Article reprint source: Qiu Xiangyu
OpenAI launched the "GPTs" service at its first developer conference, allowing users to create "their own version of ChatGPT" according to specific needs. GF Securities said that after the launch of GPTs & GPT Store, AI applications may emerge faster and the ecosystem is expected to prosper further. Alibaba Group CEO Wu Yongming said at the 2023 World Internet Conference Wuzhen Summit that we are at the switching node from traditional computing to AI computing, and AI computing will eventually take over all computing resources. Wu Yongming revealed that Alibaba is about to open source a large model with 72 billion parameters. This will be the largest open source large model in China.
Image source: Generated by Unbounded AI Financing
OpenAI plans to get new funding from Microsoft, adding to $10 billion in investment this year
According to a report by the Financial Times on November 13, OpenAI plans to seek new financial support from Microsoft to realize the vision described by its CEO Sam Altman, that is, to create computer software that is "as smart as humans." Sam Altman said in an interview that his company's partnership with Microsoft "works very well." He said that OpenAI expects to raise more funds from the tech giant and other investors over time to keep up with Microsoft's development and create more complex artificial intelligence models.
Zhipu AI seeks new round of financing with a valuation of 20 billion RMB
According to a report by Titanium Media on November 10, AI big model unicorn company Zhipu AI is conducting a new round of financing with a valuation of 20 billion yuan. Zhipu AI responded that it would not comment on this news. It is reported that on October 20, Zhipu AI announced that it has raised more than 2.5 billion yuan in financing so far this year, and investors include Meituan, Ant, Alibaba, Tencent, Xiaomi, Kingsoft, Shunwei, Sequoia, Hillhouse and other institutions. At that time, Caixin reported that the company's valuation had exceeded 10 billion yuan.
German AI startup Aleph Alpha receives over $500 million investment from Schwarz Group
According to a report by the Financial Times on November 6, German artificial intelligence startup Aleph Alpha received more than $500 million in investment from Schwarz Group, the owner of the Lidl supermarket chain. Other investors include the venture capital arm of Bosch, Europe's largest automotive parts supplier, according to the statement. People familiar with the deal said that about a quarter of the funds came from equity investment, and the rest consisted of research grants and joint business development commitments, which would not dilute shareholder interests.
"Yangtuo Education" received 230 million yuan in strategic financing, and the funds will be used to promote the company's comprehensive AI strategy
According to a report by 36Kr on November 7, "Alpaca Education" recently announced that it has obtained strategic financing of 31 million US dollars, about 230 million RMB, and the investor of this round of financing is Hunan Incremental Capital. This round of financing will be used to promote the company's comprehensive AI strategy and further iterate the company's vertical model, intelligent customer service, digital human teacher and other research and development. Among them, the company's vertical model is a professional language learning model developed by Alpaca Education for the language training industry; intelligent customer service is an intelligent question-and-answer system created by Alpaca Education using AI technology, which can quickly and accurately answer students' questions; digital human lecturers are virtual teachers created by Alpaca Education using virtual reality and AI technology, which can provide students with a more vivid and vivid teaching experience.
Lenovo invested in Haina AI's affiliated company, holding 8.77% of the shares
According to Securities Times e Company, Qichacha App shows that recently, Haina AI's affiliated company Beijing Qunxing Shining Technology Co., Ltd. has undergone industrial and commercial changes. The shareholder has added Lenovo (Tianjin) Partnership, a small and medium-sized enterprise development fund, with a shareholding of 8.7719%. At the same time, the registered capital has increased from about 4.9084 million yuan to about 5.3804 million yuan. According to its official website, Haina AI is committed to empowering the human resources industry with AI technology and has launched AI interview services.
AI startup Ajiaxi completes nearly 100 million yuan in Pre-A+ round of financing
According to a report by 36Kr on November 8, Chengdu Ajiaxi Intelligent Technology Co., Ltd. announced that it has completed a Pre-A+ round of financing of nearly 100 million yuan. This round of financing was led by Shenneng Chengyi and followed by Sichuan Development Academician Fund.
