On February 14, 2013, Jidu "ROBOVERSE Sanlitun Experience Center Kick-off Meeting" was held in Beijing. The official release of Jidu Digital Universe Experience Plan ROBOVERSE (Auto Robot Universe) will create a series of innovative digital experiences for users. On the same day, the experience centers in Beijing, Shanghai and Shenzhen were officially launched. After the event, Xia Yiping, CEO of Jidu, accepted an interview with the media. The following is the interview transcript.

Q: I would like to ask Mr. Xia what he thinks about the core car sales problem. Then we have started the price reduction war for Tesla, BYD and other brands this year. Will it be more stressful for Jidu to start large-scale delivery this year? As for our new experience store, we have opened similar large stores like Benz and Weilai before, which cost a lot of money. What is the reason why we still choose this model now. Then what is the difference between the overall definition of this store and other brand centers.

A: Two questions. The first question is about the overall market this year. I think we have always been under pressure for Jidu. We have been under pressure since the first day of its establishment. From the beginning of the pressure, we felt that we could design the car, to the beginning of the pressure, we gradually changed to whether we could build the car. Now there are people who say what the smart car can do, and how the smart car can do. When you build the car, everyone starts to say whether you can sell it.

I think there will always be pressure. In fact, it is not pressure. I think it is an internal driving force. For Jidu, everything we do is to sell the car well. First of all, we still want to make a good brand. We need to have good products and better experience. I think that if we do these three parts well, at least you will have a foundation to sell the car well.

So we have been focusing on these three things for a long time in the past. At the same time, we are also a 0-1 brand, including products in 0-1. We have experienced many things in 0-1. As you know, Jidu is a company that has been established for less than two years.

In the past, we have been experiencing things. We are doing 0-1, brand 0-1, and even we know that our logo and all our brands will not register and apply until the end of 2021, so we don't have much energy. We have to think from a different perspective and dimension in the process of 0-1.

In the past, I also said that building a car is just a part of the iceberg on the sea. In fact, how to do a good job in your store, channel, sales, experience and even after-sales service? In fact, this thing may invest more energy than building a car. In fact, we also know that these challenges are very big for us. We used to be down-to-earth and do this step by step. I believe that many things revolve around its so-called underlying logic. When we have a good product and a good design, of course, we have multiple dimensions to think about a good product.

Whether it is design or control intelligence, or its quality and good experience, we should also think about what kind of experience this brand can bring to our target users. In the past, we have been doing these things. These things must be in the process of 0-1. There is no way to do them according to the time and date that may be very formal in the past. So sometimes we often do things by leaps and bounds, but this leaps and bounds is also in the current environment, because the process of 0-1 really can't follow the conventional way.

So there must be some pressure on us to sell cars, including the fact that we saw the price reduction of Tesla. I will not extend these things today. When I can share them with the media teachers later, I personally still think that the core problem of Tesla's price reduction is that the product strength has declined very seriously.

You can think about it. Maybe three or five years ago, including me, I was also the earliest owner of Tesla. I spent tens of thousands of money on the FSD at the earliest time, but I haven't used it yet. But in fact, you will find that the expectations given in the past are very high, but this kind of expectation is still very poor in the intelligent experience.

My network is often disconnected and the experience is very poor. As a smart car, it is not a product with core competitiveness in China now. So I think the price reduction is the return of its own product value, not the high premium of its intelligent sales. I think it is normal. I have never worried about this matter. I firmly believe that our products are in many aspects, It must be beyond them.

If you take the smart car as a very complete product, we are very confident to surpass it in terms of design, space and internal intelligent experience. Therefore, I am not particularly worried that its price reduction will have a relatively large impact on our future car market.

For example, for 8295, we all know that the great power of the chip is AI. It is the first time to introduce AI with great computing power into the cockpit. But if it is not an AI company behind it, how to use AI, you don't know how to transplant the cloud big data model to the AI chip. What do you say you use the chip for.

Second, we chose 8295 unswervingly at that time. It is also our belief that 8295 can create a new and better experience. So we were determined to do it. As you know in the industry, after we chose the 8295, China's next generation of models will probably follow us. Because we are the first model in China, basically the models behind us have chosen the 8295. Basically, you can see that the mainstream of smart car tracks are choosing the 8295 chip and Orin chip as their next smart car parking capability.

So we are still very confident. I think we are the starter of 8295, and we can use the capabilities of 8295 chips to the utmost. Because we are a technology company, not just to build cars, so our great mission is to better land technology in cars and let people use it. I am not particularly worried, but you said that the market will have challenges because of the overall economic problems or the external environment. There will be challenges. I also believe that this challenge is very difficult for us. I can only say that I prepared the whole team for the war, but I don't know what will happen in this war. When we really come out, we will go to the battlefield to try.

I think this is a matter of two dimensions, which is why we do this press conference today. We didn't spend a lot of money. I can tell you that, like the press conference on June 8, we used 1/20 of the industry's money to achieve the effect, which is beyond the effect of everyone's press conference.

