Welcome to the 【AI Daily】 column! Here is your guide to exploring the world of artificial intelligence every day. Each day, we present you with the hottest 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. ByteDance launches the Doubao visual reasoning model: prices as low as 0.003 yuan per thousand tokens
At the Volcano Engine FORCE conference, the president of Volcano Engine, Tan Dai, unveiled the Doubao visual understanding model, which demonstrates exceptional content recognition and reasoning capabilities by combining text and image information. The significant price reduction of the new model gives corporate users more confidence in their digital transformation. The daily token usage of the Doubao model has exceeded 4 trillion, indicating strong market demand and application potential.
【AiBase Highlights:】
🖼️ The newly launched Doubao visual understanding model can simultaneously process text and image information, enhancing content recognition and reasoning capabilities.
💰 Since May this year, the industry price of the Doubao model has decreased by 99%, making it easier for corporate users to adopt this technology.
📈 The daily token usage of the Doubao model has surpassed 4 trillion, growing more than 33 times, showing its market demand.
2. OpenAI releases the full o1 model API: 60% cost reduction and new advanced visual processing capabilities
During a continuous 12-working-day live event, OpenAI released the o1 model API to developers on the 9th day and announced a significant upgrade to the real-time API, supporting WebRTC technology. From the release date, OpenAI will provide access to the o1 API for developers at usage level 5. The updated o1 model API has achieved a 60% reduction in thinking costs compared to the previous preview version and added advanced visual processing capabilities. Additionally, the cost of audio processing with GPT-4o has also decreased by 60%, and the price of the mini version has dropped tenfold. The o1 model API also introduces a parameter called “reasoning_effort,” allowing developers to control the time the model takes to think before answering questions. Furthermore, OpenAI has launched a new preference tuning technology that enables large models to adapt more accurately to users' personalized styles through direct preference optimization algorithms. OpenAI's weekly active user count has exceeded 300 million, with users sending over 1 billion messages to ChatGPT daily. Compared to August of this year, when the weekly active user count had just surpassed 200 million, this shows the rapid growth of OpenAI's user base. These updates and data releases not only showcase OpenAI's technological advancements in the field of artificial intelligence but also reflect its widespread influence among global users.
【AiBase Highlights:】
🚀 The o1 model API is released, supporting WebRTC technology to enhance real-time interaction capabilities.
💰 Costs have been reduced by 60%, and new advanced visual processing capabilities have been added to enhance user experience.
📈 Weekly active users exceed 300 million, indicating rapid growth in OpenAI's user base.
3. Ideogram launches batch image generation tool: Say goodbye to cumbersome operations, generate large-scale creative images with one click
The AI image generation platform Ideogram has recently launched a batch image generation tool aimed at simplifying the image generation process by allowing users to upload spreadsheet files. Users can pre-fill prompts and settings in a CSV file, and Ideogram will automatically generate images based on this information. This innovation significantly improves the work efficiency of professional designers and creative individuals, reducing the tedious process of inputting data line by line. This feature is currently only available to Ideogram Pro users, showcasing the immense potential of AI in the design field and intelligent creative methods.
【AiBase Highlights:】
🚀 The batch generation tool allows users to upload a table containing prompts, simplifying the image generation process.
🖼️ Users only need to download a template, generate prompts, and upload a CSV file to automatically generate images.
💼 This feature is currently only available to Ideogram Pro users, providing designers with an efficient creative experience.
4. Jidream AI launches poster generation feature: Transform static posters into dynamic ones with one click
Jidream AI introduced a new poster generation feature at the Volcano Engine FORCE conference on December 18, 2024. The release of this technology marks a significant advancement in the field of image generation. Users only need to input a simple description, and the system can quickly generate creative posters, greatly simplifying the time and skill requirements of traditional design. Additionally, the new dynamic poster generation feature provides content creators with richer display options, especially suitable for social media and advertising, effectively attracting audience attention and enhancing marketing effectiveness.
【AiBase Highlights:】
🌟 Users can quickly generate creative posters with just a description, simplifying the creative process.
🎥 The new dynamic poster generation feature makes the presentation of works more vivid, suitable for social media and advertising.
📈 Jidream AI considers users' personalized needs, offering flexible content generation options to assist brand promotion.