Ajiaxi is an artificial intelligence platform and product service provider focusing on the AIoT (intelligent Internet of Things) field. This round of financing will promote and explore the implementation of AI solutions in the fields of energy, photovoltaic lithium batteries, new materials, 3C, etc.
Stability AI completes nearly $50 million in financing, Intel participates in the investment
Stability AI raised nearly $50 million in the form of convertible notes in its deal with Intel, which closed in October, Bloomberg reported on Nov. 10, citing people familiar with the matter.
Precision components AI manufacturing "Liangjiang" completed 10 million yuan angel round financing
According to a report by 36Kr on November 13, the end-to-end AI autonomous manufacturing factory for precision components, "Liangjiang", has recently completed an angel round of financing of more than 10 million yuan. The investors in this round of financing are Jiuhe Venture Capital, and individual angel investors in related industries also participated. This round of financing will be used for core technology research and development, promoting the intelligent transformation of the industry, team building and operation, and business expansion.
AI company "Monte Carlo" completes angel round financing, Yusu Capital invests
According to a report by DoNews on November 12, Monte Carlo received angel round financing from Yousu Capital (Youzu Network). Monte Carlo is a high-tech Internet company focusing on artificial intelligence. The company is committed to providing vertical industry artificial intelligence solutions, providing users with highly simulated intelligent bodies and intelligent credit data products, and helping users solve problems such as game development and financial risk control.
Ozone, the maker of cloud-based AI collaborative video editor, raises $7.1 million in seed funding, with YC participating
According to TechCrunch, Ozone, a maker of cloud-based AI collaborative video editors, announced that it has received $7.1 million in seed round financing and will launch an open beta. Investors in this round of financing include Y Combinator, NEA, General Catalyst, LGVP, Bluewatch Ventures, Polymath Capital Partners, Pioneer Fund, SOMA Capital and Olive Tree Capital, as well as the founders of Dropbox, Tinder and Lightricks. The new funds will be used to build the platform and add more features, including additional artificial intelligence features expected to be launched in the coming months.
Large Model
MediaTek releases Dimensity 9300 AI chip, supports terminals running large models with up to 33 billion parameters
According to Babbitt News, on November 6, MediaTek released the Dimensity 9300 flagship 5G generative AI chip, which supports terminals to run AI large language models with 1 billion, 7 billion, 13 billion, and up to 33 billion parameters.
Alibaba CEO Wu Yongming: China's largest 72 billion parameter model will be open sourced soon
According to a report by 36Kr on November 9, the "Internet Entrepreneur Forum" of the 2023 World Internet Conference Wuzhen Summit was held on the morning of the 9th. Alibaba Group CEO Wu Yongming delivered a keynote speech. Wu Yongming said that we are at the switching node from traditional computing to AI computing, and AI computing will eventually take over all computing resources. The digital infrastructure that supports AI must be an infrastructure that integrates AI capabilities. The deep integration of AI and cloud computing will become an important driving force for the iteration of cloud computing. Wu Yongming revealed that Alibaba is about to open source a large model with 72 billion parameters. This will be the largest open source large model with the largest parameter scale in China.
vivo open source blue heart large model BlueLM-7B series
According to the news from Webmaster Home on November 9, vivo AI recently officially open-sourced the BlueLM-7B base model and 7B chat model, as well as the 32K long text base model and chat model. vivo said that it will also open-source the 13B model and the 7B-vl model that supports multimodality in the future.
Adobe Research develops new AI model that can transform a single 2D image into a high-quality 3D model in just 5 seconds
According to VentureBeat on November 10, a team of researchers from Adobe Research and the Australian National University recently developed a breakthrough artificial intelligence (AI) model that can convert a single 2D image into a high-quality 3D model in just 5 seconds. Their research paper "LRM: Large Reconstruction Model from Single Image to 3D" details this breakthrough, saying that the model will bring revolutionary changes to industries such as games, animation, industrial design, augmented reality (AR) and virtual reality (VR).
China Telecom will open source its 100 billion parameter model in April next year, with 12 industry models to be put into trial commercial use in the first batch
According to a report by Titanium Media on November 11, China Telecom's AI team announced a timetable for open source, which will open source the 10 billion parameter model by the end of this year and the 100 billion parameter model in April next year. All underlying codes will be open source. In addition, based on the basic large model, China Telecom further demonstrated the layout of the industry's large model.
