Welcome to the AI Daily section! This is your daily guide to exploring the world of artificial intelligence. Every day, we bring you the latest hot topics in the AI field, focusing on developers, helping you understand technology trends, and discover innovative AI product applications.
Fresh AI Products Click to Learn More: https://top.aibase.com/
1. Shocking the AI World! Llama 3.1 Leak: The 405 Billion Parameter Open Source Giant is Here!
Llama 3.1 has been leaked! This open-source model with 405 billion parameters has caused a stir on Reddit. Although it has not been officially released, it has already caused a sensation in the community. Llama 3.1 surpasses GPT-4o, becoming the SOTA in open-source models. The model supports multiple languages, uses rich training data, and emphasizes security.
AiBase Highlights:
🔥 Llama 3.1 leaked, with 405 billion parameters, surpassing GPT-4o, becoming SOTA.
💡 The model supports multiple languages, including English, French, German, etc., performing excellently in multilingual dialogues.
🛡️ The research team emphasizes security, adopts a multi-faceted data collection method, and introduces boundary prompts and adversarial prompts.
Details Link: https://pastebin.com/9jGkYbXY#google_vignette
2. Luma AI Launches "Loops" Feature: Easily Create Infinite Loop Videos
Luma AI's latest "Loops" feature brings endless creative possibilities for content creators and digital marketers. Users can create seamless, continuous video loops with simple operations, enhancing creative efficiency, attracting audience attention, and reducing production time and costs.
AiBase Highlights:
🌟 Users can create infinite loop videos through text descriptions, images, or keyframes.
🎥 The "Loops" feature solves the problem of AI-generated video discontinuity, making it easier for content creators to produce compelling works.
🔍 Luma AI commits to responsible AI development, plans to implement watermarks and attribution systems, ensuring content transparency.
3. Stability AI Releases Open-Source Audio Generation Model Stable Audio Open
The Stability AI team has launched the open-source audio generation model Stable Audio Open, capable of generating up to 47 seconds of 44.1kHz stereo audio. The model adopts an open weight design, uses legally sourced audio data for training, ensuring data ethics. The advanced technical architecture generates high-fidelity and diverse audio. Performance evaluations show it is comparable to top industry models, with high-quality audio generation capabilities.
AiBase Highlights:
🎧 Stable Audio Open supports generating up to 47 seconds of 44.1kHz stereo audio.
📝 The model training only used audio data licensed under Creative Commons, ensuring data legality and ethics.
🔍 Stable Audio Open's audio generation quality has been verified, with high fidelity and diversity.
Details Link: https://top.aibase.com/tool/stable-audio-open-demo
4. Tomato Novel AI Training Protocol Sparks Strong Backlash from Online Writers
This article explores the dispute between artificial intelligence and author rights triggered by the Tomato Novel platform, where online writers strongly oppose the AI training supplementary agreement, fearing that AI technology development will squeeze the survival space of human creators. The industry is divided between supporters and opponents of AI applications in literary creation, sparking debates on creative efficiency and the value of original works.
AiBase Highlights:
🤖 The AI training supplementary agreement has sparked strong opposition from online writers, allegedly infringing on author rights, posing a potential threat.
💡 Discussions on AI applications in literary creation are widespread, with supporters believing it can enhance creative efficiency, while opponents worry about diluting the value of original works.
⚖️ Calls for platforms to balance AI technology applications and author rights, establish reasonable profit distribution mechanisms, and maintain the healthy development of online literature creation.
5. Baidu Health Launches Intelligent Entity Family, Introduces Lingyi Open Platform and Other Large Model Applications
Baidu Health introduced a series of large model application products at the 2024 Industrial Ecosystem Conference, aiming to address the "impossible triangle" problem in the healthcare industry, enhancing the accessibility, affordability, and quality of medical services. Through AI technology, Baidu Health helps doctors and patients obtain and utilize medical resources more efficiently.
AiBase Highlights:
👩⚕️ Baidu Health launches a health intelligent entity family, providing services such as AI accurate doctor search, AI medical report interpretation, AI medication assistant, etc., supporting multimodal interaction.
💻 Online medical Copilot for both doctors and patients, providing online medical services.
⚕️ CDSS improves the diagnostic capabilities of primary healthcare institutions, reduces misdiagnosis and missed diagnosis, and has been implemented in over 4,000 primary healthcare institutions.
6. Baidu Smart Cloud Launches Financial Intelligent Entity Application "ZhiJin"
Baidu Smart Cloud introduces the financial intelligent entity application "ZhiJin," enhancing the business efficiency and revenue of financial institutions by deeply integrating AI large model technology with core business scenarios in the financial industry. ZhiJin applications cover four core scenarios: wealth management, asset evaluation, business compliance, and over-the-counter transactions, providing intelligent and efficient business auxiliary tools.
