[Industrial Internet Weekly is a featured product released by TMTpost of Titanium Media, which will integrate the cutting-edge trends, blockbuster policies and industry research reports in the fields of enterprise-level services, cloud computing and big data this week. 】
Domestic information
Joe Tsai: Nvidia's supply outage has affected the cloud business, and China will develop its own high-end GPUs
Chairman and co-founder Joe Tsai of Alibaba recently gave a podcast to Nicolai Tangen, Chief Investment Officer of Norway's sovereign wealth fund. In the interview, talking about the situation in chips and semiconductors, Tsai Chongxin pointed out that in October last year, the United States imposed very strict restrictions on the ability of companies such as Nvidia to export high-end chips to each company in China. As a result, they abandoned the Entity List method and added the whole of China to the list. "I think we've really been affected. In fact, we've actually said publicly that this really impacts our cloud business and our ability to deliver high-end computing services to our customers. So in the short term, maybe in the medium term, that's a problem. But in the long run, China will develop its own capabilities to make these high-end GPUs. ”
How big of a problem is the lack of chips? Tsai said, "It's a big problem. It's something that everyone is trying to solve. But in the short term, he thinks people have stockpiled up before the restrictions are lifted. So for now, in the next year or 18 months, the training of large language models can still continue, taking into account people's inventory. He argues that training requires more computing power than applications, which people call inference. So when it comes to reasoning, there are a variety of options. There is no need to have high-power and high-end chips like the latest models of Nvidia.
Zhou Hongyi claims to be an open-source believer: he announced that he will open-source the 360 Zhibrain 7B model and support 500,000 words of long text input
Zhou Hongyi, the founder of 360, revealed that 360 Zhibrain 7B (7 billion parameter model) will be open-sourced, supporting 360k (500,000 words) long text input. Zhou Hongyi said that some time ago, the length of the text of the large model industry volume, 1 million words, "will soon be the standard". "We plan to open source this capability and set it at 360k mainly for the sake of lip service. He also calls himself a "believer in open source" and believes in the power of open source.
According to reports, the 360 intelligent brain long text ability has been settled in the large model product "360AI browser". Zhou Hongyi also talked about the advantages of the small model: he believes that the small model is fast, the user experience is good, and the single machine and single card can run, which has a higher cost performance.
Large factories no longer require employees to set OKRs
In 2013, Byte introduced OKR for the first time in China, and then major Internet companies followed suit, and even many traditional industries began to absorb this management essence from Silicon Valley.
Starting in 2021, the OKR Improvement Department of the Byte Management Research Institute, which specializes in the use of OKRs in companies, no longer mandates employees in their departments to set OKRs. In February 2023, ByteDance issued an all-staff letter, changing bi-monthly OKRs to quarterly OKRs. The management explained that it is because Byte's business is relatively mature now, and the bi-monthly changes are not obvious, and the review cycle can be longer. But what is reflected behind it is the "cognitive refresh" of the role of OKR from top to bottom.
In addition to bytes, many large manufacturers have also begun to stop writing OKRs. In the context of weakening growth expectations, OKRs may no longer be a panacea.
Alibaba Cloud fully implements AI code writing internally, and 20% of the code in the future will be written by Tongyi Lingcode
Alibaba Cloud is fully implementing AI programming internally, using Tongyi Lingcode to assist programmers in writing code, reading code, checking bugs, and optimizing code. Alibaba Cloud also assigned a formal employee ID number to Tongyi Lingcode - AI001.
"In the future, 20% of the company's code will be written by Tongyi Lingcode, but programmers are still the core of R&D, and they will have more time to focus on system design and core business development," Alibaba Cloud said. ”
Baidu released Comate 2.0, an intelligent code assistant, which has written a quarter of the internal code
Baidu officially released BaiduComate 2.0, an intelligent code assistant, which is free for individual developers. According to reports, Comate2.0 can interpret the enterprise and personal private domain code base, and at present, all functions including single/multi-line continuation, multiple recommendations, comment generation code, repair suggestions, long function splitting, and chat questions have been used by all developers. It is understood that BaiduMate has mastered more than 100 mainstream development languages and has written a quarter of Baidu's internal code. In addition to Baidu, BaiduComate has joined more than 10,000 companies such as Himalaya, iSoftStone, and Shanghai Mitsubishi Elevator, with an enterprise code adoption rate of nearly 50%.
