Nvidia releases blockbuster new products; Kingdee releases the large GPT model of the sky; Daxing Airport introduces automatic driving; Cyberspace Administration of China: Face recognition requires personal consent丨Daily events...


02463d94c13e69faf0c4c4f3caca7874.png

02ba39505801b279611d91561ff0ea26.png




‍Data intelligence industry innovation service media

——Focus on digital intelligence and change business


Entreprise's news

  • Google releases AI code editor IDX: Simplify development environment configuration with cloud virtual machine

On August 9th, Google announced the "Project IDX" project, which is officially announced as a code editor that incorporates AI technology. Officials say that one of the pain points that developers are currently facing is that it is difficult to configure the technical environment and stack to run a specific project. The Project IDX project runs on a cloud-based Linux virtual machine, and provides framework templates such as Flutter, Angular, React, and Next.js, and allows developers to configure freely to meet individual development needs. Project IDX integrates a real-time preview while the user's application is running. Currently, this is limited to the preview web app, but will expand to Android emulators and even iOS emulators soon.

  • Shengzhe Technology won the bid of Qinghai Artificial Intelligence Computing Center Project with RMB 117 million

According to news on August 9, Beijing Shengzhe Technology Co., Ltd. (SENSORO) announced that the company won the bid for the Qinghai Kunlun Intelligent Computing Power Center Project, with a total project value of about 117 million yuan. It is understood that the successful completion of the project also means that the construction of the first artificial intelligence computing center on the Qinghai-Tibet Plateau is about to start.

  • Microsoft Azure AI text-to-speech service upgrade: new male voice and more language support

According to news on August 9, Microsoft Azure provides enterprises with many cloud-based services and functions, one of which is Azure AI text-to-speech, which uses machine learning and artificial intelligence to convert text input into lifelike speech. Microsoft announced an upgrade to Azure AI text-to-speech. Not only added more language support, but also added a male voice - Ryan Multilingual. Additionally, both new voices feature automatic language prediction, which automatically recognizes the input language based on the input text and adjusts speech output accordingly, eliminating the need for manual tagging.

  • The Harbin Institute of Technology Shenzhen team entered the multi-modal large model, and the self-developed "Ruoyu-Nine Heavens" topped the OpenCompass list

Recently, the Harbin Institute of Technology (Shenzhen) Computing and Intelligence Research Institute team, relying on the school's Harbin Assets Management Co., Ltd. to transform results, established a multi-modal large-scale model research and development enterprise - Shenzhen Ruoyu Technology Co., Ltd. Ruoyu Technology's first multi-modal large-scale model "Ruoyu-Nine Heavens" entered the OpenCompass multi-modal large-scale model list for the first time.

bf7ea159e9d3cd18cfc26f32e2ead871.png

  • China Telecom Chairman Ke Ruiwen: In-depth research on the general large-scale model will see the results in the second half of the year

According to news on August 9, at China Telecom's 2023 first-half performance briefing, Ke Ruiwen, chairman of China Telecom, said: "China Telecom is doing in-depth research on the general-purpose large-scale model, and we will see the results in the next few months of this year. Some new achievements have been released in the field of large models."

Ke Ruiwen said that in terms of models, China Telecom will become the main force of industry models in the future. As for the general large model, the company released a deep semantic large model not long ago, and will further study the general large model in the future. "We study the general large-scale model to better understand the general large-scale model, and then make a good industry large-scale model to give full play to China Telecom's computing power advantage."

  • Foreign media: Disney forms task force to explore artificial intelligence

On August 8th, Walt Disney set up a task force to study artificial intelligence and how it can be applied to the entire entertainment group. The group, formed earlier this year ahead of the Hollywood screenwriters' strike, hopes to develop AI applications in-house and forge partnerships with start-ups, three sources said. A Disney spokesman declined to comment.

  • Kingdee Releases Sky GPT Large Model

At the "Kingdee 2023 Global Visionaries Conference" held in Shenzhen on August 8, the Kingdee Cloud Cosmic GPT model was officially released, and the industry's first financial model was launched first. Kingdee said that it will continue to launch AI assistants with richer content covering supply chain, manufacturing and other fields.

  • Nvidia's upgraded large-scale model chip will be put into production next year

On August 8, Nvidia CEO Jensen Huang, wearing a classic leather jacket, released a new generation of GH200 Grace Hopper super chip at the top computer graphics conference SIGGRAPH 2023, and announced that it will be put into production in the second quarter of 2024. The GH200 is specially designed for artificial intelligence large-scale models and used to expand data centers around the world. Huang Renxun said that the era of generative AI (AIGC) is coming, and NVIDIA will continue to vigorously promote the determination of generative AI development.

