Swift Tiger, Wise Dragon: China's Projects and Prospects in the Generative AI Race / ITech content

Contents:

• Grand plans for 2030
• Achilles and the tortoise: the race with the USA
• Why ChatGPT was not invented in China
• China's weaknesses in the AI ​​race: the opinion of the Western press
• Triple shortage: experience, hardware and data
• China is ahead of the rest in AI regulation
• A "hurricane of models" from the Middle Kingdom
• Some neural networks from the "hurricane" models"
• To summarize

This article discusses why ChatGPT was not developed in China and how local entrepreneurs perceived its emergence as a challenge. This article examines why the Chinese leadership openly declares its goal of achieving global dominance in the field of artificial intelligence. It also analyzes the steps China is taking to overtake the United States in the quantity and quality of neural networks. Sergey Karelov, PhD and expert in interdisciplinary studies and strategic analytics, has made a significant contribution to understanding the specifics of the AI ​​industry in the East. His experience and knowledge, as well as his work on the author's channel "Little-known interesting", allow for a deeper analysis of the trends and characteristics of artificial intelligence development in Eastern Europe and Asia. Studying Sergey's views helps understand the unique aspects and challenges this industry faces in Eastern countries. Great plans for 2030 China's artificial intelligence development strategy aims to achieve world leadership status in this field by 2030. The official announcement was made in 2017, when the State Council of the People's Republic of China published the "Next-Generation Artificial Intelligence Development Plan." This plan includes initiatives and measures aimed at stimulating research, technology development, and the implementation of AI across various economic sectors, which will allow China to strengthen its position on the global stage in the field of artificial intelligence. By 2030, China's research and development in artificial intelligence should reach a global level, allowing China to become a major innovation hub in this field. The country is expected to demonstrate significant achievements in the application of intelligent economy technologies and the formation of an intelligent society. This will secure China's confident position as a leading national center of innovation and economic power, promoting the development of new technologies and the optimization of existing processes. The Next-Generation Artificial Intelligence Development Plan, presented by Stanford University, focuses on key aspects and directions that will shape the future of AI technologies. The main goal of this plan is to create more powerful, safe, and ethical artificial intelligence systems capable of effectively interacting with humans and solving complex problems. Key elements of the plan include research in machine learning, the implementation of principles of transparency and explainability of algorithms, and the development of an interdisciplinary approach to the creation and application of AI. Stanford University emphasizes the need to develop ethical standards and regulations to ensure that advances in artificial intelligence benefit society and do not pose security threats. Investments in education and training also play a key role in this plan, which will ensure the sustainable development of AI technologies and their integration into various areas of life. The Chinese economy is developing within the framework of five-year plans similar to those used in the USSR. Important goals for the development of artificial intelligence (AI) were included in the 14th Five-Year Plan for 2021–2025. These goals are also reflected in long-term strategies extending through 2035. The development of AI in China is aimed at strengthening the country's position in the global technology race and increasing the efficiency of various economic sectors.

Dr. Lee Kai-Fu, former president of Google China and author of "AI Superpowers," emphasizes that a significant event that spurred public attention in China to artificial intelligence was the victory of the AlphaGo neural network over world Go champion Lee Sedol. This victory not only demonstrated the potential of AI but also served as a key catalyst for further research and investment in AI technologies in the country.

The enormous popularity of Go in China has played a significant role in its spread. In 2016, a landmark match took place in which the neural network of the British-American startup DeepMind defeated Go master Lee Sedol with a score of 4:1. The event attracted over 280 million viewers nationwide, further confirming the interest in this ancient game and its modern technologies.

China's drive to become a global leader in artificial intelligence should not be viewed solely as a reaction to AlphaGo's impressive victory. In fact, China's AI ambitions are underpinned by carefully considered and rational strategies. Chinese policy in this area includes investments in scientific research, technology development, and infrastructure, underscoring the country's seriousness.

Due to historical reasons, China missed out on the previous industrial revolutions that began in the 18th and 19th centuries in England and the United States. Currently, China's leadership is actively seeking to participate in the "Fourth Industrial Revolution" based on modern information technology. Artificial intelligence, as a relatively new and dynamically developing field, offers unique opportunities for achieving significant breakthroughs. This makes China a key player in the global race for AI innovation.

China strives to take its "rightful place" in the world through advances in artificial intelligence. The government recognizes that AI development can be a key driver of economic growth, helping to eliminate regional inequalities and improve social governance. Investments in AI not only stimulate the economy but also provide more effective solutions to social problems, ultimately strengthening China's position on the global stage.

Chinese leaders are actively considering the use of intelligent systems in the military. The use of modern algorithms and technologies allows them to achieve a strategic advantage over potential adversaries. The integration of artificial intelligence into military operations opens up new horizons for increasing the efficiency and effectiveness of actions, which is an important aspect in the context of global security.

China's task in the coming years is to take a leadership position in the field of artificial intelligence by 2030. This plan is pragmatic and is not based on vain ambitions or a desire to surpass the United States to prove its power. China focuses on the practical aspects of AI development, such as investing in technology, training specialists, and creating innovation ecosystems.

Chinese philosophy, based on Confucian principles, has for millennia emphasized the need to simplify life. In Chinese culture, progress is viewed as increasing fairness in society. In today's environment, the development of artificial intelligence can be a key factor in achieving this fairness. AI can optimize processes and improve access to information and resources, which in turn contributes to the creation of a more equal society. Thus, Confucian philosophical ideas find new application in modern technological realities, emphasizing the importance of harmony and fairness in the course of progress.

They formulate clear and specific goals, which allows the government to effectively oversee the process of their implementation. This also facilitates the coordination of the efforts of a large number of people united to achieve these goals. This approach ensures transparency and simplifies progress monitoring, which ultimately leads to more successful task completion.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He hosts his own channel, "Little-Known Interesting," where he shares unique knowledge and analytical materials. The channel's content covers a wide range of topics, helping to broaden horizons and deepen understanding of complex issues. Follow Sergey Karelov's work to receive up-to-date information and analysis in the field of science and technology.

Achilles and the Tortoise: The Race with the USA

The United States is China's main competitor in the quest for leadership in the field of artificial intelligence. Sergey Karelov argues that in the near future, no country, including European states, India, and Russia, will be able to seriously compete with China. The only question that remains is how China and the United States will share the top positions in this technological race.

By 2030, China will likely become the second-largest power in artificial intelligence, although its gap to the leader will be minimal. The country is expected to actively develop AI technologies, which will allow it to significantly strengthen its position on the global stage.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He is the host of his own channel, "Little-Known Interesting," where he shares unique knowledge and analytical materials. On the channel, Sergey examines current topics, linking various fields of science and practice, allowing viewers to gain a deeper understanding of complex issues and modern trends. His approach to analysis and research helps develop critical thinking and expand horizons of understanding.