Jidu has always been a company that hopes to use technology to change efficiency. We never spend money for the purpose. The purpose of our Sanlitun store is very simple. Many people may spend tens of millions on the opening ceremony, but our investment in Sanlitun store is very small. You can see that our cooperation with artists is actually very simple. In fact, it is related to the brand.

We don't want to learn from others. We don't want to put some stools in the store to create an atmosphere and let the children drink coffee. This is not what we want to do, because what others do is not what I want to do. We want to connect future target users, because the whole world is changing.

We were the first to think about why we should do the meta-universe press conference. You see, we have done a lot of things that others don't understand. I am sometimes asked by many people in the forum. I think sometimes this kind of try, you don't try new things out, you don't do the first wave, you don't want to be brave to try this thing, how do you know the effect.

Whether we go to the Meta Universe press conference, we go to digital marketing or we go to digital collections, many people don't understand it and think it is a lottery. Digital collections because I am also a technical background, I play digital currency Web3 for a long time, and digital collections are mainly guided by scarcity.

So at that time, we wanted to build a way to do this with scarcity matching permissions. After our constant attempts, we found some problems. In fact, we sometimes felt that others understood it, but found that many consumers and the general public did not understand it. Later, we also made a lot of reflections. Maybe the next time we do this thing, we might change a way to make everyone understand it better, But you don't know these things unless you try.

I think from my point of view, we are such a company. We dare to make some innovations and breakthroughs. So when we build this experience store, we don't really think about spending tens of millions of dollars like Benz and BMW. Our store spends a lot of money, including you see why we cooperate with artists and these IPs, To a large extent, we have been focusing on the brand products.

So our brand experience store is more about extending from the perspective of brand. From the beginning of brand establishment, we talked about that we are an AI company. You may go back to our early press conference. We talked about how we connect the physical world with the digital world. In fact, the whole set of concepts has not changed since we started the company.

This store is also the core of our brand. We also hope to attract our TA better and give them a completely different experience in what way. So you see, I have been talking with them today at the press conference. We are also constantly changing the changes of the retail industry and adapting to the changes of the times, because we think digital things, The digital experience must be the future, including some experiences like digital twins and digital symbiosis.

When you have an idea, how to make use of existing technology to realize it and make users understand it, we should actually take an intersection of it. In fact, we also try our best to reflect such things in the store.

So you can see that we are not like other brand stores, but simply put up some boutiques here, some coffee there, a rest area here, and what area there. I feel like I have experienced it. Maybe this is the experience of the previous generation of retail stores. We think that the new generation of retail stores must not be like this. What is the direction of the new generation of retail stores? We have been exploring, and we have not wasted much, Maybe you also heard from outside that a store like Weilai spent a lot of money. We really spent a little money to do this. We are still thinking about the impact of the change of technology digitization on the future brand promotion and even brand communication.

As for why we open this store, we have two dimensions. The first dimension is that the brand should be recognized by others. As for why we are in Sanlitun and Taikouli, you may not know that we are the first car brand store in the core area of Sanlitun and Taikouli. We are also competing with many competitors to win this store in the process. Sanlitun is not really concerned about how much you pay, It is more concerned about whether your brand can bring special influence to Sanlitun.

So we are also competing with many brands, even all the luxury brands and high-end brands you can think of now, and finally they chose us. Of course, our pressure is also very heavy. We also hope that through such a cooperation, we can bring a better and different car brand experience to Sanlitun.

So you can have a look. It's really different. We still think a lot about the whole brand research. Of course, this kind of store is not opened on a large scale. It is more in some core business circles. We have appropriate positions to match the brand, so we will do this. Most of the stores are more brand-oriented, including our stores in Shanghai and Shenzhen, In the future, we will continue to land in all parts of the country. In fact, those stores will accept part of our understanding of the digital world in terms of experience, but not every store will have such a big digital experience as the Sanlitun store. This is not what we need to do.

Anyway, we are different. We don't want to simply learn from Benz, BMW and Weilai. We still have our own ideas and internal brand demands.

Q: I want to ask you about the application of ChatGPT. There are two main questions about the combination of Wenxin and our car. Question 1: What functions will it implement? Do we have a specific timetable for launching these functions now? The second question is how it will be presented, because we can see that ChatGPT may be typing now, but it is certainly impossible for us to type in the car. What form will we take, what kind of interaction will it bring to our intelligent cockpit experience?

A: As for the time point, I can only say that we are still working hard to connect with their team, because we also know that Baidu may release it in the next March or so, and we will only do our application after the release. This time point is the same.

You just talked about products. I played ChatGPT early. I played it when I came out of DayONE. I forgot when I registered, because I especially like to do it. You may know by paying attention to my circle of friends that I may have sent something a year ago. At that time, I took it to see what it was.

