Collection of big model daily reports on October 26
[Collection of big model daily reports on October 26] Douyin has launched a new public beta version of “Jichuang” AI one-stop platform to comprehensively improve the creative efficiency of creators! ; AI models also pay attention to speed, efficiency and economy. Medium and large enterprise customers and developers feel that OpenAI is no longer good, and they are flocking to Microsoft. Let’s wait and see the developer conference next month…; Zuckerberg: Next year AI will become Meta’s largest investment area, and manpower will be concentrated on AI; using MEG to reconstruct the human brain imaging process, Meta releases blockbuster research! ; Microsoft, OpenAI, etc. invested US$10 million and appointed directors of the “Frontier Model Forum”; publishers blocked AI web crawlers and Direcqt developed AI chatbots
Douyin has launched a new public beta version of the “JiChuang” AI one-stop platform to comprehensively improve the creative efficiency of creators!
Link: https://news.miracleplus.com/share_link/11191
Recently, Douyin launched a public beta version of “Jichuang”, a one-stop intelligent creative production and management platform. Creators can use this platform to improve the creation efficiency of videos, graphics, and live broadcasts with the help of various new functions provided by AI. The details of the 3 major functions and 6 small projects are as follows: 1) Video creation: intelligent filmmaking, AI video script (AI video creation artifact, script survival, digital human, intelligent filmmaking are all available) 2) Graphic and text creation: product card tool , Graphic and text tools (quickly generate product card pictures, high-quality graphic and text content) 3) Live broadcast creation: AI live broadcast background, AI live broadcast script (supports one-click generation of live broadcast room background, live broadcast script) and the above functions, creators can rely on AI in one stop Ways to complete all kinds of videos and graphics on Douyin. “Jichuang” can be said to be a very useful AI tool.
AI models also focus on speed, efficiency and economy. Medium and large enterprise customers and developers feel that OpenAI is no longer good and are flocking to Microsoft. Just look at the developer conference next month…
Link: https://news.miracleplus.com/share_link/11192
According to the latest report from The Information, OpenAI’s enterprise sales are under pressure as customers in the AI field are looking for cheaper solutions. Large companies such as Salesforce and Morgan Stanley, small and medium-sized companies including Wix, and independent development teams are all looking for A more affordable alternative. According to Microsoft’s recently disclosed financial report for the first quarter of 2024, cloud services such as Azure grew by 29%, and the number of organizations using OpenAI services on the Azure platform reached 18,000. When customers purchase OpenAI models through Azure, Microsoft will retain these Most of the revenue, this number is up from 11,000 in August last year. In addition to providing OpenAI models, Microsoft Azure also provides other open source cutting-edge models, including Microsoft’s own, Meta and Huggingface models. Microsoft has even persuaded some customers who previously purchased OpenAI software directly from startups to also purchase it from Azure. These customers can Access the same OpenAI software from either source for the same price, but they may be incentivized to choose Microsoft if they already have existing computing contracts with Microsoft and can bundle those contracts with the OpenAI software.
Zuckerberg: AI will become Meta’s largest investment area next year, and manpower will be concentrated on AI
Link: https://news.miracleplus.com/share_link/11193
Zuckerberg is trying to lead the company to the forefront of the AI (artificial intelligence) craze. On October 25, local time, during the fiscal year 2023 third quarter earnings call, Meta CEO Zuckerberg emphasized that he believed that generative AI-related technologies will make the way people use various applications more meaningful. In the future, Meta may even use AI to directly generate content for users based on their interests. Zuckerberg said that AI will help creators who use Meta’s major applications improve content quality and production efficiency, and as time goes by, the proportion of content generated by AI participation in user consumption will become larger and larger. .
Using MEG to reconstruct the human brain imaging process, Meta releases blockbuster research!
Link: https://news.miracleplus.com/share_link/11194
Global social and technology giant Meta (the parent company of Facebook, Instagram, etc.) announced a major study on its official website. It developed an AI model through MEG (magnetoencephalography) to decode the imaging process of visual activity in the human brain, and The paper was published. It is reported that this is a non-invasive neuroimaging technology that can detect thousands of brain activities per second, and can reconstruct in real time the images the brain perceives and processes at every moment. It provides an important research basis for the scientific community to understand how the brain expresses and forms images. From the perspective of application scenarios, this technology can better understand and control the neural network actions and neurons of AI models such as ChatGPT and Stable Difusion, improve content output accuracy and reduce risks, laying the foundation for the evolution to AGI (General Artificial Intelligence). Looking at the larger goal, it will accelerate the clinical research and development process of “brain-computer interface” and help those who have suffered brain injuries and lost the ability to speak.
Microsoft, OpenAI and others invested US$10 million and appointed directors of the “Frontier Model Forum”
Link: https://news.miracleplus.com/share_link/11195
On October 26, Microsoft, OpenAI, Google and Anthropic issued a joint statement on their official website, appointing Chris Meserole as the executive director of the “Frontier Model Forum” and investing a US$10 million security fund to enhance generative AI products such as ChatGPT and Bard. Safety. This statement is an important manifestation of the four major technology companies’ commitment to the development of safe, reliable, and stable generative AI. It is also part of the official “AI Security Commitment” signed with the government, which is crucial to the healthy development of global generative AI.
Honor Magic 6 officially announces support for self-developed 7 billion device-side AI large model
Link: https://news.miracleplus.com/share_link/11196
Today, Honor Magic 6 was also officially announced at the Qualcomm press conference, announcing that it will be equipped with Snapdragon 8 Gen3. In addition, Honor Magic6 has always attached great importance to the field of AI and has developed many AI functions. This time, thanks to the soaring AI performance of Snapdragon 8 Gen3, it can achieve more powerful capabilities. Zhao Ming said that Honor Magic6 can handle large language models with up to 7 billion parameters. The most important thing is that such AI processing is completely realized through the mobile phone side. All tasks and data do not need to be uploaded to the cloud, and the entire process is run locally, which can protect user privacy information to the maximum extent.
Publishers block AI web crawlers, Direcqt develops AI chatbot
Link: https://news.miracleplus.com/share_link/11197
As more media publishers start blocking AI-powered web crawlers from accessing their content, a startup called Direqt is partnering with them to provide customized chatbot solutions. Understanding that messaging apps were taking up more and more of consumers’ time, Direqt saw an opportunity to build a dedicated chatbot platform for publishers. These chatbots enable publishers to reach readers directly and increase content engagement.