Welcome to the 【AI Daily】 section! Here is your daily guide to exploring the world of artificial intelligence. Every day, we present you with hot topics in the AI field, focusing on developers to help you gain insights into technology trends and innovative AI product applications.
Fresh AI products click to learn more: https://top.aibase.com/
1. OpenAI officially launches Sora, allowing ChatGPT Pro users to generate unlimited videos for up to 20 seconds
OpenAI has recently released its new AI video generation software, Sora Turbo, which allows users to generate various videos from text or static images. The software provides different generation limits and resolution options for ChatGPT Plus and Pro users. Although Sora Turbo performs excellently in video generation, there are still some content generation limitations and challenges, especially when compared to other competitors.
【AiBase Highlights:】
🌟 Sora Turbo is officially launched, supporting text and image generation of various videos, covering most countries and regions.
🎥 Users can easily generate and manage videos on the new interface, with a plot outline feature for smoother editing transitions.
⚠️ Sora Turbo has strict content generation restrictions aimed at preventing the creation of realistic portraits and violent content.
Details link: https://sora.com/
2. Zhiyuan AI launches free multimodal model GLM-4V-Flash: Improved image processing accuracy
Beijing Zhiyuan Huazhang Technology Co., Ltd. has launched its first free multimodal API - GLM-4V-Flash, aimed at enhancing image processing accuracy and lowering the entry barrier for developers. This model supports multiple languages and includes advanced image processing features such as image description generation and visual question answering, providing precise solutions for specific industries. 
【AiBase Highlights:】
🌐 GLM-4V-Flash is the first free multimodal API, supporting 26 languages and lowering development barriers.
📊 It includes advanced features such as image description generation, classification, and visual reasoning, applicable across multiple industries.
🚀 The model has shown significant benefits in social media, education, beauty, and other fields.
Details link: https://www.bigmodel.cn/console/trialcenter
3. Tencent Cloud AI Code Assistant launched, built on a hybrid large model
The AI Code Assistant launched by Tencent Cloud aims to help programmers enhance development efficiency by predicting and providing code suggestions. This tool utilizes a hybrid large model to deeply understand code context and provide accurate code completion suggestions, surpassing traditional keyword matching methods. It adapts to programmers' coding styles and has demonstrated strong coding assistance capabilities in several key scenarios, such as generating regular expressions, quickly creating front-end pages, and clearly interpreting complex code.
【AiBase Highlights:】
⚙️ The AI Code Assistant provides accurate code completion suggestions by deeply understanding code context, significantly enhancing development efficiency.
📈 This assistant learns programmers' coding styles, offering customized code completion that aligns with personal habits.
🔍 Through the hybrid large model, the AI Code Assistant exhibits strong capabilities in various scenarios, including generating regular expressions and quickly adapting to new interface specifications.
4. Keling AI API V1.5 model adds standard std mode, V1.0 model adds motion brush
Beijing Kuaishou Technology Co., Ltd. recently launched the Keling AI API V1.5 model standard mode and the "motion brush" feature for the V1.0 model. These updates aim to enhance user experience and increase the flexibility and efficiency of artistic creation. The V1.5 model offers excellent results and fast processing speed, providing users with a cost-effective choice, while the new feature in the V1.0 model allows users to specify motion trajectories for characters or objects in images, leading to more precise motion control and vivid expression.
【AiBase Highlights:】
✨ The V1.5 model standard mode provides excellent results and fast processing speed, enhancing user experience.
🖌️ The new "motion brush" feature in the V1.0 model allows users to specify motion trajectories for precise control.
🌟 The new features enrich Keling AI's capabilities, bringing innovative possibilities for visual art creation.
5. Shusheng · Wanxiang multimodal large model InternVL 2.5 open-source performance rivals GPT-4o
Shanghai AI Lab has launched the Shusheng · Wanxiang InternVL2.5 model, which has achieved over 70% accuracy on multimodal understanding benchmarks, making it the first open-source model comparable to commercial models like GPT-4o and Claude-3.5-Sonnet. The model enhances performance through chain-of-thought reasoning techniques and demonstrates strong scalability and multidisciplinary reasoning capabilities across multiple fields.
【AiBase Highlights:】
🚀 The InternVL2.5 model has achieved over 70% accuracy on multimodal understanding benchmarks, demonstrating outstanding performance.
📈 Through chain-of-thought reasoning techniques, the model has achieved a 3.7 percentage point performance improvement, showcasing strong scalability.
🌐 The open-source nature allows researchers and developers to freely access and use the model, promoting the development of multimodal AI technology.
Details link: https://www.modelscope.cn/collections/InternVL-25-fbde6e47302942
6. Swift Ventures launches AI Company Index clarifying AI investment standards
Swift Ventures has launched a new AI company index aimed at helping investors identify publicly traded companies genuinely investing in AI technology. The index analyzes thousands of data points and finds that, despite companies frequently mentioning AI in financial reports, very few are making substantial investments. Currently, 90 tracked companies excel in AI research and talent density, with annual growth rates significantly surpassing the market average.
【AiBase Highlights:】
📊 The index tracks about 90 companies, scoring them based on AI research investment, talent density, and AI revenue.
