Welcome to the AI Daily section! This is your daily guide to exploring the world of artificial intelligence. Every day, we bring you the hottest topics in the AI field, focusing on developers to help you understand technology trends and discover innovative AI product applications.
Fresh AI Products Click to Explore: https://top.aibase.com/
1. ChatGPT Edu Version Launched: Supports GPT-4o, Custom GPT, Data Analysis
The ChatGPT Edu version is introduced, offering a multi-functional ChatGPT for university campuses to enhance the quality of learning and teaching efficiency for students and teachers. Several top universities have already integrated ChatGPT into their education, with the primary user base being 18-24 year olds. Features include GPT-4o support, data analysis, custom GPT, and more, providing higher message limits and voice support. The security mechanism is robust, with rich practical applications.
AiBase Highlights:
🚀 The ChatGPT Edu version supports GPT-4o, custom GPT, and data analysis, aiding students and teachers in improving learning efficiency.
🔍 Several top universities have adopted ChatGPT in education, with 18-24 year olds being the main user group, demonstrating strong learning needs and acceptance capabilities.
💡 Rich practical applications include end-of-term reflective assignments, community service, language training, and deep application in the education sector.
2. Kuaishou Launches Self-Developed Text-to-Image Large Model Product "Kolors"
Kuaishou's self-developed text-to-image large model, "Kolors," is now open to the public, offering a new AI image creation experience. This large model supports both text-to-image and image-to-image functionalities, suitable for AI-generated images and AI image customization. Users can easily experience cutting-edge technology through the "Kolors" WeChat mini-program or web version.
AiBase Highlights:
🔍 "Kolors" large model has a parameter scale of over a billion, with data sources including open-source communities, internal Kuaishou builds, and self-developed AI technology integration.
🎨 "Kolors" covers common millions of Chinese entity concepts, providing broader and deeper support for image creation.
🧠 "Kolors" introduces reinforcement learning and reward model technology to address the issue of poor performance in handling long and complex semantic text inputs in large text-to-image models.
Product Entry: https://top.aibase.com/tool/kuaishouketudamoxingkolors
3. Baidu Netdisk Launches AI-Generated Comic Avatar Function
Baidu Netdisk has introduced an amazing new AI feature, allowing users to transform into the main character of a childhood anime with just one photo before Children's Day. This feature is highly personalized, immersing users in the wonderful animated world, and the fast processing speed enhances the experience.
AiBase Highlights:
🎨 Personalized transformation: Users can become anime characters by uploading photos, experiencing unique styles.
🚀 Fast processing: Baidu Netdisk's AI function processes quickly, generating beautiful anime photos in just a few seconds.
🌟 Diverse effects: Not only can users transform into anime characters but also simulate classic animation effects, meeting different user needs.
4. Claude 3 Opens Third-Party API for Business Process Automation
Anthropic's new feature allows users to interact more conveniently with Claude and automate various tasks through structured API calls. Claude can also process images and integrate into real-time applications, providing businesses with smarter and more efficient solutions.
AiBase Highlights:
🔍 Users can have Claude automatically perform multiple tasks through text queries.
🔍 Users can achieve routine operations and question answering through structured API calls.
🔍 Anthropic's new features enable Claude to process images and real-time applications.
5. Novita AI Open-Sources Animate Anyone Project, Upload a Photo to Synthesize Animation
Novita AI has open-sourced the Animate Anyone project, allowing users to synthesize animations by uploading just one photo. This technology brings new possibilities and opportunities for animation production, enabling users to quickly create stunning works.
AiBase Highlights:
👉 ViViD can naturally transfer clothing onto video characters.
👉 Both skirts and pants can be freely replaced, meeting various clothing try-on needs.
👉 The official has only released demos and papers, with the code not yet disclosed.
Project Page: https://top.aibase.com/tool/vivid
Paper Address: https://arxiv.org/pdf/2405.11794
6. Alibaba and USTC Jointly Launch Virtual Try-On Technology ViViD for Easy Video Clothing Replacement
Alibaba and the University of Science and Technology of China have jointly launched the ViViD framework, revolutionizing virtual try-on experiences by enabling real-time clothing replacement in videos, solving the challenges of temporal consistency and image quality, and enhancing the try-on effect.
AiBase Highlights:
👗 Advanced technology: ViViD, based on diffusion model technology, enables real-time clothing replacement in videos, generating natural and realistic effects.
🔧 Three core components: Clothing encoder, pose encoder, and temporal module work together to extract clothing details, encode poses, and maintain temporal consistency.
🌟 Innovative feature fusion: Introducing an attention-based feature fusion mechanism optimizes the integration of clothing semantic information, enhancing the try-on effect to meet user needs.
Details Link: https://top.aibase.com/tool/vivid
7. Perplexity Launches Page Creation Feature for Rapid Generation of Professional Documents
Perplexity AI has introduced its latest feature, Perplexity Pages, aimed at helping users quickly generate professional-level documents, enhancing productivity for content creators and challenging traditional knowledge base platforms. The tool quickly generates content, supports high customization, media content insertion, information verification and source management, and sharing and search optimization.