The course of this competition is reminiscent of the famous ancient Greek paradox "Achilles and the Tortoise." The Chinese Achilles persistently tries to catch the American tortoise, but, having reached the position where it was previously, it notices each time that its rival has managed to take several steps ahead, once again maintaining the lead in the race. This paradox illustrates the concept of an endless process, where even the fastest participants face difficulties in reaching the goal, which makes the competition even more exciting and intense.

The creation of the American language model GPT-4 by OpenAI and its innovative implementation in ChatGPT serve as a shining example of a significant breakthrough in the field of artificial intelligence. This step has allowed OpenAI to set new standards in natural language processing, leaving competitors significantly behind. GPT-4 not only improves the quality of user interactions but also opens up new possibilities for application in various fields, such as education, business, and scientific research.

Of course, I am ready to help you with text editing. Please provide the source text to be corrected.

The tortoise continues to move, running a few centimeters before Achilles reaches its starting position. However, it's worth noting that the gap between them is narrowing. China still lags behind the United States, but the speed at which it is catching up is faster than the pace at which the Americans are advancing. This observation highlights the dynamics of international competition and technological advances that influence the economic and strategic position of countries.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He is the host of his own channel, "Little-Known Interesting," where he shares unique knowledge and analysis of current topics. On the channel, Sergey explores various aspects that unite sciences and offers unconventional solutions to complex issues. His work helps to expand the horizons of understanding and opens up new perspectives in research.

The sphere of artificial Intelligence in China receives significant financial support from the government. In 2022, China ranked second in private investment in AI, amounting to $13.4 billion, while the United States led with $47.4 billion. This underscores the active development of AI technologies in China and the growing interest in this field from private investors.

In 2021, the size of the Chinese artificial intelligence market was $23.2 billion, and it is projected to grow to $60 billion by 2025. By the late 2020s, China will account for up to 30% of the total number of billion-dollar tech startups, known as unicorns. This data underscores China's importance in the global tech ecosystem and its potential for innovation and startups in the field of artificial intelligence.

The country is implementing nationwide initiatives aimed at developing local manufacturing, such as the "Made in China 2025" program. China is also actively recruiting highly qualified IT specialists through the "Thousand Talents Plan." A plan has also been developed to collect and store big data, which will facilitate advances in machine learning. These projects aim to strengthen the country's position in the global economy and develop innovative technologies. China has made significant progress in artificial intelligence, as evidenced by a number of formal indicators, such as the number of scientific publications and patent applications. Since 2018, the country has not only caught up with the United States but also taken a leading position globally. In 2018, China filed 30,000 patent applications in the field of AI, 2.5 times more than the US. Furthermore, Chinese scientists published 28% of the world's total papers on artificial intelligence. These data highlight China's rapid rise as a key player in AI technology development.

Despite significant quantitative advances, the quality of most research still leaves much to be desired. Many do not meet the high standards established by the best scientific papers.

China does indeed lead in the number of articles and patents. However, a closer analysis of this data reveals that the United States surpasses China in the quality of scientific research and development. This is confirmed by the high citation rates and influence of American publications on global science.

Analyzing the top publications, especially the top quartile, we see that the United States significantly outperforms China in the number of high-quality, in-demand, and cited scientific papers. If we consider only the top 10% of papers, it becomes clear that the United States dominates by a significant margin. This confirms the high standards of American science and research, as well as their influence on the international scientific community.

The United States is a leading country in the field of artificial intelligence, several years ahead of other countries, including China. However, China is gradually closing this gap and is increasingly developing its AI technologies every year. This process is creating a competitive environment that could impact global leadership in artificial intelligence.

Sergey Karelov is a specialist in interdisciplinary research and strategic analytics. He is the host of his own channel, "Little-Known Interesting," where he shares unique knowledge and analysis of current topics. In his content, Sergey strives to uncover non-obvious aspects and trends, allowing audiences to gain a deeper understanding of complex issues. Экспертный подход Сергея делает его контент ценным ресурсом для тех, кто интересуется последними исследованиями и аналитическими данными в различных областях.

Почему ChatGPT изобрели не в Китае

Изучив планы развития искусственного интеллекта, опубликованные в Китае до появления ChatGPT, можно заметить, что акцент в них делался не на генеративном интеллекте, а на других направлениях data science. В частности, в этих планах упоминаются системы, предназначенные для умных городов, технологии машинного зрения, распознавание голоса, развитие беспилотных транспортных средств, а также применения в медицине и образовании. Эти направления подчеркивают стратегический подход Китая к внедрению искусственного интеллекта в ключевые сферы жизни, что может существенно повлиять на дальнейшее развитие технологий и их интеграцию в повседневность.

Генеративный интеллект используется для создания текстового, визуального, аудио и видеоконтента по запросу пользователей. К числу известных платформ относятся ChatGPT, Claude, Bard и GigaChat, а также нейросети для генерации изображений, такие как Midjourney, DALL-E, Stable Diffusion, Kandinsky и «Шедеврум». Эти технологии активно развиваются и находят применение в различных сферах, включая маркетинг, искусство и образование, что делает их важными инструментами в современном цифровом мире.

Появление ChatGPT произвело значительные изменения в китайской технологической индустрии, вынудив местные компании пересмотреть свои стратегии и сосредоточиться на разработке собственных аналогов этого популярного чат-бота. Специалисты в области технологий отмечают, что ChatGPT кардинально изменил правила игры в секторе, и многие компании оказались неготовыми к таким быстрым трансформациям. Это создает новые вызовы и возможности для китайских разработчиков, стремящихся адаптироваться к современным реалиям и конкурировать на глобальном уровне.

В один миг все ранее обещающие технологии уступили место большим языковым моделям. Теперь компании стремятся успеть адаптироваться к новым условиям и интегрировать эти модели в свои процессы. Без разработки уникальных языковых решений они рискуют утратить конкурентные преимущества и оказаться на обочине цифровой экономики. Это может привести к серьезным последствиям для бизнеса, включая потерю рыночных позиций и, в крайних случаях, крах.

Фрагмент статьи в журнале «Китайский предприниматель» представляет собой анализ текущих тенденций в бизнес-среде Китая. В последние годы наблюдается рост интереса к китайскому рынку среди иностранных инвесторов. Это обусловлено не только экономическим развитием страны, но и улучшением бизнес-климата.

Ключевыми факторами, способствующими этому росту, являются инновационные технологии, поддержка со стороны правительства и расширение инфраструктуры. Китай становится центром для стартапов и высоких технологий, что открывает новые возможности для предпринимателей.

Важным аспектом является также развитие электронной коммерции, которая продолжает набирать популярность и трансформирует традиционные модели ведения бизнеса. Компании, ориентированные на онлайн-продажи, имеют большие шансы на успех в условиях быстро меняющегося рынка.

Таким образом, предприниматели, желающие выйти на китайский рынок, должны учитывать его динамичную природу и тенденции. Понимание местной культуры и бизнес-практик играет ключевую роль в достижении успеха.