GDS announced its entry into the Tokyo market in Japan, and will build a data center campus with a total IT capacity of 40MW
GDS and Gaw Capital announced that they have reached a cooperation agreement to jointly invest in the construction of a data center park with a total IT capacity of 40MW in Tokyo, Japan, with a total area of 10,969 square meters, which is scheduled to be put into operation by the end of 2026 and will become the largest data center cluster in Fuzhong City after completion. The signing of this agreement marks GDS's further expansion of its international business territory to Northeast Asia after its layout in Southeast Asia.
Alibaba invested $12.5 billion in fiscal 2024 to repurchase a total of 1.249 billion common shares
Alibaba announced on the Hong Kong Stock Exchange that it repurchased a total of 524 million ordinary shares (equivalent to 65 million American depositary shares) for a total price of US$4.8 billion for the quarter ended March 31, 2024. For the financial year ended March 31, 2024, a total of 1,249 million ordinary shares (equivalent to 156 million American Depositary Shares) were repurchased at a total price of US$12.5 billion.
Mobvoi passed the listing hearing: annual revenue of 500 million, or the first share of Hong Kong AIGC
Mobvoi, an AI company, has passed the listing approval of the Hong Kong Stock Exchange and obtained the IPO filing of the Securities Regulatory Commission, and is expected to become the first stock of AIGC in Hong Kong.
According to the prospectus, Mobvoi's revenue in 2021, 2022, and 2023 will be 398 million yuan, 500 million yuan, and 507 million yuan respectively. Mobvoi's main business is generative AI and voice interaction technology, with revenue of 500 million yuan in 2022. The company's revenue is mainly derived from AI software (including AIGC solutions and AI enterprise solutions) and AIoT hardware. In recent years, the revenue of AIGC solutions has grown rapidly, with a compound growth rate of more than 300% from 2021 to 2023. The gross profit margin of Mobvoi in the past three years was 37.5%, 67.2% and 64.3% respectively. The company has completed multiple rounds of financing, and its shareholders include Sequoia Capital, Zhen Fund, Google, etc.
Liu Yuhong of Tencent: Tencent's hybrid model is actively developing multimodal models and applications
Liu Yuhong, Vice President of Tencent Cloud and Head of Tencent Hybrid Model, said that the multimodal model will reconstruct the content industry, and applications based on multimodality will also flourish. He introduced that the technical architecture of Tencent's hybrid model has been upgraded to a hybrid expert model (MoE) architecture, with a parameter scale of trillions of yuan, and is good at handling complex scenarios and multi-task scenarios. At present, multimodal models are being actively developed to further strengthen the capabilities of Wensheng diagrams, Wensheng 3D and Wensheng/Tusheng videos. At the same time, we will explore applications such as digital intelligent humans, voice dialogue virtual humans, and embodied intelligent robots based on large-scale models to improve efficiency and experience.
117 generative AI services have been recorded
The Cyberspace Administration of China (CAC) updated the filing information of generative AI services in March, adding 23 new companies, including Lenovo's Tianxi model, iQiyi's AI dialogue product, Didi's Didi Chuxing model, Honor's mobile phone test, and the PC version of YOYO Assistant. As of March 2024, a total of 117 generative AI services in mainland China have completed filings.
JD Cloud Live Room "Hard" Alibaba Cloud: Real-time price comparison, up to 15% lower
At 7 p.m. on March 31, Lao Luo sold Alibaba Cloud live for the first time. At the same time, the JD Cloud live broadcast room will conduct real-time price comparison for the Alibaba Cloud products brought by Lao Luo, and the price will be lower on the basis of the price of Alibaba Cloud products, up to 15% lower. In addition, users who place an order for price comparison products in the live broadcast room can renew the price in the live broadcast room for another 3 years.
