Welcome to the AI Daily column! This is your daily guide to exploring the world of artificial intelligence. Every day, we bring you the hottest topics in the AI field, focusing on developers, helping you understand technological trends and innovative AI product applications.
Fresh AI Products Click to Learn More: https://top.aibase.com/
1. OpenAI Launches ChatGPT Advanced Voice Mode with Five New Voice Styles
OpenAI has announced the launch of a new advanced voice mode for ChatGPT Plus and Team users, offering a personalized communication experience. Users can choose from five voice styles and speeds, supporting up to 50 languages, enhancing the fluency and personalization of voice communication. This new feature makes ChatGPT more widely applicable in fields such as education, law, business, and healthcare, providing users with a better experience.
AiBase Highlights:
🎤 Advanced voice mode: Supports up to 50 languages, providing a personalized communication experience.
🎶 Customizable interaction: Users can choose from five voice styles and speeds, making personalized communication more flexible.
🌍 Wide application scenarios: Voice mode is widely used in education, law, business, and healthcare, enhancing user experience.
2. Google Gemini 1.5 Upgrade: Performance Surge, Price Cut in Half
Google today announced the release of the new upgraded Gemini model series, including Gemini-1.5-Pro-002 and Gemini-1.5-Flash-002. This update not only significantly enhances performance but also offers a surprising price reduction, which is sure to create a stir in the AI development community. The Gemini 1.5 upgrade brings higher performance and lower costs to developers, along with more choices and flexibility.
AiBase Highlights:
✨ Significant price reduction, performance improvement, and increased development efficiency
⚙️ New Gemini models have been comprehensively enhanced in mathematics, long text processing, and visual tasks
💡 Gemini 1.5Pro's long text processing and multimodal capabilities open up new application scenarios
3. Alibaba's New Technology MIMO: Turn a Flat Image into an Anime Hero in an Instant
The MIMO technology developed by the Alibaba Group's Intelligent Computing Research Institute has completely revolutionized the way animated characters are created. With just a static image and simple motion instructions, it can be transformed into a manipulable virtual character, whether it's a real person, cartoon, or anthropomorphic character. MIMO is simple and efficient to operate, eliminating the need for multi-angle shooting or separate character training, integrating 2D video information and 3D spatial modeling. It has a wide range of applications, can extract complex movements, and achieve real-scene interaction, lowering the threshold for animation production and enhancing the realism and immersion of animations.
AiBase Highlights:
🎨 Innovative tool MIMO revolutionizes animation character creation, transforming simple static images and motion instructions into manipulable virtual characters.
🔄 MIMO is simple and efficient to operate, eliminating the need for multi-angle shooting or separate character training, integrating 2D video information and 3D spatial modeling.
🌐 Wide application range, capable of extracting complex movements and achieving real-scene interaction, lowering the threshold for animation production, and enhancing the realism and immersion of animations.
Details link: https://menyifang.github.io/projects/MIMO/index.html
4. Xunfei Spark API Major Upgrade: Lite Version Permanently Free, Max Version Offers 100 Million Tokens for Free
Xunfei Open Platform announces a significant upgrade to the Xunfei Spark API platform, including the Spark Max model and 4.0Ultra model, which improves key performance such as generation speed, logical reasoning, creative capabilities, and internet search. The upgraded models perform faster and more accurately in logical reasoning, generate articles with logic and practicality, support long text task processing, maintain the same price, and offer a free Lite version and promotional activities. The performance has been comprehensively upgraded, surpassing the internationally leading GPT-4Turbo.
AiBase Highlights:
🚀 Improve generation speed and key performance, including logical reasoning, creative capabilities, and internet search.
💡 Models perform faster and more accurately in logical reasoning, providing detailed reasoning processes.
📝 New models generate articles with logic and practicality, support long text task processing, maintain the same price, and offer a free Lite version and promotional activities.
Details link: https://xinghuo.xfyun.cn/sparkapi
5. Google's New Voice Cloning Technology: Just a Few Seconds of Audio Sample for Voice Cloning
In today's rapidly advancing technology, Google researchers have proposed a zero-shot voice conversion technology to help people with speech loss regain their voice memories. This technology has zero-shot capabilities, supports multi-language voice conversion, and demonstrates strong adaptability and practicality. Through a short audio sample, it successfully synthesizes the voice of a specific speaker, greatly enriching the possibilities of voice communication.
AiBase Highlights:
🎤 Zero-shot voice conversion technology: No need for a large number of samples, helps people with speech loss regain their voices.
🌍 Language capabilities: Achieves voice conversion between different languages, enriching the possibilities of voice communication.
🗣️ Application for specific speakers: Successfully synthesizes the voice of a specific speaker through a short audio sample, demonstrating the technology's adaptability and flexibility.