First of all, you should understand it this way. Suddenly, you find that we were talking about robots before. Many people don't understand why we talk about robots. You can understand the idea that we might put the previous generation closer to something like ChatGPT through 8295. But I didn't expect that after ChatGPT came out suddenly, it pushed the natural communication of robots to another height. We suddenly found that ChatGPT could be a mobile phone, a projector, or any form of existence, as long as there is a very strong person behind it who can understand you, whether it is TXT or voice, such a form exists.

So for the cockpit, I think the biggest change is when you use voice in the car. Now you may think that it is still a directive thing. In the future, it must be a matter that is more close to the communication between people. It can answer many of your questions and solve them in the car. Because I always think that voice interaction is one of the most efficient ways of in-car interaction. Any interaction is not as direct as voice. For example, "Rapids" is very popular recently. If you are in the car, whether you use gesture interaction or button interaction, you need to operate many steps to find iQIYI and be able to click to see it. But you need to say a word in voice. This car will open iQIYI and help you call out the movie you want to watch. You only need one command to interact with each other. No interaction is more efficient than this.

If it is gesture interaction, it will be very troublesome. The habit of human thinking is not so constructed. The habit of human interaction must be that he does not need to think. In fact, there is a very important design principle in Apple's design, that is, how to understand people's way of thinking, that is, what you think should be the next step when you see this thing, then I will make this interaction happen according to the next step you think.

Just like you want to watch the fifth episode of "Rapids", you don't mean that I should first think about what the gesture is, or where iQIYI is. This is impossible to happen when driving. It is too complicated for this interaction.

So I think ChatGPT may bring about the efficiency improvement of this kind of interaction in the car, which can make many of our past interaction scenarios more comprehensive. When we didn't have ChatGPT in 8295, we were able to do what we see is what we get (and offline). With ChatGPT, it will expand a lot of scenarios, but I want to sell a lot of specific functions. The next time we really have a product, we can let you experience it.

But it must be in the car. It must be voice. In fact, I was talking with a teacher about this matter just now. You can see that the voice we do now is basically a sentence you say. For example, I want to open the window. In fact, after I understand it in the voice, I will extract the matter of opening the window and then implement it. For ChatGPT, it may be that I want to open the window. I just have to throw the whole sentence to you. As for the result of your return, I don't need to do further research. In the past, there was a layer in the middle to simulate this. It was more like a strategy, more like a person to do something based on these scenarios. By processing the core words and then doing the action, the middle layer may become very vague for ChatGPT. When our products come out, we will definitely communicate with you. Please look forward to it.

Q: I have two questions. First, the Sanlitun experience store you just mentioned is different from other brands. You seem to pay more attention to the service image. Can you explain in detail what is different? In addition to the different ways of using funds, each brand can also reflect its value proposition. What's the difference between Jidu's store and other stores? What's the difference between the way we spend money and other brands? I don't think the cost of the cooperation between the three bodies is too low, such a hot IP.

A: It's impossible to talk about this. Sometimes the matter of resources depends on when you take it. We talked about many things earlier, and we already talked about many things before they got hot. We have talked with Sanlitun for a long time, and we didn't start talking today. For example, if we do a press conference, we will look at the number of articles, traffic, live broadcasts or clicks. We have many criteria to measure. We will know, for example, under such a cost, what effect we have achieved. In fact, we are very clear that each company must be different. For example, we rarely do commercial advertising, basically not. We still feel that we can do a good job of the brand. When your brand is improved, you will find that others are willing to cooperate with you. If we really don't have any brand power, you may not know that Sanlitun is not based on your more money and less money. It is based on your brand power, your creativity and your influence. We compete with nine brands in this store. It's not money at all. They don't care about money at all. They care about brands. Can you bring enough influence to Sanlitun and the tone of your brand.

Our brand's inherent pursuit of both the digital world in the future and the AI technology capability, in fact, to a large extent, I think the whole industry is moving towards this trend now.

As for what you say is different, you can only go to the store to experience it. After a look, you may feel that the difference is still quite big. Because our store is called brand experience store, which is not for sales. It's very simple that you used to put tens of millions of advertisements outside, but you will find that if you use this money to make brands and experience, of course, there is not so much money. The effect you get may include brand promotion is completely different.

Q: There is another question that I would like to ask. Since last year, automatic driving has changed. Before, everyone has gone to L3 or even L4 to conquer. This year, more people may go to L2. Smart cockpit or more practical or easier to achieve these things. What do you think of this change? Is it a combination of difficulties or a more down-to-earth look at the current mature technology realization path or consumer demand. Including the impact on our future planning.

A: Your question is very good, because I have been thinking about it recently. In fact, we have questions about the extent of concentration and your ability. We are still down-to-earth, because there is no need to say what we can't do, because we have to think about it from the beginning. I don't know if you have seen a map online. For every company, there is such a map, One is what consumers want, the other is what technical capabilities can achieve at present, and the other is what your capabilities can achieve. In fact, these three have a certain intersection.

