At the 2024 Zhongguancun Forum Annual Conference Future Artificial Intelligence Pioneer Forum, Vidu, the first pure self-developed Wensheng video model in China, which fully benchmarked Sora, was unveiled.
Vidu's appearance, without warning, was "expected".
In 2024, a battle of the gods around artificial general intelligence continues, and in the year that AI continues to rewrite, giants will focus on training and iterating basic models, and startups will focus on the "last mile", on the basis of giant models, optimize all walks of life through specific scenarios. In short, from ChatGPT to Sora, the new round of shockwaves triggered by AI on the global market is still violent and profound.
Not bubbles
Artificial intelligence is still the absolute protagonist.
At the 2024 Zhongguancun Forum Annual Meeting, artificial intelligence is also one of the veritable hot words. For example, the first "Artificial Intelligence Theme Day" was held, and a large number of original cutting-edge achievements were released in a centralized manner, and Vidu was one of them.
On the opening day of the forum, at the "Investment Beijing Conference", Zhou Hongyi, founder of 360 Group, also shared his new thinking on artificial intelligence, saying that the artificial intelligence model is not a bubble, but will lead the industrial revolution.
Zhou Hongyi mentioned that 2023 will be the mainland's "100 model war", which solves the problem of large models from scratch.
2024 is the "year of scenes" for large models, if 10,000 companies look for 10,000 scenarios and develop 10,000 large models, then the large models will really enter the thousands of industries in the mainland, and a new round of industrial revolution will take the lead in China.
Similarly, at the Future Artificial Intelligence Pioneer Forum, He Zhengyu, chief technology officer of Ant Group, also mentioned that the emergence of large models is only a small spark in the new world of AI, and only when it is integrated into innovative applications in thousands of industries can a hundred flowers bloom.
"The present is the 'best of times' for everyone, and the possibility of general artificial intelligence has not been seen for thousands of years, but the current technological progress means that the birth of general artificial intelligence is no longer out of reach. During the Future Artificial Intelligence Pioneer Forum, Wang Xingxing, founder and CEO of Unitree Technology, told a reporter from Beijing Business Daily.
In Wang Xingxing's view, the large language model that is most commonly mentioned at present can empower the "workflow" with very strong generalized language capabilities and significantly improve production efficiency, which also means that it contains huge business value in the field of to B.
However, in terms of commercialization, the field of to C may be better, "conservatively estimated that at least a few million to tens of millions of people are using large language models every day, and the usage rate is very high." Wang Xingxing said.
According to the recently released "China AIGC Application Panorama Report", the scale of China's AIGC application market is expected to reach 20 billion yuan in 2024 and trillion yuan by 2030, becoming an important growth point in the global AI field. Among them, B-end products have become the main force, accounting for eighty percent of the revenue of the entire large model market, and half of the C-end products are still free.
A wave of embodied intelligence
Associated with the ultimate goal of general artificial intelligence, enterprises are competing for large language models at the same time, and embodied intelligence is "emerging".
At the Future Artificial Intelligence Pioneer Forum, Tang Chao, deputy director of the Management Committee of Zhongguancun Science City and deputy head of Haidian District, released Haidian District's "Three-year Action Plan for Building a National Embodied Intelligent Innovation Highland".
In fact, back in May 2023, Nvidia CEO Jensen Huang mentioned that "the next wave of AI will be embodied intelligence, that is, intelligent systems that can understand, reason, and interact with the physical world." ”
Wang Xingxing even judged that embodied intelligence is likely to surpass large language models in promoting new quality productivity.
"Real humanoid robots can greatly improve productivity, such as being able to work in factories, farming in the fields, or replacing humans in more dangerous jobs. The biggest difference between it and the previous automation equipment is that the embodied intelligence is 'universal', it is not limited by the production line, and can be seamlessly transferred to another job, thus greatly reducing the investment in rebuilding the factory. Wang Xingxing said.
In 2023, the "Guiding Opinions on the Innovation and Development of Humanoid Robots" issued by the Ministry of Industry and Information Technology clearly stated that humanoid robot technology has great development potential and wide application prospects, and has become a new highland for scientific and technological competition, a new track for future industries, and a new engine for economic development.
Industry organizations predict that by 2026, the global humanoid robot market will exceed 2 billion US dollars, and by 2030, it will exceed 20 billion US dollars, of which the Chinese market will reach 5 billion US dollars.
Compared with large language models, embodied intelligence also has a prominent advantage - physical alignment and solving illusions. Wang Xingxing said that traditional large language models have hallucination problems because the data used for training is virtual and has no physical definitions to constrain them. Embodied intelligence, or humanoid robots, can better repair this "model" with the logic of the physical world precisely because of their contact with the physical world in the process of work.
However, Wang Xingxing also mentioned that the premise of all this is that the development of embodied intelligence has "reached a critical point".
"Artificial Intelligence +"
The "bigness" of the large model lies not only in the parameters, but also in the imagination space that empowers thousands of industries.
According to the "Panorama Report on AIGC Applications in China", there are seven major tracks worth paying attention to in AIGC applications, including marketing, retail, education, film and television, office collaboration, and games and medical fields with relatively mature technology, clear business models and considerable revenue.
During the two sessions of the National People's Congress this year, "artificial intelligence +" was written into the government work report for the first time, which sparked heated discussions. At that time, Wei Kai, executive deputy director of the Artificial Intelligence Research Center of the Chinese Academy of Information and Communications Technology, once analyzed that artificial intelligence emphasizes the technology itself, while "artificial intelligence +" emphasizes the integration with industries and scenarios.
In Wei Kai's view, this fusion may not be a simple addition, but a chemical reaction, or even a "nuclear reaction". On the one hand, this combination of advanced productivity and industry can improve the efficiency of various industries, and at the same time, such a combination will also create a huge market space for the development of the artificial intelligence industry.
And together with "artificial intelligence +", it appeared in the government work report for the first time, as well as "new quality productivity".
As a strategic general technology with high innovation content and leading the future, artificial intelligence is also known as an important engine to promote the development of new quality productivity and the core driving force of new quality productivity.
Cheung Qingjie, Head of Digital Enablement and Artificial Intelligence, KPMG China, told Beijing Business Daily that emerging technologies, as powerful engines leading social progress, are changing the face of the world at an unprecedented rate.
Among them, artificial intelligence, as one of the most representative and forward-looking scientific and technological trends, is gradually penetrating into all fields of human society, such as finance, medical care, manufacturing, transportation, education, security and other fields, with the help of AI technology to achieve technological innovation and landing applications.
"These changes have not only promoted the digital transformation of enterprises, optimized the structure of the industrial chain, but also greatly improved production efficiency. Behind this, the comprehensive rise and integration of new hardware, new algorithms, and new data has also injected strong momentum and potential into the development of artificial intelligence. Zhang Qingjie said.
He Zhengyu also mentioned that there is still a long way to go from the explosion of technology to the implementation of the industry, and then to the realization of changing people's lives. From the emergence of intelligence to the emergence of applications, AI needs to solve three core problems: reliability, economy, and ease of use.
Taking economy as an example, He Zhengyu believes that this determines whether the application of large models can really enter thousands of households, and then become a new quality of productivity, changing the way of human production, life and even thinking.
Beijing Business Daily reporter Yang Yuehan / text Jia Congcong / drawing Zheng Yuhan / cartography
