Big data industry innovation service media
——Focus on data·Changing business
Entreprise's news
Intelligent programming assistant based on SenseTime's self-developed large language model is open for public beta
On December 7, SenseTime announced on its public account that Raccoon, an intelligent programming assistant based on SenseTime’s self-developed large language model, will be open to public beta from now on! Code Raccoon covers software requirements analysis, architecture design, code writing, software testing and other aspects to meet users' various needs such as code writing, data analysis, and programming learning. In practical applications, it can help developers improve programming efficiency by more than 50%.
Google launches new AI model Gemini
In the early morning of December 7, Google launched Gemini, a new artificial intelligence model that it claims is the largest and more powerful. Gemini will include three different kits: GeminiUltra, Gemini Pro and Gemini Nano. Gemini will power the generative AI chatbot Bard. Bard competes with Microsoft-backed OpenAI's ChatGPT.
ByteDance starts a new round of shareholder buybacks at $160
On December 6, it was reported that ByteDance would buy back $5 billion worth of options from investors, and its IPO plan has not yet been determined. It is understood that the repurchase price is US$160 per share, which is the same as the ByteDance employee option repurchase price in October this year, corresponding to a valuation of approximately US$268 billion. Compared with the US$300 billion valuation at the time of shareholder buyback in September 2022, the valuation has dropped by approximately 10% in one year.
Meta launches standalone AI image generator
On December 6, Meta launched a new, independent AI image generator——Imagine with Meta, which allows users to Description to create the image. According to reports, the new artificial image generator is powered by Meta’s existing Emu image generation model and can create high-resolution images based on text prompts. It's currently free for English-speaking users in the U.S. (unknown if fees will follow), and each prompt generates four images.
TCL Huaxing releases the world's first large-scale model of semiconductor display vertical field
On December 7, during the TCL CSOT Global Display Ecosystem Conference, TCL CSOT released the world's first large-scale model in the semiconductor display vertical field - Xingzhi X-Intelligence. Officials stated that the large model was developed by TCL Huaxing, Zhipu AI and the National New Display Technology Innovation Center, and cooperated with Tsinghua University. It has now been privatized and deployed.
AMD launches new high-performance AI chip
On December 6, AMD announced at the "Advancing AI" event held in San Jose, USA, that it will launch a new AI chip series Instinct MI300, including models such as MI300A and MI300X, as well as corresponding software and services. MI300A has begun mass production, and MI300X has begun shipping. AMD CEO Su Zifeng said that the storage performance of MI300X is better than NVIDIA's H100, and compared with the H200 previously revealed by NVIDIA, its storage performance specifications are also better than the H200 using HBM3e.
Alibaba will pay annual dividend of approximately US$2.5 billion
Alibaba announced on the Hong Kong Stock Exchange on the morning of December 6 that it will provide ordinary share holders and American depositary share holders registered as of the close of Hong Kong time and New York time on December 21, 2023, regarding the 2023 financial results. The first annual dividend was paid in 2017, in the amount of US$0.125 per ordinary share or US$1.00 per American depositary share, payable in US dollars. The dividend payment date for holders of ordinary shares is expected to be on or about January 11, 2024, and the payment date for holders of American depositary shares is expected to be on or about January 18, 2024. Dividends total approximately $2.5 billion.
McDonald's China APP launches Hongmeng native application development
According to news on December 7, McDonald’s China reached a Harmony cooperation agreement with Huawei on December 6, officially announcing that the McDonald’s China APP will launch Harmony native application development based on HarmonyOS NEXT. At present, Meituan, Qunar, Sina, DingTalk, Ant Group, Xiaohongshu, 58 Group, Bilibili, Amap, etc. have started the development of Hongmeng native applications.
AMD raises AI accelerator market size to US$400 billion in 2027
According to news on December 7, AMD CEO Lisa Su said that the size of the artificial intelligence accelerator market is expected to reach US$400 billion by 2027. The company estimated it to be US$150 billion in August; the AI chip market is expected to expand rapidly.
GTCOM and Shenzhen Data Exchange reach strategic cooperation
According to news on December 6, GTCOM Technology Co., Ltd. ("GTCOM") and Shenzhen Data Exchange ("Shenzhen Data Exchange") signed a strategic cooperation agreement. The two parties will explore the construction of a knowledge big data alliance, form a multi-party collaboration mechanism such as data owners, large model technology developers and users, and carry out cooperation around data resource utilization, assetization, capitalization and other aspects.
OpenAI CEO Altman: The board of directors is rethinking the company’s structure
On December 6, at an AI event in Vietnam, OpenAI CEO Sam Altman said that OpenAI is reconsidering its profit-making and non-profit hybrid company structure. Altman said the current structure "obviously has some flaws" and that the new board is seriously considering the best corporate structure that fits its purpose.