Разработка OpenAI стала неожиданным триумфом, который жители Китая восприняли как серьезный вызов. Эти настроения ярко отражены в цитатах из статьи одного из популярных китайских бизнес-журналов: «Либо мы создадим свой конкурентный генеративный ИИ, либо окажемся в невыгодном положении» и «Мы должны разработать его, даже если, возможно, уже поздно!» Данные высказывания подчеркивают стремление Китая не отставать в области искусственного интеллекта и стремление к созданию собственных технологий, способных конкурировать на мировом уровне.

Chinese scientists are frustrated that their own AI developments, similar to ChatGPT, have long remained in the shadows. Previously, this field was not considered promising enough. In 2021, the Beijing Academy of Artificial Intelligence presented the Wu Dao 2.0 language model, which had a record number of parameters: 1.75 trillion. This number exceeds the parameters of GPT-3.5, which ChatGPT is based on, by a factor of ten. The development of such technologies in China opens new horizons and highlights the country's potential in the field of artificial intelligence.

A year before the launch of the American chatbot Wu Dao 2.0, there was a model capable of conducting dialogues in Chinese and English, as well as creating poems, recipes, and newspaper articles. This model was sent to Tsinghua University for training, where it acted as a virtual student named Hua Zhibin.

The Chinese project failed to achieve success and worldwide fame. Local scientists have not developed an innovative learning method similar to RLHF, which is used in the OpenAI chatbot. This limited the project's capabilities in comparison with its competitors in the international arena.

Sergey Karelov believes that it is not surprising that his Chinese colleagues did not notice the emergence of the new generative AI technology in the United States. He emphasizes that few people in the world were able to appreciate the colossal potential of this innovation in advance. It is important to note that generative AI opens up new opportunities for various industries, and its impact on the future of technology cannot be overestimated.

When the ChatGPT developers realized the scale of the breakthrough achieved with their model, they encountered unexpected emergent properties that were not originally intended. For about six months, they tried to comprehend the phenomenon they had created and understand its consequences. This discovery became an important milestone in the development of artificial intelligence and its application in various fields.

No one could have predicted such a breakthrough in the field of artificial intelligence. Therefore, Chinese companies cannot be blamed for failing to keep up with the latest trends. In fact, the entire world found itself in a similar situation and only later began actively developing in this direction. Nevertheless, Chinese specialists and enterprises quickly responded to the changes and began actively investing in AI technologies to close the gap.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He hosts his own channel, "Little-Known Interesting," where he shares unique knowledge and analysis of current topics. His work covers a wide range of topics, making his content valuable to viewers interested in deep and innovative approaches to studying various fields.

Read also:

The best alternatives to ChatGPT: an overview Bard, Claude, and GigaChat

In recent years, artificial intelligence technologies have rapidly developed, bringing new solutions for text creation and communication to the market. Some of the most popular alternatives to ChatGPT include Bard, Claude, and GigaChat. These tools offer unique features and capabilities, allowing users to choose the most suitable option for their needs.

Bard is an innovative platform designed with a focus on creativity and creative writing. It is ideal for writers and creative professionals seeking inspiration or help with content creation. Bard uses advanced natural language processing algorithms, allowing it to generate original ideas and stories.

Claude stands out for its ability to deeply analyze and understand context. This tool is suitable for more complex tasks, such as writing analytical reports or research papers. Claude can efficiently process large volumes of information and provide structured responses, making it indispensable for researchers and professionals.

GigaChat focuses on creating texts in a dialogue format, making it suitable for developing chatbots and customer support systems. Its capabilities enable natural conversations, ensuring a high level of interaction with customers. GigaChat also supports integration with various platforms, making it a convenient tool for businesses.

Each of these alternatives has its own unique advantages, and choosing the right solution depends on the specific needs of the user. Whether you require a creative approach, analytical capabilities, or a conversational system, Bard, Claude, and GigaChat offer high-quality text and communication tools.

China's Weaknesses in the AI ​​Race: Opinion of the Western Press

Analyzing the success of ChatGPT, Western observers put forward several reasons for the slow development of the Chinese tech sector. Their comments often highlight the contrast between the American model, based on "freedom and democracy," and the authoritarian regime in China, run by the Communist Party. These differences, according to analysts, affect the innovative potential and flexibility of Chinese technologies.

The development of significant technological products is closely linked to the system and the environment in which they operate. The lack of an open environment significantly complicates the creation of such products. China, for example, risks repeating the Soviet Union's experience, where state control became an obstacle to innovation. History demonstrates that state-led resource exploitation cannot compete with the free, natural development of the country. It is essential to create conditions conducive to innovation to ensure progress in the technology sector.

Xu Chenggang is a senior fellow at Stanford's Center for the Study of Chinese Economies and Institutions. His work focuses on the analysis of economic processes and institutional change in China. Xu conducts research that provides a deeper understanding of the dynamics of the Chinese economy and its impact on global markets. His expertise in the Chinese economy makes him a valuable resource for scholars, policymakers, and businesspeople seeking to understand the complexities of China's economy.

Experts revisited the topic of internet censorship in China. Some have expressed the view that generative artificial intelligence, with its so-called "hallucination" properties, will have difficulty complying with all Party directives. This raises important questions about information control and the capabilities of AI under strict censorship, a hot topic for research in technology and politics.

Journalists frequently cite a provocative quote from a Chinese social media post, allegedly written by an anonymous programmer: "We need to not only teach machines to speak, but also teach them not to speak." This phrase raises important questions about the development of artificial intelligence and its interaction with humanity. In the era of rapid technological progress, it is necessary to consider not only the possibilities, but also the limitations that can be useful for ensuring safety and ethics in the field of AI.

The unpredictability of chatbots, which sometimes generate meaningless or false statements, Known as hallucinations, it contradicts the Party's desire to control information online. Generative AI challenges two key Party goals: maintaining control over information and achieving leadership in artificial intelligence. This situation highlights the complexity of the current interaction between technology and information control, threatening plans to monopolize knowledge and manipulate public opinion.

Matt Sheehan is a recognized expert in the field of Chinese artificial intelligence. His research and analysis are instrumental in understanding current trends and prospects for AI development in China. Sheehan actively shares his knowledge, highlighting the impact of Chinese technology on the global market. His work is frequently cited in leading global publications such as The New York Times, underscoring his authority in this field.

Experts have noted that entrepreneurs developing communist ideas in China are unexpectedly more focused on quickly profiting from technology compared to their counterparts in capitalist America. This paradox highlights the differences in approaches to business and innovation, where Chinese companies strive to capitalize on their achievements as quickly as possible, while in the US, the emphasis may be on long-term strategies and sustainable development.

Competition between IT companies in China is significantly fiercer than in the US, which explains the cautious approach of Chinese investors to funding new technologies. In a highly competitive environment, companies prefer to invest in projects that are guaranteed to deliver an economic return. This leads to investments in technologies whose effectiveness is not clear to proceed slowly and cautiously.

