Data intelligence industry innovation service media
——Focus on digital intelligence and change business
Entreprise's news
Meituan becomes the sole shareholder of AI company light years away
Recently, according to the national enterprise credit information publicity system, Beijing Guangnianwai Technology Co., Ltd. completed the industrial and commercial change, the founder Wang Huiwen withdrew, and Tianjin Sankuai Technology Co., Ltd., a subsidiary of Meituan, became the sole shareholder. Beijing Light Year Beyond Technology Co., Ltd. was established in July 2018. Currently, Wang Huiwen is still serving as the legal representative, executive director, and manager of Light Year Beyond.
XVERSE-13B, an open source multilingual large model of Yuanxiang, is free for commercial use
On August 7th, Yuanxiang XVERSE Company announced that its high-performance general-purpose large model XVERSE-13B with tens of billions of parameters is open source, which is free for commercial use. According to reports, XVERSE-13B is a large language model that supports multiple languages independently developed by Shenzhen Yuanxiang Technology. The services generated based on it include text generation, automatic writing, data analysis, knowledge question and answering, etc.
IGN launches AI chatbot that can ask for game guides
News on August 7th, according to the latest news from the game evaluation media IGN, IGN will soon launch its own AI chatbot IGN Guides Chatbot Alpha. Once logged in, users can ask questions about the game through a chat window at the bottom of the guide page, and it will use the knowledge contained in its own website, coupled with natural language processing, to provide answers to users' specific questions. According to reports, the bot was trained based on game guides and walkthrough content written by dozens of IGN guide authors and contributors. Since this is an alpha version of the experiment, the quality of the answers may not be very good. The chat window appears at the bottom of the guide page only when logged in.
Intel Core Ultra conference official announcement
News on August 7th, according to information on Intel’s official website, Intel will hold a new innovation conference “Innovation 2023” (Intel On 2023) in San Jose, California on September 19th, US time. In this event, Intel will introduce future consumer hardware platforms, including the highly anticipated new Core Ultra series code-named Meteor Lake, as well as future roadmaps. This will be the first time that Intel’s consumer class has integrated an AI engine, just like AMD’s current Ryzen 7040 series, which can handle some low-load AI reasoning tasks more efficiently, especially image, video, and audio processing, such as dynamic backgrounds Blurring, voice noise reduction, and more. Meteor Lake will adopt the Intel 4 manufacturing process (formerly known as 7nm) for the first time, and adopt the chiplet small chip design for the first time, including CPU, GPU, SoC, and IO.
2023 Huawei Developer Conference Releases Huawei Cloud Pangu Model 3.0 and Ascend AI Cloud Service
According to news on August 6, at Huawei's 2023 Developers Conference held in Dongguan, Zhang Pingan, executive director of Huawei and CEO of Huawei Cloud, announced the launch of Huawei Cloud Pangu Model 3.0 and Shengteng AI cloud services. Pangu Model 3.0 aims to enhance core competitiveness and assist customers, partners, and developers to implement artificial intelligence and create value in various industries. At the same time, Ascend AI cloud service improves AI computing power, making it easier for industry customers to use large models. The Pangu model 3.0 architecture follows the "5 NX" three-tier structure. The base layer (L0) includes NLP, CV, multimodal, predictive, and scientific computing models, providing general-purpose functionality for specific industry applications. Pangu Model 3.0 provides parameters of different scales, ranging from 10 billion to 100 billion, to meet the needs of different customers. The Pangu model has been successfully implemented in various industries such as finance, manufacturing, pharmaceutical research and development, coal, and railways. For example, the Pangu government affairs model empowers Xiaofu, an intelligent government service assistant, to improve service efficiency. The Pangu mining model improves the safety and efficiency of coal mining operations, while the Pangu railway model accurately identifies faults and improves inspection efficiency. The Pangea weather model surpasses traditional methods of weather forecasting in accuracy and speed.
