Welcome to the 【AI Daily】 column! This is your daily guide to exploring the world of artificial intelligence. Every day, we present you with the hottest topics in the AI field, focusing on developers, helping you gain insights into technology trends and innovative AI product applications.
Fresh AI products Click to learn more: https://top.aibase.com/
1. Baichuan Intelligence Releases the Full-Scene Deep Thinking Model Baichuan-M1-preview Now Live on Baixiao App
Baichuan Company has launched the Baichuan-M1 series models today, including the full-scene deep thinking model Baichuan-M1-preview and the open-source medical enhancement model Baichuan-M1-14B. Both models excel in technological innovation and performance, particularly the Baichuan-M1-preview, which has surpassed competitors in multiple authoritative evaluations, demonstrating powerful deep thinking capabilities and medical evidence-based models, providing strong support for applications in the medical field.
【AiBase Highlights:】
🧬 Baichuan-M1-preview is the first domestic model with language, visual, and search reasoning capabilities, performing exceptionally well.
🏥 Baichuan-M1-14B outperformed larger parameter models in medical knowledge and clinical ability evaluations, showcasing strong medical capabilities.
🚀 Baichuan Company aims to inspire innovation and promote the widespread application of medical technology through the open-source Baichuan-M1-14B.
2. OpenAI Launches the First AI Agent Operator, Initially for ChatGPT Pro Users
The newly launched AI agent Operator by OpenAI is designed to assist users in performing various tasks online, initially targeting ChatGPT Pro users. This tool combines advanced visual capabilities with reinforcement learning, enabling interaction with web pages and self-correction features. The Operator is designed with a focus on security, ensuring that users maintain control when handling sensitive information.
【AiBase Highlights:】
🌐 OpenAI introduces the “Operator” AI agent, assisting users in online task execution, initially for ChatGPT Pro users.
🖱️ The Operator interacts with web pages through the browser, featuring self-correction and user control functions to ensure safety.
🤝 OpenAI collaborates with several well-known companies to meet real-world needs and plans to expand to more users in the future.
Details link: https://openai.com/index/introducing-operator/
3. HeyGen Launches Digital Human Motion Control Feature: Can Play Instruments and Dance
HeyGen's latest digital human motion control system enables significant control over virtual character movements. This technological breakthrough allows digital humans to not only perform basic micro-expressions but also smoothly execute complex physical actions such as playing instruments and dancing. By incorporating kinematic control algorithms, the motion response delay has been reduced to 12 milliseconds, greatly enhancing video production efficiency.
【AiBase Highlights:】
🎹 HeyGen's digital human motion control system enables complex physical actions, allowing smooth execution of instrument playing and dance performances.
💡 The system generates virtual characters using deep neural networks, supporting real-time generation of over 200 joint position data, showcasing biomechanical features.
🚀 Video production efficiency increased by about 47%, with dynamic scene production costs reduced to 1/8 of traditional methods, and future integration of haptic feedback simulation planned.
Details link: https://app.heygen.com/
4. Perplexity Launches Android Mobile Assistant: Can Write Emails and Book Dinners
Perplexity has recently launched a new AI assistant designed for Android users, capable of performing various tasks such as writing emails, setting reminders, and booking dinners. This assistant features multimodal capabilities, recognizing screen content and identifying surroundings through the camera, enhancing user convenience. In practical experience, the assistant's response speed and accuracy are impressive, although it is still expanding supported applications and functions, its potential is already evident.
【AiBase Highlights:】
🌟 The assistant supports various functions such as writing emails, setting reminders, and booking restaurants.
📱 It features multimodal capabilities, able to recognize screen content and identify surroundings through the camera.
🚀 Currently supports applications like Spotify, YouTube, and Uber, with functionality still expanding.