Experts point out that OpenAI, the creator of ChatGPT, was originally founded by Elon Musk as a non-profit project with multibillion-dollar funding focused on achieving key breakthroughs in artificial intelligence. At the same time, the possibility of commercializing this project has long remained uncertain.

When the kid next door publishes a groundbreaking dissertation, one should not be surprised by their intelligence. You also have a smart child, but instead of supporting their education, you ask them to earn money while they can still use their physical abilities. It's important to understand that investing in your child's education and skill development can reap significantly greater dividends in the future. Supporting their studies and pursuing knowledge lay the foundation for a successful career and personal growth.

An anonymous post that went viral on Chinese social media attracted the attention of users and sparked active discussion. It raised important social issues reflecting current societal problems. The publication quickly spread, garnering numerous comments and reactions, demonstrating the significance of the topic. Viral posts like these highlight how social media can serve as a platform for sharing opinions and drawing attention to important issues.

Sergey Karelov notes that the Chinese governance model has problems and vulnerabilities, and, in his opinion, their number continues to grow. However, he suggests viewing these shortcomings as a side effect of its advantages. Effective management of large resources in China allows society to focus its efforts on achieving specific results. This approach demonstrates how the strengths of the system can create specific challenges that require attention and solutions.

The main advantage of any society is its ability to scale efforts and establish cooperation. This makes it possible to implement grandiose projects, from the Egyptian pyramids to the International Space Station. Effective cooperation and the pooling of resources play a key role in achieving large-scale goals, demonstrating how collective action can overcome any obstacles and achieve significant results.

The Chinese governance model contributes to the achievement of high results, especially in the practical sphere. It is in these aspects that China outstrips its competitors. Local entrepreneurs operate rationally, focusing on maximizing profits. This makes Chinese businesses more efficient and competitive in the global market.

China lags in fundamental research in artificial intelligence, especially when it comes to the work of individual inventors or small teams. These advances could be significant game-changers in technology and science. Despite significant investment and infrastructure development, in the area of ​​non-practical research, China cannot yet compete with leaders who are achieving breakthroughs in AI through innovative ideas and creativity.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He runs his own channel, "Little-Known Interesting," where he shares unique knowledge and analytical materials. On the channel, Sergey examines various aspects of scientific research, emphasizing its practical application and significance for modern society. His approach to analyzing and presenting information helps viewers gain a deeper understanding of complex concepts and expand their horizons.

Read also:

Chinese firewall: a mechanism for controlling the internal Internet in China

The Chinese firewall, also known as the Great Firewall, is a complex system of internet censorship and control. The main purpose of this system is to restrict access to certain resources and information that may be undesirable to the government. The firewall provides content filtering, blocking websites and applications that do not comply with official political and social norms.

The system operates using various technologies, including IP filtering, packet analysis, and DNS denial. It not only blocks access to foreign websites but also monitors the internet activity of users within the country. This creates certain difficulties for citizens wishing to access information not controlled by the government.

The Chinese firewall also impacts the business environment. Many international companies face restrictions on access to their websites and platforms, which hinders operations in the country. As a result, many of them are forced to adapt their strategies to comply with local rules and requirements.

Thus, the Great Firewall of China is not only a tool for information control, but also a major factor influencing the internet ecosystem in China. This system continues to evolve, and its impact on the country's society and economy remains significant.

Triple Deficiency: Experience, Hardware, and Data

Practical reasons hindering the development of artificial intelligence in China include a lack of experience, hardware issues, and a shortage of high-quality training data. This assessment is consistent with the views of the State Council of the People's Republic of China, which also emphasizes key areas of development in its documents: hardware, data, algorithms, and AI implementation. Successful advancement in the field of AI must overcome these barriers, which will improve the efficiency of the technology and its application in various fields.

Chinese universities graduate thousands of IT professionals every year. Однако в сфере генеративного искусственного интеллекта наблюдается нехватка квалифицированных кадров, так как это направление ранее не считалось особенно перспективным. Потребуется время, чтобы местные ученые приобрели необходимый опыт в разработке и обучении больших языковых моделей. Важно отметить, что с ростом интереса к генеративному ИИ в стране, количество учебных программ и исследований в этой области значительно увеличивается, что может способствовать улучшению ситуации в ближайшие годы.

В мире существует ограниченное количество высококвалифицированных специалистов в области языковых моделей (LLM). В Китае к разработке активно привлекаются ветераны, которые, казалось бы, завершили карьеру и написали мемуары. Одним из таких экспертов является 61-летний Ли Кай-Фу, значимая фигура в китайской ИТ-индустрии. Он анонсировал запуск своей платформы генеративного искусственного интеллекта под названием Project AI 2.0.

Конечно, я помогу вам с редактированием текста. Пожалуйста, предоставьте исходный текст, который вы хотите переработать.

В Китае и США существует дефицит специалистов с реальным опытом в разработке и исследовании масштабных моделей. В мире насчитывается всего около 100 высококлассных экспертов в данной области. Бигтех-компании и стартапы активно ищут этих специалистов, что привело к значительному росту их зарплат. В результате, нанять таких профессионалов становится крайне сложной задачей.

Чжоу Чжифэн – известный китайский IT-предприниматель и соучредитель компании Qiming Venture Partners, активно работающей в области венчурного капитала. Он стал значимой фигурой в технологическом секторе Китая и вдохновляет многих молодых предпринимателей. В интервью, опубликованном в журнале «Китайский предприниматель», Чжоу делится своим опытом и видением будущего индустрии технологий в стране. Его идеи и стратегии по развитию бизнеса остаются актуальными для стартапов и инвесторов, стремящихся к успеху в быстро меняющемся мире технологий.

В Соединенных Штатах не остаются безучастными к попыткам Китая обойти их в высоких технологиях. В контексте торговой войны правительство США внедрило ограничения на экспорт высокотехнологичных микрочипов, предназначенных для обработки больших данных и машинного обучения, а также на оборудование, необходимое для их производства. Эти меры направлены на защиту американских технологических достижений и поддержание конкурентоспособности на мировом рынке.

Сергей Карелов отмечает, что проблемы с поставками иностранных полупроводников являются ключевым фактором, способным существенно замедлить развитие искусственного интеллекта в Китае. Эти ограничения в поставках напрямую влияют на возможности китайских компаний внедрять современные технологии и улучшать свои разработки в области ИИ. В условиях глобальной конкуренции доступ к качественным полупроводникам становится критически важным для поддержания темпов роста и инноваций в данной сфере.

Будущее зависит от жёсткости санкционной политики США и от успехов Китая в налаживании собственного производства аппаратной базы для обучения искусственного интеллекта, comparable к самым современным продуктам NVIDIA, которые ожидаются к 2030 году. Китай готов вложить от 20 до 40 миллиардов долларов в эту сферу, однако в текущий момент сложно предсказать, насколько эти инвестиции окупятся и принесут ли ожидаемые результаты.