💡 Companies investing in AI research have an average gross profit twice that of non-investing companies, indicating a positive correlation between research and profitability.
🚀 Some low-profile companies have performed exceptionally well in the AI field, with annual growth rates exceeding 50%, indicating that AI transformation has surpassed major tech companies.
7. Quantum leap in computing! Google's Willow chip completes a task in 5 minutes that would take 138 billion years on a traditional computer, leaving OpenAI astonished
Google's Willow quantum chip has achieved a groundbreaking breakthrough in quantum computing, successfully reducing computation time from 10^25 years on traditional computers to just 5 minutes, showcasing the immense potential of quantum technology. Through meticulous engineering design, Willow significantly reduces computational errors while increasing the number of quantum bits, advancing the field of quantum computing.
【AiBase Highlights:】
⚡ The Willow chip achieves below-threshold error control in quantum computing, significantly reducing error rates.
⏱️ The computation speed is astonishing, completing a task that would take 10^25 years in just 5 minutes, demonstrating the immense potential of quantum computing.
🔒 The advancements of Willow raise concerns about encryption security, particularly regarding potential threats to cryptocurrencies like Bitcoin.
8. A blessing for introverts! VR role-playing AI arrives, with Nanyang Technological University making breakthroughs in "human creation," capable of singing, dancing, interacting, and chatting with you!
A research team from Nanyang Technological University in Singapore has launched an AI technology called SOLAMI, capable of creating lifelike 3D virtual characters that support real-time interaction, voice understanding, and action response. This technology utilizes deep learning to convert users' voices and actions into a language understandable by virtual characters, providing a natural and smooth interactive experience. SOLAMI is also equipped with a VR interface, allowing users to interact face-to-face with virtual characters using VR devices.
【AiBase Highlights:】
🎮 SOLAMI is an end-to-end social visual-language-action modeling framework that enables natural interaction between users and virtual characters.
📊 The SynMSI synthetic dataset provides rich dialogue and action data for training, addressing data scarcity issues.
🌐 The immersive VR interface of SOLAMI allows users to interact with virtual characters in a highly engaging manner, enhancing the social experience.
Details link: https://solami-ai.github.io/
9. X officially announces the launch of the new AI image generator Aurora for all users within this week
Recently, the social network X (formerly Twitter) launched a new image generator called Aurora, trained on billions of samples, capable of generating high-quality images. Although it was initially taken down, it has now been relaunched and is set to be promoted to all users within a week. Aurora can accurately render visual details of the real world, although testing has revealed occasional issues with unnatural blending and missing details in the generated images.
【AiBase Highlights:】
✨ Aurora is a new image generator developed by xAI, featuring photo-level rendering capabilities.
🌍 It is currently available in some countries, with plans to promote it to all users within a week.
🔍 Testing has found that images generated by Aurora sometimes exhibit unnatural blending and missing details.
Details link: https://x.ai/blog/grok-image-generation-release
10. Reddit launches AI Q&A feature, but users are not impressed!
Reddit recently introduced a new feature called "Reddit Answers," aimed at enhancing user search experiences through AI-driven Q&A. However, despite the feature's ability to provide answers based on posts and comments on the platform, user feedback has been lukewarm, with many believing that improving search functionality should take priority. The feature is currently being tested among a limited number of users in the U.S. and has not yet been launched on the Android platform.
【AiBase Highlights:】
🔍 The new feature "Reddit Answers" is currently being tested among limited users in the U.S., aimed at enhancing search experiences.
🤖 This feature utilizes posts and comments on the Reddit platform to provide AI-driven Q&A services.
😟 User responses have been mixed, with many expressing dissatisfaction regarding the prioritization of search functionality improvements.
11. Tesla's Tao Lin: Committed to a pure vision approach for autonomous driving
Tesla Vice President Tao Lin reaffirmed the company's commitment to a pure vision approach in autonomous driving technology. She emphasized that only by combining cameras with visual neural networks can the company better simulate human driving habits, leading to safer and smarter fully autonomous driving. Tesla's AI4 chip is now equipped in all its sold models, significantly enhancing computing power and marking the company's readiness for fully autonomous driving from a hardware perspective.
【AiBase Highlights:】
🔍 Tesla insists on achieving fully autonomous driving through pure vision technology, believing it to be the safest and smartest solution.
💡 The autonomous driving technology employs an end-to-end large model, achieving the entire process from photon input to decision output.
📈 All sold models are equipped with the latest AI4 chip, with a fivefold increase in computing power, laying the foundation for achieving fully autonomous driving.
12. Remarkable recovery! Stability AI's new management team achieves debt-free status and triple-digit business growth in six months
Under the leadership of new CEO Prem Akkaraju, Stability AI has successfully achieved triple-digit growth and cleared all debts within six months. Akkaraju emphasized the company's healthy balance sheet and focused on the rapid development of API and licensing services. The formation of the new management team has attracted back investors who had previously left, signaling a positive outlook for the company's future.
【AiBase Highlights:】
💼 Stability AI's new CEO Prem Akkaraju stated that the company's business has achieved triple-digit growth and is now debt-free.
📈 The new management team completed the recovery within six months, attracting back previously departed investors.
🎥 Notable director James Cameron has joined the Stability AI board, reflecting renewed confidence in the industry.