Shunwang Technology and Moonlight Technology jointly launched the cloud AI APP "Moonlight AI"
The computing cloud business team of Shunwang Technology and Hangzhou Chengli Moonlight Technology Co., Ltd. launched the cloud AI APP "Moonlight AI". It is understood that "Moonlight AI" integrates a number of AI service applications, and users can experience AI text creation, AI picture creation, AI short video creation and other functions through the platform. Previously, Shunwang Technology expressed its opinion that there will be three different implementation paths for the development of cloud AI PC, cloud AI APP, office cloud AI PC and personal cloud AI PC.
Dongtu Technology released the first industrial AI controller in China
The first intelligent controller in China to realize industrial AI, the intelligent controller can convert human natural language instructions into robot motion commands through AI large models, and control the robot work by human language.
The business strategy of Inspur Hairuo large model was officially released
Inspur Cloud officially released the business strategy of Hairuo large model, and Inspur Cloud will invest 5 billion yuan to accelerate the rapid implementation of Hairuo large model in 100 cities relying on the digital innovation ecological community, continue to activate the value of data elements, and help the development of new quality productivity.
The government and enterprise affairs channel on the official website of Alibaba Cloud was launched
It is understood that in order to better serve government and enterprise customers and accelerate business value innovation, in January 2024, Alibaba Cloud will newly establish a government and enterprise business department, mainly for government, central state-owned enterprises and public utilities customers.
Overseas news
Google considers charging for AI-based searches
According to foreign media, Google is considering charging for AI-based searches, which is a major change in its business model. Google is considering various options, including adding certain AI-based search features to its premium subscription service.
SK hynix plans to spend nearly $4 billion to build its first U.S. chip factory
SK hynix plans to spend USD 3.87 billion to build an advanced packaging plant and an AI product research center in Indiana. The world's second-largest memory chip maker said it will build its first U.S. factory in West Lafayette, with mass production scheduled to begin in the second half of 2028. The plant will focus on building a production line for next-generation high-bandwidth memory chips, which are key components of graphics processors for training AI systems.
Apple is reportedly researching home robots as the next major project
Apple's team is working on advancing personal robotics in an area that could be one of the company's ever-changing "next big projects," according to people familiar with the matter. It is reported that Apple's engineers are developing a mobile robot that can follow users around their homes. Apple has also developed an advanced desktop home device that utilizes robots to move displays around. Although R&D is still in its infancy, and it's unclear that these products will eventually be released, there is growing pressure on Apple to find new revenue streams. The company abandoned its electric vehicle project in February, and the mixed reality headset is expected to take years to become profitable. With robotics, Apple can gain a greater foothold in the consumer home device space and take advantage of advances in artificial intelligence. But it's unclear what approach will be taken. According to people familiar with the matter, although robot smart displays have developed faster than mobile robots, they have been in and out of Apple's product roadmap for many years.
Musk: Will raise the pay of Tesla's AI engineers to stop OpenAI poaching
Tesla CEO Elon Musk said that Tesla has been raising the pay of AI engineers due to the war for talent. He posted on his social media site, X: "Tesla is increasing the pay of its AI engineering team (subject to progress milestones). OpenAI, the creator of ChatGPT, has been "actively recruiting engineers for Tesla, offering huge pay, and unfortunately, in some cases with success," he said. ”
Intel's total product revenue in 2023 was $47.7 billion, compared to $57 billion in the previous year
Intel's total product revenue in 2023 was $47.7 billion, compared to $57 billion in the previous year, data center and artificial intelligence revenue was $12.64 billion in 2023, compared to $16.86 billion in the previous year, Intel Foundry is expected to break even between now and 2030, Intel foundry's operating loss is expected to peak in 2024, Intel foundry adjusted gross margin target is 40%, adjusted operating margin target is 30%; Lorenzo Flores has been appointed Chief Financial Officer of Intel Foundry.