AiBase Highlights:
🚀 Quickly generate professional documents, saving time and effort.
🔧 Highly customizable to meet different needs.
📸 Media content insertion enhances document appeal.
Details Link: https://top.aibase.com/tool/perplexity
8. Midjourney to Release V6.5 Version, Web Version to be Available to Everyone Soon
Midjourney is set to release the V6.5 version, which will bring a significant improvement in image quality, and the web version will also undergo a major update. Despite challenges in video model development, the team is confident that continuous efforts will achieve greater breakthroughs.
AiBase Highlights:
🚀 Significant improvement in image quality, possibly consistent with the V7 version, with improvements in coherence, skin, hands, and body representation.
💻 The web version will no longer rely on Discord, providing a better user experience.
💡 Introducing a style space explorer and updating the exploration page, considering offering subscription discounts to attract more users.
9. Suno's 3.5 Version Model Open to Everyone, Capable of Producing 4-Minute Songs
Suno's latest 3.5 version model is open to all users, featuring the ability to produce 4-minute songs, 2-minute song extensions, and improved song structure. Suno has also launched a new feature that can transform any sound into music, bringing new possibilities to music creation. The company has secured $125 million in funding, solidifying its leading position in the AI music field. Suno demonstrates strong innovation capabilities and a leading position.
AiBase Highlights:
🎵 Produce 4-minute songs and 2-minute song extensions
🎶 Transform any sound into music, creating new possibilities
💰 The company has secured $125 million in funding, solidifying its leading position
Details Link: https://top.aibase.com/tool/suno-ai
10. You.com Launches Custom Assistant Feature
You.com introduces a custom assistant feature, allowing users to create personalized AI assistants using top-tier language models like GPT-4o, Llama3, and Claude3, aiming to enhance productivity in complex work tasks. This feature makes powerful language models more accessible and adaptable to individual needs, providing a customized AI assistant experience.
AiBase Highlights:
⭐️ Custom AI assistants are designed to enhance productivity in complex work tasks
⭐️ You.com is committed to providing accuracy and real-time information, offering more relevant and reliable responses through online access
⭐️ The impact of the technology is profound, with custom AI assistants having the potential to transform knowledge work in fields such as healthcare, finance, and education
11. Cartesia Releases Low-Latency Voice Generation Model Sonic, Aiming to Recreate ChatGPT's Real-Time Voice Chat?
Cartesia's Sonic low-latency voice generation model has garnered widespread attention for its fast inference speed and ultra-low latency. Sonic can generate voice in real-time with realistic emotions and expressive abilities, and users only need to provide a 10-second recording to mimic the speaker's voice characteristics. Cartesia's goal is to create real-time intelligent systems, and they have introduced an innovative SSM architecture, achieving initial progress.
AiBase Highlights:
🚀 Sonic model latency is only 135 milliseconds, suitable for chat applications.
😊 Sonic demonstrates human emotions and expressive abilities, making conversations more natural.
🔧 Users can adjust parameters such as pitch, speed, and emotion to customize voice output.
Details Link: https://top.aibase.com/tool/carteisa-sonic
12. Gartner Predicts AI Chip Revenue to Reach $71.2 Billion in 2024
According to Gartner's forecast, global AI semiconductor revenue will grow by 33% to reach $71.2 billion by 2024. This trend will drive computers to generally possess AI capabilities, and corporate computer purchases will completely shift to AI computers. AI processing will primarily occur in data centers, with the value of accelerators expected to reach $21 billion.
AiBase Highlights:
📈 Expected AI semiconductor revenue growth of 33% to $71.2 billion by 2024
💻 22% of computers are expected to have AI capabilities by 2024, with corporate computer purchases completely shifting to AI computers by the end of 2026
🏭 AI processing will primarily occur in data centers, with the value of AI accelerators in servers reaching $21 billion by 2024
13. Google Outdone! High-Fidelity 3D Avatars So Real They're Scary, Girls Winking and Raising Eyebrows Seamlessly
The NPGA algorithm proposed by a research team from the Technical University of Munich, University College London, and other institutions has attracted widespread attention for generating high-fidelity 3D avatars with lifelike expressions that are hard to distinguish from real humans. The technological innovation lies in using Gaussian point clouds to generate 3D human shapes and introducing a neural parametric head model to capture subtle facial expression changes, enhancing realism.
AiBase Highlights:
⭐ High-fidelity 3D avatars: NPGA generates realistic 3D avatars with rich expressions, close to real humans.
⚙️ Innovative technology: Using Gaussian point clouds to construct 3D human shapes, improving rendering efficiency and realism.
😲 Neural network model: Utilizing a neural parametric head model to capture subtle facial expression changes, simulating real human expressions.
Details Link: https://tobias-kirschstein.github.io/nersemble/