Maybe technology can be realized, but if your ability is not good, you can't maximize the middle. I think there was a problem in the past. People regarded automatic driving as a marketing gimmick. Everyone said they were good at automatic driving, and even let the media teacher test drive and even talked a lot about it. But no one in the store said that I would give you an experience of our automatic driving at the first time, and no one did so. As you know, the unique selling point is supported by the ability of the product. It is not a marketing point. It is a problem for you to market something you cannot do.

Tesla also has a problem. In China, we have too high expectations, but for some reason, we also have problems in the United States. People may expect different things. However, we still hope to use it out of the box. In fact, we are also under great pressure because the time is too short. When our first generation just started to run, it was actually a waterfall development in the middle. From December 2021, we ran on our Simu Car 1 for the first time. Let's start with the smallest one. After that, in fact, in the first half of 2022, we have the real 8295 and Orin A samples. So a lot of software developed on 8155 and xavier was transplanted to our 8295 and Orin, and then we started to run. First, we ran the hardware with our entire vehicle architecture. After that, we really started to generalize in this software level. We started running in Beijing at the earliest, then went to Shanghai, and then you may see our video at the Guangzhou Auto Show, That's how we made a simple adaptation in seven days in Guangzhou.

In fact, one of the problems that we have been solving is how to generalize automatic driving, but we have never thought about it. I think we have been very down-to-earth. What we are telling you today is that we have done it now, but I am a bit clean, which means that this product is less than 100%, and I will not give you experience.

So many people have been saying on the Internet when they can experience it. It's not really that we want to sell it. It's really that I have been telling you that if our products, whether it's smart cabin or smart car, can't be delivered to the user in the first time, it will be a failure in my opinion.

So we have always been strict with the team. We have always felt that it is not yet time, so we will not take it out. I just talked about L2 and L3. I think this is a very rational return. We have never said that we will be L4 as soon as we come out. I am thinking about one thing recently. The relationship between usage and automatic driving at the user end. After people get on the car, they are driven by robots 100% of the time, but this is a very challenging thing.

But you will find that we used to claim that no matter who does NOP or NGP in China, you will find that it is not a 100% standard car. Some models that claim to be self-driving, even the entry-level model with a price of 300000, do not have self-driving and laser radar, which is problematic. Let's make a hypothesis in the past. If a person drives 20000 kilometers of cars a year on average, you can think about how much of the so-called NOP can be used in the 20000 kilometers of driving in China with the current market capacity of each company? It's less than 10%. You spend tens of thousands of dollars. It's something you can't use even 10% of a year. This is determined by its ability. So we do it, not to say that we should be 100% at the beginning.

For example, you can analyze that a person drives 20000 kilometers a year, about 90% of the journey takes place in the city, about 7% - 8%. Maybe the parking scene accounts for 1% - 2%, which is good. But parking is a high-frequency scene, and the rest may be high-speed, about 7% - 8%. Even high-speed is not good. You charge users tens of thousands of yuan, do you think it is reasonable? Unreasonable. What we need to do is not to do 100% at once. What we need to do is to increase the utilization rate by 60% first. Then we will continue to iterate, 60%, 70%, 80% and 90%, which is the path we want.

Why can we do 60% or 70% in the first place? It is because our start is different. Maybe Tesla L1, L2, L2+, L2++, L3, and finally L4. Although our start is different, we use Apollo's experience in perception, vision, and ODD (automatic driving limited operation range), and its core capabilities, let us actually, to a large extent, when we start, It is an infinitely close L2++. It's just that we iterate on it. That's why the logic behind it is different. So we don't think that I will experience the product every two months. We are very clear about what we can do now.

It was also in Guangzhou that time. I also personally experienced the automatic driving of our products in the city of Guangzhou. I am still very clear about the automatic driving ability of our products. Next, our biggest challenge is generalization, how to provide in multiple cities, and how to provide it quickly.

For example, you also know that the establishment of automatic driving maps in open cities, such as the sudden opening of the next city, how can I quickly land that function in that place within the time after its opening, is the problem we need to solve now, so what we need to solve now is a generalization problem, not the problem of making it run in a specific path, which is not the problem we need to solve, This is the problem that we have solved since December 2021. Generalization is how to replicate a city in the whole city if we transform a city into a grid, and how to replicate it in a square kilometer of a grid. This is what we are doing now, and how to quickly replicate between cities. Because if a technology cannot be generalized, every city depends on a large number of high-precision maps and data to do this, which is not a good way to quickly commercialize.

So we are still relatively rational, and don't particularly boast to you about what we can't do. I can assure you that everything we talk about is what we can do. At this stage, we can do it, and present it to the big family in a product-based and regional way. It is absolutely not to say that we sell tricks and gimmicks here, give you an expectation, and then the results will be delivered at the end, The relationship between buyer show and seller show.

