Qi Lu, YC China's No. 1 Employee

*Note: These are Jeffrey Ding’s translations. Jeffrey Ding is a researcher at the Future of Humanity’s Governance of AI Program. As a Boren Scholar, he directly enrolled in classes with Chinese undergraduates at Peking University, and interned at a Chinese law firm. Others are welcome to share excerpts from these translations as long as my original translation is cited.


Author: Meng Xiaobai

Source: 36Kr (popular tech news portal, think: MIT Tech Review of China)

Original Mandarin: https://36kr.com/p/5148299.html 

Qi Lu is different from the Qi Lu you know.

When he said that he would no longer be a professional manager in the future, most people speculated that he would either go into entrepreneurship or investment. Qi Lu decided to kill two birds with one stone.

Today, Y Combinator, the world's most reputable startup accelerator, has officially entered China; Qi Lu will serve as founder and CEO of YC China and the head of YC’s global research institute. The accelerator investment organization established in 2005, has incubated more than 1,900 startups whose total valuation numbers $100 billion.

For more than a decade, although startups served by YC have come from all over the world, they have only operated in Silicon Valley. Entering China constitutes the first overseas business expansion team established by YC, which will be formed by China's talents (By China), will service China's entrepreneurs and economic development (For China), and become part of Chinese society (Of China).

Qi Lu, the No. 01 employee of this new aircraft carrier, is also this ship’s helmsman.

The addition of Qi Lu represents the determination of YC, as a Silicon Valley company, to be completely localized in China.

YC China's mission is consistent with the other side of the Pacific: using technological innovation to drive social change; and allowing the fruits of progress to be enjoyed fairly by the widest possible population. Qi Lu believes that China is about to enter an era of surging technological innovation, and he, as well as the YC China he leads, must be a part of it.

YC China's new business will include 1) business incubation, 2) talent training, 3) scientific research and 4) public good. In Qi Lu’s plans, these will all operate in an unprecedented way. What he wants is not just new technology. Technology is only the capacity to change society; he wants to establish a new ecosystem that supports new technology to change society. That is the correct way to open up his new worlds.

This is obviously a long and difficult road, and he may need a lot of helping hands. However, in this age of abundant anxiety and heavy fog, there is someone who says, "I want to set off for a beautiful new world. Do you want to come with me?" - What a wonderful beginning to a story.[a]

To open the story, we met with the protagonist. We chatted about his life choices, vision, plans, and some interesting bonus topics.

*****The Interview is divided into sections with headings and a short summary of the topics discussed at the beginning of each sections*****

Taking into account “the time is right, the terrain is favorable, and the people are friendly,” I will not regret it[b].

Actually, most of the things Qi Lu has experienced in the past few months most people have experienced - transitions in professional life, what is right thing to do?

In order to make a decision that he would not regret, he sought recommendations from his best friends as well as an in-depth mutual understanding with his future boss. Putting pen to paper to write characters and draw diagrams analyzing his situation, writing down his own perception of the world, and his thoughts about the future. He examined himself and and fortified his faith.

36Kr: Many BAT-level companies have also sent you job offers.[c] Why did you choose to join YC in the end?

Qi Lu: I made the decision based on "the time is right, the terrain is favorable, and the people are friendly." The wisdom of the ancestors[d][e] is truly useful. Considering all the factors, YC is my only remaining choice.

The first “time is right:” I am almost 57 years old.

What I needed to consider is what can happen in the next 5 to 10 years? It is no longer suitable for me to continue to do large-scale, high-intensity work in large companies.

The second “time is right”: we are currently on the eve of large-scale technology-driven new epoch of innovation, which makes me very excited.

A new round of technological innovation driven by China and the United States will revolutionize the world's industries and society, so we need to design new mechanisms and environments to promote and embrace such challenges. The cradle of innovation is the early-stage ecosystem. I think that the four aspects of capital investment, talent development, scientific research and commercialization innovation opportunities in this ecology are in urgent need of comprehensive reform.