Qualcomm will jointly develop open source RISC-V architecture with NXP, Bosch and other companies
On August 6th, according to an official statement, Qualcomm will cooperate with NXP, Nordic Semiconductor, Infineon and Bosch to develop and promote the open source RISC-V architecture for chip design. The five companies will jointly invest and establish a company in Germany. The initial application focus will be on automobiles, and the new company will expand its business to mobile and Internet of Things chips in the future. Previously, Qualcomm had hinted that at some point it might prefer the RISC-V architecture over the Arm architecture for its products, including Snapdragon smartphone chips.
OpenAI will update ChatGPT
On August 7th, OpenAI announced that it will make a series of small updates to ChatGPT, aiming to improve the user experience. After the update, ChatGPT will add hint examples to help new users understand how to use it. In addition, a new reply suggestion function will be added to provide users with conversation suggestions based on the chat content. Plus users will default to the GPT-4 version for conversations, rather than the current GPT-3.5 version. In addition, ChatGPT will also support uploading multiple files, and users can implement this function through the code interpreter. The updated ChatGPT will also add login status retention and keyboard shortcuts.
Apple may hold an autumn conference on September 13
According to reports on August 7, Apple's annual iPhone event is expected to be held on Tuesday, September 12 or Wednesday, September 13 this year. As is customary for Apple, pre-orders for the new iPhones will begin a few days later on Friday, September 15, with an on-sale date a week later on September 22.
Hanvon Technology: It is expected to release the large model of Hanvon World in October
According to the news on August 7, Hanvon Technology, a domestic leader in OCR recognition, stated on the interactive platform that the company is currently working on developing industry-specific models. More blockbuster new products on the C side. In addition, Hanvon Technology has launched virtual AI robot projects such as digital employees, and is actively researching and exploring the multi-joint control of biped and quadruped robots, forward and reverse motion of robots, and the software and hardware structure of robots.
The Zidong Taichu model is the third largest shareholder of Hanvon Technology, and the model was developed by the Institute of Automation, Chinese Academy of Sciences. Hanvon Technology is deeply analyzing how to better transform the application results of the "Zidong Taichu" multi-modal large model, especially in the application of "medical imaging" and other professional fields.
Microsoft Windows stops supporting Cortana and focuses on generative AI
According to news on August 7, Microsoft will stop supporting the standalone Cortana application on Windows from August 2023 and switch to generative AI. Cortana will still be supported in some Microsoft products such as Outlook Mobile, Teams mobile, Microsoft Teams Display, and Microsoft Teams Rooms. This move means that Microsoft will focus on generative AI, and plans to introduce new AI features in Windows 11, such as voice commands to operate PCs, AI-driven Bing search, and more. With this decision, Microsoft hopes to provide a smarter, personalized user experience and maintain a competitive edge in the ever-growing field of artificial intelligence.
Qunar's cooperation with Huawei involves businesses such as AI and smart full-scenario
According to news on August 7, the cooperation between Qunar and Huawei in the field of tourism has been further expanded. The two parties will create more value for consumers through cooperation involving Huawei's terminal cloud service application ecosystem business, browser and cloud space business, and AI and smart full-scenario business. The content of the cooperation includes that functions such as Qunar's travel reservations, travel guides, and scenic spot tickets will be connected to Huawei's intent framework, and the layout of terminal service pages will be optimized, and the function of flexibly switching cities will be improved.
The goal of the cooperation between the two parties is to improve the end user experience and create a more convenient and value-added service experience. The cooperation also includes cooperation in artificial intelligence technology to provide users with a smart experience. This cooperation shows the deepening and upgrading of the relationship between the two parties, and also shows Qunar's expectation and preparation for the tourism market.
Toyota partners with self-driving company Pony.ai to mass produce robot taxis in China
August 7 news, according to foreign media reports, self-driving company Pony.ai and Toyota said they are cooperating with the goal of producing a large number of "unmanned robot taxis". GAC Toyota will provide Pony.ai with approximately $139 million in funding as part of the partnership.