Q: Just now, President Xia mentioned high-precision maps. I want to ask a lot of car companies have been lifting perception and lightening maps since last year. What do you think of this problem? Does Jidu have a plan or idea for the subsequent high-end play?

A: It is right to emphasize perception and ignore map. These are two dimensional issues. China has Chinese characteristics. Many technologies are not achieved overnight, but are essentially technological realization.

In fact, back to just drawing three circles, technology can achieve what users want and what you can achieve. The intersection of the three is very important. The greater the intersection, the stronger the product power.

In fact, we all know that we are taking a purely visual route, because once you rely heavily on high-precision maps, it is no different from ROBOTaxi. Relying on high-precision maps can quickly achieve some automatic driving ability, which is a way of using hormones and can be achieved in a very short time, but it is unfavorable for a company's long-term technological development.

Tesla has been able to go to the present because of his adherence to this route. But I also think that this is also the possibility that Musk could do OpenAI, and he really got a lot of ability and inspiration from it. Sometimes we are also thinking about this matter. It is true that there is still a certain gap in our understanding of some things.

We also follow this direction. This time, we have a plan that is not so dependent on high-precision maps. But from the long-term technical route, we still need to go through a period of time before we can completely get rid of the dependence on high-precision maps. This is a process, and the technical direction is right.

With regard to high-end car sales, the specific marketing methods will not be mentioned. Although we have been doing this, we have never felt that we are a high-end brand. We have always believed that we are a technology brand. You also know that many so-called high-end brands now sell the technology attribute as a luxury attribute, and need to add a lot of prices. We are just a troublemaker in this market, and our marketing ideas and selling ideas will be different from them. We think we are a smart car company, so we pay more attention to the intelligent experience when we deliver it to users. Our car has been equipped with dual laser radars since the entry version. All automatic driving or intelligent hardware equipment are standard in the whole series. We first provide an intelligent car. As for the simple or luxury, you can choose. Everyone understands differently. But you can clearly see that many brands will sell intelligent configuration as a gimmick, and sell it very high.

In the past, in the car company, everyone thought that the door and frame could not be sold for money. There was no market price and no user would pay for a door. So we are all looking for the things with the highest premium of intelligence to sell, because intelligence can sell a higher premium. We regard many of the capabilities of intelligence as those only available in the luxury high-end version, and sell them at a high price.

We make this thing a standard configuration. If we want to decorate it better, we should choose a car with a better interior. It takes 3.9 seconds to accelerate, so we should choose the performance version.

But first of all, we should do a good job of the car, so you can expect that our handling is still very good.

Q: We found that ChatGPT technology is also very popular in China. Today, we also announced that Wenxin Yiyan's technology will be on board. I have a question. Today, Wenxin announced that there was a plan for the car to be launched in Jidu. Today is just the release of the boat along the river. Or is it because the technology of ChatGPT is too hot recently. Now it is a hot spot. If it is the latter, will there be any rush after getting on the car.

A: Good question. When ChatGPT was released, we found that there were many applications around ChatGPT. I think this is actually a return to your understanding of Baidu. Recently, we have a perception that the matter of automatic driving is too simple. There is an article on the Internet about ChatGPT. If Google or any company wants to do this, the cost investment is no less than that of a new car company.

Therefore, this is still related to Baidu's accumulation in AI in the past. Baidu is not because ChatGPT is very popular suddenly. Today, we can give you a ChatGPT, which needs to be accumulated. This requires data, computational power, and continuous work on the model. Especially the model layer needs a lot of fusion capabilities, so it is a relatively general AI.

We have been doing a lot of vertical AI before. We use AI to solve driving and AI to solve things in various fields. Now in the case of large models, the models are integrated. The model layer is a general layer, which is not something you can do if you want to do it, but requires long-term investment and technical precipitation, step by step. Today, we announced to Wenxin that we want to do this, which is more of the accumulation of Baidu for so many years. It is not to clap our heads and say that I will get on the car tomorrow, but to catch up with the hot air of ChatGPT. This is in the era of AI. Companies like Baidu will have such ability only after long-term accumulation.

Don't worry too much about our cooperation, because in the era of ChatGPT, we may not rely on cars, but on the strength of the model behind us. ChatGPT now has many applications connected to it. In essence, it is very simple for it. It just gives it a complete language description. Finally, it returns the answer to you. Everyone should know that no matter you ask it to write an article, write a piece of code or do anything, it is very simple in nature. Its processing power is at its brain end, and its dependence on the front end is very small.

A media person interviewed ChatGPT online some time ago. This form can exist anywhere, such as in a small speaker or in a car. This experience will be in our life in the future. It will be a new interaction in the future, including the current search.

I don't think it needs us to do many things in essence. It greatly simplifies the front end. In the past, for example, the voice recognition in the car used to be very heavy. But with ChatGPT, the efficiency of generating applications on it has become higher. In essence, you can imagine that it is a model layer, providing countless open capabilities for the upper layer. We are just an access, and the ability to apply it. Access is very simple. As long as it is done later, it is very simple for me. I will throw a word to you and you will give me the answer.

