering the usage threshold for novice users. The integrated Python environment and automatic update features allow users to easily import existing projects, while the built-in manager and automatic resource import functionality further enhance usability. With various templates and customizable nodes, users can achieve personalized creations, marking the popularization of AI drawing tools.

image.png

【AiBase Summary:】

🖥️ The ComfyUI desktop version supports Windows and Mac systems, providing comprehensive operation guides to reduce the difficulty for novice users.

🔄 The integrated Python environment and automatic update features allow users to easily import existing projects for a smooth experience.

✨ The built-in ComfyUI manager and customizable node features allow users to extend program functions for personalized creations.

Details link: https://comfyorg.notion.site/ComfyUI-Desktop-User-Guide-1146d73d365080a49058e8d629772f0a#aab1bb6f65b64b0d82a85a947781cd60

8. Amazon Launches New AI Video Model, Reducing Dependence on Anthropic

Amazon has recently developed a new generative artificial intelligence model capable of processing images and videos, aimed at improving the efficiency of video search and management. This model can not only identify objects and scenes in videos but also understand contextual information, providing smarter search results. This innovation will reduce dependence on external suppliers, enhance Amazon's competitiveness in the AI market, and may also positively impact its cloud computing services.

image.png

【AiBase Summary:】

🤖 Amazon's newly developed AI model can process images and videos, enhancing video search efficiency.

🔍 This model can understand the context of video content, providing more accurate search results.

📈 Amazon's in-house development reduces dependence on Anthropic, strengthening market competitiveness.

9. Kimi Partners with Tsinghua University to Open Source Mooncake Inference Architecture, Enhancing AI Inference Efficiency

In the context of rapid development in artificial intelligence, Kimi has collaborated with Tsinghua University's MADSys laboratory to launch the KVCache-based Mooncake inference system, aimed at addressing high inference loads and reducing inference costs. This system significantly improves inference efficiency through innovative architecture and was officially released in June 2024. The Mooncake project is open-sourced on GitHub, aiming to build a standard interface for high-performance memory semantic storage, inviting more companies and research institutions to participate and promote advancements in AI technology.

image.png

【AiBase Summary:】

🌟 Kimi and Tsinghua University jointly released the Mooncake inference system, significantly enhancing AI inference efficiency.

🔧 The Mooncake project is open-sourced on GitHub, aiming to build a standard interface for high-performance memory semantic storage.

🤝 We look forward to more companies and research institutions joining to advance AI technology.

Details link: https://github.com/kvcache-ai/Mooncake

10. Baidu Wenku Launches "AI All-Net Search": Structured Answers with Chart Presentation for Clearer Responses

Baidu Wenku recently launched the "AI All-Net Search" feature, utilizing advanced AI technology to enhance the efficiency and accuracy of information retrieval, helping users quickly obtain core points in an age of information overload. This feature integrates a wealth of document resources, particularly excelling in the academic field, and can provide structured answers through natural language processing technology, improving the readability of information through visualization. This innovation not only offers convenience for researchers and professionals but also showcases Baidu Wenku's continuous progress in AI applications.

image.png

【AiBase Summary:】

📚 Integrating over 1.4 billion professional documents across multiple fields to ensure the authenticity and professionalism of information.

🔗 In close cooperation with Baidu Academic, covering 95% of domestic academic journals, especially suitable for academic research.

🖼️ Introducing visualization features to transform information into easy-to-understand graphics, enhancing users' reading experience.

11. Video Editing Made Easy! Google DeepMind Releases Revolutionary AI Model for Movie-Quality Effects

The Generative Omnimatte AI model launched by Google DeepMind brings special effects production from the film industry into the lives of ordinary users. This model can decompose videos into multiple layers, easily achieving separation of objects and backgrounds, even handling occluded parts. Users can create various stunning visual effects with simple operations, making video editing unprecedentedly easy. However, this technology is still in development, with some minor issues to be resolved.

image.png

【AiBase Summary:】

✨ This AI model can perfectly separate objects from backgrounds in videos without the need for green screens or depth information.

🪄 The core model "Casper" can accurately erase objects and their shadows in videos, leaving the background intact.

🚀 Users can easily achieve creative effects such as teleportation, changing motion speed, or even reversing time.

Details link: https://gen-omnimatte.github.io/

12. Google Launches New Online Chess Game

Google has recently launched an innovative online chess game that allows players to customize the appearance of chess pieces using generative artificial intelligence. Users can input descriptions based on personal preferences, and the system generates themed pieces for opponents, adding to the fun of the game. Although the game features are relatively simple, the operation experience is smooth and offers various difficulty settings, it still lacks some common features, such as historical game reviews. Google also plans to launch a new chess robot integrated into the Gemini platform to enhance the user experience of playing against AI.

image.png

【AiBase Summary:】

🕹️ Google launches a chess game with AI-customized pieces, allowing players to design the appearance of pieces according to personal preferences.

♟️ The game features are simple, supporting three difficulty settings but lacks common features like historical game reviews.

🤖 A new chess robot will be launched on the Gemini platform, allowing players to compete against AI by inputting moves.

Details link: https://labs.google/genchess

13. Making Games Great Again! Elon Musk's xAI to Establish AI Game Studio

Elon Musk, the world's richest person and a seasoned gaming enthusiast, recently expressed concerns about the current gaming industry on social media, announcing that his company xAI will establish a new AI game studio aimed at revitalizing the gaming industry. He criticized many game studios being acquired by large companies, leading to interference from commercial and political factors in game content. Musk calls on developers to focus on the quality and fun of games rather than catering to woke culture. He hopes to push the gaming industry back to its essence through technological innovation and AI applications, providing a higher quality gaming experience.

image.png

【AiBase Summary:】

🎮 Musk announces that xAI will establish a new AI game studio to revitalize the gaming industry.

📉 He criticizes large company acquisitions leading to game content being influenced by commercial and political factors.

✨ Musk calls on developers to focus on game quality and fun, avoiding woke preaching.

14. Man Arrested for Using AI to Fabricate BYD Acquisition of NIO Rumors for Traffic

Recently, rumors about BYD acquiring NIO have sparked widespread attention online. After investigation, the police found that the rumor's creator was a 29-year-old man who used AI technology to synthesize false information to increase traffic to his social media account. Both BYD and NIO responded to this news, emphasizing its falsehood and urging the public to remain rational. Ultimately, the man was detained by law enforcement for disturbing social order.

image.png

【AiBase Summary:】

🔍 Rumor creation: 29-year-old man Xie used AI technology to fabricate false information about BYD acquiring NIO.

🚓 Hefei Public Security Bureau quickly intervened, confirming the rumors and legally detaining Xie.

❗ Both BYD and NIO responded to the false information, urging netizens not to believe or spread rumors.

15. OpenAI Applies for Trademark Protection for Its Inference Model o1

OpenAI has recently submitted a trademark registration application to the U.S. Patent and Trademark Office, marking an important progress in its intellectual property protection. This application involves its newly launched inference model o1, indicating OpenAI's strategic positioning in the market. Although the application has not yet been approved, OpenAI preemptively applied for related trademarks in Jamaica and has achieved preliminary victory in the trademark dispute with Guy Ravine, further solidifying its brand rights.

image.png

【AiBase Summary:】

🌟 OpenAI applied for the "OpenAI o1" trademark to protect the intellectual property of its new inference model.

📅 OpenAI preemptively applied for related trademarks in Jamaica, showcasing its market strategy.

⚖️ OpenAI achieved preliminary victory in the trademark dispute with Guy Ravine, continuing to protect its brand rights.