Toyota's investment follows a roughly $400 million injection into Pony.ai in 2020. Under the partnership, Toyota will provide an undisclosed number of electric vehicles to Pony.ai, which will be equipped with its self-driving technology and "robotaxi network platform." Although this is a sizable investment, Pony.ai has raised more than $1 billion in funding since its inception.
Sailisi Automobile established AITO Wenjie business group
Recently, according to sources, Cyrus Automobile has recently carried out a new round of organizational structure adjustments, and newly established the AITO Wenjie business group. It is understood that the purpose of this adjustment is to further improve management efficiency, respond quickly to decision-making, focus on core business, and promote innovation and development. Previously, in order to strengthen the joint business of Sailisi Automobile and Huawei to develop more deeply and closely, the two parties decided to establish the "AITO Wenjie Sales and Service Joint Working Group", which will be fully responsible for marketing, sales, delivery, service, channels, etc. from July 1. End-to-end closed-loop management of services.
TSMC issued NT$15.9 billion in corporate bonds for new construction and expansion of plant equipment
On August 7, according to Taiwan media reports, TSMC announced the main issuance conditions of the company’s fourth phase of unsecured ordinary corporate bonds. The total issuance amount is NT$15.9 billion (approximately RMB 3.6 billion). , Class B and Class C 3 types. The use and utilization plan of the proceeds raised is to build and expand plant and equipment.
Nvidia H100 rises to $45,000? Some companies have already started to use it for mortgage financing
Recently, some media found that the price of Nvidia H100 has reached as high as 45,000 US dollars by querying the eBay website, an increase of more than 10% from the price of 40,000 US dollars in April this year, and the supply of goods has also decreased significantly compared with the first half of the year. Nvidia insiders said that the price of the e-commerce channel is not controlled by Nvidia, and "is a private online resale." It is learned from the industry that the price of GPU is still rising. As the "hard currency" of artificial intelligence infrastructure, overseas start-ups have begun to use GPU for mortgage financing. This emerging avenue of financing underscores the value of such hardware in the capital-intensive AI "arms race."
Policy/Report
Chengdu: In the first half of the year, the scale of artificial intelligence industry reached 42.4 billion yuan, a year-on-year increase of 36.9%
Recently, the official website of the Chengdu Municipal Government announced that the scale of the city's artificial intelligence industry in 2022 will be 61.6 billion yuan, ranking seventh in the country, the industry's growth rate will rank first in the country, and its comprehensive strength will rank sixth in the country. In the first half of this year, the industrial scale reached 42.4 billion yuan, a year-on-year increase of 36.9%. The recently issued "Several Measures for Accelerating the Innovative Application of Large-scale Models and Promoting the High-quality Development of the Artificial Intelligence Industry in Chengdu" pointed out that it is necessary to promote breakthroughs in core technologies, focus on the research and development of large-scale models in industries such as medical care, finance, commerce, and transportation, and at the same time promote the independence of the computing power industry. Controllable, so that more market players can "carefully calculate".
IPO
Hua Hong Company will officially land on the Science and Technology Innovation Board, which will be the largest IPO this year
On August 7, the leading wafer foundry Hua Hong Company was listed on the Science and Technology Innovation Board. Hua Hong’s IPO raised 21.203 billion yuan, which is the largest A-share IPO this year and the third largest IPO in the history of the Science and Technology Innovation Board. It is understood that Hua Hong is the world's leading foundry company with unique crafts, and it is also the foundry company with the most comprehensive coverage of featured craft platforms in the industry. According to the 2021 global wafer foundry company ranking released by IC Insights, Hua Hong Company ranks sixth and mainland China ranks second. At the same time, the company is also the world's leading and the No. 1 specialty craft wafer foundry company in mainland China.