  • Hongqi's third-generation L4 Robotaxi passed the autonomous driving road test license exam with full marks

On August 8, Yiren Yiche learned from Hongqi officials that Hongqi's third-generation L4-level Robotaxi passed the road test license exam for autonomous driving in the Beijing Intelligent Networked Vehicle Policy Pioneer Zone with full marks. The success of this test also marks the official launch of Hongqi's third-generation L4 robotaxi. Hongqi's L4 robotaxi adopts the industry's first "vision + solid-state radar" autopilot solution, and maintains an update frequency of one iteration per year. On the basis of the previous generation, the third-generation L4 robotaxi has been upgraded from the E-HS3 model to the E-HS9 model, adding a redundant backup system, a remote control system, and vehicle-road cloud collaboration functions to realize a comprehensive software and hardware system for autonomous driving. upgrade.

7da52ef8b9131b6618551066c7b050be.png

  • Lingnan shares: the subsidiary has signed a long-term cooperation with Shenzhen Cultural Exchange in the direction of digital economy and big data development

On August 8, Lingnan shares stated on the investor interaction platform that the company's subsidiary Hengrun Group actively responded to the national cultural digitalization strategy and promoted the development of digital transformation. At present, a long-term cooperation has been signed with the Shenzhen Cultural Exchange. The two parties will be close to the market demand and cooperate vigorously in the direction of digital economy and big data development.

  • Zhejiang Shibao: The company's smart steering can be applied to self-driving cars

On August 8, Zhejiang Shibao said in response to investor questions on the interactive platform that the company mainly provides steering system assemblies and key component products for automakers, and that the company's smart steering can be applied to self-driving cars.

  • Haitian AAC: It is speeding up the construction of the data service capability system related to the layout of large models

On the evening of August 8, Haitian Ruisheng said in an agency survey that the company, as a leading company in the field of training data, is accelerating the construction of a data service capability system related to large models, including further integrating the value of existing data sets and expanding data sources , and the use of large model technology to enhance the company's data service capabilities, etc. In the construction of large-scale high-quality data sets, the company has made staged progress and released high-quality projects such as "Large-model Chinese 10-million-round dialogue data set" and "Large-model multilingual speech data set" at the Global Digital Economy Conference Optimized AI datasets.

  • Kuaishou established a new technology company in Beijing with a registered capital of US$1 million

According to the news on August 9, Beijing Evergreen Exploration Technology Co., Ltd. was established with a registered capital of 1 million US dollars. The legal representative, executive director, manager and financial director are the co-founder of Kuaishou Yang Yuanxi, and the supervisor is the co-founder of Kuaishou. Xin, the business scope includes technical services, technology development, software development, etc. Shareholder information shows that the company is wholly owned by Innovation Exploration Hong Kong Co., Ltd.

  • Huawei signed contracts with nearly 40 platform companies such as Xiaohongshu and Ctrip to carry out multi-faceted cooperation

According to news on August 9, Huawei recently held a signing ceremony for Hongmeng ecological cooperation with nearly 40 application and service providers. Sexual services bring new experiences to users, enabling partners based on AI large models brings new traffic, shares new opportunities brought by full-scenario service distribution, and jointly builds a Hongmeng ecosystem of openness, mutual trust, and benefit sharing.

Partners participating in this signing include: Xiaohongshu, Ctrip, 58.com, JD Health and other application and service providers.

  • Tsinghua University's Fianbi Intelligence launched ToolLLM to improve the ability to use large model tools

According to the news on August 9, Tsinghua University’s Mianbi Intelligence cooperated with several research institutions to launch the ToolLLM tool learning framework, which has been added to the OpenBMB large model tool system. The ToolLLM framework fine-tunes the dataset ToolBench by creating high-quality tool learning instructions, trains a ToolLLaMA model with tool usage capabilities, and automatically evaluates tool learning performance through ToolEval.

The introduction of this framework aims to promote the better application of various tools in open source language models and improve the reasoning ability in complex scenarios. The ToolLLaMA model performs well in performance evaluation, its generalization ability surpasses traditional methods, and it also achieves remarkable results in tool retrieval. Overall, the ToolLLM framework provides important impetus and support for tool learning capabilities of large models.

d213e5bae995f645fc078410d02b7dc6.png

financial report

  • Rivian hit $1.12 billion in second-quarter revenue

On August 8, U.S. electric vehicle manufacturer Rivian announced that its revenue in the second quarter was US$1.12 billion, a year-on-year increase of 207.69%; its net loss was US$1.195 billion, compared with US$1.712 billion in the same period last year.

