On April 20, 2018, Source Code Capital’s 2018 Annual Code Class Conference themed “Opening Sources, Decoding the Future” convened in Beijing. Amongst the attendees were Peng Zhang, the Founder and President of GeekPark; Dr. Hongjiang Zhang, Venture Partner of Source Code Capital; and the Chairman of Googol Technology and Songshan Lake Xbot Park, Professor Zexiang Li from Hong Kong University of Science and Technology. Together, they discussed the hottest topics of the moment: the state, trends, difficulties of technology entrepreneurship, and how it can be better integrated with commerce.

“As entrepreneurs, we should develop technology from scenarios and not the other way around. Nowadays, China’s economy, manufacturing industry, and the Internet industry have become massive in scale, and many problems may be solved by Internet and AI; this is the perspective we should assume. Incidentally, this touches upon what was stated earlier. Scientist-entrepreneurs tend to develop technologies first and then try to fit them into scenarios, but a better way is to develop technologies from scenarios and industry needs.”

CEO Quotes

• As of right now, AI is very important to us as a tool. It has infiltrated every industry, and in particular, the manufacturing industry. However, remember, AI is but one tool amongst many.
• In the last 30 years, industry and technology have both seen great changes, but our education model has remained relatively the same; it is still the one used in a command economy.
• When schools encourage different teaching methods, students will cease to be monotonous. The individuals you need will be diverse in their talents. This is where our education system needs improvement.
• In the process of finding talented individuals, companies have often chosen the ones that are interested in business over the ones that are academically accomplished.
• If candidates are interested in the business of your company, then they will try to incorporate technology into it in every way possible. This is the kind of person that companies need the most.
• If AI is manufactured solely for the sake of having AI, then it will not have any practical use besides wasting investors’ money and creating one app after another. It is very dangerous to view AI as a panacea.
• Let’s think about which areas of our business could be digitalized, software-ized, and systematized. This will be our key to overtake on the racing track and an opportunity for establishing systems in the future.

From the left: Peng Zhang, Dr. Hongjiang Zhang, and Professor Zexiang Li

The full transcript of the panel discussion is as follows: 

Peng Zhang from GeekPark: The two honorable guests of this panel are both involved with technology. One of them is a Venture Partner at Source Code Capital, Dr. Hongjiang Zhang, and the other is the Chairman of Googol Technology and Songshan Lake Xbot Park, Professor Zexiang Li from Hong Kong University of Science and Technology. Both are scientists that I deeply respect. Today we would like to explore three topics: how should growing companies break the current status quo in the technology industry? How should one hold onto talented individuals in order to promote the growth of the industry? Lastly, what are the leading trends in technology? Professor Zexiang Li is widely respected in the industry; DJI had his support and encouragement in the early stages of its development. Now, you had some new ideas for Songshan Lake Xbot Park. Could you please share with us these ideas and their origin?

Professor Zexiang Li: The establishment of Songshan Lake Xbot Park was a gradual process over a long period of time. I returned to the Hong Kong University of Science and Technology (HKUST) in 1992. My research interests have always been in robotics and automation. Hong Kong University of Science and Technology was established 26 years ago; it is a modern research university. Hong Kong has had a number of universities, but they have always prioritized education over research. The challenge the university faced at the time was whether it would be possible to establish in China, a modern Hong Kong university that concentrates on research. After 20 years, HKUST has become an accredited and world-renowned research institution. We can say that our goal has been reached. Many among our staff have had a similar trajectory to me: they completed high school and their bachelor degrees in China, received their masters and doctoral degrees in Europe or North America, and wanted to do something back home. At last, they came to HKUST.

Our original goal was to establish a modern system of research at HKUST. However, if research were all we had in mind, then we could have done it in the U.S. HKUST has a favorable condition – it is right beside Shenzhen and Pearl River Delta. These places have been the hubs of manufacturing activities for the last 20 years. I was wondering if it would be possible to combine the manufacturing industry of Pearl River Delta with research and education. Later, my colleagues and I established in Shenzhen our first company, Googol Technology. This was the first motion-controlled company in China. I gained many precious experiences in the process of establishing this enterprise, such as what kind of students and abilities are needed to establish a strong foothold on the frontier of the manufacturing industry.

1. Scientist-Entrepreneurs

Peng Zhang from GeekPark: In the last two years I have discovered a trend of researchers becoming entrepreneurs. How should we assess whether or not they have become competent entrepreneurs?

Professor Zexiang Li: We have noticed that those who have been outstanding in school, students and researchers alike, might not be suited for entrepreneurship. So, our assessment should look at this question: which qualities are necessary for students and researchers to be excellent in both areas? Our rationale for establishing Songshan Lake Xbot Park was not simply to create a robot industry; more importantly, it was to collect experience and data for education.

