The Future is Here: Key Themes at Wuzhen World Internet Conference - AI, AIGC, and Computing Power
-
On November 10, the 2023 Wuzhen Summit of the World Internet Conference officially concluded, marking its tenth anniversary.
Looking back at past conferences, internet leaders from companies like Alibaba, Tencent, Baidu, and ByteDance have often been a special group at summits. Their keynote speeches, filled with 'golden quotes,' and their off-stage gatherings have drawn significant attention. People eagerly anticipate piecing together the future of the internet from their words.
With the passage of time, the internet landscape has seen its rise and fall. Many well-known first-generation internet figures have gradually retreated to the background. Today, their successors stand at the forefront, shaping the industry, yet they lack the same level of widespread public recognition.
At this year's World Internet Conference, only a few veteran faces from domestic internet companies, such as Robin Li, Zhou Hongyi, and Zhang Chaoyang, remained active on the front lines. The discussions continued to focus on hot topics like artificial intelligence, computing power networks, and the digital economy, which remained the key themes of the event.
It is reported that the 2023 World Internet Conference Wuzhen Summit will be held in Wuzhen, Zhejiang from November 8 to 10, with the theme 'Building an Inclusive, Beneficial, and Resilient Digital World – Working Together to Build a Community with a Shared Future in Cyberspace'.
Under this theme, more than 1,800 guests from 126 countries and regions interacted and exchanged ideas. The topics not only revolved around grand themes like digital communication and cultural exchange but also included discussions on current hot topics such as artificial intelligence, e-commerce, computing power networks, and the digital economy.
With the emergence of large models and the rise of AIGC (Generative Artificial Intelligence), 'Artificial Intelligence' naturally became the most frequently mentioned buzzword at this year's Wuzhen Summit.
As one of China's earliest entrants into the AI field, Baidu's founder, chairman, and CEO Robin Li publicly disclosed last year that the company has invested up to 100 billion yuan in artificial intelligence research and development over the past ten years.
At the Internet Entrepreneurs Forum held on November 9th, Robin Li pointed out that the hallmark of humanity entering the AI era is the emergence of a large number of AI-native applications, rather than the proliferation of large models. He believes that for entrepreneurs, 'competing' in large models is 'reinventing the wheel,' stating that 'many companies are still unnecessarily purchasing hardware and conducting training.'
Robin Li believes that in the AI era, super applications are where the opportunities lie. This is analogous to the mobile internet era, where there were only two operating systems, Android and iOS, but many successful applications emerged.
He stated that in the mobile era, China relied on its strong market advantages to produce numerous super applications such as food delivery and ride-hailing. In the AI era, Chinese developers will undoubtedly be able to create many heavyweight applications. The AI model ecosystem built on foundational models is the basis for prosperity, and a thriving AI-native application ecosystem will drive a new round of economic growth.
According to reports, over the past few months, Baidu has been restructuring its various product lines with AI applications. For example, the new search service no longer just provides links but offers services through generated images and copywriting.
At the plenary session of the summit opening ceremony, Alibaba Group CEO Wu Yongming frequently mentioned artificial intelligence, predicting that 'the intelligent era has just begun, with all changes still in their very early stages, presenting enormous development opportunities.'
Related topics like AIGC were also widely discussed during this conference.
Charles Zhang, founder, chairman, and CEO of Sohu, believes that AIGC is developing rapidly, but it should not be underestimated in the next ten years nor overestimated in the next one or two years. In the long term, he maintains an optimistic attitude, while in the short term, there's no need for anxiety. "At the very least, I still think video is more important than AIGC right now," he said.
Zhang Chaoyang also reminded practitioners, "If your financial strength is limited and cash flow is not ideal, it might be risky to invest all your funds into seizing AIGC opportunities. You still need to keep your feet firmly on the ground."
China Literature Group, with its extensive collection of authors and works, has positioned AI+IP as a strategic focus. In July, the company unveiled its self-developed large-scale model for the online literature industry, named 'Yuewen Miaobi,' which is primarily applied in areas such as creative assistance, IP development, and the globalization of online literature.
During the Wuzhen Summit, Hou Xiaonan, CEO of China Literature, stated that AI translation is overcoming capacity and cost limitations, demonstrating excellent performance in both quality and efficiency. In his view, AIGC is not an independent industry but a technological foundation that will foster a thriving application ecosystem.
In the field of large models, Tencent naturally wouldn't be absent. On September 7, 2023, the company officially unveiled its Hunyuan large model to the public.
According to Jiang Jie, Vice President of Tencent Group, the company's Hunyuan large model is currently undergoing internal testing across more than 180 business units. Capabilities such as "text-to-image" and "text-to-video" are significantly improving creative efficiency in various design and advertising material productions.
"The development of artificial intelligence is driven by computing power as its core." At the recent Huawei Connect conference, Meng Wanzhou, Huawei's Vice Chairman, Rotating Chairman, and CFO, stated that large models require substantial computing power. The scale of computing power determines the speed of AI iteration and innovation, as well as the pace of economic development. The scarcity and high cost of computing power have become the core factors constraining AI development.