U.S. covert operation exposed: ASML was asked to relocate to the United States
Media reports have focused on the dissatisfaction of ASML, a Dutch manufacturer of high-end lithography machines, with its restrictive stance in its country. It is reported that ASML is considering moving its headquarters out of the Netherlands due to restrictions on foreign employees and obvious conflicts over exports to China.
This move by ASML has caused great concern for the Dutch government. The Dutch government, fearing that ASML's departure would have a major impact on the country's high-tech industry, urgently launched a secret project called Operation Beethoven to keep the important company.
Previously, it was widely believed that ASML's relocation destination would be France. However, according to tech media semiwiki, the United States has been secretly trying to persuade ASML to move to the United States, codenamed "Roll Over Beethoven".
Apple's market value has wiped out more than $300 billion this year, and investors are speculating that the worst may be over
Apple has lost more than $300 billion in market value this year, ceding the position of the most valuable American company to Microsoft. As the S&P 500 rose, Apple's return in the first quarter lagged the index by 21 percentage points, the worst performance since 2013, the data showed. But Apple's share price has fallen far enough to approach the key support level of $165, which was the low it hit when the broader market bottomed out in October last year.
OpenAI launched a registration-free ChatGPT feature, and the demand for ChatGPT Enterprise is growing dramatically
OpenAI announced that users can use and experience ChatGPT's new features without registration. The aim of the initiative is to make AI accessible to any interested user, rather than requiring a complicated registration process to use tools like GhatGPT.
OpenAI's COO said that despite the competition, demand for ChatGPT Enterprise has grown dramatically, with more than 600,000 people currently signing up for the version, compared with about 150,000 in January.
Microsoft rose 1.67% and intends to sell Teams and Office products separately worldwide
Microsoft rose 1.67% to $427.765. Microsoft has announced that it will sell its chat and video app Teams separately from Office products globally. Starting April 1, customers can continue their current license agreement, renew, renew, or switch to a new offer. Six months ago, the company began spinning off the two products in Europe to avoid facing EU antitrust fines. A Microsoft spokesperson said, "It would also address the issues raised by the European Commission and provide greater flexibility for multinational companies when they want to standardize cross-regional procurement." ”
Microsoft and OpenAI plan to invest $100 billion in data center projects
According to sources, Microsoft and OpenAI are working on plans for a data center project that could cost as much as $100 billion, including an artificial intelligence supercomputer called Stargate, which will launch in 2028.
AWS plans to invest $148 billion over the next 15 years to build data centers
According to Bloomberg, AWS plans to invest $148 billion over the next 15 years to build data centers around the world to meet the demand for cloud services in areas such as artificial intelligence.
In addition to expanding its existing data center facilities in Virginia and Oregon, AWS plans to expand into other U.S. states and overseas countries such as Saudi Arabia and Malaysia.
Stability AI发布免费音频生成工具，可创作AI歌曲
Stability AI released Stable Audio 2.0, an audio generation model that now allows users to upload their own audio samples, then convert them with prompts and create AI-generated songs for free.
Stability AI资金耗尽，无法支付租用的云GPU账单
According to foreign media, Stability AI's popular text-to-image generation model, Stable Diffusion, is out of funds and cannot pay for the GPU cluster fees needed to train large models. As of October last year, the company had only $4 million left in reserve.
According to detailed reports citing company documents and dozens of people familiar with the matter, the British model-building company's extremely high infrastructure costs have depleted the company's cash reserves, leaving it with only $4 million as of October last year. In addition, $54 million is required for salaries and operating expenses. And its estimated revenue in 2023 is only $11 million. It owes tens of millions of dollars on bills to AWS, Google, and CoreWeave.
The fundraising for Stability AI was also unsuccessful, with Intel promising to invest $50 million, but only $20 million has been paid. CEO Emad Mostaque revealed on social media late last month that he had resigned, with COO Shan Shan Wong and CTO Christian Laforte serving as interim co-CEOs.
