“You can limit yourself, only your own knowledge.”
Editor’s note: This article is from WeChat public account “Tencent University” (ID: tencent_university), author of Tencent University. The original topic “One of my experiences in robotics: Be sure to get rid of mobile phone thinking | Dialogue Cheetah Mobile Fu Sheng”
The Internet is changing rapidly, and entrepreneurs frequently switch tracks. Some people have returned in the new environment, but some people have quickly taken root in the changes, and they have smashed new horizons again and again. Fu Sheng belongs to the latter.
The term “transformation” runs through Fu Sheng’s 10 years of entrepreneurship.
In 2010, Fu Sheng became the CEO of Jinshan Network (later renamed Cheetah Mobile). When the company was killed by 360 killings, Fu Sheng made a decisive decision and drastically cut the original product line, leaving only Kingsoft Internet Security and Jinshan Guardian. During the 3Q war, Jinshan drug tyrants rose from 7 million to 12 million, and Jinshan Guardians rose from 2 million to 10 million. Fu Sheng led Jinshan Chong to the death line and transformed from a traditional software company to an Internet company.
In 2012, the mobile Internet wave came, facing fierce competition from the giants. Fu Sheng took a different approach and led the company to go out to sea. In the unfamiliar overseas market, the product matrix was laid out from 0, and the business income grew at a high speed. Two years later, the company went smoothly. Listed on the New York Stock Exchange.
Afterwards, Fu Sheng stepped on the beat and completed the layout in four areas: information flow, live broadcast, short video and light game. Among them, his short video product Musical.ly was acquired by today’s headline to become the overseas version of Tik Tok; the live broadcast product LiveMe in the US market has become the “net red manufacturing machine” in the United States.
In 2016, Cheetah anchored the AI robot field and invested in the intelligent service robot company Orion Star Sky, starting again on the vast new track.
Fusheng, who can always turn around Qiankun at a critical moment, what challenges will he face in the new robot field, and what kind of thinking?
“CEO is coming” in the fifth quarter of the second quarter, by Tencent.com editorial board Long Binghua dialogue Cheetah mobile CEO Fu Sheng.
Click below to see the full conversation –
The following is an excerpt from the essence of the interview.Multi-program benefits.
About going to the sea:Finding the edge market,Copying existing models and delving into overseas markets
龙兵华: Telling the story of the successful leopard going to sea?
Fu Sheng: The space left by the giants in the security field is small, but we found that good security software overseas is done by individuals, so realize that this may be a huge opportunity.
I have a methodology: target-path-resources. Focus on the goal first. If the goal is likely to be achieved, then the lack of ability or the unfamiliarity of the field can make up for it. Instead of doing something based on how many resources there are. After realizing that the goal of going to sea may be reached, we will visit overseas developers and users, try to make mistakes with small teams, and gradually increase resources, and finally launch products in the United States and Europe on a large scale.
Entrepreneurs need to find opportunities that are not valued by the giants but may develop in the future.
龙兵华: How to find this opportunity?
Fu Sheng: is essentially insight. The best way is to have insight into the industry, market, and geography, while still grasping the micro experience.
Insight determines how far you can see. You know what you are doing today, not for today, but for tomorrow and the day after tomorrow, you will be particularly determined and full of mission. Insight is not an inspiration, but a formula that can be deduced. Just because the business environment is extremely complex, the formula is not easy to derive, and because there is no chance to repeat the experiment, which method is better and cannot be verified twice. .
Fu Sheng is recording the scene at the show
龙兵华: Do you think there is still a chance to go out to sea today?
Fu Sheng: Going to the sea must be a huge opportunity. From the United States to Japan, many companies are developing their economies to a certain extent. When local opportunities are reduced, they begin to do globalization. The same is true for Chinese entrepreneurs.
I have three suggestions for Chinese companies going to sea:
One, Try to find a marginal market with large growth potential. The competition in Europe and America is too fierce. Southeast Asia and even Africa are worth considering.
Second, Copy the familiar Chinese model, Don’t make innovative models in unfamiliar markets. So far, many of China’s product models have been leading the world, as long as the model can be copied to other regions, it is already a great opportunity.
