Welcome to the AI Daily section! Here, you'll find your daily guide to exploring the world of artificial intelligence. Each 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.
Discover fresh AI products by clicking here: https://top.aibase.com/
1. iFlytek: iFly Star Fire Lite API Permanently Free
iFlytek announces the official free opening of the iFly Star Fire API capabilities, reducing developer calling costs and promoting the process of AI technology becoming more accessible.
AiBase Highlights:
🚀 Reducing developer calling costs, Pro/Max API fees as low as 0.21 yuan/10,000 tokens, with 1 token equivalent to 1.5 Chinese characters.
💡 iFly Star Fire Lite API is permanently free, developers only need 2.1 yuan to call the "iFly Star Fire 3.5 Max" to generate content.
🌍 iFlytek fully opens API capabilities, gathering global developers, and promoting the application of AI technology in various fields.
2. Kai-Fu Lee: The Insanity of Large Model Price Drops is a Lose-Lose Situation, Inference Large Model Costs Will Decline 10x Annually
Kai-Fu Lee believes that the gap between Chinese and American large models has narrowed to 6 months, showcasing the pursuit and achievements of Chinese technological strength. However, he expresses concerns about the price war in the large model market, believing that insane price cuts could lead to a lose-lose situation in the short term. He emphasizes that price cuts are not the only path to commercial success, and 01 Worlds will continue to launch the highest-performing large models.
AiBase Highlights:
🚀 The gap between Chinese and American large models has narrowed to 6 months, showing the pursuit and achievements of technological strength.
💰 Concerns arise from the price war in the large model market, with insane price cuts potentially leading to a lose-lose situation in the short term.
🔮 01 Worlds plans to launch models of various sizes, committed to being the best in China, emphasizing that large models perform excellently in complex problems.
3. Baidu Launches Comate Code Knowledge Enhancement 2.0
As an intelligent code completion and recommendation tool based on Baidu's Ernie large model, Comate Code Knowledge Enhancement 2.0 boasts three major advantages: comprehensive retrieval and Q&A, web-based online retrieval, and automatic generation of call code and test cases. It provides comprehensive support for code writing, learning, interviews, and testing, accelerating knowledge acquisition, improving programming efficiency, and generating code that aligns more closely with business needs.
AiBase Highlights:
⚙️ Comate Code Knowledge Enhancement 2.0 is based on Baidu's Ernie large model, real-time parsing of developers' programming intentions, and recommending the best code snippets and library functions.
🔍 Comate 2.0 supports comprehensive retrieval and Q&A, helping developers obtain specific business explanations in open-source frameworks, enhancing programming skills.
💻 Comate 2.0 can automatically generate call code and test cases, greatly improving code quality and efficiency.
4. Microsoft Releases the Latest Large Model Series Including Phi-3 Vision
Microsoft has released the Phi-3 series of large language models, showcasing powerful performance and flexibility, providing new tools and possibilities for researchers and developers. These models have unique advantages in specific fields, supporting different context lengths and different sizes of parameters.
AiBase Highlights:
🔍 Phi-3 Medium 14B model boasts strong performance and competitive edge, surpassing other large models such as Command R+104B and GPT3.5.
🔍 Phi-3 Small 7B model has potential application value in resource-limited scenarios, outperforming Mistral 7B and Llama 38B.
🔍 The Phi-3 series models were trained using 4.8T tokens over 42 days, supporting long text data processing, introducing new tokenizers and weight compatibility.
Details link: https://top.aibase.com/tool/phi-3-vision-128k-instruct
5. Windows 11 Introduces AI-Based Copy and Paste Features
Microsoft has introduced new advanced paste features in Windows 11, utilizing AI technology to enable instant conversion of clipboard content. This feature allows users to convert clipboard content into plain text, Markdown, JSON, and other formats, and even perform summarization, translation, and code generation, enhancing work efficiency.
AiBase Highlights:
⭐ Windows 11 introduces new advanced paste features, utilizing AI technology for instant conversion of clipboard content.
⭐ PowerToys' Advanced Paste feature allows users to convert clipboard content into plain text, Markdown, JSON, and other formats, and even use OpenAI for summarization, translation, and code generation.
⭐ Using this feature requires adding an OpenAI API key and purchasing corresponding OpenAI credits for conversion operations.
6. Adobe Lightroom Adds AI Magic Eraser Feature
Adobe Lightroom has recently launched new generative AI tools, including object removal and AI lens blur effects, providing users with a simpler and more convenient photo editing experience. These new features make background blurring more natural and make it easier to remove unwanted objects or people. Adobe continues to apply the Firefly AI model to creative software, bringing users more convenience and possibilities.
AiBase Highlights:
🎨 The "Generative Remove" feature can easily remove unwanted objects or people from photos.
📸 The AI lens blur tool can apply various blur effects with one click, enhancing the naturalness of background blur.
📱 New features make photo editing with Lightroom simpler and more convenient, providing a friendlier experience for new users.
