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 innovative AI product applications.
Discover new AI products by clicking here: https://top.aibase.com/
1. viva releases Sora-based video generation model supporting 4K resolution
viva has recently launched a video generation model based on the Sora architecture, which is free to use and has a significant positive impact on ordinary people's ability to quickly produce video content. The model supports text-to-video, image-to-video, 4K resolution upscaling, and automatic prompt optimization. It excels in video generation, producing high-quality vertical screen videos, close to Google's Veo model. This democratizes video content creation, allowing anyone to create high-quality videos.
AiBase Highlights:
🚀 The first video generation model based on the Sora architecture, currently available for free
🎬 Supports 4K resolution, text, and image-to-video, likely the most significant motion video generation model currently
📱 Supports automatic optimization of prompts, text-to-video in 5 seconds, image-to-video in 4 seconds
Product experience link: https://top.aibase.com/tool/viva
2. Coze launches Web SDK for quick embedding of bots into web pages
Coze, ByteDance's AI chatbot development platform, has introduced a Web SDK, offering users a convenient way to embed bots into web pages, expanding the application scenarios of chatbots. Coze provides a rich set of features and advantages, including unlimited scalability, a wide range of data sources, persistent memory capabilities, flexible workflow design, and more, offering users greater creative space.
AiBase Highlights:
🚀 Unlimited scalability: Coze offers a rich set of plugin tools to help bots perform a wider variety of tasks.
📚 Rich data sources: Users can manage and store data, enabling bots to interact with their own data.
🔐 Persistent memory capabilities: Supports AI memory of important parameters, enhancing interaction consistency and personalization.
Details link: https://www.coze.com/docs/developer_guides/web_sdk?_lang=en
3. Sony warns over 700 companies against using its music data to train AI models
Sony Music Group has warned over 700 companies against using music data for training large AI models without permission, emphasizing respect for the intellectual property rights of songwriters and recording artists. This move reflects the importance placed on intellectual property and the regulatory control over the use of music data in AI models.
AiBase Highlights:
⭐️ Sony warns over 700 companies against using music data to train large AI models without permission
⭐️ AI model manufacturers must respect the intellectual property rights of songwriters and recording artists
⭐️ Sony Music Group is one of the largest music companies in the world, with extensive music copyright resources
4. Google launches 3D generation model CAT3D for 1-minute 3D scene creation
CAT3D, launched by Google, has made significant progress in the field of 3D reconstruction. It can quickly generate 3D scenes, support multi-view input, achieve high-quality 3D capture and real-time rendering, and has structural advantages crucial for the 3D reconstruction pipeline. The emergence of CAT3D will transform industries such as virtual reality, game development, and architectural design, offering users more realistic and interactive experiences.
AiBase Highlights:
✨ Quick generation: CAT3D can create an entire 3D scene in one minute, faster than existing methods
🔍 Multi-view support: CAT3D not only supports single image input but also handles multiple image inputs, generating richer and more detailed 3D scenes
🌟 High-quality 3D capture: Utilizing a multi-view diffusion model, it generates highly consistent new views of the scene
Details link: https://top.aibase.com/tool/cat3d
5. Google releases technical report on Gemini 1.5 detailing improvements in the Gemini 1.5 Pro model architecture
Google's Gemini 1.5 technical report details the performance characteristics and architecture of the Gemini 1.5 Pro and Gemini 1.5 Flash models, showcasing the latest advancements in the field of multi-modal large models, providing new directions for future AI technology development.
AiBase Highlights:
🚀 Gemini 1.5 Pro and Gemini 1.5 Flash models have significantly improved performance, with longer context understanding and stronger reasoning capabilities.
💡 Gemini 1.5 Flash is a lightweight variant, improving efficiency and reducing model service latency, optimizing the use of tensor processing units in multi-modal functions.
🔍 Gemini 1.5 excels in cross-modal long-context retrieval tasks, achieving near-perfect recall, and improving long document QA, long video QA, and long-context automatic speech recognition.
Details link: https://storage.googleapis.com/deepmind-media/gemini/gemini_v1_5_report.pdf
6. Open AI's internal strife season 2 timeline and perspectives
This article reports on the recent internal conflicts and personnel changes at Open AI, which have attracted widespread attention in the industry. The disputes surrounding the safety and development speed of AI models reflect two major themes in AI development: safety and efficiency. Promoting the development of AI technology under the premise of ensuring safety is a question that the entire industry needs to consider.