AiBase Highlights:
🔍 Wealth Intelligence Consultant: Provides a super assistant for financial wealth advisors, offering 7x24-hour professional consultation and intelligent recommendation services for products, assisting in precise marketing.
📊 Asset Intelligence Evaluation: Utilizes large model capabilities to collect, process, and organize massive information, assisting financial analysts in completing market analysis, investment strategies, and other research reports.
⚖️ Compliance Intelligence Judgment: Provides intelligent and highly available compliance management tools, reducing compliance risks in financial business.
Details Link: https://top.aibase.com/tool/baiduzhijin-jinrongzhinengti
7. Musk Announces the Official Launch of the World's "Strongest" AI Training Cluster
Musk announces that xAI is conducting the world's strongest AI training in the Memphis super cluster, equipped with 100,000 Nvidia H100 GPUs. This move has attracted media attention but also raised concerns about the progress of Musk's projects. Meanwhile, xAI faces pressure from competitors and needs to launch innovative models to maintain competitiveness.
AiBase Highlights:
🌟 xAI launches the world's strongest AI training cluster, equipped with 100,000 Nvidia H100 GPUs.
⚡ Musk plans to train the "strongest AI" by December 2024, with the cluster providing a significant advantage.
🚀 xAI faces pressure from competitors like OpenAI and Google, needing to launch innovative models to maintain competitiveness.
8. LensGo AI Launches FaceSync Feature
LensGo AI's new FaceSync feature can synchronize the user's performance video with the selected image or video, creating a new visual experience. Users can freely express themselves, become any character, and tell any story, with creativity unconstrained by appearance. The FaceSync feature is still in the testing phase, with LensGo AI continuously optimizing to ensure stability and expressiveness.
AiBase Highlights:
✨ Users can record performance videos and synchronize them with selected images or videos, creating a new visual experience.
🎭 FaceSync accurately captures the user's facial expressions, voice, and lip movements, mapping them onto the target image or video.
🚀 FaceSync feature in testing phase, LensGo AI continuously optimizing to meet user expectations and professional standards.
9. From One Photo to 3D Model in Just 30 Seconds! Aiuni AI: A Project Based on Unique 3D Open Source
In the digital era, Aiuni AI, based on the Unique3D open-source project, achieves the goal of generating high-fidelity 3D models from any RGB image. Its core technology, Unique3D, utilizes deep learning and diffusion models to improve 3D modeling efficiency. In the future, Aiuni AI plans to expand its application areas, such as virtual reality and augmented reality, bringing more possibilities to various industries.
AiBase Highlights:
🌟 High Fidelity: Unique3D-generated 3D models are highly accurate, consistent with real objects.
⏱️ High Efficiency: From input image to 3D model takes only 30 seconds, significantly improving efficiency.
🌐 Strong Generalization Ability: Unique3D can handle various input images, generating high-quality 3D models.
Details Link: https://top.aibase.com/tool/aiuni
10. Nvidia's Stock Soars as New AI Chip Complies with U.S. Export Rules
Nvidia's stock surged in Monday's early trading, primarily driven by the news that the upcoming Blackwell chip's version for the Chinese market complies with U.S. export control requirements. Market confidence in Nvidia's successful response to export control challenges has strengthened, with positive stock performance. The company has achieved significant results in technological innovation and market demand, showing good development momentum.
AiBase Highlights:
📈 Nvidia's stock soared in Monday's early trading, influenced by the news that the new Blackwell chip complies with U.S. export control requirements.
💡 Nvidia plans to cooperate with Chinese tech company Inspur to produce the "B20" chip compliant with export controls, expected to ship in the second quarter of 2025.
🚀 Due to strong demand for the Blackwell platform, Nvidia's stock has more than doubled since the beginning of the year, with a Monday increase of over 4%, reaching $122.76.
11. Cohere Raises $550 Million in Funding, Valuation Reaches $5.5 Billion
Cohere, a renowned large language model platform focusing on AIGC, has recently successfully raised $550 million in a D round of financing, with a valuation reaching $5.5 billion. This marks the company's strong performance and broad prospects in the market.
AiBase Highlights:
🌟 Lead investor is PSP Investments, a Canadian pension investment management company, with participation from Cisco, Fujitsu Japan, AMD Ventures, and other well-known institutions, indicating market recognition and support for Cohere.
🚀 The founding team includes Aidan Gomez, Ivan Zhang, and Nick Frosst, all of whom have worked at Google Brain and Cortex, with strong technical capabilities and deep backgrounds, laying a good foundation for the company's development.
📝 Cohere provides various enterprise services, including large language model customization, semantic search, text creative writing, long text summarization, etc., helping enterprises improve work efficiency and save costs.
12. Google AI Weather Model NeuralGCM Makes New Breakthrough: Computational Efficiency 100,000 Times Higher Than Other Models