According to public information, computing power is a new form of productivity that integrates information processing capability, network transmission capacity, and data storage capacity. Many fields are related to the processing power of computing, including scientific research, data analysis, artificial intelligence, game development, and more.
Therefore, beyond AI and AIGC, the topic of computing power has become a focal point of discussion at the internet conference. On the morning of November 9, the 'Computing Power Network Collaborative Innovation Forum' was held at the Wuzhen Cloud Boat Guest Center, where industry leaders gathered to envision the future of computing power networks in the AI era.
At the meeting, Liang Baojun, Vice President of China Unicom, mentioned that computing power is playing an increasingly significant role in promoting economic development. "The Global Computing Power Index Evaluation Report shows that for every one-point increase in the computing power index, it can contribute 3.5‰ to the digital economy and drive 1.8‰ growth in GDP."
Yang Jie, Chairman of China Mobile, stated that there is currently an explosive growth in the common demand for computing power across all sectors of society. It is projected that the total global computing power will grow at an annual rate of over 50% in the next five years.
Yang Jie believes that to further unleash the potential of AI, it is necessary to promote deep integration of industry, academia, research, and application, strengthen the collaborative innovation of high-value data, high-performance computing, and high-quality algorithms, accelerate breakthroughs in key technologies and the implementation of industrial applications, so that AI can not only 'compose poems' but also 'get things done'.
However, due to NVIDIA's monopoly on high-end GPU chips, computing power shortages are constraining the development of some AI companies. According to public data, NVIDIA holds 80% of the discrete GPU market share and up to 90% of the high-end GPU market share.
To alleviate computing power anxiety, some participants proposed solutions. Hou Shengli, Senior Vice President and Chief Technology Officer of Cisco Greater China, mentioned that the impact of AI on computing power networks is mainly reflected in enhancing computing capabilities, requiring faster chip speeds to meet demand; and increasing network bandwidth requirements, evolving from 10G and 100G to 400G and 800G.
Meanwhile, against the backdrop of China's large model competition, the intelligent computing power scale shows tremendous growth potential. According to IDC estimates, China's intelligent computing power scale is expected to achieve a compound annual growth rate (CAGR) of 52.3% from 2021 to 2026.
Wu Yongming, from the perspective of a cloud computing practitioner, points out that the entire industry is undergoing significant changes. The importance of AI computing is surpassing traditional computing, and model-driven AI computing will soon become the cornerstone of the digital world, eventually taking over traditional CPU-dominated computing resources.
Cybersecurity is another 'key term' at the Wuzhen Summit.
On November 8th, Zhou Hongyi, founder of 360 Group, posted on Weibo that he was stopped by several confident Gen Z university students in Wuzhen who handed him resumes for information security-related positions. "My first reaction was to ask if they had included their phone numbers on the resumes," Zhou said. "I've often received resumes from very talented people before, but they usually forget to include contact information."
Beyond chance encounters with resume submissions, Zhou Hongyi naturally discussed cybersecurity-related topics at the conference. He stated that the future of security should be a new type of digital infrastructure, serving as a productive service industry to ensure the high-quality development of industrial digitization.
"Today, we are expanding Software as a Service to 'Security as a Service', turning security into infrastructure and public service platforms. This allows cities and enterprises to save costs and effort while achieving tangible security results," said Zhou Hongyi.
Looking ahead, Zhou Hongyi believes that in the next decade, the security industry must achieve self-breakthroughs and upgrades, developing security into new digital infrastructure and public service platforms to promote the democratization of digital security.
Multiple internet industry leaders have shared similar viewpoints. Cybersecurity expert Luo Hongwei believes that initially, cybersecurity primarily focused on the security of underlying technologies and infrastructure. With technological advancements and expanding applications, cybersecurity has gradually extended to more specific business scenarios and application layers.
Regarding the currently trending AIGC (Artificial Intelligence Generated Content), Wu Yunkun, President of Qi An Xin Group, pointed out that the most significant security risk brought by generative artificial intelligence is the risk of data leakage. Any algorithm or application like generative AI essentially requires large amounts of data, which may be highly important. Therefore, throughout the entire application process, data security must be ensured.
Secondly, there are issues of algorithms and ethics. 'Generative artificial intelligence involves many aspects related to law, systems, policies, and ethics. If these are not properly addressed, they may bring new risks.'
It's worth mentioning that Zhou Hongyi also shared his insights on the development of the internet industry: ten years ago, the hot topics were business models, user experience, traffic, click-through rates, etc.; now the focus of industrial internet has shifted to hardcore technologies, large models, autonomous driving, and more.
When discussing the shift in the roles of internet companies, he mentioned that a decade ago, these companies were full of ambition and saw themselves as the main players. Now, even giants like BAT are willingly taking on supporting roles, serving industries by shifting their focus to the ToB and ToG sectors, which requires a change in mindset.