5. Yuanxiang Launches Intelligent Digital Human Platform "Yuanxiang Daily Broadcast"
Shenzhen Yuanxiang Information Technology Co., Ltd. has launched the intelligent digital human platform "Yuanxiang Daily Broadcast," providing innovative solutions for brand presentation and content production with its high natural customization capabilities and real-time interaction features. The platform's integrated tools allow users to quickly set up live broadcast spaces and achieve audience interaction through self-developed large models, significantly enhancing user experience.
【AiBase Highlights:】
🎥 The Yuanxiang Daily Broadcast platform provides one-stop live broadcasting tools, allowing users to quickly set up professional live broadcast spaces.
🗣️ Through voice cloning technology, users can easily customize personalized digital human images and voices.
📈 The platform has been widely applied across various industries, significantly increasing clients' sales conversion rates.
6. 300 Times Volume Reduction! Hugging Face Launches SmolVLM Model: Compact and Intelligent, Runs on Mobile
Hugging Face's SmolVLM model leads a new trend in AI technology with its compact size and outstanding performance. This model not only runs on mobile devices but also outperforms its predecessor Idefics80B, which requires large data center support, marking significant progress in practical AI deployment.
【AiBase Highlights:】
🌟 The SmolVLM model can run on mobile, outperforming the 300 times larger Idefics80B model.
💰 The SmolVLM model helps businesses significantly reduce computing costs, achieving processing speeds of 16 instances per second.
🚀 The technological innovation of this model enables small businesses and startups to launch complex computer vision products in a short time.
Details link: https://huggingface.co/blog/smolervlm
7. China Unicom Releases the Yuanjing Thinking Chain Large Model: Performance Surpasses GPT-4
China Unicom has recently launched the Yuanjing Thinking Chain large model, marking an important advancement in the field of artificial intelligence. This centrally enterprise open-source general thinking chain large model demonstrates exceptional slow thinking and multi-scenario reasoning capabilities, outperforming currently the best general language models, such as OpenAI's GPT-4, in multiple evaluations.
【AiBase Highlights:】
🚀 The Yuanjing Thinking Chain large model is the first centrally enterprise open-source general thinking chain large model from China Unicom, featuring strong slow thinking and reasoning capabilities.
📊 In mainstream evaluation rankings, this model outperformed OpenAI GPT-4 and other top language models, demonstrating its competitiveness.
🔍 The model achieves task and difficulty adaptation, enhancing response efficiency and accuracy, and has been successfully applied in various fields.
Details link: https://github.com/UnicomAI/Unichat-32B-c1.git
8. Enthusiasts! A Foreign Software Engineer Purchases OGOpenAI.com Domain and Redirects to DeepSeek
Recently, software engineer Ananai Arora purchased the domain OGOpenAI.com at a very low price and redirected it to the Chinese AI lab DeepSeek. DeepSeek's groundbreaking advancements in open-source AI have attracted widespread attention, with its AI models outperforming OpenAI's o1 model in certain benchmark tests. In contrast, OpenAI has appeared relatively cautious in releasing powerful models, facing criticism from the industry.
【AiBase Highlights:】
🌐 Ananai Arora purchases the OGOpenAI.com domain and redirects it to DeepSeek.
📊 DeepSeek's AI models outperform OpenAI's o1 model in certain benchmark tests.
🔍 OpenAI has faced criticism for failing to release powerful models, facing scrutiny from the industry.
9. OpenAI CEO Announces ChatGPT Free Version Will Introduce o3-mini, Doubling Efficiency!
OpenAI CEO Sam Altman announced that the free version of ChatGPT will upgrade to the new o3-mini model, aimed at enhancing user experience and meeting daily needs. Paid users will receive more opportunities for use, boosting productivity. This move not only promotes the democratization of artificial intelligence but also provides millions of users access to cutting-edge technology while ensuring the value experience for paid users. 
【AiBase Highlights:】
🌟 Free users will enjoy the newly upgraded o3-mini model, offering faster response speeds.
💼 Paid users will receive more opportunities to use o3-mini, enhancing productivity.
📈 OpenAI is committed to promoting the democratization of artificial intelligence, ensuring more users enjoy cutting-edge technology.