Welcome to the 【AI Daily】 column! Here is your daily guide to exploring the world of artificial intelligence. Every day, we present you with the hottest topics in the AI field, focusing on developers to help you gain insights into technological trends and understand innovative AI product applications.
Fresh AI Products Click to Learn More: https://top.aibase.com/
1. Keling AI Unveils AI Model Feature Late at Night!
Keling AI has recently launched a revolutionary AI model feature, marking a significant transformation in visual presentation in the e-commerce industry. The new version of the model has shown substantial improvements in semantic understanding and lifelike effects, especially for Asian portraits. Users can quickly generate high-quality AI model images through simple settings, seamlessly integrating with the AI dressing function, greatly enhancing the convenience and efficiency of clothing displays.
【AiBase Highlights:】
🌍 The AI model feature combined with Keling 1.5 allows users to quickly generate high-quality model images, enhancing e-commerce visual presentations.
👗 The AI model integrates seamlessly with the AI dressing function, enabling users to directly generate images of models wearing new clothes, making it user-friendly.
🎥 The AI video feature allows models to dynamically showcase clothing, reducing costs and increasing efficiency, helping small businesses compete with big brands.
2. Reports Indicate ByteDance's AI Investment Approaches Total of BAT, Budget to Double Next Year!
Reports suggest that ByteDance is increasing its investments in the field of artificial intelligence, with capital expenditures expected to reach 80 billion yuan in 2024, nearly equivalent to the total of Baidu, Alibaba, and Tencent. By 2025, this figure is set to double to 160 billion yuan, focusing on the procurement of AI computing power. Founder Zhang Yiming is personally involved in talent recruitment to achieve the goal of "general artificial intelligence."
【AiBase Highlights:】
💼 Huge capital expenditure: ByteDance's AI capital expenditure will reach 80 billion yuan in 2024, close to the total of BAT.
🚀 Budget doubling plan: Expected capital expenditure will increase to 160 billion yuan in 2025, focusing on AI computing power investments.
🧠 AI talent strategy: Zhang Yiming is personally recruiting AI talent, pursuing the goal of "general artificial intelligence."
3. ChatGPT Experiences Second Major Outage This Month
OpenAI's AI services faced an interruption on December 27th, affecting ChatGPT, the Sora video generation model, and some API calls. While some services were restored by 6:05 AM, the chat history loading feature remained limited, and a full recovery timeline is yet to be determined. OpenAI stated that the disruption was caused by upstream providers, and Microsoft Azure also reported power issues at its data centers.
【AiBase Highlights:】
⚠️ OpenAI's AI services faced an interruption on December 27th, affecting multiple functionalities.
🔧 Partial recovery occurred by 6:05 AM, but chat history loading features remain limited.
🔌 The disruption is related to power issues at Microsoft Azure's data centers.
4. Baidu Wenxin Kuaima Launches Accessibility Version
Baidu AI has recently launched an accessibility version of Wenxin Kuaima, aimed at helping visually impaired IT professionals overcome challenges in programming and improve work efficiency. This tool converts code into natural language, making it easier for visually impaired individuals to understand and modify code. The release of Wenxin Kuaima not only reflects technological advancement but also showcases Baidu's efforts in promoting social inclusivity, providing substantial support for special groups.
【AiBase Highlights:】
🛠️ The accessibility version of Wenxin Kuaima uses natural language processing technology to help visually impaired individuals better understand and write code.
👩💻 This tool allows users to describe requirements in natural language, quickly generating runnable code, greatly enhancing development efficiency.
🌍 The release of Wenxin Kuaima reflects Baidu's commitment to promoting social inclusivity and the warmth of technology.
5. Xiaomi's Surge OS 2 AI Assistant Super Xiao Ai to be Opened to Official Version Users for the First Time
In the upcoming annual summary live stream, Xiaomi President Lu Weibing will introduce significant updates to Xiaomi's Surge OS 2, particularly the first opening of the Super Xiao Ai feature. This update not only enhances user experience but also showcases Xiaomi's innovation in the AI assistant field. Super Xiao Ai, as an evolved version of Xiao Ai, possesses memory capabilities and security features, aiming to provide users with an intelligent and convenient interactive experience.
【AiBase Highlights:】
✨ The Super Xiao Ai feature is being opened to official version users for the first time, marking a significant upgrade for Surge OS 2.
🔒 All data is end-to-end encrypted, ensuring user information security and privacy protection.
📅 Super Xiao Ai has memory capabilities, helping users keep track of schedules and personal preferences.
6. Dazhong Dianping Beta Tests AI Recommendation Tool "Dianzi" for Smart Restaurant and Dish Recommendations
Meituan Dazhong Dianping has launched the AI smart tool "Dianzi" to provide users with personalized dining recommendation services. By integrating multiple features such as "Store Helper," "Dish Pairing," and "Scenic Tour Companion," users can easily input their needs and quickly find suitable restaurants and recommended dishes. Additionally, "Dianzi" also has an AI writing feature to assist users in crafting high-quality reviews.
