理论上成立的事情，就要想办法成为现实（2015 中国企业领袖年会演讲）

What Holds Up in Theory Should Be Made Real (2015 China Business Leaders Annual Meeting Speech)

> By Zhang Yiming · speech in the Young Leaders session of the 2015 China Business Leaders Annual Meeting · iceo.com.cn (website of the organizer, China Entrepreneur magazine) · 2015-12-08 (speech given 2015-12-06) · [Original link](http://www.iceo.com.cn/com2013/2015/1208/300534.shtml) ([archived](https://web.archive.org/web/20160807080543/http://www.iceo.com.cn/com2013/2015/1208/300534.shtml))
> The source is a live shorthand transcript; apparent mishearings are kept as written, with notes.

The forum's theme happens to be the Young Leaders Summit, so let me talk about issues to do with young people. Speaking of young people, I don't know whether I still count as young now, but ten years ago I certainly did. I've been thinking: suppose the me of ten years ago were sitting in the audience, what could I say that would help him most?

On my thoughts about start-up theory and the start-up environment of recent years: the rocket belongs to an entrepreneur I respect; it's SpaceX launching a rocket. [Translator's note: the transcript has 斯贝斯, a mishearing of SpaceX; the PEdaily transcript makes clear that he was showing a picture of a SpaceX launch and speaking of Elon Musk.] Go back ten years and he had a very bitter time: the ventures he started kept failing. He said that if something can hold up in theory, if in theory it can be done best, then you should find a way to make it. This reminds me of a saying in China's start-up scene, the "downgrade theory": you don't need to do anything so hard; you can use the internet to sell sesame cakes or raise pigs, and improve the efficiency of some traditional industries. I think that's good business and of course there's no problem with it, but it shouldn't stop there. Five or six years ago, when I met investors, I'd always hear this: if you said you wanted to do something technically very challenging, they'd say technology doesn't matter; what they wanted to know was what your model was and how you'd make money.

The industry as a whole doesn't have much feel for innovative, challenging things, and that creates a problem. Different people can do different things, and we do need the internet to change every industry, but China's young people also need the chance and the confidence to do innovative, challenging things. In this era, China's entrepreneurs and China's young people can come to that conclusion [the downgrade theory], but there should also be an "upgrade theory": doing things that hold up in theory but haven't yet appeared in reality. [Translator's note: the bracketed words follow the PEdaily transcript ("not only a 'downgrade theory' but also an 'upgrade theory'"); the organizer's transcript leaves "that conclusion" unexplained.]

Between holding up in theory and not yet appearing in reality lie countless entrepreneurs. From the earliest social networking sites in 1999 to Facebook's IPO in 2012, countless entrepreneurs came in between. [Translator's note: the PEdaily transcript names GeoCities as the 1999 starting point.] It existed in theory, a social platform was needed, but it hadn't yet appeared in reality. There are many, many examples like this.

Back to my own industry. In the 1990s there was a concept: in theory there should be an internet newspaper that gives you, every day, a paper of the content you most ought to see. [Translator's note: the PEdaily transcript names the concept as "Daily Me".] Since the concept was put forward, many companies have tried it one after another. From 2008 I kept watching this direction, whether such a product could appear. Since the internet appeared, first there were directories and portals: whether Yahoo, hao123 or the big portals, they organized information by category browsing. Then search engines appeared, organizing information by keyword. Around 2008 I felt strongly that search engines alone weren't enough. For example, I'm interested in something over a long time but don't think to search for it actively, and there may be no new content in the information I'm interested in.

At the time, in 2006, I was working on train-ticket search. You could search for train tickets with your search conditions, but there were never any tickets that met my conditions. I had other things to do then, and I wanted the information to find me. So I wrote a program that would text me as soon as a ticket meeting my needs appeared, and half an hour after I went out, I'd bought the ticket. I felt it shouldn't be people looking for information, but information finding people according to their needs. [Translator's note: the transcript literally says "it shouldn't be information finding people", a slip; the PEdaily transcript has "it shouldn't be people looking for information".] In the process of information finding people there should be many features. In 2010 lots of follow-type products appeared: you follow a person or a topic, and that content keeps coming to you.

Distributing information isn't only about people and topics. You may like reading what someone writes about technology but not what he writes about Chinese medicine; on the same topic, you may like reading what experts write but not what novices write. In America content is ranked by popularity, with popular content at the top.

In theory there should be a way to put social features, location features, popularity features, features of a user's reading history and features of following authors all together for recommendation. This is something that holds up in theory but hasn't yet appeared in reality.

I thought it existed in theory, so why hadn't it appeared yet? I kept thinking with this question in mind. In the second half of 2011 I noticed that smartphone shipments kept rising: 2011 equaled the previous three years combined, and people were more and more used to getting information through the phone rather than the newspaper. First, the phone screen is small; one screen can hold only four or five items. Second, the phone is carried with you, a device that uniquely identifies a person. Third, the phone has many sensors that know your location and speed, and it's connected in real time: while you're using the phone, the phone is also paying attention to your actions. So I felt that what existed in theory now had a chance to appear in reality. In 2011 machine learning, big data and cloud computing were also on the rise.