Three, spend more time studying overseas markets. When the cheetah just went out to sea, many people didn’t even have a passport, and they didn’t know what the United States was like, but we pulled dozens of middle layers to the United States to communicate with local companies. Continuously dispatching employees, let us have a very deep understanding of the local format.
About transformation to be a robot:From security software to artificial intelligence,essence is not faulty
Long Binghua: From the beginning of security software, to the field of AI (artificial intelligence) and robotics, why is there such a transformation?
Fu Sheng: After the launch of Cheetah Mobile in 2014, I have judged that the era of mobile Internet giant monopoly has come. It is very difficult to subvert the market, so I am looking for a new track. I was thinking about several standards at the time:
First, the new track must conform to the inheritance of capabilities. The most central ability of Cheetah is to make a tool that makes the user’s device better. And robots, in the traditional scene, to replace some people doing things, such as labor services, this essence is still a tool. Therefore, the core competence is equally applicable in AI and robots. There is no capacity fault, but only more capabilities are needed.
Second, mobile Internet competition is getting worse and stronger, and you need to find a new outlet. The future demand for robots will become larger and larger, but it is still invisible.Become an opportunity point. The mission of the company is to be as close as possible to the future, to stand on the tide, and to grow at a high speed in the future.
Third, so far, everyone has different definitions of robots. Defining a product on the new track is an opportunity that every entrepreneur has dreamed of.
It is these three logics that allow us to enter the AI robotics field.
Long Binghua: For this new business, have you set up a new company?
Fu Sheng: Invested in a company. Cheetah’s main revenues are all in mobile applications. It is easy to turn a department into a lab by setting up a department to do products that don’t see returns in the short term. In this way, the department has no substantial pressure to make results, and it is difficult to get resources in the company.
龙兵华:CEO is not strong enough?
Fu Sheng: The previous R&D investment is a loss in the listed company, which will have a great impact on the company’s financial report. So setting up a new company is a better choice. It seems that it is because of this that we have built a complete AI team.
p>
CEO is coming” recording scene
Long Binghua: What are the lessons in the two years of artificial intelligence?
Fu Sheng: I think this track is right. The popularity of the robot industry is on the rise, and our products are really able to solve problems for users in some scenarios. This is the core point.
Although the first batch of products looked like To B, it was actually To B To C. Regardless of whether the product is placed at the cheetah front desk, library or municipal hall, in the final analysis, users need to go forward and ask questions that need to be solved. This is essentially the thinking logic of the Internet.
So we put forward the slogan of “a really useful robot”, this big logic is definitely right. But the small logic made a lot of mistakes, such as we used to think of hardware as a mobile phone.
龙兵华: What is mobile thinking?
Fusheng: For example, mobile phones are mass-produced at least 1 million units, so the requirements for the entire mold and structure, and the production of thousands and tens of thousands of robots are definitely different; Second, the mobile phone is very small, it is necessary to prevent falling, the structure is tight, and the input cost is much higher than that of the robot. However, the robot has high requirements on the sensor, and has high adaptability to the scene, and the requirement for rapid iteration is much higher than that of the mobile phone. For example, a mobile phone from prototype to mass production has a cycle of one year, but robots need to have faster speeds to meet user needs… These are the pits we have stepped on.
About competition:Samsung is not necessarily an advantage
龙兵华: As you said, the AI robot market is so big, the Internet giant will definitely enter, what do you think of competition?
Fu Sheng: First, I think the giants don’t think that the robotics field is reliable at this point in time. Everyone thinks that this matter is not reliable now. This is a good opportunity. If everyone thinks that it is reliable, they are all in the game. In fact, it is very difficult to have a chance.
龙兵华: They may not think so.
Fu Sheng: And the chain of robotics is particularly long, hardware and software, AI, 2B, even if the giant wants to enter the game, it is not easy to set up a department alone. In order to study the robot, I have lost a layer of skin myself, and the process and cost of various mold opening are almost clear.
Long Binghua: Do you believe in the innovation mechanism of big companies?
