**Summary:**Baidu launched Wenxin Yiyan Professional Edition today, priced at 59.9 yuan/month, and the preferential price for continuous monthly subscription is 49.9 yuan/month. The basic version of Wenxin Yiyan, which has been previously available to users, is still free to use. It is reported that compared with the basic version based on Wenxin Large Model 3.5, the professional version based on 4.0 has stronger model capabilities and image generation capabilities, and supports various plug-ins, which is suitable for users who need to use Wenxin Yiyan for code programming, copywriting, painting design and other professional work needs. vivo today released its self-developed AI “Blue Heart” model and released a large model matrix, including five large models with different parameter scales of 175 billion, 130 billion, 70 billion, 7 billion, and 1 billion, and open-sourced 7B large models.
![]()
Image source: Generated by Unbounded AI
Hot Info:
Zhongshan Hospital Affiliated to Fudan University and Kuang-Chi Huiyu jointly released the Guangyu medical model
On October 31, at the Yunqi Conference held in Hangzhou, Zhongshan Hospital affiliated to Fudan University and artificial intelligence company Kuang-Chi Huiyu released the jointly developed multimodal medical model “Guangyu Medical Large Model”.
The large model is a self-developed large model with the deep participation of Zhongshan Hospital, customized for medical scenarios, and based on the scale of tens of billions of parameters and trillions of tokens pre-trained corpus of Kuang-Chi Huiyu. Based on Zhongshan Hospital’s medical experience, medical knowledge and medical data resources, Guangyu Medical Model includes medical experience in offline diagnosis and treatment scenarios, medical literature guidelines, evidence-based knowledge base, medical information terminology, knowledge graph, etc.
According to reports, the performance of the Guangyu medical model in the general ability assessment, Chinese knowledge, mathematics, reading comprehension, logical reasoning and other aspects are significantly ahead of the open source large model LLaMA2-70B developed by Meta. In the USMLE medical industry evaluation, the Guangyu medical large model is comparable to GPT-4 in capabilities.
Samsung plans to launch smartphones powered by generative AI technology
On November 1, it was reported that Samsung Electronics announced on the third-quarter earnings call that it plans to launch a smartphone integrating generative artificial intelligence (AI) technology next year, which will build AI technology into the device without connecting to an external cloud. “We are working on AI technology as a new experience in our devices,” Samsung said, adding that they “plan to offer this technology starting in 2024.”
Co-founder of DeepMind: The AI industry should not blindly pursue a “move fast, break the mold” development model
DeepMind co-founder Demis Hassabis recently expressed concerns about the development of the AI industry in a recent interview, warning that, unlike traditional tech giants, the AI industry should not blindly pursue a development model of “moving fast and breaking the mold”. It is reported that “move fast and break the mold” is the slogan proposed by Mark Zuckerberg.
Google launches generative AI bug bounty program with up to $30,000
In a blog post published over the weekend, Google announced that it would expand the scope of its bug bounty program to include bugs and vulnerabilities found in generative AI systems, as reported on October 31. The program will cover Google’s own generative AI products, such as Google Bard, as well as Google Cloud’s Contact Center AI, Agent Assist, and more. Eligible vulnerabilities will be eligible for bounties of up to $30,000.
Li Weike releases the first all-in-one AI AR glasses “Li Weike Meta Lens S3”
Recently, the 2023 China VR/AR Industry Expo was held in Shanghai. At the meeting, Li Weike officially released its first all-in-one artificial intelligence AR glasses “Li Weike Meta Lens S3”. According to reports, the product integrates Li Weike’s self-developed AI large model system, which integrates music, telephone, navigation, photography and chat, sports data recording and other functions for outdoor cycling and sports scenes.