Pan Tianshu, general manager of China Telecom's government and enterprise information service business group, said that China Telecom's first batch of trial commercialization of 12 industry big models, including the Xingchen education big model, governance big model, government service big model, emergency big model, medical insurance big model, transportation big model, housing and construction big model, financial big model, Shennong No. 1 big model, travel big model, tourist guide big model, and mining big model.
Amazon's large model Olympus has 2 trillion parameters, surpassing GPT-4
According to a report by Wall Street Journal on November 10, Amazon is training its second large language model, internally codenamed "Olympus", which is likely to be launched in December this year. The Olympus large language model has a scale of 2 trillion (2000B) parameters, exceeding the parameter scale of GPT-4 (GPT-4 has about 1 trillion parameters). Amazon plans to connect "Olympus" to the Alexa voice assistant on online retail stores, Echo and other devices, and provide new features for the AWS platform.
application
OpenAI launches user-customized version of ChatGPT and releases preview version of GPT-4 Turbo
According to a report by IT Home on November 7, OpenAI held its first developer conference today and launched a service called "GPTs" for its ChatGPT, allowing users to create "their own version of ChatGPT" according to specific needs. OpenAI said that they launched these features to give users "some control" over ChatGPT, such as companies can create a special assistant that only employees can access, and parents can also create a problem-solving tool specifically to teach children how to solve tricky math problems.
OpenAI also launched a preview version of GPT-4 Turbo, which supports 128k contexts, has an input price that is three times cheaper than GPT-4, doubles the rate limit, and has a knowledge base updated to April 2023. In addition, JSON Mode has been added, and multiple function call capabilities have been updated. Compared with the previous GPT-4, it has stronger performance, cheaper price, and lower development cost. In addition, OpenAI has also launched a new feature called JSON Mode, which ensures that it is easier for developers to make models call JSON and xml content, making it easier for models to return consistent output results, thereby achieving repeatable output, which has a certain effect on controlling model behavior and writing model unit test content.
OpenAI launches Assistants API, allowing developers to customize AI assistants for their own applications
According to IT Home on November 7, at today's OpenAI first developer conference, OpenAI launched the "specially built AI tool" Assistants API, which can use "additional knowledge" to help developers build AI assistants in their own applications.
OpenAI said that the Assistants API provides three functions: "Code interpreter", "Retrieval" and "Function calling" to help users handle "heavy lifting". The Assistants API is now open for beta, and users can experience it on the Assistants Playground homepage without writing any code.
In addition, OpenAI has launched the "Copyright Shield" feature to alleviate people's concerns about the potential infringement risks of generative AI. OpenAI said that if a user is required to make a legal claim for "copyright infringement" due to AI-generated content, OpenAI will intervene and defend the customer, and will also fully reimburse the expenses incurred in the process. The relevant feature is applicable to ChatGPT Enterprise and all API users, but the premise of exemption is that the relevant user "has no intention of infringement" and "cites the source" under responsible AI use.
Meituan's first AI chat product "Wow" is launched, which can communicate with multiple AI characters with different personalities
Babbitt News: On November 7, a "Wow" app developed by Shanghai Sankuai Shengxingou Technology Co., Ltd. has been launched on iOS and Android platforms. It is reported that the company is 100% controlled by Shanghai Hantao Information Consulting Co., Ltd., a subsidiary of Meituan, which means that "Wow" is Meituan's first AI chat product.
"Wow" is defined as an AI friend community for young people. After logging in, users can talk to 29 AI characters with different personalities through voice or text communication, such as Socrates, Daji, Fusu, and psychological counselors. Asking the same question to different AI characters will also give corresponding answers based on different personalities.
Alipay launches AI photo function, which can generate personal photo albums by taking selfies
Babbitt News: Alipay has recently launched a free AI photo album function. You can enter the mini program by searching for "AI Photo Gallery" on the homepage. The main information of the mini program is Zhixin (Shanghai) Network Technology Co., Ltd., a subsidiary of Alipay. According to the official introduction of the mini program, one selfie can generate a personal photo album with one click. It is currently in a limited-time free stage, and the styles that can be generated include space explorers, magazine photos, ancient style, etc.
