Robin Li’s speech made people suddenly understand the misunderstandings encountered by enterprises in the field of AI.

Today, the Xili Lake Forum attracted attention from the technology and business fields

This forum, hosted by the Ministry of Science and Technology and the Shenzhen Municipal Government, gathered many heavyweight guests. Participating in the keynote speeches were Baidu founder Robin Li, Turing Award winner Martin Edward Herman, Silicon Valley Artificial Intelligence Institute founder Piero Scaruffi, etc.

When artificial intelligence is booming, Robin Li put forward two calm thoughts

The first cold thought is that there are many large models in China, but there are very few AI native applications developed based on large models. 238 large models have been released in China in October, while the number in June was 79, which is equivalent to a three-fold increase in four months. But how many AI native applications are there in China? In contrast, abroad, there are already thousands of AI native applications.

The second cold thought is that the value of dedicated large models is very limited. There is a very strange phenomenon. Many companies and even cities are buying cards and hoarding chips, hoping to train their own special large models from scratch. As everyone knows, the large model refined in this way does not have the ability to emerge intelligently. The industrialization of large models should be based on large models nesting small models. Small models are quick to respond and have low costs, while large models are more intelligent and can provide assurance.

At the scene, Robin Li also shared a point of view: "To embrace the AI ​​era, we need top leaders to drive it."

We agree with this view. It can even be said that from digital transformation to intelligent upgrade, most companies have fallen into a misunderstanding, that is, they lack a leader to fully grasp and promote it. Why do you say that?

The content that needs to be rewritten is: 01

Why is it called the top project?

The opportunities brought by large models and generative AI are as big as the industrial revolution. This has become a consensus. Every company and every organization is thinking about how to embrace this new era and how to take advantage of it. new technologies to enhance their competitiveness.

Just like the acceptance process of any new thing, the concept of AI native must be accepted by end consumers and startups first, followed by small and medium-sized enterprises, and finally large enterprises. This is because large companies are naturally conservative, insensitive to new things, and unwilling to take risks.

Therefore, if large companies want to embrace the AI ​​era, they need to be driven by CEOs.

Li Yanhong said frankly that after meeting many companies, he found that this opportunity was valued at all levels, but the nature of the problem was not understood deeply enough, so the CEO handed it over to the person in charge of IT to handle this task

IT leaders and engineers are deceived every day by the so-called "shock release", "epic update", "iPhone moment", "explosion" and other claims. They all want to make a basic model themselves, or follow the Evaluation methods spread on the Internet to select a large model with high scores

Everyone thinks that this is embracing the AI ​​era, but they don’t know that the large model itself not only does not produce any value, but also causes a huge waste of company resources and social resources.

To put it more bluntly, only the CEO will have the final result.

In a company, only the CEO will pay attention to whether the new technology has a positive impact on the key indicators of the business

For example, for the Internet industry, does the large model have a positive impact on your DAU, user duration, retention rate, and monetization efficiency? For all enterprises, has the large model made your costs lower, your revenue and profits higher, and your growth faster?

Senior managers of small companies are usually responsible for all aspects of work, which makes it easier for them to develop native applications that suit their needs. In contrast, the division of labor in large companies is clearer, and CEOs must actively lead this change

In Baidu, this is how it is practiced...

02

Baidu’s internal artificial intelligence (AI) plan

"At Baidu, we have resolutely restructured each of our product lines to be AI-native, bringing you an unprecedented intelligent experience." Robin Li said.

In May, Robin Li publicly stated that Baidu will be the first company to reconstruct all its products. We must use AI thinking to create AI-native applications.

On October 17, at the Baidu World Conference, Robin Li introduced various products after AI reconstruction in one breath, including search, Ruliu, maps, network disks, libraries, etc.

At the Xili Lake Forum, Robin Li highlighted three examples.

The first example is about Baidu’s new search function

New search has the characteristics of ultimate satisfaction, recommendation stimulation and multiple rounds of interaction. When the user asks a question, the new search no longer provides a bunch of links, but generates text, pictures and dynamic charts through the understanding of the content. Wait for multi-modal answers to allow users to be satisfied in one step.

The second example is Baidu New Library.

In the past, preparing for a speech would take several days to produce speech notes and PPT. Now, Baidu Wenku can generate a 20-page PPT in one minute, including chart generation, format beautification, etc., at almost zero cost.

Xinwenku has also realized the transformation from a content tool to a productivity tool. Because of this kind of native transformation, the payment rate of the library has been significantly improved. This is what I call the role of AI in promoting key business indicators.

The third example, intelligent code assistant Comate.

You must know that the greater room for imagination lies in the fact that large models have spawned AI native applications that have never been seen in the past. Baidu has incubated some new AI native applications, such as the intelligent code assistant Comate.

Baidu currently has tens of thousands of engineers, and 20 of the 100 lines of code added every day are generated by artificial intelligence, and this proportion is growing rapidly. Through human-machine collaboration, this native artificial intelligence application helps us greatly improve R&D efficiency

Robin Li also said: "I think the changes brought about by AI native applications have just begun."

03

Let’s welcome the arrival of the native era of artificial intelligence

Robin Li believes that AI native applications should meet three basic conditions: first, the ability to interact with natural language; second, full use of the understanding, generation, reasoning, memory and other capabilities of large language models; third, the interaction of each application is not consistent. There are more than two levels of menus, so interactions will be simpler.

Shen Nanpeng once said that from steam engines to the Internet, every technological revolution is marked by the application and industrialization of new technologies, and digital intelligence represented by AI technology is no exception.

Whether it is Baidu’s internally restructured products: New Search, New Ruliu, New Map, New Netdisk, New Library...

It is also Baidu Eco’s “Wenxin Cup” entry project: chatDOC (tools), Nolibox (design assistant), Buysmart.AI (e-commerce shopping guide assistant), Gennie AI (two-dimensional community), Huanliang Technology (AI New materials), null hypothesis (AI medicine)...

Through Baidu, we can see such a scene: from the basic layer to the model layer to the application layer, China's generative AI industry chain has been formed. Startup projects represented by AI Native are rapidly emerging in China.

Not long ago, Baidu launched a large model plug-in platform. Both individuals and enterprises can use this platform to quickly turn their own data and capabilities into plug-ins. Such as legal assistant, resume assistant, decoration assistant, travel, office and other assistant plug-ins.

API is the main way for AI native applications to call large models. In real industries such as manufacturing, energy, electric power, chemical industry, and transportation, they will become an extremely important starting point for large models and native AI applications, and will also become the main position to promote the integration of data and reality.

A new wave of artificial intelligence entrepreneurship and innovation is sweeping across thousands of industries and is being implemented in various industries and scenarios.

"To this day, whether it is China or the United States, I think the best AI native applications have not yet appeared. Just like the mobile era gave birth to mobile-native applications like WeChat, Douyin, and Uber Likewise, in the AI ​​native era there will definitely be excellent AI native applications developed based on these large models."

Robin Li said: "I have talked so much about large-scale models and AI native applications. The purpose is to hope that everyone can take action to use, understand, and experience them, and actively participate in the innovation of AI native applications. Come and work together to create an AI-native era with hundreds of flowers blooming and infinite possibilities."

What could be more exciting than this? Come and welcome the arrival of the native era of artificial intelligence together.

The above is the detailed content of Robin Li’s speech made people suddenly understand the misunderstandings encountered by enterprises in the field of AI.. For more information, please follow other related articles on the PHP Chinese website!