5. Coze 1.5 officially launched: Supports multimodal capabilities and allows immediate experience of the new Doubao model
Coze introduced the brand-new version 1.5 at the Volcano Engine FORCE conference, marking significant progress in the field of AI application development. This version supports a GUI building interface, allowing users to easily create and publish various application forms, greatly lowering the development threshold. Meanwhile, Coze 1.5 enhances multimodal capabilities, supporting the latest Doubao model, providing rich templates and solutions to help developers improve efficiency, and has attracted over 1 million active developers.
【AiBase Highlights:】
🖥️ Coze 1.5 supports a GUI building interface, allowing users to publish various application forms with one click, lowering the development threshold.
🌐 Multimodal capabilities have been significantly enhanced, supporting Doubao visual understanding, music, and image generation models, expanding the scope of AI applications.
📊 Offers a vast array of high-quality templates covering multiple business scenarios, improving development efficiency, attracting over 1 million active developers.
Details link: https://www.coze.cn/docs/guides/vlm
6. ByteDance: Doubao video generation model will officially open services in January 2025
At the 2024 Volcano Engine FORCE conference, ByteDance showcased the new upgrades of the Doubao model family, with daily token usage exceeding 4 trillion, showing significant growth. The conference introduced the visual understanding model and upgrades to multiple models, enhancing the comprehensive task handling capabilities of the Doubao general model pro. Additionally, the Volcano Engine announced the release of the veOmniverse + Doubao 3D generation model supporting AIGC creation and declared that the Doubao video generation model will officially open services in January 2025, marking a deep development of large model technology.
【AiBase Highlights:】
🌟 The daily token usage of the Doubao model exceeds 4 trillion, growing more than 33 times, indicating widespread application.
🛠️ The newly released veOmniverse + Doubao 3D generation model supports high-fidelity 3D asset generation and editing, enhancing AIGC creation capabilities.
📅 The Doubao video generation model will officially open services in January 2025, and users can make reservations to experience it.
7. ByteDance Volcano Engine launches global AI search: supports multimodal search
At the 2024 Volcano Engine FORCE conference, ByteDance launched a global AI search service aimed at improving the accuracy of enterprise recommendations and information discovery by integrating various information and needs. This service relies on a powerful AI search engine, supporting multimodal understanding, quickly processing massive content, and providing real-time hot answers to enhance user experience. Meanwhile, the Volcano Engine also introduced a large model memory solution to help clients build efficient memory systems, which is an important direction for large model development.
【AiBase Highlights:】
🌐 The Volcano Engine global AI search integrates scenario-based search, enterprise private domain information, and networked Q&A services, enhancing the accuracy of information recommendations.
⚙️ The AI search engine utilizes the technology of the Doubao model family, supporting multimodal understanding of text, images, audio, and video, suitable for various application scenarios.
💡 The large model memory solution combines context caching and RAG technology to help clients build effective memory systems, enhancing the memory capabilities of large models.
8. WeChat launches "Author Reading Voice" new capability
The "Author Reading Voice" feature launched by the WeChat platform allows public account authors to voice their articles with personalized voices, enhancing the interactivity and personalization of the reading experience. Authors need to download the "Subscription Account Assistant" app to record their voices, replicating their personal tone and emotions for use in public accounts. This feature is currently in a gray testing phase and has not been fully opened; WeChat encourages creators to be patient. This move marks an important advancement for WeChat in enhancing user experience and meeting creator needs, and is expected to enrich the content presentation forms of public accounts.
【AiBase Highlights:】
🎧 Authors can use personalized voices to voice their articles, enhancing interactive experiences.
📱 Downloading the "Subscription Account Assistant" app is necessary to record voices and replicate personal styles.
🔄 This feature is currently in a gray testing phase and has not been fully opened.
9. NVIDIA releases generative AI supercomputer: only $249, performance improved by 1.7 times
NVIDIA's Jetson Orin Nano Super is a generative artificial intelligence supercomputer aimed at developers, priced at $249, with significant performance improvements suitable for various AI application scenarios. The device has improved generative AI performance by 1.7 times and also shows significant advancements in memory bandwidth and computational power. Jensen Huang emphasized that this device provides outstanding computational performance for developers at a lower cost, showcasing its broad application potential in smart cities, agriculture, and robotics, marking an important step in the popularization and application of AI technology.
【AiBase Highlights:】
🚀 Performance boost: The generative AI performance of Jetson Orin Nano Super has improved by 1.7 times, with memory bandwidth increased by 50%.
💰 Affordable pricing: The device is priced at $249, making it suitable for developers and lowering the barriers to AI technology.
