Welcome to the AI Daily column! Here is your daily guide to exploring the world of artificial intelligence. Every day, we present the hottest content in the AI field, focusing on developers to help you understand technical trends and innovative AI product applications.
Discover the latest AI products click to learn more: https://top.aibase.com/
1. The Strongest Open-Source Model Arrives! Llama3.1 Leads with 405B Parameters, Surpassing GPT-4o
The competition between open-source and closed-source AI has never ceased. The release of Meta AI's Llama3.1 model marks the maturity of open-source AI, heralding the arrival of a new era. Llama3.1 outperforms closed-source models, matching their performance and opening up a new era where open-source AI could become the industry standard.
【AiBase Summary:】
🚀 Llama3.1 surpasses GPT-4o and Claude3.5Sonnet with 405B parameters, achieving superior performance.
💡 Meta optimizes the training stack significantly, scaling the model's computing power to 16,000 H100 GPUs, enhancing performance.
🔍 Llama3.1 boasts extended context length, multilingual support, and superior performance, demonstrating advantages in multiple aspects.
Details: https://ai.meta.com/blog/meta-llama-3-1/
2. Kering AI Video Launches Paid Plan, Ending Free Generation Experience
Kering announces the end of its free generation experience and introduces a multi-tiered paid membership plan, offering users a more diverse range of service options. The new membership system includes non-members, gold members, platinum members, and diamond members, with a limited-time 50% discount. Kering offers affordable pricing and high cost-performance, catering to different user needs. The shift in business model demonstrates maturity and will play an important role in the AI video generation field.
【AiBase Summary:】
🔑 Kering introduces a multi-tiered paid membership plan, including non-members, gold members, platinum members, and diamond members.
💡 Limited-time 50% discount, with affordable pricing for each membership plan, platinum members can generate 300 videos for an average of 133 yuan per month.
💼 Kering retains the policy of daily free points, meeting the creative needs of daily users.
Details: https://top.aibase.com/tool/keling-ai
3. Kuaishou's AI Video Product Kering International Version 1.0 Launches
Kuaishou announces the official launch of its AI video product Kering International Version 1.0, which allows users to obtain 66 free points daily without a Chinese mobile number, sparking global enthusiasm. Kering has quickly risen to prominence, changing the AI video landscape and marking an important step for domestic AI models to enter the international market. Paid user experience is enhanced, with reasonable pricing to maintain competitiveness.
【AiBase Summary:】
🌐 Kering International Version 1.0 officially launches, allowing users to obtain 66 free points daily without a Chinese mobile number.
🚀 Kering quickly rises to prominence, attracting over 500,000 users to apply for beta testing, demonstrating significant market potential.
💰 Kering offers reasonable pricing, with a minimum monthly fee of 33 yuan, supporting the generation of 66 short videos, removing watermarks, and maintaining competitiveness.
Details: https://klingai.com
4. AI Video Editing Wonder Vozo: Rewrite Scripts and Re-dub with Prompts
Vozo is a groundbreaking video editing tool that revolutionizes content creation using artificial intelligence. It offers intelligent video rewriting and re-dubbing, text-driven voice editing, multi-character lip-syncing, and other features, providing users with unprecedented video editing and adaptation capabilities. Additionally, Vozo supports professional video translation and automatic video optimization, greatly simplifying the internationalization process of content.
【AiBase Summary:】
✨ Intelligent video rewriting and re-dubbing: Vozo uses AI technology to quickly transform classic videos into viral marketing videos, which can be completed in just a few seconds.
🎙 Text-driven voice editing: Users can modify video commentary by editing text, with a variety of accents and tones available to enhance video content expressiveness.
👥 Multi-character lip-syncing: Vozo can achieve natural and smooth lip-syncing for multiple characters in a video, enhancing the viewing experience.
Details: https://top.aibase.com/tool/vozo
5. PixVerse V2 Video Model Released, Supports Direct Generation of 8-second Videos
PixVerse has released the V2 version of its video model, which can directly generate videos up to 8 seconds long, improving image quality and details. This tool allows users' creativity to be limitless, with precise control over details, strong continuity, and rich scenes. It is simple to operate, meeting various style needs, enabling users to create a world in just 8 seconds.
【AiBase Summary:】
✨ Supports generating videos up to 8 seconds long, improving image quality and details.
🎥 Infinite creativity, capable of shooting scenes like aliens dancing at a square or dinosaurs riding bicycles.
🌍 Simple operation, meeting various style needs, allowing users to create multiple scenes in 8 seconds.
Details: https://top.aibase.com/tool/pixverse
6. Tencent PhotoMaker V2 Upgrades, Personalized Character Image Customization Completed in Seconds, No Need for LoRA Training
Tencent recently released the PhotoMaker V2 version, revolutionizing the efficiency and quality of AI personalized character image customization. Through ID embedding technology, it achieves rapid high-quality personalized image generation without a cumbersome training process. It enhances the authenticity of image IDs and improves users' generative control capabilities, increasing efficiency by nearly four times. It opens up new possibilities for individual users and professional creators, simplifying the process of creating personalized image content. The advancement of AI technology will change the creative industry, giving rise to new forms of artistic expression and business models.
【AiBase Summary:】
🚀 Innovative ID embedding technology achieves rapid high-quality personalized image generation without a cumbersome training process.