During Tmall Double 11, merchants called AI more than 1.5 billion times
According to a report by Cailianshe on November 8, during the preparation and sales period for Double 11 this year, merchants on the Taobao Tmall platform have begun to use AI tools on a large scale, and the number of AI calls has exceeded 1.5 billion times.
Tencent Miying Digital Medical Imaging Platform has fully opened more than 20 medical AI engines
According to a report by 36Kr on November 8, the 2023 World Internet Conference Wuzhen Summit was held in Wuzhen. At the World Internet Leading Scientific and Technological Achievements Release Event, Tencent Health President Wu Wenda introduced that the "Tencent Miying Digital Medical Imaging Platform" has fully opened more than 20 medical AI engines. At present, the "Tencent Miying Digital Medical Imaging Platform" provides medical AI research services for 17 universities and research institutes, 23 public hospitals and 33 technology companies.
Samsung launches Galaxy AI, Galaxy S24 series phones will be equipped with edge AI functions
According to IT Home on November 9, Samsung announced the launch of Galaxy AI, bringing the edge AI big model to the new Galaxy flagship phone.
According to reports, the AI Live Translate Call feature will soon provide personal translation services to users with the latest Galaxy AI phones. Since it is integrated into the native call function, there is no need to use a third-party application to translate. When the user makes a call, the audio and text translation will be displayed on the phone in real time. In addition, since it is a large model running on the phone side, the call translation is completely local and users do not need to worry about privacy issues.
Galaxy AI is said to be launched early next year, coinciding with the regular release of the next generation of Galaxy S series flagship phones (expected to be the Galaxy S24 series).
CodeGPT, a code generation tool, has become a new favorite among developers and can efficiently generate standard code
According to a report by Webmaster Home on November 9, CodeGPT is an AI code generation tool that uses the GPT-3 language model and is popular among developers. The tool is a plug-in for Visual Studio Code that uses the GPT-3 language model to generate code, translate languages, write various types of content, and answer questions. CodeGPT is said to be able to generate efficient and compliant code by training the code of a large number of real projects.
NetEase Vice President Liu Jie: NetEase expects to provide 100,000 new digital economy jobs such as AI painters by the end of the year
According to a report by The Paper on November 9, Liu Jie, Party Secretary and Vice President of NetEase Group, mentioned at the Digital Yangtze River Delta Development Conference on the Co-construction and Sharing of Digital Civilization in the Yangtze River Delta Integration that NetEase Group expects to provide 100,000 new digital economy jobs by the end of the year. Liu Jie mentioned that based on its years of accumulation in the field of game artificial intelligence, NetEase has developed a human-computer collaboration platform for intelligent body programming. The artificial intelligence tools provided by the platform allow users to perform human-computer collaboration tasks such as art creation and mechanical operation through gamified operations. The platform has created new professions such as excavator remote drivers, AI painters, and AI expression binders, and is expected to provide 100,000 new digital economy jobs by the end of the year.
TikTok tests its first AIGC tool Dreamina for text and image content creation
According to a report by Tech Planet on November 10, Douyin's Jianying is testing an AIGC tool called "Dreamina", which belongs to the field of literary image creation. Users can generate four AI-generated creative images based on a piece of text. These images are generated from multiple dimensions such as abstraction and realism. In addition, users can also modify the text and images, such as the size ratio of the generated image, and the type of image template. The template types are divided into two categories: general and animation. The generated images can be saved or regenerated. According to sources, the tool will be used for Douyin's content creation of pictures, texts or short videos in the future to enrich Douyin's content library in AI creation.
Runway introduces Motion Brush, which lets you set specific areas of an image in motion
On November 11, Runway, a generative AI video startup, announced the launch of the Motion Brush feature, which allows users to select specific areas of an image and set them to motion. The feature will soon be available on the Gen-2 model.
