Sticking to AI native applications, Robin Li redefines large models

"Without rich AI native applications built on the basic model, the basic model has no value.", "The future will be generated by us together.". On October 17, at the annual Baidu World Conference 2023, founder, chairman and CEO Robin Li redefined the future of large models - "We will all enter the era of AI native, a human-machine world through prompt The era of interaction." Robin Li showed off Wenxin 4.0 and AI native applications at the scene. He was the first to cross the line when the industry was still "rolling out large models" and rolled out more than a dozen AI native applications, which is already a "fast" People take two steps.”

Baidu Wenxin is officially upgraded to version "4.0"

Baidu GBI can do anything quickly

As the opening speech of the 2023 Baidu World Conference, Robin Li announced that Baidu Wenxin has been officially upgraded to version 4.0. Compared with the previous version, the new version has achieved significant improvements in the four major capabilities of understanding, generation, logic and memory. These capabilities are the basis for the survival of all AI native applications. Similarly, without rich AI native applications built on top of the basic model, the basic model has no value. At present, the comprehensive level of the Wenxin large model is not inferior to that of GPT4. This is Baidu's most powerful large model at present, and it has achieved a comprehensive upgrade of the basic model.

The new search after the reconstruction of the Wenxin large model meets three characteristics: ultimate satisfaction, recommendation stimulation, and multiple rounds of interaction. “It can be said that generative AI and search are a match made in heaven.” Robin Li’s on-site demonstration showed that although search still started from that simple box, the next step has surpassed and reconstructed the original search model. It not only improves search quality, but also breaks through the application scope of search and expands product boundaries. Such AI interactive search will lead an intergenerational change in the search industry.

As the first generative business intelligence product in China, Baidu GBI. For users of Baidu GBI, the intuitive feeling is - they can do everything quickly! So, how does Baidu GBI achieve one step ahead of others? Robin Li’s in-depth analysis believes that it subverts traditional data analysis in three aspects: First, traditional BI can only be operated by professionals, while GBI can directly understand the president’s questions and execute them in real time; second, GBI provides a convenient access method. Enterprises can access data, ask questions and analyze any data, and no longer need to manually analyze across databases and tables; thirdly, GBI also has learning capabilities, and enterprises can inject professional knowledge in the industry to make it an industry expert.

According to Robin Li, Baidu GBI can shorten the work of data analysis and report writing that takes business analysts more than ten days to the level of minutes; Ruliu meeting minutes can be generated with one click, greatly reducing the number of meetings workload, shortening the meeting time by 40%.

Reconstructing Thousands of Industries

Large models are becoming an important driving force for new industrialization

Robin Li believes that the most thoroughly reconstructed product at present is Baidu Library. In Robin Li’s view, Baidu Library will completely subvert our office and creative habits. "We can do this because Baidu Wenku already has 1 billion high-quality documents. Based on the understanding of these contents, we can write truly practical and clear manuscripts, which will not be pure fictions or general empty talk. Robin Li proudly said: "Based on the four major capabilities of the Wenxin model, it is no exaggeration to say that Baidu Wenku has fundamentally revolutionized its own life and completed the evolution from a 'content tool' to a 'productivity tool'." Not long ago, on the first day of the new version of Baidu Wenku, more than 2 million people experienced the new AI features. In the era of big models, Baidu Library is the best starting point for content production.

The new network disk after AI reconstruction-Baidu Network Disk "Yun Yiduo" is the first personal cloud intelligent assistant in the industry and even the world. As of October 17, the number of users of Yunyiduo has reached 20 million. It is an AI intelligent assistant based on personal cloud. In Robin Li's view, Baidu has defined network disk in the past ten years; in the next ten years, Baidu will use AI to redefine network disk.

Another product with a longer history, Baidu Maps, AI-native thinking has also reconstructed this 18-year-old product. This is also the world's first AI-native map product. The reconstructed Baidu Map is not only a navigation, but also a travel guide. The more you use it, the smarter it becomes and the more it understands you. And when the large model is combined with the smart cockpit, the more the AI ​​interacts with it, the better it can understand your needs. There is no need for complicated button operations anymore. You can control everything in the car with your voice, allowing the smart cockpit to have many capabilities such as logical reasoning, strategic planning, and content generation, even with chattering, multi-person instructions, interwoven voices, and continuous dialogue. situation, it can also understand the different needs of each person and meet them at the same time.

"Plug-ins are a special kind of AI native application. It is also the AI ​​native application with the lowest threshold and the easiest to use." According to Robin Li, Baidu is vigorously promoting the deep integration of digital technology and the real economy. Baidu's large model technology has been applied in real industries such as manufacturing, energy, electricity, chemical industry, and transportation. It uses intelligent inspection to reduce safety hazards, intelligent dispatching to shorten delivery cycles, intelligent decision-making to improve enterprise efficiency, and intelligent information control to solve traffic congestion. …Large models are becoming an important driver of new industrialization.

AI native era is coming

The technology industry ushered in the "Baidu moment"

"The AI ​​native era is coming. Without rich AI native applications built on the basic model, the basic model will have no value. The development of native applications is an inevitable trend in the development of the AI ​​industry." Teach everyone how to make AI native After application, Robin Li came to this conclusion.

Not only is it leading in basic large models, Robin Li also made an accurate judgment on the next step of AI large models by releasing Wenxin large model 4.0, which is comparable to GPT-4: application is the real value of the model. The release of a number of AI native applications at this Baidu World Conference demonstrates Baidu's rapid action: becoming the first technology company to use AI to reconstruct products.

Baidu’s rapid actions are inseparable from Robin Li’s forward-looking judgment. As early as March 16, when Wen Xin Yi Yan App was released, he gave an accurate prediction: the next step should be the application of large models. . In the past six months, Robin Li has emphasized the importance of application in large models on multiple occasions: he has defined three conditions for native AI applications: natural language interaction, full use of the four capabilities of understanding and generating logical memory, and interaction of no more than two people. level menu; he also emphasized at the entrepreneurship competition that entrepreneurs "do not make sense by rolling out large models, but rolling out applications is where the real value lies."

Under the leadership of Robin Li, in the past few months, Baidu has "reconstructed every product with great enthusiasm, and continuously deepened its understanding of AI native applications." - This is how Baidu has many AI native applications in the world 's appearance.

Robin Li is not only the leader of Baidu, but also the leader of China’s AI. Over the past ten years, he has been China's most staunch AI investor. Not only did he take the lead in establishing the Deep Learning Research Institute more than ten years ago, but for many years, regardless of the ups and downs of AI development, Robin Li has persisted in researching and investing in AI. Baidu R&D intensity is nearly 20%, ranking first in the country.

It is precisely because of Robin Li’s long-term investment in AI that Baidu has cultivated a group of endogenous AI talents such as Wang Haifeng, Wu Tian, ​​and Wu Hua, making Baidu a highland for AI talents in China. As early as 2019, it released the Wenxin model Version 1.0, and today brings version 4.0 and AI native applications.

As the best representative of the long-term growth of artificial intelligence in China, Robin Li is leading Baidu to stand on the top of the wave and usher in the "Baidu moment" in the technology industry.

The above is the detailed content of Sticking to AI native applications, Robin Li redefines large models. For more information, please follow other related articles on the PHP Chinese website!