During the Industry Development Main Forum of the 2024 World Artificial Intelligence Conference, Baidu founder Li Yanhong delivered a speech. He emphasized that in the era of artificial intelligence, developing "super-efficient" applications is more important than chasing "super-apps" with a daily active user count (DAU) of 1 billion. Li Yanhong believes that we should transcend the thinking patterns of the mobile era and avoid falling into the "super-app trap," recognizing that the definition of success should not be limited to the number of users.
Li Yanhong proposed that agents are the simplest form for developing AI applications and the direction he is most optimistic about. He pointed out that open-source large models have their value in academic research and teaching, but in commercial competition, closed-source models are more applicable due to their efficiency and cost advantages. He explained that open-source models require a larger parameter scale to compete with closed-source models, which leads to higher inference costs and slower response speeds. Additionally, using open-source models may result in the inability to share computing power and benefit from the continuous upgrades of the foundational models.
He also mentioned that after Baidu shifted from open-source models to closed-source models for novel creation, the usability and quality of the generated novels significantly improved. Li Yanhong believes that the value of large models lies in their performance in specific applications, not just the models themselves.
Regarding the importance of AI applications, Li Yanhong gave examples of how AI applications in the logistics industry have significantly improved efficiency through large models processing orders. He also mentioned that Baidu's Wenxin Rapid Coding is gradually penetrating the code generation field and is already generating about 30% of the code internally with AI.
Li Yanhong predicts that with the enhancement of foundational model capabilities, developing agents will become increasingly simple, and in the future, there will be millions of agents forming a vast ecosystem. He used Baidu's college entrance examination AI agent as an example to illustrate the huge potential of agents in practical applications.
Finally, Li Yanhong discussed the issue of whether AI will replace human work. He believes that AI is currently more of an auxiliary tool rather than a replacement. The development of AI technology has created new job opportunities such as data annotators and prompt engineers. He emphasized that AI will always just be a tool, with the purpose of meeting human needs, enhancing human capabilities, and making life better.