For example, the scene of driving is actually the same. I was in the auto industry from 2008 to 2014, when it was the era of the first wave of smart cars. When we were working in the call center, we encountered a particularly interesting thing. Someone often drove to call the call center for no reason, and talked with the people in the call center, because he drove very boring, just to find someone to chat with him on the road (this is a real thing).

So people are the same. When you are driving, you say ChatGPT will tell me a joke, or you talk with him. Sometimes it is just a person who is driving, and he feels that there is another person in the car to chat with me. And this kind of chat is not very mechanical, it is a very natural way of interaction.

So, I don't think there are too many requirements for us in essence, but that as long as the back-end model is OK, it is actually very simple for us to access.

Q: In fact, you just talked a lot, including the media teacher. For example, the market pressure is relatively high this year, and everyone is reducing prices. I want to ask a particularly straightforward question. It is in this case that we may have confidence in our products, but can we really achieve that we do not reduce prices when everyone is reducing prices? This is a question of product strategy and pricing strategy.

Another thing is automatic driving. You just talked a lot about it. We are all looking forward to the specific performance of centralized automatic driving. But one thing is that we have already found many products of the new force of car building. The actual consumption intention of consumers is not as strong as you think. Even the FSD of Tesla in foreign countries, in fact, the purchase rate is not very high, At the very least, it did not meet its expectations. I want to know how we look at such a situation in view of a series of situations, because this may be our big selling point, and whether we will be affected by these things. thank you.

A: There are two questions, which are hard to say because it involves company secrets. First of all, I have to tell you about pricing. Our price is very surprising. I can't reveal it now. Even with the current price, it is also very competitive compared with many products.

This is due to our cost control. I think this is the core competitiveness. Just like Tesla has dropped so much this year, it still has a high gross profit rate, which is due to its core competitiveness. I won't elaborate on this. When you look at our car's ability and follow our car's price, you will see.

The second problem is that although Jidu is an intelligent company, our goal is to make money through the positioning of software, but first of all, we must ensure our hardware strength. Whether we will make money on a large scale by software or not, our premise is to first do the business of making money on cars. This is a bottom line. If we can't do this, for example, if I go to fool others, hardware can't be sold, It is all earned by software. It can cheat people in the past few years, but it may not be good now.

The second question, which I have already answered, is still not up to the product strength. In essence, it is still lack of technical capabilities, so it cannot provide users with the best experience.

As I said just now, if you buy something worth tens of thousands of yuan, you may only spend 5% to 7% of your time using it in China a year, which is impossible. Now we have a misunderstanding. User-oriented automatic driving must not be measured by the take-over rate, no matter how many scenarios.

Second, it is not safe enough. I can't throw away the matter of security. I think you should take over when it's time to take over. It's not necessary to emphasize the take-over rate. It's about experience. I should let you take over when you take over. But I can guarantee that your whole experience is the best, and I can guarantee your safety.

We have not yet established users' confidence in automatic driving. Even when the car began to sell, not everyone thought the car was safe. This is a process.

Now you have some misunderstandings. The same is true for the whole market. One person or one enterprise needs to come forward and tell you what is a good standard for user-oriented automatic driving experience. I don't think we will brag about what we need to do now. We will not brag about the L4 as soon as we start. The laws and regulations do not support it. The product experience is not the best. Why should we brag about it. What we need to do now is how to cover more scenes, so that users can cover at least 70% of the scenes in the average driving process of 20000 kilometers a year.

As for 70%, we just said that reducing the take-over rate does not mean that 100% of the take-over should be done. In the special and complex scenario, I think the take-over should be done. It is user safety-oriented, but we should do a good job in the experience, and do not make the user take over. This is what we need to do.

What we said in the past is a little too idealistic, which leads us to think what we should do now. In fact, we are very down-to-earth. We just think that we should solve the scene coverage first. If the coverage rate is less than 70%, why do you sell it to others at this price, who will use your things? Tesla I think it is too expensive to sell it, 10000 dollars is too exaggerated.

Q: I have a question. In fact, Jidu has been emphasizing efficiency since its establishment. In fact, when we observed last year, we also found that, like the first generation of car building new force, Weixiaoli, they have experienced some organizational problems, and have also experienced some adjustments with the expansion of such personnel and the increase of vehicle types.

Maybe I want to change my thinking. You just said that using technology to change efficiency. If you look at the organizational structure, has Jidu ever had some reference models in this management, and because our shareholder relationship is complex, including the supply chain and so on, and the rhythm of ensuring production, how do we improve it through the management of organizational efficiency.

A: You are talking about two questions. The first is our internal organizational capacity; The second is foreign cooperation.