Dr. Hongjiang Zhang: What Zexiang said was very good. The rate of failure of capable researchers turned entrepreneurs, especially those who were well-versed in theories, is usually higher than entrepreneurs who had worked for a relatively long period of time in companies. There are a few elements of failure worth inspecting, such as the motive and ideas behind establishing an enterprise; if it is something they enjoy doing, and if it is something they are good at and have prepared for. The worst of all is to establish an enterprise as a way to follow a trend. Recently, AI has become a fad. This has led many professors and researchers to believe that they could become billionaires if they turned to entrepreneurship; that they could be like the successful few, running from one panel to another, decorated in fame. This type of people would run into great risks as entrepreneurs.

In many AI companies, some capable researchers completely ignored application scenarios before they attempted entrepreneurship. Many here today are investors. So much money was blindly invested into this AI industry. Enterprises which should have fallen already, are still alive because of generous and “blind” funding. This phenomenon has created an adverse effect on the industry.

I agree with Zexiang’s views. The talents demanded by entrepreneurship are different from those suited for research and technology development work in universities. In the business of entrepreneurship, one should reflect upon questions such as what kind of product is meant for the future? How are you different from your competitors? How big of a role does your technology play in your product, and in which areas? A finished product often represents a 90% success of the technology applied in the eyes of researchers. However, we who actually work with technology know that technology only allows you to take the first step. Whether or not you can take the second or the third step depends on an amalgamation of different factors, such as your ability to adapt to changes in the market and borrowing capacity. I believe that researchers and professors should seriously reflect upon the last point before they attempt at entrepreneurship.

2. An Urgent Call for A System of Education that Breeds Diverse Talents

Zhang from GeekPark: Not all scientists are naturally suited for entrepreneurship; they might need to compensate for the things that the traditional education system could not offer or find sensible business partners. Many have said that China’s education system is relatively backward in comparison to European and American systems. At least in the area of commercial innovation, it is so. Which methods should be used to establish a system that encourages innovation? In the next 20 or 30 years, the mission of nurturing people will be very important. Our generation might have already seen the glimmers of a new system, but without generations of outstanding individuals in the future, those glimmers will fade. How could individuals be cultivated then?

Professor Zexiang Li: We can see that entrepreneurship is becoming increasingly linked to technology, especially with the kind that was developed gradually over time. Nowadays, it is very difficult to establish an enterprise without good means of technology. In the last 30 years, industry and technology have both seen great changes, but our education model has remained relatively the same – it is still the one used in a command economy. The students we chose were able to be admitted into good schools because good schools are rare resources. The problem is whether or not the qualities of the ones admitted are pertinent to the standards of the industry and entrepreneurship. Moreover, whether or not the things they learnt in school will be used towards establishing an enterprise or solving modern engineering problems persists to be a concern as well. We have seen many problems and have been consulting different solutions, but we have yet to decide on one that is universally approved and can be widely implemented.

Dr. Hongjiang Zhang: I think that the lack of a standard or universal plan helps cultivate talents needed in all areas, which include entrepreneurship, engineering, and research. The strength of American and English universities is diversity. The U.S. has comprehensive universities such as Ivy League schools, their undergraduate and graduate programs are exceptional. There are also small arts and science colleges that only admit around 200 to 300 students per year. These colleges concentrate on developing students’ ability to learn and think. On the other hand, state universities place stress on students’ ability to retain the core knowledge of each academic subject. When schools encourage different teaching methods, students will cease to be monotonous, and the individuals you need will be diverse in their talents. This is where our education system needs improvement.

Peng Zhang from GeekPark: Many believe that in the future, China’s technology industry will play an important role on the world stage as it has been developing rapidly. This development is the process of using technology to overtake on the “racing track.” Will there be possibilities for new education models? Today we might not be as good as the others, but if we merely imitate them, won’t we be behind once more in 20 years?

Dr. Hongjiang Zhang: Regarding this point, Zexiang has expressed some very good opinions, and he has done a fine job.

Professor Zexiang Li: I am testing the waters in this area. Songshan Lake Xbot Park is, in fact, a base for practical application. We use technology and robot entrepreneurship as tools to investigate the necessary types of quality, ability, and study method that will enable our students to progress further on the path of robot or technological innovation. Through this, we have gathered a high volume of primary sources, from which we then give feedback to the departments responsible for course structure design at school. We are currently a base where students can apply their knowledge to real-world situations.

3. The Phenomenon of Recruiting, Nurturing and Keeping Talents in AI Companies

Peng Zhang from GeekPark: The question of where to find future talents in technology is probably on the minds of most companies and CEOs that are here today. Many technology companies state that in the future, they will need individuals skilled in the field of AI. However, even AI companies could no longer find recruits in the last two years. They had to go into universities and train students themselves. Both of you come from research backgrounds. In your opinions, where can we find those skilled in technology? How should we attract those people? Could you give us some advice on how to attract the professionals in your circle?