Baidu launched Wenxin Yiyan Professional Edition, priced at 59.9 yuan/month
On November 1, it was reported that Baidu launched the professional version of Wenxin Yiyan today, priced at 59.9 yuan/month, and the continuous monthly preferential price was 49.9 yuan/month. The basic version of Wenxin Yiyan, which has been previously available to users, is still free to use. It is reported that compared with the basic version based on Wenxin Large Model 3.5, the professional version based on 4.0 has stronger model capabilities and image generation capabilities, and supports various plug-ins, which is suitable for users who need to use Wenxin Yiyan for code programming, copywriting, painting design and other professional work needs.
vivo released the “Blue Heart” AI model, with a maximum parameter size of 175 billion
As reported on November 1, vivo today released its self-developed AI “Blue Heart” large model and released a large model matrix, including 175 billion, 130 billion, 70 billion, 7 billion, and 1 billion large models with different parameter sizes, and open-sourced 7B large models.
Canva launches AI tools for education, free for K-12 students and teachers
Australian AI startup Canva announced its foray into education less than a month after launching its AI-powered Magic Studio, as reported on November 1. Last week, Canva launched its AI tool, Classroom Magic, the version of Magic Studio designed for teachers and students, and made available for free to all students, teachers, and districts in K-12.
It is reported that Classroom Magic has introduced a series of AI features, including “Magic Write”, which allows students and teachers to quickly select actions from drop-down menus, including summarizing text, expanding short text into a longer format, rewriting text, and changing the tone, among others. Magic Animate allows students and teachers to turn static text into dynamic text and automatically add transitions to presentations. In addition, Canva has launched the “Canva Shield” project, which aims to ensure the safe use of AI tools in education. The project includes features such as advanced educator controls, automated moderation, masking terminology, and reporting options to ensure content safety.
Microsoft releases a major update to Windows 11 that includes Copilot, an artificial intelligence assistant
On Tuesday (October 31), local time, Microsoft announced on its official website that it had begun rolling out the major version update “23 H2” of the Windows 11 operating system to IT professionals, including the long-awaited artificial intelligence assistant “Copilot”.
It is understood that Microsoft Copilot is a generative AI tool that has similarities with ChatGPT, which has previously exploded around the world. Microsoft said last month that Copilot, as an “everyday AI companion,” will be able to work seamlessly across “all of the apps and experiences you use most.” With large language models (LLMs), users can simply use “natural language” to give instructions, and Copilot can help write the text of an email, answer questions, automate actions on Windows, and more. Compared to Microsoft Cortana, Copilot is significantly more powerful.
Amazon deploys more than 750,000 bots to develop AI
Amazon announced on October 31 that it has deployed more than 750,000 robots across its global warehouse and distribution network, most of which are robotic arms or wheeled machines, and that the company has begun testing human-like robotic solutions at its robotics R&D sites to perform repetitive tasks and work with employees. Digit, the company’s bipedal robot, which the company recently began testing, is able to pick up and lift items for faster warehousing, helping to deliver orders 25% faster.
Good Article:
“Google AlphaFold Model Ushers in a Major Breakthrough!” Predictability of biomolecules and ligands》
On November 1, DeepMind, an AI research institute under Google, released on its official website the latest technological progress of the protein structure prediction model AlphaFold: it has significantly improved the prediction accuracy and expanded the coverage from proteins to other biomolecules, including ligands (small molecules).
“Large models continue to lose money, and technology companies are eyeing AI intelligent learning machines”
**It is no coincidence that Baidu and Alibaba have successively launched AI intelligent learning machines. **With the maturity of technology and the vigorous promotion of policies, education informatization has increasingly become a new trend in the education industry. The large-scale model technology that technology companies are actively deploying just meets the needs of the era of education informatization.
However, it is worth noting that, on the whole, due to the high price and the value of teaching has not been widely recognized by parents, the current penetration rate of pan-AI terminals is not high. This may also mean that it is difficult for AI intelligent learning machines to become an ideal terminal for large models of technology companies to open up the business closed loop.
“OpenAI Infiltrates Hacker Group Chat! The theft of ChatGPT was replaced with “Meow Meow GPT”, netizens: Absolute legend”
When ChatGPT is “hacked” by hackers, how will OpenAI respond?
Cut off the API and don’t let them use it? No, no, no.
The approach taken by these geeks can be described as a sword bias - a backhand hit of “Infernal Affairs”.