I have always been very strange. I always face such problems. From the first day, everyone asked, "Will there be any problems with your cooperation with Baidu, and will the two shareholders kill us?"? After all, we still have dreams to do this.

So in our cooperation with them, we still have to pay attention to our own problems. We ourselves, including many people I recruited. Ideal recently published an article. Anyway, I didn't want to write these things in the past. For example, my entrepreneurial experience in the previous period, our thinking about organizational structure and organizational ability is missing, so I have spent a lot of time on organizational ability since the first day of concentration, and so it is now. Our team is also very clear that in the construction of organizational capacity, because we have a culture inside, which is strong OKR-oriented, and is a very matrix enterprise. If teachers are interested in understanding our organizational culture, we can do a special session later. From the first day on, our company has always said that it is not determined by position, but by person. You can do as much as you can in the company. We are a very matrix organization.

Matrix organizations need to have a very strong penetration ability in terms of culture, including the organization's OKR in the early stage. We have a very strong model, from strategy to objectives, to the decomposition of OKR objectives, to the alignment of organizational structure, and to the ability of people to deliver these objectives. In fact, all things of the company are based on this.

So we have paid close attention to our organizational ability since the first day. In addition, we will carry out organizational adjustment at each stage, and it is very fast. Maybe we will finish the adjustment tomorrow if we want to do it today. The development of our company is the same, from a team of 20 people to a team of 200 people, and then to a team of 2000 people, including each stage. Our growth is very fast. From the beginning, the OKR of the whole company is to make one car, to start making two cars, to start making two cars, and to plan two cars. My organizational ability must be changed. So you must change quickly, and you can change as you say.

I have been fighting with our team's analogues in different battles. Sometimes the cavalry may be in front, sometimes the infantry may be in front, and sometimes the artillery may be out first. The same is true of starting a business. When your strategy changes in different periods, your organizational ability must change.

Therefore, our company always takes strategy as the bottom thinking. Once the strategy changes, OKR changes, and OKR changes, we should immediately adjust the organizational structure to support our OKR delivery. Next is people. If people are not right, they will not be able to do it.

So along the way, our organizational structure has basically changed dynamically. I have always written my own thoughts to the company's employees. It is the same principle to treat the company as a product. When you find that there is a problem with this product anywhere, you should quickly modify it and continue to iterate this product. We have been practicing this all the time. In each period, including after the Spring Festival this year, the organizational structure will also change with the adjustment of this year's strategic objectives and OKR.

This is also true of myself. In the first stage, I am fully involved in everything, from the establishment of the company to the brand to the product and technology. Basically, this is what I focus on. Because the company focused on these three aspects at the earliest time, first the brand, then the product definition, and then the technology planning.

In the later stage, our company had three relatively important committees at an early stage: the first, the product committee; Second, the technical committee; Third, the brand committee.

At the later stage, when the project starts to rise slowly, it will grasp some things that are partial to the project and partial to the whole vehicle technology. Then it will be found at the back end that when the manufacturing is going to rise, it will start to add the manufacturing skills, and add them little by little. It has changed since now. From the beginning, the whole team has made one car, to the beginning, the whole team has made two cars, and then the two cars are being developed at the same time. There are also two new cars being planned and synchronized, and the ability requirements are different. So you must constantly and quickly iterate your organizational structure to support your strategic development.

Therefore, it is difficult for me to say this efficiency. I have always told the team that others have gone faster than us. If we just follow others' path, we will probably not succeed. Because with our current resources and the number of people, it is very good that you can achieve 80% of others. Only if we improve our efficiency ourselves, every penny we invest or every person we recruit, his contribution in the organization is beyond the efficiency of other enterprises, can you catch up with others, otherwise you will always be only 80% of others.

So we all have a sense of crisis inside, and everyone has a deep understanding of this aspect. We feel that if we are not efficient, we cannot catch up with others in the next few years. Let's not talk about surpassing others. We catch up with others first. After all, others have a great advantage over us. So we have always been very concerned about the organization.

Q: Now we think that compared with some mainstream new forces, from the product level and the service level, are there any special labels? For example, in the past few years Tesla said it might be smart, and some cars said it might be space or home. Do we have any special labels.

Another is that we are now promoting. I think there are many elements of science fiction and art. Personally, I feel that it is not particularly grounded from the perspective of consumers. So will we have some choices that are more suitable for actual users in the future promotion or in the actual implementation of products, that is, real experience, which is very intuitive.

A: Good question. I found that when communicating with media teachers, everyone particularly likes to pay attention to brand issues. We all think that Weilai did it. Why didn't you do it? I have been telling our team that if we just bring the framework made by others to us, we will probably not do it well.

I would like to ask you what the theme of today's press conference is. In fact, there are practical and pragmatic aspects to brand. What you see is retreat, but there are still many practical aspects. The practical aspect is growth, which is simple sales leads.