Investment and Financing
Ossogna completes multi-million dollar Pre-A round of financing
Recently, Shanghai Orthogonal Technology Co., Ltd. Orthogonal Technology (hereinafter referred to as Orthogonal Technology) announced the completion of multi-million dollar Pre-A round of financing. The investors include Linear Capital, Yunqi Capital, and Fengwu Capital. This round of financing is mainly used to accelerate new product development and promote global commercialization. Founded in 2021, Osogna is an enterprise dedicated to creating a new digital design software platform and auxiliary design hardware products.
Jiaan Intelligent completed the A round of financing of nearly 100 million yuan
On August 7, Jiaan Intelligent announced the completion of a new round of A-round financing, with a financing amount of nearly 100 million yuan. The main purpose of this round of funds is to develop new products and improve delivery efficiency. Founded in July 2018, Jiaan Intelligent is a company focusing on the R&D, production and sales of robotic intelligent grinding and polishing workstations and core components.
Yuanlian Technology received tens of millions of RMB A round of financing led by Huaying Capital
On August 7, Shenzhen Yuanlian Technology Co., Ltd. (hereinafter referred to as "Yuanlian Technology"), a provider of integrated infrastructure for intelligent manufacturing, announced the completion of the A-round financing of tens of millions of RMB. This round of financing was led by Huaying Capital, and Hualuo Capital Follow up. This round of funds will be mainly used for industrial smart machine software and hardware research and development and marketing. Yuanlian Technology is committed to building a localized intelligent manufacturing platform with industrial security edge cloud, serving the automation, digitalization, and intelligent transformation and upgrading of manufacturing enterprises.
other
AI brain implant restores sensation and movement to patients
News on August 6, according to the latest news from the US "Popular Science" website, the artificial intelligence brain implant technology developed by the Feinstein Institute of Bioelectronic Medicine at Northwell Health Center in the United States has been able to help paralyzed patients regain their senses and movements.
Keith Thomas suffered a car accident in 2020 which resulted in injuries to his C4 and C5 vertebrae and a total loss of sensation and movement below the chest. Now, however, Thomas can move his arm freely and can feel his sister holding his hand. Chad Bouton, developer of the technology and principal investigator of the clinical trial, said it was the first time a paralyzed human's brain, body and spinal cord had been electronically linked together to restore long-lasting movement and sensation. When a study participant wants to move his arm or hand, "supercharging" his spinal cord and stimulating his brain and muscles can help reestablish contact, provide sensory feedback and facilitate recovery.
New Acoustic Attack AI Model: Steal data from keyboard taps with 95% accuracy
A team of researchers from a UK university has trained a deep learning model that can steal data from keystrokes recorded using a microphone with 95 percent accuracy. Such an attack is said to seriously compromise the data security of the target, and it may leak people's passwords, discussions, messages, or other sensitive information to malicious third parties. To train the algorithm, the researchers collected training data by repeatedly pressing 36 keys on a MacBook Pro and recording the sound each key made. They then processed the recordings to extract waveforms and spectrograms for each key, providing identifiable differences.
New AI can recognize objects in dark environments
News on August 7th, recently, American scientists developed a new type of visual sensor Hadar, which combines thermal imaging and artificial intelligence technology, can penetrate the dark world of darkness, accurately identify objects in the environment, and even perceive the texture of objects , color, etc., just like seeing objects during the day.
The technology could not only be used in self-driving cars and robots, but could also be further developed for use in military, earth science, medical and wildlife detection.
Aletheia: Nvidia will significantly expand AI GPU production
According to the "Science and Technology Innovation Board Daily" on the 7th, a report released by the research institution Aletheia predicts that by 2024, the AI server market will double to 135 billion US dollars, which is 4.5 times the size of 2022. In AI servers, GPU accounts for the highest cost. Aletheia predicts that by 2025, the GPU market will grow three times, compared to 2023. In order to meet the needs of the AI server market, Nvidia plans to significantly increase its AI GPU production capacity. It is estimated that by 2024, its AI GPU production capacity will increase by 2.5 times. In addition, Nvidia's data center business revenue compound annual growth rate will reach 85% from 2023 to 2026.