  • Sony's first-quarter net profit reached 217.5 billion yen, a year-on-year decrease of 17%

According to news on August 9, Sony released its financial report for the first quarter ended June 30, achieving revenue of 2.96 trillion yen, a year-on-year increase of 33%; operating profit of 253 billion yen, a year-on-year decrease of 31%; 217.5 billion yen, a year-on-year decrease of 17%. The company still predicts that the annual operating profit will be 1.17 trillion yen, and the market estimate is 1.24 trillion yen; the annual net profit is expected to be 860 billion yen, which was previously expected to be 840 billion yen; the annual net sales are expected to be 12.20 trillion yen , the market estimates 11.81 trillion yen.

Policy/Report

  • Cyberspace Administration of China: The use of face recognition technology to process face information should obtain the individual's individual consent or obtain written consent in accordance with the law

On August 8th, the State Internet Information Office publicly solicited opinions on the "Regulations on the Safety Management of Face Recognition Technology Application (Trial) (Draft for Comment)", which mentioned that only when it has a specific purpose and sufficient necessity, and Face recognition technology can only be used to process face information when strict protection measures are taken. To achieve the same purpose or meet the same business requirements, if there are other non-biometric identification technology solutions, the non-biometric identification technology solutions shall be preferred. The use of face recognition technology to process face information should obtain the individual's individual consent or obtain written consent in accordance with the law. Except for those that are not required by laws and administrative regulations to obtain personal consent. Users of face recognition technology who use face recognition technology in public places, or who store face information of more than 10,000 people, shall file a record with the network information department at or above the municipal level within 30 working days.

  • Applications such as enterprise scientific research and autonomous driving will not be affected by the implementation of the interim measures for generative artificial intelligence service management

According to the news on August 9, Luo Fengying, deputy director of the Network Security Coordination Bureau of the Central Network Information Office, stated at the 11th Internet Security Conference of ISC2023 that the interim measures for the management of generative artificial intelligence services, which will be officially implemented on August 15, will The scope of the law is limited to generative artificial intelligence, which is the familiar large language model, and other artificial intelligence applications such as autonomous driving will not be affected; moreover, the interim measures only regulate the situation of providing services to the public, and enterprises, scientific research institutions, universities, etc. Scientific research is not included in the supervision. During the drafting process of the Measures, the promotion of the development and application of new technologies such as artificial intelligence has always been regarded as an important starting point and end point of the work, hoping to support and promote the development of the artificial intelligence industry. Zhou Hongyi, the founder of the 360 ​​Group who attended the meeting, said later that unless the application of similar media attributes or social mobilization attributes needs to pass a security inspection, if it is a TO G or TO B application, it is a vertical and professional application for the government or enterprises. Yes, there is no security check procedure.

  • Two departments: organize and carry out the application of "5G+smart tourism" application pilot project application work

According to news on August 9, the General Office of the Ministry of Culture and Tourism and the General Office of the Ministry of Industry and Information Technology issued a notice on organizing the application of "5G + smart tourism" application pilot projects. The content of the pilot project includes but is not limited to the following main directions: (1) 5G+smart tourism services, promoting the scale development of application scenarios such as 5G+4K/8K ultra-high-definition video, 5G smart tour guide, and 5G+VR/AR immersive experience. (2) 5G+smart tourism management, exploring the landing application mode of 5G high-definition video surveillance, 5G unmanned equipment automatic driving patrol, 5G Beidou positioning and other services. (3) 5G+ smart tourism marketing, promote the integrated development of new businesses such as 5G new calls, 5G news, 5G panoramic live broadcast and tourism destination marketing, and cultivate new media marketing methods such as 5G interactive live broadcast and 5G+AR live broadcast. (4) 5G+ smart tourism products, strengthen the R&D and industrialization process of 5G integrated application with cruise ships, passenger ropeways, amusement facilities, ice and snow equipment and other tourism equipment. Increase the supply of tourism products such as 5G+ rural tourism, ice and snow tourism, and cruise and yacht tourism. Promote the deep integration of digital products such as 5G-based AI cameras, VR/AR terminals, and wearable devices with cultural and tourism enterprises and institutions, and promote the large-scale promotion of 5G+ smart tourism products.

Investment and Financing

  • Taiyi Chuanxin completed the Pre-A round of financing

Recently, Shenzhen Taiyi Chuanxin Technology Co., Ltd. completed the Pre-A round of financing of several million yuan. The investors in this round include Zhuode Investment, Tianbo Culture Group, etc. Founded in 2021, Taiyi Chuanxin is a technical service provider in the field of smart civil aviation construction. Its services include AI passenger services, airport automated security operations and other service products using cutting-edge technologies.