Сергей Карелов — эксперт в междисциплинарных исследованиях и стратегической аналитике, а также ведущий авторского канала «Малоизвестное интересное». Его работы охватывают широкий спектр тем, что позволяет глубже понять взаимосвязи между различными областями знаний. На канале Сергей делится уникальными исследованиями и аналитическими материалами, которые привлекают внимание зрителей и способствуют расширению их кругозора.

Современные модели, такие как GPT-4, требуют для своего обучения мощные вычислительные системы, включающие десятки тысяч высокопроизводительных видеопроцессоров (GPU) или нейропроцессоров (NPU). Наиболее совершенные из них, такие как NVIDIA A100 и H100, производятся по технологическим процессам 7 нм и 4 нм соответственно. Эти процессоры были включены в список товаров, запрещённых к экспорту в Китай, что создает сложности для разработки и внедрения передовых технологий в этой стране.

Sergey Karelov emphasizes that training models for the new generation of processors requires significantly more resources than previous versions. While the US is confident that its hardware advances will give it a competitive advantage for the next two to three years, the situation in China is less optimistic. This highlights the importance of investing in technology and R&D to maintain market leadership in processors.

Despite statements by Chinese scientists about the creation of a breakthrough technology for the production of chips with an accuracy of 1 nm or less, in practice the main chip manufacturer in the country, the Shanghai-based SMIC, is still China is facing difficulties in mass production of processors using 7-nm technology. This technology is inferior to modern imported samples, which already use 4-nm and 3-nm processes. Thus, despite ambitious plans, the Chinese semiconductor industry has not yet achieved the desired level of competitiveness in the global market. Local entrepreneurs remain skeptical about media reports that China will soon begin mass production of chips comparable in quality to foreign counterparts. As a result, they prefer to purchase imported chips as a reserve. This cautious attitude towards Chinese technology is due to the desire to ensure stability and reliability in their business processes. The situation in the semiconductor market remains tense, and therefore many companies choose proven solutions to minimize risks. The price of the A100 chip has increased by more than 50% since the end of last year, reaching 100,000 yuan from the previous 50,000 yuan. According to our data, only 30,000 units of these processors remain in stock. The US ban on A100 sales to China has forced businesses to seek alternatives, resulting in high demand for less powerful NVIDIA A800 chips, which are not subject to restrictions. However, even these chips are often out of stock.

Chen Wei, Chairman of Qi An Xin Technology, noted the importance of innovative technologies for business development in the modern world. In an interview with Chinese Entrepreneur magazine, he emphasized that the company's success depends on the constant implementation of new solutions and adaptation to changing market conditions. Qi An Xin Technology is actively working on cybersecurity projects, which helps build trust with customers and partners. Wei also emphasized the need to invest in research and development to remain competitive in the high-tech sector.

The graph shows the impact of US trade restrictions on the development of the Chinese semiconductor industry. American company Intel, South Korean Samsung, and Taiwanese TSMC have successfully mastered the mass production of microchips using advanced 4- and 3-nanometer process technologies. At the same time, the Chinese company SMIC continues to remain at the 7 nanometer level, which indicates a slowdown in its technological progress and limited capabilities in the face of fierce competition.

Rework the text, maintaining the main topic and avoiding unnecessary elements. Optimize the content for SEO and add additional information where necessary. Avoid using emojis and unnecessary characters. Do not include structured sections such as 1, 2, 3, or *. Please provide clear text.

Read also:

The Russian company Baikal Electronics plans to begin production of processors designed for artificial intelligence. This step is aimed at developing domestic technology and reducing dependence on foreign manufacturers. The new processors will be optimized for machine learning and deep learning algorithms, which will significantly improve the performance and efficiency of AI applications. Baikal Electronics aims to take a leading position in the market by offering competitive solutions for various sectors, including industry, healthcare, and fintech. Developing indigenous AI processors will be a key step in creating a high-tech infrastructure in Russia.

The challenges of training data for artificial intelligence are complex and multifaceted. China has long been a leader in big data collection. In his book, Lee Kai-Fu asserts, "If data is the new oil, then China is the new Saudi Arabia." This comparison underscores the importance of data in the modern technological world and shows how it is becoming the foundation for AI development. However, despite the abundance of data, questions arise about its quality, accessibility, and ethics, making the topic of AI training even more pressing and requiring in-depth analysis.

The assertion of a high level of control in China is due not only to the centralized policy of collecting citizens' digital traces, but also to the size of the population, the widespread use of internet-connected devices, and the widespread implementation of video surveillance systems. This combination of factors makes China one of the countries with the most advanced system of monitoring and analyzing data on citizens.

Check out additional materials:

A digital footprint is the information we leave on the Internet during our activities. Every online activity, whether visiting websites, posting on social media, or using apps, creates a unique user profile. Understanding your digital footprint is important because it can impact your reputation, security, and privacy. There are two main types of digital footprints: active and passive. An active digital footprint includes information you intentionally create, such as messages, photos, and comments. A passive digital footprint consists of data collected without your input, such as cookies and browsing logs. Managing your digital footprint is becoming an important aspect of online security. It's recommended to regularly review your privacy settings on social media, use tools to remove irrelevant information, and be mindful of what you post. Building a positive digital footprint can help advance your career and enhance your personal brand. Remember that everything you do online leaves a trace, and it's important to monitor how that trace is formed.

The emergence of large language models (LLMs) has challenged China's dominance in artificial intelligence. The main challenge is that LLMs require high-quality text datasets. However, such datasets are in short supply in any language other than English. This limitation could impact the development of Chinese technologies in natural language processing and artificial intelligence.

The Chinese language occupies a unique position in the global context. Currently, approximately 56% of the world's websites are in English, while only 1.5% of content is in Chinese. This underscores the dominance of English as the international language of communication and science. More than 90% of scientific publications, including significant research by Chinese scientists, are also written in English. Thus, English remains the primary means of communicating knowledge and information globally, while Chinese, despite its importance, has not yet achieved the same level of adoption online or in academia.

Developers of Chinese language models often face the need to train their algorithms using low-quality texts. These are primarily dialogues extracted from instant messaging apps and social media. Such texts may contain errors and demonstrate a disregard for literary norms, which hinders the training process and reduces the overall effectiveness of the models. It is important to consider that the use of high-quality texts helps to improve the accuracy and adequacy of language models.

China is experiencing a shortage of high-quality Texts for training, which has become a source of complaint among local researchers in this field. This is a difference between Chinese laboratories and their Western counterparts. When attempting to use high-quality academic literature, it becomes obvious that most of these materials are written in English.

Jeff Dean is an associate professor at George Washington University and the founder of ChinAI. His research and projects in artificial intelligence and machine learning play a significant role in the development of technologies and their application in various fields. Jeff Dean actively participates in scientific discussions and shares his knowledge, which contributes to a broader understanding of artificial intelligence as an important tool for solving modern problems.