Details link: https://google.github.io/tacotron/publications/zero_shot_voice_transfer/
6. Shengshu Technology's Video Generation Model Vidu Opens API
At the Baidu Cloud Smart Conference, Shengshu Technology announced that its video large model Vidu has officially opened its API and connected to Baidu Smart Cloud's Qianfan large model platform, becoming the first video large model. Vidu has leading advantages in high dynamics, multi-style, and extreme reasoning, solving the problem of consistent video model generation, and is expected to accelerate video creation in the film, animation, and advertising industries.
AiBase Highlights:
🚀 Vidu opens API and connects to Baidu Smart Cloud's Qianfan large model platform, becoming the first video large model.
💡 Vidu has leading advantages in high dynamics, multi-style, and extreme reasoning, solving the problem of consistent video model generation.
💼 Vidu is expected to accelerate video creation in the film, animation, and advertising industries, reducing costs, improving efficiency, and stimulating innovative thinking.
7. Big Leap! Director James Cameron of Titanic Joins Stability AI Board
James Cameron's joining of the Stability AI board has caused a sensation in the film industry, integrating AI technology with film art to open up innovative ways of storytelling. Cameron's collaboration with the Stability AI team reshapes the future of visual media, which is highly anticipated.
AiBase Highlights:
📽️ Cameron Joins Stability AI: The legendary Hollywood director joins the board of the artificial intelligence company, bringing a significant victory for the company.
🤖 Combination of AI and CGI: Cameron believes that the integration of generative AI and CGI will drive innovation in storytelling.
🌟 Powerful Industry Alliance: Stability AI introduces the former president of Facebook, enhancing the company's industry influence.
8. Report: Anthropic's Revenue Expected to Surpass $1 Billion This Year, with a Year-on-Year Growth Rate of 1000%!
Anthropic is an artificial intelligence startup company, and it is expected that this year's revenue will reach $1 billion, with a growth rate of 1000%, showing a strong demand for AI technology. 60% to 75% of the company's revenue comes from third-party API usage, indicating a high market dependence on its technology. Competitor OpenAI plans to raise $6.5 billion, with a valuation of $150 billion, and the competition in AI is intensifying. AI technology is constantly reshaping the future of various industries.
AiBase Highlights:
🌟 Anthropic is expected to surpass $1 billion in revenue this year, with a year-on-year growth rate of 1000%.
🤖 60% to 75% of revenue comes from third-party API, indicating a high market dependence on its technology.
💰 OpenAI plans to raise $6.5 billion, with a valuation of $150 billion, and the AI competition is becoming more intense.
9. HuggingFace Launches HuggingChat Native macOS Client
HuggingFace's latest launch of the HuggingChat native macOS client brings a seamless and intuitive advanced AI conversation experience to macOS users, supporting the use of language models locally, and integrating practical functions such as Markdown, web browsing, and code syntax highlighting. Users can quickly start the application through simple installation steps and enjoy powerful AI chat capabilities at any time.
AiBase Highlights:
🚀 HuggingChatOS client provides macOS users with a seamless and intuitive advanced AI conversation experience.
💻 Users can easily install HuggingChat by visiting the Releases section of the GitHub repository, downloading the latest HuggingChat-macOS.zip file, and unzipping it to use.
🔑 Users can quickly start the application through the program folder or by using the shortcut key ⌘ + Shift + Return.
Details link: https://github.com/huggingface/chat-macOS
10. Beware! Hackers Exploit ChatGPT Vulnerability to Implant False Memories and Steal User Information
Recently, security researcher John Ribeiro discovered a vulnerability in ChatGPT that could allow hackers to implant false information and malicious commands in the user's long-term memory. Although OpenAI has released partial fixes, users still need to be vigilant against prompt injection attacks that may be caused by untrusted content.
AiBase Highlights:
🛡️ ChatGPT vulnerability allows hackers to implant false information into user memories
💻 Vulnerability exploits long-term memory function to permanently steal user input data
🔍 Users need to regularly check stored memories to prevent the implantation of false information
Details link: https://embracethered.com/blog/posts/2024/chatgpt-hacking-memories/
11. Baidu Baike 4.0 Upgrade: Second-level Deployment, 95% Training Efficiency, 99.5% Effective Training Duration
The Baidu Baike computing platform 4.0 upgrade has improved multi-chip hybrid training capabilities, with effective training duration exceeding 99.5%, significantly enhancing the efficiency of computing power use. After the upgrade, deployment is done in seconds, with 95% training efficiency and 99.5% effective training duration, greatly improving deployment efficiency and shortening the business launch cycle. The speed and cost of model inference are optimized, with efficiency in long text inference more than doubled, meeting market demands.
AiBase Highlights:
✨ Multi-chip hybrid training capabilities improved, effective training duration exceeds 99.5%
⚙️ Second-level deployment, 95% training efficiency, 99.5% effective training duration, improving deployment efficiency
💡 Optimize model inference speed and cost, efficiency in long text inference more than doubled
12. Baidu AI Code Assistant Wenxin KuaiMa Upgrade: Introduces Enterprise-level Code Architecture Explanation and More