【AiBase Highlights:】
🔍 Provides personalized dining recommendations, simplifying the decision-making process for users.
🍴 Smartly recommends restaurants and dishes based on user reviews, enhancing the ordering experience.
🖋️ The AI writing feature assists users in crafting reviews, improving content quality and efficiency.
7. Zhipu AI Releases Open Source Agent Task Model CogAgent-9B: Predicts Actions Through Screenshots
Zhipu AI has launched CogAgent-9B, an open-source agent task model designed to predict users' next actions based on screenshots. This model is trained on GLM-4V-9B and has a wide range of application scenarios, including personal computers and mobile devices. Compared to previous versions, CogAgent-9B shows significant improvements in GUI perception and reasoning accuracy, supporting bilingual interactions in Chinese and English.
【AiBase Highlights:】
🖼️ CogAgent-9B predicts the next GUI action through screenshots and user commands, suitable for various devices.
📈 Compared to the first version, CogAgent-9B shows significant improvements in accuracy, versatility, and generalization, supporting bilingual interactions.
🔧 The model's output includes descriptions of thought processes and next actions, demonstrating advantages in GUI positioning and operation.
Details link: https://github.com/THUDM/CogAgent
8. 2024 Douyin Live Governance Report: Introduces Audio Review Large Model to Strengthen Governance of Violations
The 2024 live governance report released by Douyin Group showcases its governance philosophy and achievements in the live streaming field. By introducing an audio review large model and improving host management mechanisms, Douyin aims to build a safe and trustworthy online live streaming environment. The report mentions that the implementation of the host health score system has significantly reduced violation rates, while collaboration with law enforcement has strengthened the crackdown on offline illegal activities.
【AiBase Highlights:】
🔍 Douyin launches the "Health Score" system, with 1 million hosts penalized, and 76% of low-scoring hosts proactively adjust their content, significantly reducing violation rates.
🤖 The introduction of the audio review large model gradually achieves automatic identification and handling of violations, enhancing governance efficiency.
🛡️ Upgraded mechanisms for protecting minors, with a rapid refund system effectively safeguarding minors and reducing violations through coordinated urban governance.
9. Beijing Adds 11 Newly Registered Generative AI Services
According to news from the "Internet Information Beijing" WeChat official account, as of December 27th, Beijing has added 11 newly registered generative artificial intelligence services, bringing the total number of registered services to 105. These new services include Qingying AI, Xiaomi edge-side text, and Xiaomi Surge image, among others. All launched generative AI applications or features must prominently display the registration status of the services used, including model names and registration numbers, to ensure transparency and compliance.
【AiBase Highlights:】
📝 11 new generative AI services added, bringing the total to 105.
📱 New services include Qingying AI, Xiaomi edge-side text, and Xiaomi Surge image.
🔍 Launched applications must display the registration status of the services used to ensure transparency and compliance.
10. Li Auto's AI Application "Li Auto Classmate" App Officially Launched
Li Auto has launched the "Li Auto Classmate" app, now available on the Apple App Store and Xiaomi App Store, aimed at providing users with intelligent interactive services. The app is based on Li Auto's self-developed large model and offers various features such as object recognition, knowledge Q&A, and translation. CEO Li Xiang emphasized that true artificial intelligence should be able to autonomously use various devices and meet user needs. The launch of Li Auto Classmate not only enhances user experience but also promotes the development of artificial intelligence technology.
【AiBase Highlights:】
🛠️ The Li Auto Classmate app is officially launched on iOS and Android platforms, providing rich intelligent interactive features.
👶 Many users' children have solved learning problems through interactions with Li Auto Classmate, showcasing its educational value.
🚀 The launch of the Li Auto Classmate app paves the way for the popularization and development of artificial intelligence technology.
11. OpenAI and Microsoft's Secret Agreement Exposed, AGI Priced at $100 Billion
According to the latest report, Microsoft and OpenAI have a profit-based specific definition for artificial general intelligence (AGI), believing that AGI is only achieved when OpenAI's AI system can generate at least $100 billion in profit. Although OpenAI expects to lose billions this year and does not anticipate profitability before 2029, this agreement may allow Microsoft to continuously use OpenAI's technology for the next decade.
【AiBase Highlights:】
💰 Microsoft and OpenAI define AGI as an AI system capable of generating at least $100 billion in profit.
📉 OpenAI expects to lose billions this year and does not anticipate profitability before 2029.
🤖 Microsoft may have access to OpenAI's technology for the next decade or longer.
12. Overkill! Anthropic: Most Tasks Do Not Require Complex AI Agents
Anthropic's latest analysis points out that the scenarios in which AI agents are used in practical applications are not common, and in many tasks, simple prompts are often more effective. The company emphasizes that while AI agents can handle complex tasks, they require more regulation and computational power. Teams are advised to start with basic prompts and gradually increase complexity, focusing on simplicity in design, transparency in processes, and thoughtful user interface design.
【AiBase Highlights:】
🤖 Most tasks do not require complex AI agents.
🔍 Start with basic prompts and only increase complexity when necessary.
🛠️ Customer service and software development are the areas with the most potential for AI agents.