TSMC’s self-developed chat AI tGenie has been put into operation
According to reports on December 6, Taiwan Semiconductor Manufacturing Co., Ltd. has recently developed a chat AI tool "tGenie". This AI has been launched in May this year and claims to have helped TSMC save NT$100 million (approximately NT$2,270). Ten thousand yuan) outsourced translation costs. This AI is built on the NVIDIA chip platform. In the future, many foundries are expected to follow TSMC's lead and bring this type of generative AI into the production process.
In addition, TSMC also plans to use AI to reduce wafer process costs and improve real-time monitoring of processes in different factories. Simultaneously detect abnormal behaviors on millions of process control drawings, find the causes through AI comparison, and quickly repair the machine.
Google announces new AI processing chip and cloud 'supercomputer'
On December 6, according to technology news website Engadget, Google has announced the creation of its most powerful TPU (formally known as Tensor Processing Unit), Cloud TPU v5p, an artificial intelligence supercomputer from Google Cloud. Cloud TPU v5p is an artificial intelligence accelerator, training and serving models.
Cloud TPU v5p is an improvement on previous versions such as v5e and v4. According to Google, TPU v5p has twice the floating point performance (FLOPs) and four times the FLOPs performance scalability per rack compared to TPU v4. In addition, it trains LLM models 2.8 times faster than TPU v4 and embeds dense models 1.9 times faster than TPU v4. The new artificial intelligence supercomputer includes an integrated system with open software, performance-optimized hardware, machine learning framework and flexible consumption model. The idea is that this merger will increase productivity and efficiency compared to looking at each part individually.
SK Hynix sets up new department to take charge of artificial intelligence chip business
According to Yonhap News Agency on December 7, South Korea's SK Hynix, the world's second largest memory chip manufacturer, said on Thursday that it will establish a new department called AI Infra to be responsible for artificial intelligence semiconductor-related business. This is the company's next update. Part of a strategy to focus on high-end chips in high demand. This new division will integrate high-bandwidth memory (HBM) capabilities and functions dispersed within the company, and will lead the development of artificial intelligence technologies such as new-generation HBM chips to find and develop new markets. Kim Juseon, currently head of global sales and marketing, will lead the new unit.
financial report
C3.AI’s second-quarter revenue fell short of expectations
According to news on December 7, C3.AI’s second fiscal quarter adjusted loss per share was US$13 million, analysts expected a loss of US$0.18; second fiscal quarter revenue was US$73.2 million, analysts expected US$74.3 million; maintaining full-year operating income The revenue forecast is unchanged at US$295 million-320 million, while analysts expect US$308.2 million; third-quarter revenue is expected to be US$74 million-78 million, while analysts expect US$77.7 million. It is reported that C3.ai is an artificial intelligence company. Founded in 2009 and listed in December 2020.
Douyu: Adjusted net profit increased 180% year-on-year in the third quarter
On December 7, Douyu released its unaudited financial report for the third quarter of 2023. In the third quarter, Douyu achieved revenue of 1.359 billion yuan, a year-on-year decrease of 24.4%; gross profit was 192 million yuan, corresponding to a gross profit margin of 14.2%; under non-GAAP, the platform’s net profit was 71.9 million yuan, a year-on-year increase of 1.80% %, a month-on-month increase of 17.2%. In terms of user data, in the third quarter, the average MAU of Douyu mobile terminal increased to 51.7 million from the previous quarter, and the number of paying users was 3.9 million.
other
The EU is close to reaching a landmark bill to regulate ChatGPT and other artificial intelligence technologies
According to market news on December 7, the European Union is close to reaching a landmark bill to regulate ChatGPT and other artificial intelligence technologies. Representatives from the European Commission, European Parliament and 27 member states reached a compromise in a meeting that began on Wednesday afternoon and lasted several hours, bringing the parties closer to agreeing on a wider piece of legislation known as the Artificial Intelligence Act. 》) reached a formal agreement.
Gansu issues first autonomous driving demonstration application license
According to news on December 6, the Gansu Provincial Department of Transportation, the Gansu Provincial Department of Industry and Information Technology, and the Gansu Provincial Public Security Department jointly issued Gansu’s first autonomous driving demonstration application license, and announced Gansu’s first autonomous driving passenger transport demonstration application road and freight demonstration Application roads. This marks the official launch of autonomous driving demonstration application work in Gansu.
6G will be commercially available around 2030
On December 6, according to CCTV News, Wang Zhiqin, leader of my country’s 6G promotion group and vice president of the China Academy of Information and Communications Technology, said that 6G technology is actually a new technology for the intergenerational update of 5G. Mobile communications are generated every decade, so it is oriented to Looking at 6G, its commercialization time is basically around 2030, and its standardization time will be in 2025.