Daily Interactive releases large-scale model application AITA smart crowd selection tool
According to Titanium Media on November 10, at the 2023 World Internet Conference, data intelligence service provider Daily Interactive (Getui) released the big model application AITA smart crowd selection tool, which combines big model capabilities with the company's digital intelligence capabilities to achieve AI intelligent targeting of TA. Through simple natural language dialogue, the tool can generate target audiences that meet the needs of brand advertisers, helping brand advertisers connect with consumers.
policy
Shanghai releases 11 measures to promote the innovative development of AI big models: to create benchmark big model products and services
According to Babbitt News, Shanghai Municipal Economic and Information Commission and other departments recently issued a notice on the issuance of "Shanghai Measures to Promote the Innovation and Development of Artificial Intelligence Big Models (2023-2025)", introducing 11 measures to promote the innovation and development of large-scale pre-training models and accelerate the creation of a world-class artificial intelligence industry cluster. The measures are as follows: 1. Implement a big model innovation support plan; 2. Establish a big model testing and evaluation center; 3. Implement a big model intelligent computing power acceleration plan; 4. Build a smart chip software and hardware collaborative ecosystem; 5. Co-build and share corpus data resources; 6. Implement a big model demonstration application promotion plan; 7. Promote the application of scientific and intelligent big models; 8. Create a big model innovation highland for enterprises and talents; 9. Promote the construction of a big model application ecosystem; 10. Establish a normalized service mechanism; 11. Increase investment.
Guangdong: Build a complete autonomous and controllable AI software and hardware ecosystem, and develop a general AI model with hundreds of billions of parameters
According to AI New Intelligence, on November 13, the People's Government of Guangdong Province issued the Implementation Opinions on Accelerating the Construction of a Leading Innovation Area for the General Artificial Intelligence Industry. It is mentioned that by 2025, the scale of intelligent computing power will be the first in the country and the world, the general artificial intelligence technology innovation system will be relatively complete, the high-level application scenarios of artificial intelligence will be further expanded, the scale of the core industry will exceed 300 billion yuan, and the number of enterprises will exceed 2,000. Develop high-end training and reasoning chips with universality and programmability, computing chips with multi-modal and multi-precision computing, and explore the research of new reconfigurable and integrated computing and storage architectures. Develop basic software such as efficient and easy-to-use open source artificial intelligence chip compilers and tool chains, support the adaptation of autonomous artificial intelligence chips and domestic general servers, and build a complete autonomous and controllable artificial intelligence software and hardware ecosystem.
Large companies
Reuters: Baidu has placed an AI chip order with Huawei worth about 450 million yuan, which is expected to be delivered before the end of the year
According to a Reuters report on November 7, two people familiar with the matter said that Baidu ordered artificial intelligence chips from Huawei this year to cope with the export restrictions on AI chips imposed by the United States on China. The sources said that Baidu ordered 1,600 Huawei Technologies 910B Ascend AI chips for 200 servers, with a total order value of about 450 million yuan (61.83 million U.S. dollars), and Huawei will deliver all the chips by the end of this year.
Source: Apple is using a large model to transform Siri and plans to make it a standard feature of iPhone 16 and subsequent models
Babbitt News, on November 9, X (Twitter) user "Revegnus" broke the news that Apple is using a large language model (LLM) to completely transform Siri into the "ultimate virtual assistant" and is preparing to develop it into the "most powerful killer AI application". He said that Apple is actively carrying out this comprehensive development work, and the first product is expected to debut at WWDC 2024 and is planned to become the standard configuration of iPhone 16 and subsequent models.
Nvidia confirms: It will "customize" three chips for China and officially launch them on November 16
According to a report by Yinshi Finance on November 10, in response to the previous news that Nvidia will launch three new and improved AI chips for the Chinese region, Nvidia insiders said that the news is true and will be officially launched on the 16th of this month. In addition, Nvidia's US headquarters responded: "Thank you for contacting us, we have no comment on this news yet."