AiBase Highlights:
🔍 Open AI's internal conflicts have attracted attention, with disputes over the safety and development speed of AI models.
💼 Key personnel changes, such as Ilya and Jan leaving, have sparked industry discussions.
⚖️ Community opinions are divided, with some believing a balance between efficiency and safety should be maintained, while others think excessive concern for safety is unnecessary.
Details link: https://www.chinaz.com/2024/0520/1617697.shtml
7. ElevenLabs launches Audio Native to automatically convert web content into podcasts
ElevenLabs' new service, Audio Native, is an embedded audio player that can automatically generate high-quality human voice narration for web content, helping to convert content into podcast format. Users can listen to real-time generated voice narration without waiting, increasing audience engagement. It also supports multi-platform integration and flexible content management, allowing users to customize the player's appearance and track audience engagement.
AiBase Highlights:
🔊 Automatically generates high-quality human voice narration, real-time voice content generation
🎛️ Embedded audio player easily integrates into any web page, supports custom appearance
📊 Multi-platform support, provides audience engagement tracking and flexible content management
Details link: https://elevenlabs.io/blog/audio-native/
8. Free AI illustration library PictoGraphic offers over 40,000 images
PictoGraphic is a platform offering a free AI-generated illustration library with over 40,000 images and SVG files. It provides designers with an intuitive and easy-to-use interface to quickly find or create illustrations that meet their needs. Users can customize generated illustrations, adjust colors, and start downloading and generating illustrations without needing a credit card.
AiBase Highlights:
🎨 Rich illustration library: Offers over 40,000 images and SVG files in various styles and concepts, meeting designers' diverse design needs.
🖌️ Customizable generated illustrations: Users can generate custom illustrations based on artistic styles in seconds with text prompts, easily creating new illustrations.
🎨 Color customization: Allows users to adjust the colors of illustrations directly on the platform, enhancing creativity and saving time. Supports direct color adjustments to ensure consistency with the design scheme.
Details link: https://top.aibase.com/tool/pictographic
9. Hollywood agency CAA offers solutions for managing AI representations to prevent misuse
Hollywood's top agency, CAA, has partnered with AI technology company Veritone to launch a digital asset management solution aimed at protecting celebrities' AI representations from misuse. They have established a virtual media storage system, "theCAAvault," to help celebrities store digital assets such as names, images, and voices, ensuring legal use and protecting rights.
AiBase Highlights:
💡 CAA partners with Veritone to provide digital asset management solutions, protecting celebrities' AI representations from misuse.
💡 CAA establishes the virtual media storage system "theCAAvault," where celebrities can store digital assets such as names, images, and voices.
💡 CAA's goal is to help celebrities ensure legal use and protection of their rights through ownership of their digital representations.
10. Washington Post adds AI audio features
The Washington Post has recently introduced AI-generated audio features for political and policy news briefings, adding a new reading experience. This move has not only attracted a large number of users but also opened up new advertising channels.
AiBase Highlights:
🎙️ Audio features added: The Washington Post has added AI-generated audio features for three political and policy news briefings, providing a new reading experience.
📊 Audio user growth: The Post's platform has 4 million audio plays daily, with 90% coming from the app, and the play count continues to grow.
🔊 Advertising support: Solventum and PhRMA are the launch sponsors for this week's briefings, which for the first time include AI-generated audio ads.
11. Snapchat plans to invest $1.5 billion annually in artificial intelligence
Snapchat's developer, Snap, has announced plans to increase investment in artificial intelligence and machine learning, adjusting its advertising business and user feedback, and increasing investment in machine learning, AI, and augmented reality features. Snap is collaborating with Amazon and Google for cloud computing, planning to invest 84 cents per daily active user per quarter for infrastructure, with an annual investment of approximately $1.5 billion.
AiBase Highlights:
🔍 Snapchat increases investment in AI and machine learning, adjusting its advertising business and user feedback.
🚀 Investing in machine learning, AI, and augmented reality features, closely integrated with the advertising business and user feedback.
💡 Collaborating with Amazon and Google for cloud computing, investing 84 cents per daily active user per quarter for infrastructure.