🌍 Wide applications: Supports various power consumption scenarios, suitable for smart cities, agriculture, and robotics.
10. OpenAI states: No plans to launch Sora API yet, video generation demand exceeds expectations
OpenAI recently announced that there are currently no plans to launch its video generation model Sora's API, due to user demand far exceeding expectations. Sora can generate realistic videos based on text or images, but due to the surge in user applications, OpenAI has had to pause new user registrations. CEO Sam Altman apologized for this and emphasized that solving this issue will take time. Meanwhile, competitors like Google and AWS have launched their own video generation APIs, putting OpenAI under market pressure, and future strategies are closely watched.
【AiBase Highlights:】
🌟 OpenAI has stated that there are no plans for the Sora API launch due to demand far exceeding expectations.
📈 Registration for Sora is temporarily closed due to a surge in user applications, and the CEO has apologized for this.
🤖 Competitors like Google and AWS have launched video generation APIs, putting pressure on OpenAI.
11. AI "magically modifies" pets dancing to go viral online: curiosity and absurdity become traffic passwords
Recently, AI-generated pet dancing videos have gone viral on Douyin, showcasing a perfect blend of absurdity and humor. The cats and dogs in these videos instantly transform into dance masters, delivering a strong visual impact and drama. While some viewers feel uncomfortable with this peculiar visual experience, it undoubtedly challenges our inherent perceptions of animal images, demonstrating the infinite possibilities and creativity of AI technology.
【AiBase Highlights:】
🎉 AI-generated pet dance videos have quickly gone viral on Douyin, with views reaching 880 million.
😹 The pets in the videos showcase surreal dance moves, breaking traditional perceptions and providing a strong visual impact.
🤖 These videos are not only a demonstration of technology but also a new dimension of entertainment and creativity, challenging people's understanding of animal imagery.
12. AI pet Moflin becomes popular on Xiaohongshu without needing to be fed
Moflin is a new type of AI pet that has quickly gained popularity on Xiaohongshu due to its cute appearance and emotional interaction features. Users share their interaction experiences with Moflin through videos, attracting many online viewers. Although Moflin cannot replace real pets, its emotional companionship meets the needs of modern people, becoming a new consumer trend. The emotional simulation and personalized interactions of Moflin give it substantial premium potential in the market, making it a new type of emotional companion product.
【AiBase Highlights:】
🐾 Moflin is an emotionally interactive AI pet with a cute appearance, attracting a lot of online attention.
💰 Priced at 2832 yuan, it quickly sold out after launch, showing strong market demand.
❤️ Moflin interacts with users through emotional simulation, meeting the need for emotional companionship.
Details link: https://www.moflin.com/
13. Boston Dynamics lays off 5% of staff due to financial pressure and urgent need for transformation
Boston Dynamics recently announced a 5% layoff, affecting about 45 employees across nearly all departments. The company is facing severe financial pressure, and although its robotic products like Spot and Atlas have gained market attention, business development has not met expectations. CEO Robert Playter pointed out that the rate of capital consumption exceeds revenue growth, necessitating urgent operational optimization for sustainable development. In a fiercely competitive market environment, Boston Dynamics needs to address pressure from companies like Tesla, making transformation a priority.
【AiBase Highlights:】
🦾 Boston Dynamics lays off 5% of staff, affecting about 45 employees across nearly all departments.
💰 The company faces issues of rapid capital consumption and urgently needs to streamline operations for sustainable growth.
🤖 With increasing market competition, Boston Dynamics must respond to pressure from companies like Tesla and is struggling to convert media attention into profits.
14. Hundreds of OpenAI employees are set to gain $10 million through private stock sales
Recently, OpenAI announced a $1.6 billion stock buyback for SoftBank, allowing hundreds of current and former employees to potentially gain up to $10 million from this transaction. This news has garnered widespread attention, especially for those who joined the company early, as they may achieve financial freedom. The stock sale not only motivates employees but also strengthens the trust relationship between the company and its investors, showcasing OpenAI's potential and value as an innovative company.
【AiBase Highlights:】
💰 Hundreds of current and former OpenAI employees will have the opportunity to gain up to $10 million through the stock buyback.
📈 OpenAI's $1.6 billion stock buyback proposal for SoftBank has attracted widespread attention.
🤝 This stock sale not only motivates employees but also enhances the trust relationship between the company and its investors.