💡 Enhances the authenticity of image IDs, improves users' generative control capabilities, and increases efficiency by nearly four times.
🎨 Opens up new possibilities for individual users and professional creators, simplifying the process of creating personalized image content.
Details: https://top.aibase.com/tool/photomaker-v2
7. Peking University's New 3D Scene Generation Framework HoloDreamer
HoloDreamer is a new type of 3D scene generation framework that generates immersive, complete 3D scenes through user text descriptions. This innovative development will promote research on text-driven 3D scene generation, providing new solutions for virtual reality, gaming, and film industries.
【AiBase Summary:】
🌟 HoloDreamer is a new type of 3D scene generation framework that generates complete 3D scenes through text descriptions.
🖼️ The framework includes two core modules: stylized panorama generation and enhanced panorama reconstruction, ensuring high quality and consistency of the generated results.
🚀 HoloDreamer performs excellently in experiments, surpassing previous 3D generation methods, providing new solutions for virtual reality and gaming fields.
8. Adobe Introduces More Generative AI Features for Illustrator and Photoshop
Adobe has recently introduced a new round of generative AI features aimed at enhancing the creative workflow efficiency of design software Illustrator and Photoshop. These updates include Illustrator's generative shape fill tool and Photoshop's image generation capabilities, providing designers with more creative freedom and convenience. Although some creative professionals are concerned, Adobe emphasizes that generative AI is intended to enhance creative expression rather than replace human creativity. These updates are the latest achievements since Adobe launched the Firefly model last year, hoping to bring more innovative opportunities to the creative industry.
【AiBase Summary:】
✨ Illustrator adds a generative shape fill tool that can fill shapes through descriptive text; Text to Pattern feature generates scalable custom vector patterns.
🖼️ Photoshop updates introduce generative image capabilities, supporting higher image detail enhancement; new selection brush tool for easy selection and separation of specific objects.
🔍 Although Adobe emphasizes that generative AI aims to enhance creative expression, some creative professionals still question the potential reduction in job opportunities due to automated design.
9. Kuaishou: Adding AIGC Content "Author Declaration" Will Not Affect Traffic
Kuaishou platform has introduced a new feature called "Author Declaration," encouraging creators to add a declaration when publishing AI-generated content to reduce misunderstandings. Creators need to follow the operation guide manual and can add an author declaration to mark AI-generated content. Adding a declaration does not affect the traffic of the work unless it violates laws or rules. Updating the Kuaishou app to the latest version can solve the problem of the identification operation entry. Kuaishou hopes to cooperate with creators to jointly create a real and interesting community environment.
【AiBase Summary:】
🚀 Creators can add an author declaration when publishing AI-generated content
💡 Adding a declaration does not affect the traffic of the work unless it is in violation
🔧 Updating the Kuaishou app to the latest version can solve the problem of the identification operation entry
10. Kuaishou Kering Team's Open-Source Project LivePortrait Goes Viral, Turning a Photo into a Dynamic Video, Instant Transformation from a Middle-aged Man to a Young Girl!
Kuaishou Kering Team's open-source human video generation framework LivePortrait has caused a sensation in the field of artificial intelligence, demonstrating the huge potential of AI technology in video generation. The project has received 7.5K stars on GitHub, becoming one of the hottest AI projects at the moment. LivePortrait not only animates static images but also accurately replicates expressions onto different individuals, with a wide range of applications.
【AiBase Summary:】
🔥 LivePortrait has received 7.5K stars on GitHub, becoming a hot AI project
💡 LivePortrait can accurately replicate expressions onto different individuals, suitable for various styles
⚙️ LivePortrait generates at a speed of 12.8 milliseconds per frame, having an advantage in the video generation field
Details: https://top.aibase.com/tool/liveportrait
11. Baichuan Intelligence and Tianjin University Jointly Launch "Sibyl System" Agent Framework, Topping the GAIA Complex Task Ranking
The Sibyl System agent framework jointly launched by Baichuan Intelligence and Tianjin University has achieved first place on the GAIA Leader Board, highlighting its exceptional capabilities in complex task execution. The GAIA evaluation scheme reveals the capabilities defects of existing models, providing directions for model improvement, requiring AI to have reasoning, multi-modal understanding, web browsing, and tool use abilities. The Sibyl System framework design features include a human-like browser interface for retrieval, question-answering instead of dialogue, and the introduction of a "jury" mechanism, realizing the transformation of the model from "fast thinking" to "slow thinking."
【AiBase Summary:】
✨ Sibyl System tops the GAIA Leader Board, demonstrating exceptional capabilities in complex task execution
⚙️ The GAIA evaluation scheme reveals the capabilities defects of existing models, providing directions for model improvement
🔍 Sibyl System framework design features include a human-like browser interface for retrieval, question-answering instead of dialogue, and the introduction of a "jury" mechanism
Details: https://arxiv.org/pdf/2407.10718
12. Microsoft VALLE-2 Model Achieves Zero-Shot Text-to-Speech Technology
Microsoft's VALLE-2 model achieves speech synthesis at a human-equivalent level, considered a milestone in the TTS field. The model demonstrates astonishing instant imitation capabilities, improving decoding stability, accelerating the inference process, simplifying data requirements, and surpassing the previous VALLE model, sparking deep thinking about the ethical and responsible use of technology.