Maybe you don't know, because our core is a relatively strong internet gene, and you also know that there are many automobile companies that cooperate with Baidu Geely. Don't forget this. This is a company with a large user base, and iQIYI. We have too many ecological companies.

I still want to talk about this matter. These are two things. Just like you think of me as a household enterprise in the future. You just said that the ideal is to have a home, but don't define yourself as a family. Do you still have a home after you make a small car? This is very difficult. Don't define yourself to death. The brand concept has never changed since the beginning. In fact, it is a refinement of our upper class. These things have not changed since the first day of our establishment, and the system has been strictly implemented.

Including some of our products today, which are at the level of brand retreat, but pragmatic, you may be surprised to find that we are a company with hundreds of thousands of users who have not yet launched its products. This is totally different. I think we need to be both pragmatic and pragmatic, but we can't say something that we do under the water every day. For example, how many customers we get today and how many users we get from which channel tomorrow.

This is the problem of business competition, but the rise of our brand is what I must do, otherwise how can I influence others at the level of brand. The brand has two aspects. It has spiritual aspects, which is what we talk about. It also has practical aspects, which is how many registered users you have and how many sales leads you have. You also know the approximate proportion of the conversion from sales leads to final sales orders in this industry. You can get 10 million. We often say that based on the sales volume of next year, how many registered users we need to get this year, and why do I set a limit on yourself? You can get 20 million registered users this year is the best. You can sell cars next year and the year after.

So I think this is a matter of two levels. That is to say, I don't think we should look too much at how to do it and how we do it. We think we still have our own ideas, that is, our ideas have not changed since the first day of the establishment of the brand. We will tell you that from the first day, this brand has never changed from our construction of automobile robots to natural communication, free movement, and self-growth, and then to the following USPs. We have been following this strategy.

Previously, we suffered from the failure of products, which is our biggest problem. Because cars are still what you see and what you get. If you can experience it, it will be easier for you to talk about it. So you will think that we were quite modest in the past. In the past, I can only be modest and have no practical way to show this thing to you. But I also hope you understand that this is the only way for a brand to go from 0 to 1. If I don't do it, I really start to do this when I go public. I can't sell it well. Nobody knows your brand. You don't have enough registered users. You can't sell it.

In fact, in the end, I think it is still the same. In the end, everything we do is a final result oriented by business demands. I can only talk about this matter in this way. This is my style. Everyone who has attended many of my press conferences knows it.

Q: Hello, I think the last two questions are that you said that Jidu has actually encountered difficulties in channel management. Our 01 and 02 products have also come out this year. From today's point of view, which link do you think is the biggest challenge. We still want to know what support Baidu has provided in the process of cooperation with Baidu, and what is the approximate communication frequency of Baidu.

A: First, everything is a challenge for me. Whether it is supply chain management, channel construction or brand construction, it is difficult. To be honest, if you want to ask me, I can only say that our focus is different at different stages.

For example, in terms of manufacturing and supply chain, the first car is still easy for us, because we also have great support from Geely. Our first car is also produced in a factory in Hangzhou Bay. We have transformed the entire production line before the New Year. Our first car will be launched at the end of this month or early next month, very fast.

The same is true for the supply chain. In fact, we have a supply chain team of our own, and we also work closely with the Jidu supply chain team, including the logistics team. In fact, we also have a relatively advanced idea. We started to sort out some problems in the supply chain very early, including a lot of communication work on capacity production and marketing planning to ensure that we can achieve our goals this year.

You said that every link is challenging, but I never felt that these things are difficult. I always felt that this is a problem, just solve it. For example, the problem of cooperation is that the two teams of Jidu and Baidu work together every day. We now have many teams stationed in Jidu factory. Our logistics, our branches and all our teams are stationed in Jidu factory.

Our teams work together. There is no communication frequency. Everyone is integrated. You can imagine that the two teams work together completely and work together every day. Sometimes we even pull the two teams together for closed development without you and me.

I think our cooperation has been very smooth. As you all know, it is not all our efforts that have brought Jidu to this day. It is the great support of the two shareholders. If our communication is not smooth, and we encounter the problems you think in your heart, will the two shareholders change a little.

In the last word, many people thought it was a little strange for me to do Jidu at the earliest time. I replied that it was human.

When you enter an environment, you think this is your life. If you recognize this life, you will be out of play. But if you don't recognize it, you will only work out the result. I also want to tell you today that we still have to look at the results. If we can't do what we said, we have never missed every node I told you since we were founded.

The epidemic at the Shanghai Auto Show last year was so difficult that all our teams were closed in the company. We would not say any node of the whole company. We said at that time that many people would not believe that the car would be built in 2023. I think it is still a goal oriented. I think the whole company is just a twisted rope. We all work together to do this. There is no other idea.

Have you ever encountered communication difficulties? There are always problems. Internal team communication is not smooth. Just deal with it. I have never considered this matter as a matter to think about all day, or to communicate every day. It is not as complicated as thinking. Thank you.