Fu Sheng: I believe that the innovation mechanism of big companies is based on the premise that the company regards this project as its core strategy and pushes it all out. Today, I still don’t see them making such a decision on the robot market.
龙兵华:The giants like Samsung are also in the layout, which is very advantageous.
Fu Sheng:They will have an advantage in production and design, but there will be no product experience. The robot business is similar to the triathlon. It has a hardware design sense. It is not enough to look like a personal shape. It really needs to understand what the core services provided to users are.
The real challenge of robotics is the details. In terms of “sieve theory”, every detail of the product is a pinhole, which looks inconspicuous, but if the pinholes are all over, then no water can be found. For example, the voice noise cancellation function, if the actual noise cancellation capability is less than 85%, the function is not available. For the user, saying one or two sentences in ten sentences is acceptable, but if the error rate is more than half, no one will use it again. However, the details of the customs will be fatal.
How do you view missing content recommendations? It’s a competency model, not a cognitive problem
Dragon Binghua: You have tried to do content recommendation, but later the layout of the robot, it can be said that the blue sea that missed the content recommendation and short video, you think it is recognized The problem is still a problem with the ability model?
Fu Sheng: is a competency model. I have long been optimistic about the headline content recommendation mode, but at that time Cheetah Mobile developed particularly well. From a business perspective, the security software has a large search span, so I did not seriously consider this.
After hesitating for a year or two, I felt that I had to do the content, and I started to form a team to make overseas headlines, News Republic (global mobile news service operator), LiveMe (live platform), and also invested in Musical.ly (music class). Short video community app). Later, I realized that a company could not sell two of its contents, and sold most of its content business. In terms of financial income, it still brought back a lot of income, and probably got back 300 million US dollars in cash. A product like LiveMe. But did not achieve a real outbreak.
Today’s resumption, in general, is because there is no desperate to do content. The company’s performance was quite good at the time. There was no strong external debt pressure. If you did a little business, you wouldn’t succeed. When the cheetah is getting more and more firmly into the field of AI and robotics, the overall growth is getting weaker and weaker.
When the company is developing well, it calls on the team to do business outside the boundaries of capabilities. When it encounters many difficulties in the early stage, it is particularly difficult to unify the thinking. So I have a lot of ability models and organizational structures. Reflective: In the early days of doing the Internet, it is not necessary to pay too much attention to the organization itself. It is ok to use the market to drive the development of the company. Now, we often talk about strategy and organizational capabilities internally. It is very important to go to the cheetah.
Fu Sheng at the recording site
龙兵华: Why is this happening?
Fu Sheng: The environment has changed, and the stage of enterprise development has also changed. When I was doing Kingsoft Internet Security, I started from scratch and only needed to do a good job. However, when the cheetah started to make content, it faced a good development of the original business and at the same time it was necessary to open up a new battlefield. At this time, how to adjust the organizational capability is a big problem for many companies, we just exposed the problem relatively quickly. . The cheetah is also very young, only listed for four years, and I did not grow up in time.
Long Binghua: You are listed too fast.
Fu Sheng: Yes, one year after the listing, we have grown more than 100% annually, which is too fast. Including me, the whole company’s thinking about strategy and organizational capabilities is not timely. Fast money is too easy to earn, and now it is in turn to make up the class.
About learning:Change people, or the environment and self-driver
Dragon Binghua: Transforming AI and robots, many new things must be learned, how to ensure that learning ability can keep up?
Fu Sheng: On the scalp, the meeting not only listens to the frame report, but also asks every detail to make a difference. After a long period of time, naturally I will learn a lot. I will also disassemble the competing products, reflect on them, and constantly assume that I am doing the wrong thing, and then think about the correct way.
龙兵华: Sounds simple.
Fu Sheng: Test your own. I always thought that reflection is a natural ability, but later found that most people do not have the ability to reflect, because reflection is anti-human: The decision you are making now is the best solution so far. After two days, if you realize that you are not right, you will defend yourself instinct. Many people are not reflecting, but are thinking and reviewing.
I think you need to tell yourself that at every point in time, someone will make a better decision than you. People switch back and forth in a state of pride, pride, and inferiority. Press yourself and get rid of pride as soon as possible.
龙兵华: