The beginning of 2025 belongs to DeepSeek. After the launch of the R1 model, DeepSeek became popular around the world, followed by a round of “access craze” in various fields. Among them, one company is particularly special, and that is Baidu.
On February 16, Baidu announced that its search and Wenxin Agent platforms will be fully connected to the latest in-depth search functions of DeepSeek and Wenxin Model. Users can use the full-blood version of DeepSeek for free on Baidu APP and PC. Baidu has been deeply involved in large models for many years, but has not made a phenomenal model like DeepSeek, which has made Baidu controversial. This time, Baidu search quickly connected to DeepSeek, leaving the most obvious entrances on PCs and APP. This time, netizens became even more lively and joked that Baidu would join if they couldn’t beat it.
But while teasing, everyone seemed to find that the wind direction suddenly changed. The reason is that many users have found that Baidu search seems to have really become easier to use recently. For example, when we search at random, we will find that the latest review of Baidu Search has this “style”:
We are very curious why users ‘evaluations of Baidu search have begun to change. I learned from people close to Baidu thatFor a long time, Baidu AI Search has been exploring changes in the deep water of technology, trying to balance the disruptive power of traditional search genes with generative AI, including access to DeepSeek.Changes in products bring about changes in experience, which also reflect changes in user needs.
We might as well consider a question from the perspective of changes in user needs: In the post-DeepSeek era, can Baidu Search, which is busy cleaning up its chronic problems, win the headwind.
Search, which is Baidu’s core business, has changed a lot recently. On February 24, Baidu APP launched a new AI portal, integrating the model capabilities of Wenxin Model and DeepSeek, and upgrading and integrating “AI Search”,”Whole Network Information Refining”,”Pinch AI Summary”,”Digital Human Intelligence”, etc. Experience. Not only has the main button in the center of Baidu APP changed to “AI”, but the “Instant Experience AI Search DeepSeek-R1 Full Blood Version” button is placed in the center of the PC home page. Clicking can go directly to the AI search page.
The completion of this upgrade means that it will be easier for users to use AI search in Baidu, and access to the big model will be simpler and clearer. This has also led to rapid changes in user search behavior. The most significant change since then is that the search demand for subjective companionship has begun to increase significantly.
This is manifested inAfter users can easily use DeepSeek in Baidu search, their personalized interactions with search engines have increased significantly.In the traditional sense, search engines serve as toolboxes for information acquisition, and it is difficult for users to provide personalized companionship such as chatting, comfort, and encouragement. However, after connecting to DeepSeek, a large number of users began to choose to conduct subjective questions and answers on Baidu search. Casual chatting, and even some purely personal content such as destiny analysis and gossip discussions.
For example, the recent AI fire makes it easy for ordinary people to worry about whether AI will replace their jobs and lose development opportunities in the future. These concerns can be transformed into questions like this on Baidu search: “How will the social form develop in the future? Will humans not have to work because of the emergence of AI?”
It can be seen that the R1 model can analyze such grand questions in great detail and provide very comprehensive answers based on its unique “grand narrative” style. Regardless of whether these answers are correct or not, I really feel an inexplicable sense of peace after reading them. I don’t know if AI will replace me. At least I learned something from AI today.
If you don’t want to talk about such a grand issue, you can also have a simple entertainment with the AI, such as asking the AI about its horoscope next month.
It can be seen from these new user search behaviors that users are very happy to have personalized and subjective chats with AI, and even treat AI as a netizen or friend.
Further deduction, this means that users will spend a lot of their time on social platforms and media platforms instead staying in search engines.Search engines have also changed from a fast-in and fast-out tool to a long-term companionship.
Of course, search engines need to achieve more when moving towards companion tools. For example, Baidu still lacks a search + social layout, and cannot connect people through social sharing like Tencent Yuanbao. It can be seen that after the changes in search requirements, there is still a lot of work that search engines need to complete.
On the other hand, we also learned that accessing DeepSeek makes the questions raised by users in Baidu search more complex and professional, showing a shift from simply searching for information to providing suggestions, sorting data, and seeking solutions.The improvement of AI’s professional search capabilities also means that its role in personal life and work will change.
For example, a user will ask,”When will the United States impose a 25% steel tax on China? How much tax do you currently need to pay for exporting steel structure products to the United States?”
We can see that even in the face of such very professional questions, AI search can provide very detailed answers, and can also provide key factors such as action suggestions and precautions, which may far exceed the needs of the questioner and bring a lot of additional value.
At the same time, value always comes from comparison. We can see that traditional search engines or social media searches display multiple, massive, and fragmented pieces of information, requiring users to open one web page or social sharing content one by one, and then select content that is useful to them, while manually completing the recording and information organization requires at least four steps. However, AI search can be achieved in one step. After asking questions, complete and clear information can be sorted out, which greatly improves the efficiency of user needs being met by search engines.
In this way, the professionalism of search engines in our lives and work will be greatly enhanced. Its value will evolve from basic information to industry assistants and productivity tools.
In addition, users also tend to use Baidu search to create copywriting, slogans, promotional words, and various materials. Programmers, on the other hand, like to use Baidu search for code generation, and code-related needs have increased significantly. For example, we can ask AI search questions,”If I am a health blogger, can you provide me with 5 public account push themes?”
It can be seen that AI search can immediately provide detailed topic selection planning and execution plans, as well as an inspiration exploration area for further questioning, which greatly facilitates self-media creation.
The changes in user needs reflect that they pay for the most convenient way for Baidu AI search to access the full-blood R1 model, and found a way to open these new AI search capabilities in a very short period of time. In order to cope with changes in user search behavior, Baidu Search has also completed a series of upgrades and adjustments.
As users gradually get accustomed to the AI search model and Baidu’s reputation changes, we can feel that the core is the value of the AI search track itself, for example:
The positioning of search has changed from information acquisition to long-term companionship.
Search is changing from a “short, flat, and fast” source of information to a new information acquisition style and communication tool that can provide a sense of companionship and communication. In this way, search engines ‘sense of presence in life will be greatly enhanced, and their role positioning will be iterated. Both vendors and users will re-estimate the time they spend in the search.
The value of search has changed from basic information collection to professional productivity tools.
In the era of AI search, search engines will transform from information collection tools to super tools in learning and work. Whether it is information organization, data analysis, professional ability output, content creation, code assistance. These tasks will be done in search engines. The value of search will also jump from this and become a port for releasing new productivity.
The focus of search has fully returned to the user experience.
After user search behavior changes, search engines need to rethink how to retain users. At this stage, the original intention of returning to the user experience is the most important. For example, Baidu is willing to fully access DeepSeek in search engines, which can be seen as breaking down ecological barriers to meet user needs.
Just now,”Go to Baidu, you will know” returned to the opening page of Baidu APP. The logical core behind this is thatIn the AI era, the definition of “knowing” has been rewritten, and it is no longer just about obtaining information through search engines.It may be to make the search more personalized and understand you better: combine the reasoning and thinking capabilities of the large model to allow AI to more deeply understand user intentions and provide search results with thousands of people; make the search results more diverse and intelligent: It is not just text answers, but also multi-modal content such as AI summaries, videos, timelines, and digital human agents. Users can directly obtain refined results. Moving closer to and moving forward to AI has almost become a definite direction in the search field.
AI seems to be bringing search back to its original intention of user experience, back to what search should have been.