I had the chance to make something that holds up in theory appear in reality, and if I were the first to do it, I had the chance to make it arrive early. At the beginning of 2012 I founded Toutiao. At the time people asked me: lots of people have done this; if you do it too, I'm afraid it won't work, because this is the second or third wave of personalized recommendation. Some asked whether it could be done with simpler technology, or said the portals were already doing this and were already very strong, so could I really surpass them? All of these limit the possibilities with an existing framework of thinking.

In these three years, has what we imagined then, holding up in theory, happened in reality at Toutiao? Toutiao's figures: so far Toutiao has 330 million cumulative activated users and 82 million monthly active users; each user uses it more than 6.5 times a day, for more than 50 minutes. By time spent we're already one of China's top 50 mobile internet apps; among all companies that appeared after 2012, our product has the longest time spent; by DAU we're also in the top three; and by time spent per user we're second only to WeChat and Mobile QQ. [Translator's note: the PEdaily transcript has "after 2010" here, and also gives "over 33 million daily active users on mobile".] What held up in theory has now happened!

Toutiao has already gathered a huge number of users on our platform. Second, we imagined that information should be distributed in the most ideal way covering all features, with information finding people. As distributing information becomes more and more convenient, creating should also become more and more convenient, and more and more people take part in creating: this is the so-called self-media era. Toutiao is also a platform where a huge amount of content is created. We developed Toutiao Hao, which is also a form of creation. At present 27% comes from independent large media, and more than a thousand media organizations cooperate with us. [Translator's note: the transcript here is garbled (今日头条也是海量内容创造的内容); the PEdaily transcript has "Toutiao is also a platform for creating huge amounts of content; 73% of the content on Toutiao already comes from its creation platform, Toutiao Hao".]

In theory we should have the chance to become the biggest creation platform in the Chinese-speaking world, and we hope one day to drop the words "Chinese-speaking world". [Translator's note: the transcript has "we hope to finally remove the Chinese-speaking world", which makes no sense; the PEdaily transcript has "we hope one day to remove the words 'Chinese-speaking world'", i.e. to become the biggest anywhere.]

We've built a technology-intensive company that matches information and people through machine learning and personalization technology. The core of IT is information, and the core of information is organizing and matching. Information can be stored, displayed and made 3D, but the most important thing is matching: matching people and real-world locations [Translator's note: the PEdaily transcript names Uber here]; e-commerce matches purchase needs with goods; and what we do with Toutiao is match general information with people's interests. [Translator's note: the transcript has 泛资信, a slip for 泛资讯, "general information".] We're very different from the old portals or pure media. First, we've built a platform for distributing information: every day users spend more than 1.6 billion minutes reading on it, and nearly 600 million clicks distribute all kinds of content.

Second, it's a platform of users' information interests: we can know, at a fine grain, the interests of the more than 85 million users on it, including long-term, short-term and changing interests. [Translator's note: the transcript's sentence is garbled; translated by its sense as in the PEdaily transcript.] Measuring from the server side, we estimate we may be one of the unlisted companies with the most servers; this year we've grown from one or two thousand servers to six thousand. [Translator's note: the PEdaily transcript has "soon we will grow to 6,000".]

If I were to say something meaningful to the me of ten years ago, I'd say that some of us should try things that hold up in theory but haven't happened in reality. Trying such things gives young people some career advice: don't downgrade, don't speculate, and do challenging things with excellent people.

For a while the downgrade theory was popular: we'd like to reach goals that can be achieved faster in the short term, and fewer people take on more challenging things, such as electric cars and rocket launches. In this era, China's young people should have the confidence to try things others don't dare to try.

Don't speculate. Many young people choose a job like this: this company has just raised 500 million US dollars, so I'll go there; this company has found a powerful godfather, BAT has invested in it, so I'll go there. [Translator's note: the transcript has 办了豪门干爹, a slip for 拜了豪门干爹, "has taken a rich and powerful patron".] These choices don't come from their own interests, nor from the difficulty of the challenge, but from whether the company has lots of money, whether it fits today's hot concepts, whether it has strong financial backing. [Translator's note: the transcript has 处于, a slip for 出于, "come from".]

When choosing a company, you should look at whether it has organic capacity to grow: whether it can achieve the fastest growth with the least money, whether you can work with excellent people, whether the company has gathered many excellent people. [Translator's note: the transcript has 延伸的承载能力, which makes little sense; the PEdaily transcript has 原生的成长能力, "organic growth capacity", explained as growing fastest with the least money.]

Do challenging things. Many things are hot right now, but ten years ago, whether making electric cars or launching rockets, none of it was hot, and it's the same in China: when BAT were born, what they did wasn't hot, but it was very challenging. [Translator's note: the transcript has 诞生之处, a slip for 诞生之初, "when they were born".] So when young people choose a career now, they should look more at the challenge of the thing; if you have the chance to do challenging things with excellent people, you grow fastest.

What Toutiao does is very challenging, technology-intensive and future-facing. We're hiring continuously, and talent is the ceiling of an innovative company; you're welcome to recommend excellent people to us. Thank you all! [Translator's note: the transcript has 上线, probably a slip for 上限, "ceiling"; the PEdaily transcript has 战线, "front line". Translated as "ceiling", which fits the organizer's wording best.]

Copyright: © LookMinds. Our own edited text and translation, not a substitute for the original; rights in the original belong to the author and the platform. Please see the original link.