  • Boer Zhizao received tens of millions of Pre-A round of financing

On August 8, industrial vision solution provider Boer Zhizao received tens of millions of yuan in the Pre-A round of financing. The leading investor was Haolan Capital, and Yiren Capital served as the exclusive financial advisor. This round of financing will be used for further research and development of autonomous optical imaging solutions and AI technology. Bohr Zhicao was established in 2020. Combining self-developed optical solutions and AI algorithms, it mainly detects surface defects of consumer electronics products.

  • Altron Lightnian received tens of millions of US dollars in Series A financing

Recently, MogicAI, an AI full-link marketing company, received tens of millions of dollars in Series A financing, led by Cathay Fund. This round of financing will be mainly used for technical investment in intelligent algorithms and model research and development in the field of AI video to further enhance product capabilities.

Founded in 2021, Altron Lightnian is an AInative marketing company that provides corporate customers with full-link marketing solutions from marketing strategy insight, AI content production, AI smart streaming to DAM content management.

other

  • Daxing Airport introduces autonomous driving technology test project

Recently, the Office of the Beijing High-Level Autonomous Driving Demonstration Zone, together with the two units "Baidu" and "Pony.ai", successfully completed the test of self-driving vehicles getting on and off in front of the terminal building of Daxing Airport, and internal connection in public areas. Run the test, promote the self-driving vehicles of Daxing Airport to drive on the road, and explore the opening of the unmanned operation route from the Economic Development Zone to the airport.

  • Ming-Chi Kuo: The focus of recent observations on AI servers may be detrimental to the stock prices of related supply chains in the short term

On August 9th, TF International analyst Ming-Chi Kuo issued a post saying that the focus of recent observations on AI servers may be unfavorable to the stock prices of related supply chains in the short term. 1. Although the CSP laws in the past two weeks have indicated a positive view on the trend of AI, they have not clearly provided proof of the benefits of investing in AI and future AI capital expenditures. 2. Nvidia's stock price did not rise after the release of GH200. 3. Shares of Super Micro plunged after hours after the earnings report.

  • The preparatory meeting of the brain-computer interface research working group of the artificial intelligence medical device innovation cooperation platform was held

According to the news on August 9, China Academy of Information and Communications Technology, together with experts from the Medical Device Technology Evaluation Center of the State Food and Drug Administration, medical and health institutions, and enterprises and institutions, organized a brain-computer interface research working group preparatory meeting and brain-computer interface technology in the field of medical devices. The application seminar plans to apply for the establishment of a brain-computer interface research working group on the platform, effectively coordinate resources from scientific research institutions, clinical institutions, academic groups and other parties to carry out scientific research cooperation and industrial cooperation, realize resource sharing, and jointly promote scientific and technological innovation of brain-computer interface technology and healthy development. At the meeting, Jin Ruonan from the Second Evaluation Department of the Medical Device Technology Evaluation Center of the State Food and Drug Administration focused on the main concerns in the evaluation of rehabilitation devices based on brain-computer interface technology, including performance research, biocompatibility, software and network Safety, stability research, clinical evaluation path and other aspects.

  • Stability AI launches a large language model for writing code

On August 9th, Stability AI released a new open large-scale language model (LLM) called StableCode, which focuses on assisting users in generating programming language code. Stability AI, known for its stable diffusion models in the field of text-to-image generation, is now entering the field of code generation. StableCode offers three different levels: a general-purpose base model, an instruction model, and a long context window model that supports up to 16,000 tokens.

“Our goal with this model is similar to what we did for Stable Diffusion, which helps everyone become an artist,” Stability AI head of research Christian Lafort said in an exclusive interview with VentureBeat. The StableCode model can do something similar: basically, make it possible for anyone with an idea and a problem to write a program that solves a problem."

  • 75% of global enterprises consider banning ChatGPT and generative AI applications in the workplace

On August 9th, BlackBerry released a new research report showing that 75% of global companies are currently taking or considering taking prohibitive measures to limit the use of generative artificial intelligence applications such as ChatGPT in the workplace.

Of these businesses, 61 percent said the restrictions would be long-term or permanent, with data security, privacy concerns, and risks to corporate reputation being the main considerations. The report also pointed out that 83% of enterprises are concerned about unsafe applications, and worry that these applications may bring network security threats to the enterprise's IT environment.

4c92721165985e000eeedeafe7b2706a.jpeg





7b34b611612d686d7dce37ad7ee3d27d.png

9d8a71cf31570ada4a776c306e194f33.png

Guess you like

Origin blog.csdn.net/YMPzUELX3AIAp7Q/article/details/132200222