At the beginning of 2023, the AI model is like a ray of light, illuminating the unlimited potential of AI technology in promoting economic and social change. After more than a year of technical research and market baptism, all kinds of large models have sprung up, and their disruptive innovation power is eye-catching. Today, China, the United States and other scientific and technological powers are standing at the forefront of the times, leading the vigorous development of the large model industry.
Among mainland technology companies, Baidu, which was the first to release the Wenxin model, has achieved excellent results. At the recent Create2024 Baidu AI Developer Conference, Robin Li, founder, chairman and CEO of Baidu, said that the number of Wenxin Yiyan users has exceeded 200 million, and the Wenxin model has become the most advanced and widely used AI basic model in China.
At the same time, Robin Li mentioned that in the future, users of large models can become developers of large models, and large models and generative AI will completely change the developer community. "In the past, developers used code to change the world, but in the future, natural language will become the new general-purpose programming language, and if you can speak, you can become a developer and change the world with your creativity. ”
01
Wenxin Model 4.0 Tool Version was released
Empower every user to become a developer
On March 16, 2023, Wenxin Yiyan was officially released. Today, one year and one month, the number of users of Wenxin Yiyan has exceeded 200 million, the average daily call volume of APIs has also exceeded 200 million, the number of customers served has reached 85,000, and the number of AI native applications developed using the Qianfan platform has exceeded 190,000.
The pedestal model that supports Wenxin Yiyan is the Wenxin model. Compared with a year ago, the algorithm training efficiency of Wenxin's large model has increased to 5.1 times, the average weekly training efficiency has reached 98.8%, the inference performance has increased by 105 times, and the cost of inference has been reduced to 1%. The Wenxin Model 4.0 tool version officially released today provides users with a more "advanced" code interpreter function, which realizes the processing and analysis of complex data and files through natural language interaction, truly improving performance, reducing costs, and optimizing user experience.
At the same time, Robin Li once again called on the industry to vigorously develop AI native applications: "Large models themselves do not directly create value, and AI applications developed based on large models can meet real market demand."
02
Truly "out of the box"
Baidu released three major AI development tools
Over the past year, Baidu has "stepped on countless pitfalls and paid high tuition fees" on the way to develop AI-native applications. Robin Li pointed out that Baidu's goal has always been to provide users with as efficient development tools as possible to enhance social creativity. He said that from Baidu's development experience, there are three main points in developing AI native applications based on large models.
One is MoE. In the future, large-scale AI native applications will basically be MoE, that is, a mixture of large and small models, and do not rely on one model to solve all problems.
The second is a small model. The inference cost of the small model is low, the response speed is fast, and in some specific scenarios, the use effect of the finely tuned small model can be comparable to that of the large model.
The third is the agent. The agent mechanism includes understanding, planning, reflection, and evolution, which allows machines to think and act like humans, complete complex tasks autonomously, and continuously learn, iterate and evolve in the environment, leading to new applications.
In these three aspects, Baidu has launched development tools that can be used "out of the box": agent development tool AgentBuilder, AI native application development tool AppBuilder, and model customization tool ModelBuilder of various sizes.
Based on AgentBuilder, merchants can quickly create an agent in natural language by entering "prompt words", and then through further optimization, they can become a "gold salesman" who can be online 24/7. At the same time, the Wenxin intelligent twins platform connected to the agents created by merchants can also provide developers with a channel for traffic monetization, solve the worries of traffic distribution for developers, and drive agents to form a positive cycle of better quality, better traffic and greater revenue through the data flywheel of distribution-diagnosis-revenue.
AppBuilder, an AI native application development tool, greatly reduces the development threshold by encapsulating and presetting various components and frameworks required for developing AI native applications in advance. "In just three steps, developers can develop an AI-native application in natural language, which can be easily released and integrated into a variety of business environments," Robin Li said. In addition, in addition to integrating Baidu's own rich AI components, AppBuilder also supports open-source SDKs to facilitate secondary development.
ModelBuilder can customize models of any size according to the needs of developers, and further fine-tune the SFT according to the subdivision scenarios, so that developers can produce models efficiently and cost-effectively. In addition to the built-in ERNIE3.5 and ERNIE4.0 flagship general large models and several lightweight large models and vertical scene large models, ModelBuilder also supports third-party mainstream models at home and abroad, with a total of 77 models, making it the development platform with the largest number of large models in China.
"1 billion users is 1 billion developers. Robin Li finally said that the scientific and technological creativity stimulated by AI will also promote the acceleration of the AI era. "Use natural language as a universal programming language to change the world with your creativity. ”
END
Author: Wang Hejia
Editor-in-charge/Layout: Gai Beibei
Reviewer: Wang Tao, Mei Yaxin
Producer: Liu Qicheng
【Communication World New Media Matrix】
央视频 | 微软MSN | 视频号 | 微博 | 今日头条 | 百家号 | 网易号 | 搜狐 | 腾讯新闻
新浪看点 | 雪球号 | 抖音 | 快手 | 爱奇艺 | 知乎 | 哔哩哔哩 | 咪咕视频 | CSDN | 36氪
【New Media Team】Producer|Liu Qicheng Proofreader|Wang Tao, Mei Yaxin Editor|Wang Yurong, Gai Beibei Video Editor|Huang Yangyang, Lu Ruixu Operation|Lin Song
Dots to share
Point Collection