New systems and new organizational forms need to be explored and created, especially those for inter-Pacific cooperation.[f][g]

The third “time is right”: China is at a stage where a new ecosystem is especially needed.

In the past decade or so, China has done a very good job in innovation in business models and user experiences, but what really makes a huge difference is still dependent on technological innovations. With the space for innovation of the former two becoming smaller and smaller, China is in great need of large-scale technological innovation to promote a new round of progress in society.

“The time is right” also includes “the terrain is favorable” and of course, because of family reasons, I need a job that can travel frequently between China and the United States, so that I can spend more time with my loved ones.[h]

As for “the people are friendly,” in Yahoo in 1998, at work I sat by Paul Graham , the founder of YC. When Paul held the first YC summer camp in 2005, I represented Yahoo and met YC's current CEO Sam.

Those years, I really approved of the approach of YC, and I later created the Bing Fund in accordance with the YC model, which later evolved into Microsoft Ventures. This time, I talked a lot with Sam and found that our philosophy and feelings had a high degree of consistency.

Because I personally like to make choices in a structured way. There is no single factor that will affect the end result. I use all my considerations, values, and ideals to form a large funnel, and the choice that was filtered down, in my opinion, was the one that I will not regret. The harsher the conditions that make up the funnel, the less choices that pass the filter.

Therefore, apart from “the time is right, the terrain is favorable, and the people are friendly,” I also hope that my work can not only build upon my past experience, but also that I can learn new things every day. If you don't learn new things, there is not much fun. And for the things that I like to do, it is best to have an opportunity to contribute to society; it is best to make a long-term contribution.

Five years, Ten years, the Beautiful New World I See

When Qi Lu described the beautiful new world to us, he was full of hopes for the future and filled with dissatisfaction with reality. He believes that artificial intelligence-driven innovation will bring unprecedented progress and fairness to all mankind.

But at the same time, he hates to see only technological progress without the accompanying construction of a new ecosystem. What he hates even more is that today's innovation ecology is increasingly not supporting sustained technological innovation, since this will directly affect the how fast his beautiful new world will arrive.

36Kr: What type of world is the one you talk about being driven by technological innovation?

Qi Lu: Artificial intelligence technology will bring new infrastructure and computing platforms.

First, autonomous systems that can operate on their own in accordance with their intended purpose will revolutionize all industries; driverless cars are just one of the earliest applications. It will be the core infrastructure for a new generation of productivity, and this infrastructure can be deployed globally.

Second, natural language interaction technology will create new computing platforms. For humans, the most natural interaction is through natural language and vision, so mature natural language interaction technology will create ubiquitous personal assistants. This will also make society more fair. In the past, only wealthy people could afford personal assistants. Anyone in the future could have an assistant to help.

This assistant can understand through hearing, understanding through seeing, and it can also be everywhere. Then, this platform can do more things through extension, such as food, clothing, housing, and education.[i] Of course, it will take a little bit of time to realize this.

Finally, artificial intelligence technology has a very large impact on the scientific research field. I think we will usher in a Renaissance in the scientific world; because of artificial intelligence technology, more and more interdisciplinary research and applications can be developed.

What I’m doing at YC is I want to quickly push forward these visions to “land” into reality, especially in China. Because China has a large population and many industries, more people means more data. This is the main differentiating point between the artificial intelligence technology revolution and the previous technological revolution.

This instance of technological revolution is not about hammering people's knowledge into code and that’s it. For example, if you want to develop a system that can understand human dialogue, I will give you the world's smartest person or 1 million engineers and you won’t be able to develop it. Your starting point requires more than 20,000 hours of tagging data, and adding on dialects is another thousands of hours.

Data is the production capital of the new era[j], just as land in the agricultural era was the main means of production. Regardless of how much smarter people were, if there was not much land, then it was useless.[k] Moreover, in addition to (China) having lots of Chinese people, it is a unified language market, which brings more opportunities to experiment[l]. Any experimentation and iteration of a key idea will be faster. So I think China has a structural advantage in the development of artificial intelligence technology.

36Kr: Up above, you talked about data being the core driving force of artificial intelligence technology. Well, if more and more data is under the control of big companies, so the threshold for startups is getting higher and higher,[m] then what is there to be done?

Qi Lu: Actually, I have constantly been thinking about how to share data in secure and protected conditions. Labeling costs are getting more expensive, which is a big problem for startups. So from a macro perspective, for the long-term health of the early-stage ecosystem, we need to create a more open data platform[n], so that entrepreneurs can benefit from it. Doing this is like the rise of cloud computing more than a decade ago, and it can take the threshold for ventures to be very low.

I hope that there will be a data ecosystem in the future, which allows data related to people to ultimately belong to individuals[o]. He would have the right to decide under what circumstances and for what purposes to allow a company to use his data. In this ecosystem, startups can also get support from users. For example, if users are passionate about education and want to support an education company, they can openly give their own data to this company. I hope that YC will also explore this point.

36Kr: When it comes to this, I have to ask you what you think about the blockchain; some blockchain idealists are trying to create an identity-based ecosystem that allows users to have rights and interests with respect to personal data.

Qi Lu: I am optimistic in the long-term about blockchain technology, as its characteristics bring many benefits.

1. Protection of privacy and data.

2. Can bring different incentives for innovation. At present, the incentives already within the market are usually at the financial level, such as equity and bonds, but blockchain technology may bring about innovations in the aspect of incentive mechanisms in the long term.

3. More long-term, users can transfer their trust back and forth through digital forms.

The specific form of entrepreneurship is still uncertain, but blockchain technology will lower the threshold for entrepreneurship in these areas.

36Kr: I can feel that you very urgently want to speed up technological innovation and increase investment. Do you think that even companies like Google and Amazon are not investing enough in scientific research?

Qi Lu: I don’t believe it is enough.

First, let’s take an example, today's artificial intelligence technology can support the creation of new models of hospitals. In this hospital, artificial intelligence can provide data representations for all aspects and circumstances of doctors, nurses, and patients . In fact, Fei-fei Li is doing research in this area at Stanford University, but the university's investment is limited and is obviously unable to keep up with the progress of the experiment.

Second, many of the technologies that benefit our society are the results of research undertaken by American universities 20 to 30 years ago. The source of technological innovation to promote social development is long-term research investment, and what I can see is that the speed of innovation at the source has not kept pace.

The third and also the most important thing — we have to ask: Who is the biggest beneficiary of research investment? The biggest beneficiaries are those who succeed in business ventures, followed by their investors. But at the moment, they don't have a systematic way to invest in research. Currently, the United States relies on federal government tax revenues to fund research institutions, but at YC I hope to push more of the key beneficiaries to invest in research.

When I talked with many investors in private, I found that they were actually willing to put a portion of their money to do investments in scientific research, but no one has effectively organized this effort. I think that as long as such a big-picture environment exists, I can find kindred spirits to bind initially disjointed interests together.

36Kr: So, do you think that research investment methods that rely on government taxation is not good enough?

Qi Lu: I don't know much about tax-related policies, but I think that if there is more money invested in research, it will be a better world. Because innovative inventions that truly change the world are often driven by curiosity and imagination. The research and development of large companies is oftentimes task-driven. I am not saying that task-driven research is not good, but I think it would be better if there is more curiosity and imagination-driven research.

36Kr: You also just mentioned technology that promotes fairness, but there is disagreement about this point. The technological development in the past 100 years has actually been accompanied by the growing gap between the rich and the poor. How do you make sense of artificial intelligence technology bringing fairness?

Qi Lu: Promoting fairness is a component of YC’s core ideals; YC has always believed in technology-driven innovation, but at the same time seeking that the fruits of innovation are more widely and fairly enjoyed by all mankind. Artificial intelligence technology is just a force that can be used. What it can achieve requires innovation at all levels of society. In fact, YC’s research institute will also research basic income systems and universal medical insurance systems for all, hoping to create systems that integrate better with new technologies.

My Plans, Difficulties, and the Help I Need

In order to create a beautiful new world, Qi Lu will use the strength of YC China to build a new innovative ecosystem. This includes startup incubation, training, scientific research, and investment. He said that YC is an innovation engine company. What the innovation ecosystem needs to be done, YC will do. His goal is to be the engine with the greatest influence; if the entrepreneurial ecosystem needs more than just technology — also new organizational forms and new incentive mechanisms — YC will also do it.

36Kr: How will YC China set out in the beginning?

Qi Lu: Among the four business lines, the first to start is the incubator camp. The incubator camp is the first attempt at YC localization, but at the same time it draws on the core parts of YC:

1. The methodology of the incubator process and training content established by Paul Graham;

2. A network of YC alumni of over 4,000 people. Alumni will help the entrepreneurs that come afterwards, and for Chinese entrepreneurs, alumni networks around the world can set up channels for companies going overseas.

3. The YC brand: being incubated by YC is equivalent to being recognized and endorsed, and it is more easy to be favored by investment institutions.

In other aspects, YC China may develop its own entrepreneurship curriculum and hopes to influence American entrepreneurs because I believe that globalization is the best innovation method. YC China's research institute will be established under the YC global research institute and will set up collaborations with Chinese research institutions. Finally, the public welfare business line is hoping to focus on addressing the impact of new technologies on people's employment.

However, it is just the first day of the launch of YC China, and everything is still in the very early stages. I only have one person now. The most important thing at the moment is recruiting people.[p][q]

36Kr: What kind of person do you want to recruit?

Qi Lu: People who have direct entrepreneurial experience, kindred spirits who possess zeal for the undertaking of China's innovation and are willing to work hard to achieve their vision.

In order to get as many good people as possible, I hope to build a partner – part-time partner – community structure. Partners are the most important people in promoting YC China’s business development; having part-time partners is to consider that some excellent entrepreneurs still choose to start a business, but hope that they can use part of their time to help us; the community aspect is we want more innovative groups of people to have a sense of belonging through online and offline discussions and activities.

I am positioning myself as an empowerer. I hope to see people with good ideas and good visions doing bigger things on YC China's platform.

36Kr: In addition to finding good people, what other help do you need to get?

Qi Lu: Because we need to build a new innovation ecosystem, so many players with different roles can participate.

The first is Chinese investors. Because YC China will be closer to the early-stage projects, it hopes to establish favorable interactions with Chinese middle and late-stage investors. Moreover, YC China will also raise funds independently, and I hope to get support from China’s LPs[r].

Second, we will also have a lot of cooperation with big companies. YC China's investment projects may receive investment from large companies, become acquired, or provide services for large companies in the later stages.

YC China’s research institute will also establish connections with scientific research institutes in government or in companies, and it hopes to support China’s path to becoming an innovation power, receiving the Chinese government’s support.[s]

36Kr: Scaling (businesses) has always been a challenge and direction of great effort for YC. Will YC China take up the task of scaling? How can your background in artificial intelligence help the scaling of YC business?

Qi Lu: The goal behind scaling is how to maximize innovation. I think that a more engineering-based and systematic approach can push forward the scaling up of YC’s business, and YC China is definitely a platform that we can use to trial this.

However, what we need to balance is that YC China has just started, I don't want to be too ahead of myself. Because if the foundation is not done well, it will cause long-term adverse effects. Therefore, under the premise of solid foundation, we will not give up our quest for scaling mechanisms.

Regarding myself, what I have learned the most from all my development experience with artificial intelligence technologies and commercial products in the past is “the establishment of an ecosystem.” I think that the best way to commercialize artificial intelligence is to become a platform and to become an ecosystem.

This reason is very easy to understand. The core driving force of artificial intelligence is data, data is multi-dimensional, and the right of ownership (of data) is always multi-dimensional. For example, in a driverless car, who does the data collected by the event data recorder [t]ultimately belong to? Is it the owner of the car or the operator of the recorder? Or is it a city manager? And if this person is a US citizen, can the US government say that it has data ownership?

Through the establishment of an ecosystem, there is a greater degree of synergy in the interests and incentives, which is the best way to scale.

Talking about the future of Google, Einstein and Nonlinear books

Qi Lu does not like to be interviewed. He said that he is not good at it and does not enjoy it. When he went to Microsoft, he made it clear to Steve Ballmer[u], "I only do 2 public relations activities a year, the rest I ask you to take on." So during most of the interview, Qi Lu always emphasized everything about his work. It is for the mission and development of the organization.

Only when the interview was almost done, Qi Lu talked about some of his more scattered thinking and daily curiosities. I was pleasantly surprised to find that the Chinese executive who was deeply at the center of public discourse, when excited, shook in his chair, like an extremely sunny boy.

36Kr: We have observed that many Chinese companies and investors hope to continue to stimulate short-term growth and are not willing to invest in long-term development. What do you think has caused this?

Qi Lu: According to my observations, both China and America are facing this situation. There are a lot of good companies who are balancing (between short-term growth and long-term investment), have not grasped it, and lost many opportunities.

But innovative technology companies that change the world need three things: confidence, patience, and perseverance. This is a personal experience that I must emphasize. Just be sure to have faith, be patient, and be calm.

Sam has said on many occasions that 10 companies will reach Google’s scale in the next decade, and three to four of them will emerge from China. Twenty years of professional experience tells me that these three or four companies must be companies that believe in the long-term.

Of course, this is also related to some institutions. In 1855, the United Kingdom created a limited liability law, which was a very successful historical event and it enabled investment. However, this did not change the rights of investors. Now the capital market is so developed. In a few minutes, you can decide to buy or sell a company's stock. Shareholders still have the right to decide the CEO through the board of directors; but companies require long-term investments. It can’t have behavior that focuses on the short-term - this needs to be changed.

36Kr: What should a company do when it faces the innovation bottleneck after achieving some initial success? In particular, what happens when new technological innovations require reforms to the business and organizational culture.

Qi Lu: There is a methodology to this.

Geoffrey A. Moore's recent book, called Escape Velocity, is about how the more successful a company is in the past, the easier it is to die; because past successes tend to lock you in the past. How to escape the successes of the past is the methodology that he set up.

In addition, with respect to culture, it is often easy for people to learn new things, but it is often difficult to forget what they learn, just like muscle memory. The organization also has muscle memory, it can't forget the past; when the market environment changes dramatically, it thinks that it is reflecting deeply on the situation, but actually it is not reflecting deeply on the situation.

36Kr: Do you usually have time to read books?

Qi Lu: Probably 5 hours a week, 1 hour a day. There are more and more articles to read, and I can't finish it; it’s a good thing that I have a Kindle, but I usually cannot finish reading it.

Speaking of this, in fact, I have always wanted to open a small company that is in the book industry. I think the book industry needs reform. First of all, there is a problem with the incentive mechanism. In fact, most books are really worth reading for two sentences, but the publisher pays for the number of pages, so the author has to write a pile of the less valuable stuff.[v]

In addition, why are books linear things[w]? I constantly thought about this problem when I was writing my doctoral thesis.

36Kr: What does a non-linear book look like?

Qi Lu: People's communication is linear, the medium is linear, so I can only communicate my thoughts in a linear way. I very much hope that I can find a way to directly transfer my multi-dimensional thinking to you. You can also exchange your multi-dimensional ideas with me.

In terms of mathematics, human language is actually a very irrational function, and its expressiveness is very poor. For example, the light in this room is actually very rich, but I can only use less than 10 adjectives to describe it to you. Language itself is very unsatisfactory to me.[x][y]

36Kr: How do you interpret the sentence on your shirt: The Difference Between Stupidity and Genius Is That Genius Has Its Limits ----- Einstein?

Qi Lu: I actually did not notice this sentence, but I am rereading Einstein recently.

36Kr: A biography or an academic article?

Qi Lu: I’m reading everything. I want to go back and see his original ideas for society and culture. Why was he able to find the theory of relativity?

[a]I noticed this type of adulation and reverence for Qi Lu after he left Baidu - he got a lot of credit for Baidu's all in on AI strategy and recovery

[b]Reminds me of Bezos' "regret-minimization framework". Will I regret not doing something when I'm 80 years old, narrating for myself the most personal version of my life story? :)

[c]This list was long and all the big names came calling -- Xiaomi, Tencent, Alibaba, Didi Chuxing, Jingdong, Toutiao; https://www.leiphone.com/news/201805/zZchu8e9kkrWZpWM.html

[d]I believe that this is an old Chinese expression "the time is right, the terrain is favorable, and the people are friendly" 天时地利人和 by which these three factors conveyed the notion that "it was a good time to go to war" -- hence, the wisdom of ancestors

[e]The phrase comes from Sun Bin's Art of War: "天时、地利、人和,三者不得,虽胜有殃," in English: "the right timing, the favorable terrain, and friendly people; if you do not possess all these three elements in battle, you will suffer calamity even if you win."

[f]This is interesting

[g]What systems? Economic, political, or?

What organizational forms? Corporate, incubator?

[h]Support for this from an old NYT profile of Qi Lu, "(Mr. Lu, who is married and has two daughters, reserves much of the weekend to spend with his family.)" --- https://www.nytimes.com/2009/08/31/technology/internet/31search.html

[i]Also fundamentally change the way we interact with devices/apps. It's very much about making user interface disappear, which is partially we're aiming for at TwentyBN.

[j]This is a good quote.

[k]An interesting view from Ryan Avent that argues that data/AI models can be seen as the capital which is learning from humans to replace the labor. This is an "digital capitalism" evolution that is fundamentally changing the relationship between labor and capital, in which labor is being weakened but not less essential. Psoner & Weyl describes a way that we could take control of, take a right to the value of our contribution to capital: treat data as labor instead of capital, or as a collective, public, anonymized resource.


[l]Is he implying the advantage of a unified language as in terms of the consumers or the talent? English has been the lingua franca for AI research so I don't see the point. But if it's about experimenting new ideas on consumers that are language-dependent, I think there is an advantage.

[m]An interesting article from The Economist also talked about tech giants' "kill-zone" for startups: https://www.economist.com/business/2018/06/02/american-tech-giants-are-making-life-tough-for-startups

[n]Ties in well with using blockchain platforms to share data.

[o]Very interesting view, especially in the context of how China is treating data privacy. Edit: Tech Review says China chooses data over democracy:  https://www.technologyreview.com/s/611815/who-needs-democracy-when-you-have-data

[p]I have to imagine he has lined up people for key positions already

[q]I agree. His challenge is to convince them to join and channel their forces through ;) Your comment somehow reminds me of when Lincoln was just elected and was persuading Bates, Chase, and Seward to come on board.

[r]limited partners - investors in a venture capital fund (typically are institutional investors such as pension funds or insurance companies but they can also be corporations, wealthy individuals or governments)

[s]noteworthy rhetoric here

[t]think the black box recorders for airplanes but for cars

[u]Way to talk to one of the most bossy businessman :D

[v]Perhaps he should get a Blinkist subscription!

[w]I don't understand what he means here by "linear". Can someone please explain? Thanks

[x]love it when Qi Lu gets philosophical

[y]I think he's too mean to human languages. And he probably needs to take a creative writing course to expand his expressiveness in describing the lighting in the room ;)