Sergey Karelov considers this problem insignificant.

Big data technologies in China have reached a level comparable to American ones, and in some respects even surpass them. Developing machine learning methods are becoming less dependent on data volume, which opens up new possibilities for information analysis and processing. This allows Chinese companies to effectively use big data to optimize processes and decision-making.

When neural networks were given the opportunity to play chess with each other, they played more games in a single day than had been recorded in the entire history of mankind. This fact confirmed that practical experience is not so critical. This is just one of many examples. Over time, in various fields of artificial intelligence, the importance of natural data is declining, and synthetic data is becoming increasingly relevant. Synthetic data allows for the modeling of various scenarios and the training of algorithms, opening up new horizons for technological development.

English continues to be the dominant language in training large-scale language models (LLMs). However, Chinese researchers recently presented a study demonstrating that their Chinese-language models achieve results comparable to leading versions of GPTs. This demonstrates the growing influence of Chinese in the field of artificial intelligence and highlights the importance of multilingual learning to achieve more diverse and effective LLM outcomes.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics, and the host of his own channel, "Little-Known Interesting." His work covers a wide range of topics, allowing for a deeper understanding of the interconnections between various fields of knowledge. With his unique approach and analytical skills, Sergey provides valuable information and engaging ideas that stimulate discussion and the development of new concepts. The channel "Little-known interesting" is becoming a platform for the exchange of knowledge and ideas, helping to expand the horizons of its audience.

Read also:

Data: the new oil or just another hype?

In the modern world, data is becoming one of the most valuable assets comparable to oil. With the rise of digitalization and the increase in information volumes, companies and organizations understand the importance of collecting, analyzing, and using data to make informed decisions. However, the question arises: is data truly the new fuel for business, or is it just a trendy phrase without any real content?

On the one hand, data allows companies to better understand the needs of their customers, optimize business processes, and increase efficiency. Based on data analytics, it is possible to predict trends and develop innovative products. On the other hand, there are many tools and technologies on the market that promise wonderful results, but do not always live up to expectations.

It is important to note that for data to become a useful resource, it must be well-managed and analyzed. Without the right structure and strategic approach, data may be useless. Therefore, companies looking to use data to their advantage must invest in qualified people and modern technologies.

Thus, data can truly become the new "gold" for business, but only if it is used competently and an analytical approach is taken. This concept shouldn't be oversimplified—data requires serious attention and resources to deliver real value.

China Leads the Way in AI Regulation

Despite experts' assessments of China's technological lag, they unanimously acknowledge the country's leadership in regulating artificial intelligence. Chinese lawmakers have become the first in the world to recognize both the potential of generative AI and the serious risks associated with its use. This allows China not only to set AI standards but also to actively shape the global agenda for the safe and ethical use of technology.

In the United States, where disruptive technologies are actively developing, authorities seem to be ignoring existing problems. Jay Obernolte, the only member of Congress with a master's degree in artificial intelligence, expressed concern that most lawmakers do not understand what artificial intelligence is. This highlights the need for increased awareness and education among legislators about modern technologies, which are becoming increasingly important for the economy and society.

Before introducing regulations, it is important to reach agreement on the existing threats, which requires a deep understanding of artificial intelligence. Many are surprised by how much time it takes to explain to colleagues that the primary danger of AI does not come from images of evil robots with lasers. In fact, threats may lie in other aspects, such as misuse of technology, lack of transparency in algorithms, and potential ethical violations. Understanding these risks will help develop more effective strategies for regulating and safely deploying artificial intelligence.

Jay Obernolte, a representative of California in the US House of Representatives, is actively involved in the country's political life. His activities cover a wide range of issues, including economic development and social initiatives. Obernolte strives to improve the living conditions of his constituents and actively works on legislative initiatives that promote business growth and support local communities. His statements and actions are frequently covered in the media, including The New York Times, underscoring his importance in the political arena.

At the UN Security Council meeting on artificial intelligence, which took place on July 18, 2023, representatives of China demonstrated a coordinated state position on this issue, unlike their Western counterparts. Генеральный секретарь ООН Антониу Гутерриш также выразил свою неуверенность в том, какие последствия новая технология может иметь для человечества.

Практически каждое правительство и крупная корпорация на планете активно разрабатывают стратегии в области искусственного интеллекта (ИИ). Однако даже ведущие эксперты и разработчики технологий зачастую не способны предсказать, какие последствия может иметь этот революционный прорыв. Искусственный интеллект продолжает изменять бизнес-процессы, улучшать эффективность и создавать новые возможности, но его влияние на общество и экономику остается предметом обсуждения и анализа. Важно учитывать возможные риски и этические аспекты внедрения ИИ, чтобы обеспечить его безопасное и ответственное использование в будущем.

Антониу Гутерриш, Генеральный секретарь Организации Объединенных Наций, в своих выступлениях часто подчеркивает важность глобального сотрудничества и солидарности в решении мировых проблем. В его цитатах можно найти акценты на необходимости совместных действий для достижения устойчивого развития, защиты прав человека и обеспечения мира. Гутерриш настоятельно призывает международное сообщество к более активному вовлечению в процессы, направленные на решение климатических изменений и гуманитарных crises. Важность его слов заключается в том, что они вдохновляют на действия и напоминают о необходимости ответственного подхода к управлению планетой.

Китай стал первой страной в мире, которая на законодательном уровне приняла специальные правила, регулирующие работу генеративного искусственного интеллекта. Документ, известный как «Временные правила регулирования услуг генеративного искусственного интеллекта», был разработан китайскими властями за рекордно короткий срок — всего за шесть месяцев. В то время как чиновники других государств продолжали обсуждения, Китай оперативно внедрил необходимые нормативные акты, что подчеркивает его стремление к лидерству в области технологий и искусственного интеллекта. Эти правила направлены на установление стандартов и обеспечение безопасного использования генеративного ИИ, что может значительно повлиять на развитие технологий как в Китае, так и за его пределами.

Предписание из 24 пунктов, вступившее в силу 15 августа 2023 года, вводит важные требования к созданию контента с использованием искусственного интеллекта. Одним из ключевых аспектов является обязательная маркировка контента, сгенерированного ИИ, чтобы избежать путаницы у аудитории относительно подлинного автора произведения. Кроме того, правила требуют, чтобы разработчики ИИ использовали лишь данные из легальных источников, соблюдая права на интеллектуальную собственность. В случае обработки личной информации китайские разработчики обязаны получать согласие пользователей на её использование. Эти меры направлены на повышение прозрачности и защиту прав потребителей в сфере цифрового контента.

Компании, предлагающие доступ к генеративному интеллекту, обязаны обеспечивать соответствие создаваемого контента моральным нормам и исключать материалы, которые могут дискриминировать людей. Важно, чтобы системы искусственного интеллекта, разработанные в Китае, отражали патриотические ценности и поддерживали идеалы, провозглашаемые Коммунистической партией. Это требование способствует формированию ответственного подхода к использованию технологий и защите общественных интересов.