7. Music AI Startup Suno Raises $125 Million
Suno has completed a $125 million financing round, aiming to revolutionize the music industry through its generative AI platform, sparking excitement and copyright concerns. As AI is widely applied in the music field, balancing technological advancement with protecting artists' rights will become a key issue.
AiBase Highlights:
⭐️ Suno's AI platform is expected to lower the threshold for music creation, providing powerful tools for music creators, challenging traditional music industry models.
⭐️ Suno's innovative platform allows anyone to create original songs by inputting text prompts or lyrics, and AI will generate melodies, harmonies, and complete compositions based on this.
⭐️ Suno's rapid rise and large financing occurred amid intense debates about using copyrighted music recordings to train AI models without explicit consent from artists and rights holders.
8. NVIDIA's GH200 Superchip Deployed in 9 Global Supercomputers
NVIDIA's GH200 superchip has demonstrated astonishing computing capabilities in 9 global supercomputers, driving the high-performance computing industry towards AI-driven transformation. The GH200 adopts the Grace Hopper heterogeneous computing framework, integrating high-performance GPUs and versatile CPUs, with capabilities for high-performance computing and AI model processing. NVIDIA has been continuously deploying servers and high-performance computing fields in recent years, and the GH200 series has shown strong AI and high-performance computing capabilities, accelerating scientific research processes.
AiBase Highlights:
⭐ NVIDIA's GH200 superchip achieves astonishing capabilities of 200 quintillion floating-point calculations per second in 9 global supercomputers.
⭐ The GH200 adopts the Grace Hopper heterogeneous computing framework, integrating high-performance GPUs and versatile CPUs, becoming the preferred hardware for AI supercomputing.
⭐ The GH200 series chips demonstrate strong AI and high-performance computing capabilities, accelerating scientific research processes.
9. ChatGPT's Improved Data Analysis Features Begin Rollout: Significant Boost in Speed and Interactive Experience
OpenAI has launched the latest ChatGPT data analysis optimization features, enhancing processing speed, adding interactive charts and tables, making the data analysis tool more complete and powerful. Users can now more conveniently upload, analyze, and visualize data exploration, gaining deeper data insights. This integrated solution of artificial intelligence and advanced data visualization will provide significant help for data analysts and business users.
AiBase Highlights:
🚀 Speed improvement: The speed advantage of GPT-4o makes data analysis faster.
💬 Interactive charts and tables: Users can interact with data visualization components, enhancing the utility and flexibility of the tool.
📊 Direct data display: Uploaded table data is directly displayed on the interface, allowing users to intuitively view and operate.
10. FIFO-Diffusion: Generating Infinite Length Videos Without Additional Training
The FIFO-Diffusion project showcases the powerful ability to generate infinite length videos without additional training, achieving high-quality and consistent long video generation through diagonal denoising technology. This technology has significant advantages in the field of video generation, although flicker issues may occur in complex scenes, the overall effect is impressive. Future developments are expected to resolve flicker issues and improve video generation quality and stability.
AiBase Highlights:
✨ No additional training gives FIFO-Diffusion a significant advantage.
⚙️ The core technology is a unique diagonal denoising method.
🔗 FIFO-Diffusion demonstrates great potential for generating long videos.
Details link: https://top.aibase.com/tool/fifo-diffusion
11. Post-Apocalyptic Filter Makes It Hard to Tell Real from AI
This article introduces a post-apocalyptic game filter video posted by Douyin blogger "速速來電," where the AI-generated images are so realistic that it's hard to distinguish them from the real images. Netizens are热议 the video, praising its realistic effects. The blogger shared a tutorial on how to create the post-apocalyptic effects using the Meitu XiuXiu app, emphasizing the importance of the original image quality for the final effect.
AiBase Highlights:
🔍 The video shows realistic effects, making it hard to distinguish between real and AI-generated images.
📈 The video received 170,000 likes and 24,000 shares, becoming the blogger's best-performing work.
💡 The tutorial shares tips, emphasizing the importance of the original image quality for the final effect.
12. Bilibili UP Master's Creative Stunts: AI Teaching English with Puns
This article introduces how Bilibili UP masters use AI technology to synthesize celebrity voices and images to teach English, garnering widespread attention. Through puns, the video has received over 1 million plays on Bilibili. The UP master combines entertainment and education to create a unique learning experience, leveraging AI technology to increase video production frequency and fan numbers. However, using celebrity or TV series materials also carries the risk of copyright infringement.
AiBase Highlights:
🤖 Using AI technology to synthesize celebrity voices and images for English teaching, attracting a large number of fans.
🎬 The video uses puns to receive over 1 million plays, enhancing the fun and interactivity of learning.
💡 The UP master combines entertainment and education to create a unique learning experience, leveraging AI technology to increase video production frequency and fan numbers.
13. Google's "Say What You See" Game Enhances Image Reading and AI Prompt Writing