Viewpoint
Robin Li: Big models can only be considered successful if they have millions of AI applications
According to The Paper, on November 9, Baidu founder, chairman and CEO Robin Li stated at the 2023 World Internet Conference Wuzhen Summit that the ultimate goal of technological innovation is application, and the sign of mankind entering the AI era is the emergence of a large number of AI native applications, rather than a large number of large models. Large models can support the development of countless AI native applications, but only with millions of AI applications can large models be considered successful.
Hou Xiaonan, CEO of China Literature: AIGC promotes the large-scale export of online literature overseas, and the average translation cost is reduced by 90%
According to a report by Guangzhou Daily on November 9, Hou Xiaonan, CEO and President of China Literature, said at the Cross-Strait and Hong Kong and Macao Internet Development Forum of the 2023 World Internet Conference that AI translation is breaking through the limitations of production capacity and cost, and has performed well in both quality and efficiency. By establishing a special vocabulary library for human-machine collaboration, China Literature's online article translation costs have been reduced by an average of 90%, while efficiency has been greatly improved. In Hou Xiaonan's view, AIGC makes it possible for Chinese online articles to go overseas on a large scale and seamlessly connect to the global market, which is a very promising new blue ocean.
C3.ai CEO: Artificial intelligence will give rise to new tech giants
According to a report by Jinshi on November 9, Tom Siebel, CEO of application software company C3.ai, said today that the emergence of artificial intelligence does not mean that today's artificial intelligence leaders will dominate the next generation of the technology industry. "It is widely believed that OpenAI, Microsoft, Facebook or Google will win this battle. I think there is no reason to believe this at all."
Research
IDC: China's total IT investment in artificial intelligence market is expected to exceed US$38 billion in 2027
According to a report by Cailianshe on November 9, AMD and IDC jointly released the white paper "AI laptops trigger scene changes - terminal AI" today. The white paper shows that the total IT investment scale of China's artificial intelligence market is expected to exceed US$38 billion in 2027. IDC predicts that the implementation of AI applications will drive the notebook computer market to grow steadily in the next few years. In 2025, the shipments in the commercial and consumer markets will exceed 10 million and 20 million respectively, and the total market will exceed 34 million in 2027, of which AI laptops will account for 86%.
Study: Generative AI in healthcare market to reach $30.4 billion by 2032
According to a report by Webmaster Home on November 7, Allied Market Research recently released a report titled "Global Generative AI in Healthcare Market Opportunity Analysis and Industry Forecast 2023-2032". The report shows that generative AI is a revolutionary technology in the medical field, using artificial intelligence to create, generate or simulate data, images or text in the medical field. This innovative AI application has the potential to revolutionize the medical industry in many aspects, from diagnosis and treatment planning to medical research and education.
According to the report, the global generative AI in healthcare market was valued at $1.6 billion in 2022 and is expected to reach $30.4 billion by 2032, at a CAGR of 34.9%. This strong growth is driven by the widespread adoption of generative AI in medical image analysis for medical diagnosis, the efficiency of generative AI in workflows and administrative tasks, and the application of generative AI in personalized medicine.
GF Securities: After the launch of GPT Store, AI applications may emerge faster
According to a report by Jinshi on November 10, GF Securities said that after the launch of GPTs & GPT Store, AI applications may emerge faster, the ecosystem is expected to prosper further, and several trends may be seen on the application side: on the one hand, diversified applications based on GPT may emerge rapidly, and the implementation of scenario applications is expected to accelerate; on the other hand, the development threshold has been greatly reduced, which will also lead to intensified competition among homogeneous products, especially some tool-type application products with simple functions and no moat. Companies that have accumulated certain capabilities and experience in various scenario applications are closer to the user side and are expected to launch mature and high-quality AI application products faster. Pay attention to the application and implementation rhythm of domestic and foreign AI technologies in office, games, education, film and television, marketing and other sectors.
Survey: 44% of companies say AI may replace some employees in the next year
According to a recent survey conducted by ResumeBuilder.com, more and more companies believe that artificial intelligence (AI) may begin to replace some employees next year. The report surveyed 750 business leaders representing multiple industries.
According to the survey results, 44% of companies said they expect AI to lead to the replacement of employees by 2024. In addition, the survey found that 37% of companies currently using AI technology said they have already cut headcount this year because they no longer need employees.