How does Robin Li view the past year?
Baidu’s founder, chairman and CEO gave a detailed answer at the Geek Park Innovation Conference 2018.
Li first reviewed the development of Baidu. He said frankly that due to the limits of thinking in 2011 and 2012, Baidu failed to deploy a mobile Internet layout. “We were hit and affected. I was a little scared at the time,” Li said.
With the era of artificial intelligence now here, Li can relax a bit. Baidu made early plans for the AI market. After weathering problems in Baidu’s mobile Internet deployment, Li said, “We can finally welcome the days that will belong to us,”
But Li made it clear this time that he did not say slogan “All in AI”. “I tend to leave some leeway. Although I believe in AI, I don’t want you to think we’re developing a driverless car or DuerOS.”
In 2017, Li Hired Lu Qi for Baidu.
Li didn’t believe in professional managers, but now, he is pleased with Lu. The number of executives reporting directly to Li now is only six or seven, – half of the past. Li has more time to think, but there are still too many things and too many details that cause him worry.
“Sometimes, you are like a born workaholic,” Robin Li joked.
Li also explained why he chose to open in the AI era, and also his thinking on information distribution.
Li said Facebook wanted to enter China seven years ago and had negotiated with Baidu for a JV.
Finally, Li talked about some personal changes. He was also asked what he would like to leave behind for the next generation. “I want to leave a simpler and more free era for the next generation,” Li said.
What Robin Li said is recorded as follows:
On the history of Baidu
Around 2000, we launched Baidu’s first press conference. At that time, Baidu was not well-known, and the business model was different from now. The first time I saw the friends from media, I was embarrassed, and I underestimated the difficulty of explaining business to the media.
I didn’t expect Baidu could be what it is today.
In 2002 and 2003, I was told to focus on the mobile Internet, which I didn’t understand. We didn’t know how to develop the mobile Internet. This thinking lasted too long.
In 2011 and 2012, the era of mobile Internet really came. We were not ready and did not have a layout. We were sevely hit and impacted.
So, Baidu had a lot of movements, and I worked at the front of the battlefield. I thought we must stay in combat to survive in the market.
When the AI arrived, Baidu was in a better position. This time we made a plan in the early development phase. As a search engine, Baidu in essence deals with many of the problem AI wants to solve.
So it’s natural for Baidu to enter the AI field. We finally welcome the days as belonging to us.
On Management and Lu Qi
Some changes happened in management. Two years ago, a dozen people reported to me, and now the number of executives reporting directly to me is only six or seven.
As I said before, professional managers don’t run companies as well as founders. I still believe that.
I think the best managers don’t regard themselves as managers. Lu does. He regards the company as his baby.
Lu is quite reliable. He devotes a lot to the company, and to building up a positive company atmosphere.
I can say that over the past year I’ve stepped back in Baidu management. I have time to consider things that are important, but not so urgent. In 2017, I was able to stand a little further outside and look at Baidu.
Last year, the most important thing was that we updated Baidu’s mission:
To make the complex world simpler with technology.
The new mission covers all aspects of society, and is more inclusive than the previous mission, which was “Providing equal and convenient access to information for all people.” The era of artificial intelligence will change both the supply side and consumption end of business. We can do more than before. Fortunately, we have prepared a lot for this moment.
AI has changed many areas over the past year. For example, mobile phone manufacturers may say that mobile phones are AI terminals instead of mobile Internet terminals.
On the Role of the Founder
I’ve enjoyed starting my own business. As a founder, I build up the company little by little. I often wondered: what’s going on here? What role does the individual play? Is it because of the era? How much is because of luck? How much does individual effort matter? How have I changed since 18 years ago?
Unlike 18 years ago, none of Baidu’s business reports directly to me. I just manage the people and things I want.
But there are times when you are needed as a founder, such as establishing internal consensus. Only when I say it myself will people accept something is our consensus.
When Lu joined Baidu last year, I listed the problems that Baidu must solve in 2018. Today, I can unveil the first problem: the DAU problem of mobile Baidu.
DAU is a new problem for Baidu. Baidu has developed a search engine for more than 10 years, using “how many questions it can answer” as a core metric. But in the mobile Internet era, people want a searching experience that is as clear as an App. In the future, people will search mainly on the mobile Baidu App instead of the website.
The name Mobile Baidu sounded like an accessory, so we renamed it Baidu. I think it shows a shift in corporate thinking.
All in AI and Open Platform.
I do not use the slogan “All in AI”. I tend to leave some leeway. Although I believe in AI, I don’t want to give the impression we are only developing a driverless car or DuerOS.
It’s not that we don’t invest in searching. The technology behind searching is AI.
In terms of AI development, our long-term strategy is an open platform.
Having an open platform can avoid repetitive work. I don’t like others to describe open platforms as “steamrollers”. It’s not like what we made caused the other startups to collapse. Even if we didn’t create the open platform, their repetitive work ultimately would be replaced. Startups cannot escape the rules of time. We just speed up the process. The sooner you died the sooner you can reincarnate. There is no need to reinvent the wheel.
When we decided to open Apollo, we were worried it might hurt some of our partners. But communicating with them, we found they were happy about the decision. So what’s the future? We will continue to bring value to the partners.
Openness also brings brand value. Last year we went to CES, but the impact was much worse than this year. Last year, people thought we were just new competitors, while this year many people think of Baidu as partner. They treat us differently.
Apollo has been open for six months, and has attracted more than 90 partners. I believe an open platform will beat a closed platform.
On Information Flow
I think information flow is a new word, but it’s not a new concept. Facebook wanted to enter China seven years ago and negotiated with Baidu to set up a joint venture. Zuckerberg said at the time that the flow of information must come in.
The flow of information doesn’t have to be algorithm-based, but the algorithm is more efficient. Some information flows rely on algorithms, and some on social relationships. The algorithms are less sticky than the social.
I think the best distribution of information flow is both algorithmic distribution and social distribution. But I personally think algorithmic distribution should be the primary one.
It may take some time for the public to accept the algorithm. For example, Baidu App info flow knows that I like to play golf, and I am learning swimming recently. The more frequent you use it, the more accurate the algorithm is.
I don’t think there’s any fear of algorithmic distribution. As time goes on, algorithmic distribution will be more precise, more progressive and higher quality.
I think in one’s 30s, he or she has both the experience and strength to make something.
But Steve Jobs was already in his 50s when he made the iPhone, so we don’t need to limit ourselves to certain ages. The great ear is worth fighting for.
Children born after the millenium are people of smart times. I wonder what’s the difference in the smart age?
My 10-year-old daughter is already a skilled mobile phone user, and she is even more adept at using a mobile phone than I am.
Sometimes, she played my mobile phone, but some of the content I read may not be suitable for her. So I’ve been thinking, if the phone camera detects that a child is using the phone, then the phone could automatically switch to something that suits children. This is what the smart age should look like.
Sometimes, when children quarrel with parents at home, they run to the smart speaker and say, “You decide which side is right.” So those growing up in smart age are very different from us. Maybe she was born with the ability to use a cell phone and smart speaker, but she may have no idea about keyboards.
The Next Generation?
The things we take for granted may not be reasonable for the next generation.
For example, if AI can solve a security check and ID check in the airport, it will make the trip easier.
So in the age of intelligence, machines can do complicated things and people can do happy and simple things without worrying about tedious procedures.
To sum up, leave the next generation with simpler and freer times.