Intel disclosed an operating loss of $7 billion in its chip manufacturing division
On April 2, Intel disclosed in a filing with the Securities and Exchange Commission (SEC) that the company's new unit responsible for chip manufacturing, "Intel Foundry", will have revenue of $18.9 billion in 2023, down 31% year-on-year, compared with $27.49 billion in 2022, and the operating loss widened to $7 billion from $5.2 billion in the previous year.
Intel CEO Pat Gelsinger did not shy away from the losses faced by the business, and said that 2024 will be the worst year for the company's chip manufacturing business to lose money.
Amazon held a generative AI communication conference, and Chinese companies rushed to cooperate?
At the Amazon Web Services Generative AI Communication Conference, the Claude 3 series of large models released a month ago were highlighted. AWS does not deploy Claude 3 on servers in China. Like Microsoft Azure, Chinese companies can apply to call Claude 3 models deployed in other regions and complete calculations through the Bedrock service provided by AWS globally.
An Amazon source said that after the release of the Claude 3 series model, they received a large number of requests from Chinese companies for cooperation.
The AI programming language Mojo is officially open-sourced, claiming to be 90,000 times faster than Python
Modular Inc. announced the open-sourcing of Mojo's core components. Mojo, a programming language designed specifically for writing AI software, was officially released in August last year and has amassed more than 175,000 developers and 50,000 organizations to date.
AI models are often written in a variety of programming languages. Developers will implement the simplest part of neural networks in Python, a language that is easy to learn but relatively slow. The rest of the code is usually written in C++, which runs faster than Python but is also harder to learn.
Modular positions Mojo as a more convenient alternative. It offers an easy-to-use syntax, similar to Python, but can run thousands of times faster. As a result, developers can write fast AI models without having to learn complex languages like C++.
Financing mergers and acquisitions
Oneway Network received 10 million yuan in Series A financing
OneCheng Technology, an AI intelligent marketing company, received 10 million yuan in Series A financing. This round of financing is used to upgrade product technology and market expansion, continuously improve service quality and user experience, and at the same time innovate and upgrade products to help physical merchants and e-commerce enterprises with operational problems.
Aerospike, a provider of real-time vector and graph databases, has raised $109 million in Series F funding
Aerospike, a real-time vector and graph database provider based in Mountain View, California, raised $109 million in growth funding.
Backers include Sumeru Equity Partners, as well as additional investment from existing investor Alsop Louie Partners. As part of the transaction, Sumeru co-founder and managing director George Kadifa will also join Aerospike's board.
The company intends to use the funds to accelerate innovation and go-to-market capabilities across its core trading, analytics and AI solutions, including vector and graph databases.
Deep learning infrastructure developer Lambda has raised $500 million in Series D funding
Lambda announced that it has locked in a financing loan of up to $500 million. Unusually, the loan is secured by the NVIDIA GPU (Image Processing Unit) chips it owns, and is known as a "special-purpose GPU financing vehicle": secured by GPUs and backed by the cash flow they generate. This represents a significant milestone in the AI computing market, enabling the company to fund cloud deployments for thousands of users without requiring them to sign long-term contracts, Lambda wrote.
Lambda said the $500 million funding will support the company's deployment of more NVIDIA GPUs for AI developers to train, fine-tune and inference generative AI models in Lambda Cloud, a first-of-its-kind financing tool that opens up new ways to fund the deployment of tens of thousands of NVIDIA GPUs.
Riguan Chip, a digital back-end EDA company, received tens of millions of yuan in Pre-A round financing
Riguan Core has completed tens of millions of yuan in Pre-A round of financing, and the investors in this round are Industrial Capital Skyworth Investment, Production and Research Zhongxiang, etc., and Potential Energy Capital serves as the financial advisor for this round of financing. The financing funds will be mainly used for marketing and product upgrade and iteration research and development.
Founded in 2021, the company specializes in digital chip design EDA tools, especially the signoff of ultra-large-scale digital chips, to improve chip design efficiency, and is a domestic company with full-process analysis capabilities from timing constraints, timing analysis, reliability analysis, physical verification, etc.
Musk's artificial intelligence company xAI is seeking $3 billion in funding, and its valuation is rushing to $18 billion
There is news that Musk's AI startup xAI plans to raise money. It is understood that some investors close to Musk are discussing to help his company raise $3 billion. After the completion of this financing, xAI is expected to be valued at $18 billion.
According to related news, venture capital firm Gigafund and American venture capitalist Steve Jurvetson are considering investing in this round of financing.
Energy Manufacturing
China's nuclear power: the cumulative power generation in the first quarter was 51.090 billion kWh, a year-on-year increase of 2.80%
As of March 31, 2024, the company held 25 units in operation with an installed capacity of 23.750 million kilowatts, and 15 units under construction or approved to be built, with an installed capacity of 17.565 million kilowatts, with a total installed capacity of 41.315 million kilowatts. The company holds 21,340,600 kilowatts of installed capacity in operation, including 7,332,600 kilowatts of wind power and 14,008,000 kilowatts of photovoltaic power, 679,000 kilowatts of independent energy storage power stations, and 14,262,600 kilowatts of installed capacity under construction, including 3,339,800 kilowatts of wind power and 10,922,800 kilowatts of photovoltaics. As of March 31, 2024, the company's cumulative power generation in the first quarter of 2024 was 51.090 billion kWh, a year-on-year increase of 2.80%, and the cumulative on-grid electricity was 48.177 billion kWh, a year-on-year increase of 3.02%.
China Tianying: Signed an investment cooperation agreement with the Anda Municipal Government for the integration of wind, solar, hydrogen and ammonia alcohol, with a total investment of about 16.95 billion yuan
China Tianying announced that Anda Tianying and Anda Municipal People's Government signed an investment cooperation agreement on wind, solar, hydrogen storage, ammonia and alcohol integration project, with a total investment of about 16.95 billion yuan, and an installed capacity of 1.8GW of new energy, including 1.4GW of wind power generation, 0.4GW of photovoltaic power generation, 140MW/280MWh of comprehensive energy storage such as gravity energy storage, 100,000 tons/year of green hydrogen plant, 38,000 tons/year of ammonia plant, and 620,000 tons/year of methanol plant.
China's first large-scale shale gas field was commissioned in the first quarter of this year with 33 new wells
It was learned from the Fuling shale gas field of Jianghan Oilfield of China Petroleum & Chemical Corporation that the production and sales of the gas field were booming in the first quarter of this year, with a total of 33 new wells put into operation from January to March, an increase of 50% year-on-year, with gas production of 1.819 billion cubic meters and sales of 1.745 billion cubic meters, exceeding the planned output by 15 million cubic meters and the planned sales volume of 14 million cubic meters. According to the relevant person in charge of the Fuling shale gas field, since the beginning of this year, the gas field has further promoted the three-story three-dimensional development of the old Jiaoshiba area and the two-story development and adjustment of the Jiangdong and Pingqiao blocks, with an average daily increase of more than 600,000 cubic meters. At the same time, the dynamic monitoring of gas wells was carried out in an all-round way, and the comprehensive treatment of old areas was continuously strengthened, and a total of 133 wells were carried out in the first quarter, with a total of 38.8 million cubic meters of gas increased.
Sinopec plans to build 600 carbon-neutral demonstration projects
According to a news release from the Sinopec Information Office, the company officially launched the second phase of the green enterprise action plan on April 1, which will focus on four major actions: carbon reduction, pollution reduction, efficiency improvement and greening, and strive to basically build a green production system by 2028. Sinopec said that the carbon reduction action will strive to build 600 carbon neutrality demonstration projects with a carbon compliance rate of 100%, the pollution reduction action will promote the construction of a zero-waste group and environmental risk management and control, the efficiency improvement action will focus on improving mineral resources and energy utilization efficiency, and strive to achieve a 5% reduction in comprehensive energy consumption per 10,000 yuan of output value, and the green action will focus on promoting green ecological design, green technology innovation, and green digital intelligence construction.
Policies & Trends
PBOC: Continue to increase support for major strategies, key areas and weak links
The Monetary Policy Committee of the People's Bank of China held its regular meeting for the first quarter of 2024. The meeting pointed out that it is necessary to deepen the structural reform of the financial supply side and build an institutional mechanism for finance to effectively support the real economy. We will guide large banks to play their role as the main force in financial services for the real economy, promote small and medium-sized banks to focus on their main responsibilities and main businesses, support banks in replenishing capital, and jointly maintain the stable development of the financial market. We will do a good job in science and technology finance, green finance, inclusive finance, pension finance, and digital finance, continue to increase support for major strategies, key areas and weak links, and promote the acceleration of the development of new quality productivity.
National Data Bureau: Promote the construction of urban computing power networks as a whole, and promote the integration of new energy vehicles into the new power system
The National Data Bureau solicited public comments on the "Guiding Opinions on Deepening the Development of Smart Cities and Promoting the Digital Transformation of Cities". The "Opinions" proposes to further implement the action of strengthening the foundation of urban cloud networks, accelerate the construction of new radio and television networks, promote the construction of gigabit cities, and explore the development of digital low-altitude infrastructure. Promote the construction of urban computing power network as a whole, realize the efficient matching of urban computing power demand with the efficient supply and demand of computing power resources of national hub nodes, and effectively reduce the cost of computing power use. Build infrastructure for data circulation and utilization, and promote the credible and controllable flow of data between government departments, between government and enterprises, and between industrial chain links. Promote the integration of new energy vehicles into the new power system, and promote the coordinated development of urban intelligent infrastructure and intelligent networked vehicles. Accelerate the establishment of systems such as for data property rights, circulation and transactions, income distribution, and security governance, and promote the efficient circulation and use of data elements.
The National Data Bureau openly solicited the "Guiding Opinions on Deepening the Development of Smart Cities and Promoting the Digital Transformation of Cities in the Whole Region" from the public. The guiding opinions put forward the in-depth implementation of the action of strengthening the foundation of urban cloud networks, accelerating the construction of new radio and television networks, promoting the construction of gigabit cities, and exploring the development of digital low-altitude infrastructure.
Promote the reform of the market-oriented allocation of data elements, and accelerate the construction of a national integrated computing power network
The first national data work conference was held in Beijing from April 1 to 2. Promoting the reform of the market-oriented allocation of data elements is the main line of current data work. The relevant person in charge said at the National Data Work Conference that among them, accelerating the construction of the national integrated computing power network is one of the key tasks.
The agency predicts that in the next few years, the growth rate of computing power demand in mainland China may be as high as more than 30% per year. In response to the problems of insufficient supply of computing power and uneven distribution of resources, the National Data Bureau said that it is focusing on the integrated layout of general computing power, intelligent computing power and super computing power, the integration and coordination of computing power in the eastern, central and western regions, the integrated application of computing power and data and algorithms, the integration of computing power and green power, and the integration of computing power development and security assurance, etc., and actively promote the construction of a national integrated computing power network.
The Cyberspace Administration of China (CAC) has released information on the filing of generative AI services
On April 2, the official website of the Cyberspace Administration of China (CAC) released an announcement on the filing information of generative AI services.
The Cyberspace Administration of China (CAC) stated that to promote the innovative development and standardized application of generative AI services, the cyberspace administration departments, in conjunction with relevant departments, are to orderly carry out the filing of generative AI services in accordance with the requirements of the "Interim Measures for the Administration of Generative AI Services", and the information that has been filed is hereby announced. Where generative AI services with public opinion attributes or social mobilization capabilities are provided, the filing procedures may be performed through the local internet information departments, and the local internet information departments should promptly publicly release the information that has already been filed, and will periodically summarize and update it on the official website, without further announcements.
It is reported that generative AI applications or functions that have been launched are required to publicize the use of registered generative AI services in a conspicuous position or on the product details page, indicating the model name and filing number.
![The harsh reality facing homebuyers: How serious are the consequences of mortgage failure?[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)