При использовании генеративного ИИ необходимо соблюдать ключевые социалистические ценности. Искусственный интеллект не должен способствовать подрыву государственной власти или свержению социалистической системы. Важно, чтобы технологии не угрожали национальной безопасности и интересам, не наносили ущерба имиджу страны и не провоцировали отделение регионов, подрывали национальное единство и социальную стабильность. Генеративный ИИ должен избегать поощрения терроризма и экстремизма, а также недопустимой пропаганды межнациональной ненависти, этнической дискриминации и насилия. Особое внимание стоит уделять предотвращению распространения ложной, вредной информации и другого запрещенного контента.

The Interim Rules for Generative Artificial Intelligence Services were published to ensure the safe and responsible use of AI technologies. These rules aim to create a transparent and secure environment for users and developers, minimizing the risks associated with the use of generative AI. The rules emphasize the need to comply with ethical standards and protect personal data, as well as the responsibility of developers for the quality and security of the services they provide. Regulation in this area is an important step towards creating a sustainable AI ecosystem that fosters innovation and protects user interests.

Sergey Karelov believes that the introduction of new rules for regulating artificial intelligence is a significant step forward. He disagrees with the opinion of some foreign experts who fear that excessive prohibitions could negatively impact the development of Chinese AI models. At the same time, Karelov notes that approaches to regulating artificial intelligence in China and Western countries differ significantly. This difference in approach can impact the effectiveness and innovativeness of technologies, which is important to consider when analyzing the future of artificial intelligence in a global context.

In terms of regulation and policy covering the interaction of society, government, and business, China is significantly ahead of other countries. Their regulation on generative artificial intelligence is the best example, surpassing legislative initiatives in countries like the United States. This makes China a leader in technology regulation, which can influence global standards and practices in the field of artificial intelligence.

China and the United States have different regulatory priorities. China emphasizes the role of the state, which actively manages the economy and ensures control over key sectors. In turn, the United States is focused on regulating large corporations, striving to create the conditions for their effective functioning and competitiveness on the global stage. These differences in approach influence the economic strategies of both countries and shape their interactions on the international platform.

At a UN meeting, China stated that artificial intelligence has reached a level that poses an existential threat. This statement underscores the need for oversight over AI to prevent potential negative scenarios. It's important to note that as technology advances, the risk of misuse arises, making AI governance a priority for the international community. Ensuring safety and ethics in artificial intelligence must form the foundation for the continued progress and development of this technology.

The rules were adopted quickly, allowing for rapid approval by businesses and society in China. During the revision process, all ambiguities that raised concerns in the first version were resolved.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He runs his own channel, "Little-Known Interesting," where he shares unique knowledge and analytical materials. The channel covers a wide range of topics, including current trends and unexpected facts, making it a valuable resource for anyone interested in deep analysis and new approaches in various fields.

Redesigned text:

Explore further resources:

Neural Networks and Copyright: Cases, Limitations, and Uncertainties

Neural networks are increasingly becoming an important tool in content creation, raising questions about the legal aspects of copyright. Existing court cases highlight the complexity of protecting intellectual property in an environment where artificial intelligence can generate original works. The main limitations stem from the fact that copyright traditionally protects human-created works. This creates a gray area where the rights to content generated by neural networks remain unclear.

Courts in this area have already begun, and they may set precedents for the future. Importantly, many jurisdictions have not yet decided how to take into account the contribution of neural networks to the creation process. This creates uncertainty for both technology users and creators, who may face problems protecting their rights.

An important aspect is the need to adapt legislation to new technologies. Authors and developers of neural networks must consider the potential risks and limitations associated with copyright to avoid legal consequences. Discussion and analysis of current court cases will help better understand how neural networks affect copyright and what steps should be taken to protect the interests of all parties.

The European Union is also actively developing its own regulations on artificial intelligence. The AI ​​Law includes provisions similar to Chinese legislation, in particular regarding mandatory labeling of content created using AI. Among the requirements, one can also highlight more unusual ones, such as a ban on the development of systems capable of influencing the subconscious of users. These measures are aimed at protecting the rights of citizens and ensuring transparency in the use of AI technologies.

"Hurricane of Models" from the Middle Kingdom

China responded to the ChatGPT challenge with a multitude of new models. Instead of developing a single competitor to the American neural network, Chinese companies presented more than 130 different generative models. This demonstrates Chinese businesses' commitment to innovation and diversity in the field of artificial intelligence. The development of multiple variants allows for more efficient adaptation of technologies to the diverse needs of users and industries, which could significantly improve the quality and availability of generative content in the region.

Among the presented models, at least 79 have over a billion parameters. Their primary application is text generation, but there are also solutions capable of creating images, video, and audio, as well as multimodal models that combine multiple content types. This progress is impressive, especially when compared with 2020, when Chinese developers presented only two neural networks of this level. Such achievements highlight the rapid development of artificial intelligence technologies and their potential in various fields.

Sergey Karelov notes that the large number of neural network projects is not surprising. More than 6,500 companies working in artificial intelligence are operating in China, accounting for approximately 25% of the global total. Many of these companies actively respond to initiatives emanating from the Party and government, which promotes technological development and the introduction of innovation in various fields.

In recent years, approximately 2,500 new enterprises in various industries have been registered in China annually. This process is part of a strategic approach to solving global problems, which is consistent with the Chinese philosophy and governance system. China seeks to suppress competition by using the scale of its initiatives and resources. This allows the country to maintain leadership in the global market and actively develop innovative technologies, which in turn contributes to economic growth and improves the quality of life of the population.

When faced with a large, complex, and significant task, a team joins forces to complete it. This task is given high priority and provided with the necessary resources, both financial and ideological.

This is a key difference from our traditional approach to such issues. In a similar situation, we could create a single AI corporation and focus on promoting it. However, in this case, they choose to distribute resources while acting within the framework of an overall development strategy. This approach allows them to adapt and effectively develop artificial intelligence technologies, ensuring broader reach and diversity of initiatives.

Sergey Karelov is an expert in interdisciplinary research and strategic analytics. He is the host of his own channel, "Little-Known Interesting," where he shares unique knowledge and analytical materials. His work covers a wide range of topics, allowing viewers to expand their horizons and gain new insights in various fields. "Little-Known Interesting" offers in-depth analysis and interesting facts, making it a valuable resource for anyone seeking self-improvement and new knowledge.

To help navigate the diversity of local IT companies, observers distinguish two categories: "big dragons" and "little dragons." The big dragons include three leading companies, collectively known as BAT—Baidu, Alibaba Group, and Tencent. These giants account for approximately half of all private investment in AI in China, underscoring their significant influence on the technology and innovation market.

The second, larger group of "small dragons" includes companies actively developing AI technologies and cloud solutions. These include such significant players as SenseTime, Megvii, Huawei, CloudWalk Technology, YITU Technology, JD.com Incorporation, iFlytek Corporation, and Pony.ai. These companies are making significant contributions to the development of technology and innovation in China and internationally. Inspur and other members of this group are also demonstrating impressive results in high technology and digital transformation.

In China, increasing attention is being paid to the development of Baidu's ERNIE Bot chatbot, which has become a symbol of hope for the development of artificial intelligence in the country. The name ERNIE stands for Enhanced Representation through kNowledge Integration. This service, introduced in March 2023, became the first Chinese alternative to ChatGPT, highlighting China's commitment to embedding advanced technologies into everyday life and business. ERNIE Bot promises new capabilities in natural language processing and user interaction, which could significantly impact the chatbot and AI market as a whole.

On June 21, 2023, the Baidu team presented an updated version of its ERNIE neural network model, which received an index of 3.5, which corresponds to the index of the GPT model used in ChatGPT. The new version of ERNIE not only received an interface similar to ChatGPT, but also implemented a plugin system that allows the model to interact with external services and data sources. These updates make ERNIE a more functional tool for processing information and interacting with users.

The current version of ERNIE demonstrates the ability to perform most of the functions similar to those offered by the original ChatGPT, in Chinese and English. The developers claim that when interacting in Chinese, ERNIE outperforms the foreign model in a number of respects, although it is somewhat inferior when working with English.

Of course, I am ready to help you with text editing. Please provide the text itself that you want to change.

The leaders in developing their own versions of ChatGPT in China will be companies that have already created a base for developing large-scale models similar to GPT-3. Currently, Baidu, Huawei, Inspur, and Tencent are actively working on creating such models and are likely already close to the level of American developers. Competition in this field is intensifying, and Chinese companies are rapidly advancing, which may lead to significant innovations in the field of artificial intelligence.

Jeff Dean, an associate professor at George Washington University and founder of the ChinAI initiative, is a prominent expert in the field of artificial intelligence and machine learning. His research and development in this field significantly influences modern technologies. Jeff actively shares his knowledge and experience, contributing to the development of AI and its application in various fields.

Given the large number of models being developed in China, local companies are actively working on integrated platforms that provide access to multiple neural networks simultaneously. One such platform is Qianfan, offered by Baidu. This platform allows users to effectively utilize the resources of various neural networks, optimizing the processes of developing and implementing artificial intelligence in various fields.

Some neural networks from the "model hurricane"

• Tongyi Qianwen from Alibaba Group: based on the in-house developed models Qwen-7B and Qwen-7B-Chat with a size of 7 billion parameters, designed for conversational applications.
• Hunyuan from Tencent: has more than 100 billion parameters, trained using two trillion tokens. Tencent claims that the model outperforms ChatGPT in writing long texts and solving some mathematical problems, and also generates 30% fewer hallucinations than LLaMA-2.
• SenseChat by SenseTime: part of the proprietary SenseNova platform, which, in addition to the chatbot, includes an image generator, tools for creating digital avatars and 3D modeling.
• PanGu Chat by Huawei: based on its own PanGu 3.0 language model, which has versions from 10 to 200 billion parameters. The service is based on the company's own hardware - Ascend neurochips.
• ChatRhino by JD.com Incorporation: contains 100 billion parameters. Designed to solve text problems in the fields of trade, logistics, finance, and healthcare.
• SparkDesk by iFlytek Corporation: under the hood are more than 12,000 AI assistants focused on various use cases in marketing, creativity, programming, communication, and education. Built on a proprietary model called Spark.
• Tako by ByteDance: an LLM chatbot from the creators of TikTok, aimed at use within this social network. Currently available only to some users in test mode. Possibly part of a project called Grace, being tested among the company's employees.

• PanGu-Coder 2 by Huawei: designed to generate Python code from a standardized description of a task in natural language. For now, it can only "understand" tasks in English (Chinese language support is promised later).
• iFlyCode by iFlytek Corporation: an intelligent assistant for programmers in C++, Python, and Java. It can both generate code from scratch and complement what is written by a person. Based on the proprietary Spark V2.0 model.
• CodeWave by NetEase: developed by the Chinese video game manufacturer allows corporate clients to create programs without requiring extensive programming skills. Available with a paid subscription.

• ERNIE-ViLG by Baidu: a Chinese open-source analogue of Midjourney, trained on a dataset of 145 million image-text pairs and contains 10 billion parameters. It is part of the company's AI platform called Wenxin, where it is responsible for converting text into images.
• Tongyi Wanxiang from Alibaba Group: developed on the basis of its own image generation model Composer, it understands text prompts in Chinese and English, and can create images in various styles, from watercolors, oils, and Chinese painting to animation, sketches, and 3D graphics.

• Zidong Taichu 2.0 from the Institute of Automation of the Chinese Academy of Sciences: a model that can work with various types of content, including text, audio, images, videos, and 3D models.
• Qwen-VL from Alibaba Group: a multimodal variant of the Chinese open-source model Qwen-7B, capable of working with images (VL stands for vision language). Помимо поддержания диалогов на китайском и английском, способна сравнивать несколько изображений и отвечать на вопросы, относящиеся к картинкам.

Подытожим

С учетом стремительного развития искусственного интеллекта в Китае, с каждым годом мы будем наблюдать всё больше новостей о достижениях этой страны в данной области. Независимо от того, удастся ли Китаю обойти США к 2030 году, он на длительный срок останется единственным государством, способным соперничать с Америкой на равных условиях в сфере ИИ.

Завершим наш материал цитатой Илона Маска, одного из основателей компании OpenAI. Его слова могут оказаться предвестниками грядущих изменений в сфере искусственного интеллекта.

Я впечатлён мудростью и решимостью китайского народа. Когда китайцы ставят перед собой цель достичь успеха в какой-либо области, они неизменно добиваются этого. Аналогично, в сфере искусственного интеллекта китайцы способны на значительные достижения.

Илон Маск — основатель компании OpenAI, известной своими достижениями в области искусственного интеллекта. Его видение и инициативы в этой сфере оказали значительное влияние на развитие технологий и их применение в различных отраслях. Маск активно поддерживает исследования и разработку безопасного и этичного ИИ, что делает его одной из ключевых фигур в современном технологическом мире.

Переделайте текст, но оставайтесь в рамках темы. Не добавляйте ничего лишнего. Подкорректируйте текст для SEO и можете дополнить содержание. Не добавляйте смайлы или лишние символы. Избегайте разделов в виде списков. Просто предоставьте чистый текст. Читайте также.

• Новый уровень искусственного интеллекта: что такое AGI, когда он появится и каким будет
• «Худший шторм». Как китайские EdTech-компании справляются с ограничениями спустя полгода?
• Сложен ли китайский? «Если есть такая задача, освоить язык сможет любой»
• 30 мощных нейросетей для любых задач