Professor Zexiang Li: This is an interesting question. What I have observed is that after getting their investments, many newly established companies will go scouting for AI talents in prestigious schools and offering them high salaries. This phenomenon has caused many students to try and engage with these companies by turning towards AI-related majors and courses. Some put “Artificial Intelligence Major” on their resumes after having taken AI-related courses for only one or two years. As long as this is written on their resumes, pay checks, employment, and opportunities will swarm to them. This is the current situation we are seeing.

However, these CEO’s could instantly suffer the consequences of their rashness. The students that came into employment in this fashion might realize after a certain amount of time, that they are not competent their jobs. Rather than creating a vicious circle of this sort, companies should spend more energy and resources on a training model in which they and universities can combine forces in the training process.

Dr. Hongjiang Zhang: I would like to add one point. Today, I have seen many companies, including those from the foundation afflicted with anxiety. You are anxious about your companies’ inability to find AI talents. There is no need for feeling anxious. For instance, Toutiao has done wonderfully in combining AI with apps. Their business is mainly based on an algorithm. Whether it is finding the exact coordinates of people, the calculated placing of ads, or using AI for video and word editing, the application of AI is practically everywhere in their products. Considering this, they must need a great number of AI talents.

However, in the process of excavating talented individuals, they have often chosen the ones that are interested in business over the ones that are academically accomplished. If candidates are interested in the business of your company, then they will try to incorporate technology into it in every way possible. This is the type of person that companies need the most. If a company employs those who are unfamiliar with businesses for the sake of upholding an image, they will ultimately fire them. The cost of this is very high.

I used to think that large companies like Google, Facebook, Microsoft, and Amazon look for relevant majors when they recruit, but I have found that that is not the case. My two sons graduated from arts and science colleges in Ivy League schools. What I have discovered about the recruiting process of these companies, is that firstly, they recruit very early on. They usually begin hiring interns when students are in the third year of their undergraduate studies. Secondly, they stratify their recruits. In the past, investment banks hired Ivy undergrads in their third year and offered them a little over $10,000. Nowadays, be it Microsoft or Facebook, their offers all begin at $15,000 or close to $20,000. Many elite students are picked off by big companies very early on. The grabbing of talents follows a rather clear logic: these companies are trying to nurture future leaders of the company instead of mere software engineers. Their training and recruitment are tailored to this objective.

Peng Zhang from GeekPark: Professor Li, you have taught many students, and they flourish all over the world. According to your observations, is it easier for those with research backgrounds to recruit this type of scientists? Alternatively, are these scientists more attracted by recruiters with commercial and industrial perspectives, even if those recruiters do not have backgrounds in research? What is the reason for the lack of recruits, such as we have talked about?

Professor Zexiang Li: We have had much insight into this. For instance, the scope of contestants that compete in RoboMaster Robotics Competition, a contest inaugurated by Wang Tao, has expanded from fourth-year undergrads to include first years, and now high school students as well. His objective was not to screen for talents, rather, it was to create an engineering culture, which has to be done as the principal task. In the last ten years, we have hired many talented people from big companies but to no avail. Therefore, we have to start by instilling a sense of cultural identity into students, and studying is secondary. As long as they are passionate about it, they will be motivated to learn and master the relevant technologies quickly.

Peng Zhang from GeekPark: An excellent company should start seeking out talents before they enter university for long-term benefits. Enterprises will be adversely affected by a decline in the number of talents. Wang Tao is cultivating culture, and the structure of the university complements his operation. Dr. Zhang, should enterprises begin to recruit from universities as a way to drive our industry forward?

Dr. Hongjiang Zhang: To my surprise and delight, I am seeing some top-tier companies and enterprises beginning to do this. Whether it is the establishment of joint laboratories, the hosting of competitions, or the recruiting of interns, everyone is going in this direction. However, I would like to remind everyone that there are no instant rewards to be reaped here, perseverance is a must.

Why are we unable to see satisfactory results with our chips after so many years, why are we unable to make core softwares such as operating systems and databases? Some say that if we entrust these missions to China’s top-tier scientists and professors from Tsinghua University, then they will be accomplished. In my opinion, the development of operating systems and core software is dictated by engineering and ecosystems. For example, Windows is a partner with Intel and many other corporations, Android’s app store was developed by the joint effort of Android, a variety of apps, and every cell phone manufacturer out there. The problem of core systems is really a deeper issue of ecosystems.

4. Scenarios before Technology

Peng Zhang from GeekPark: Enterprises play a rather big role in this, and it needs an industry to support. If it weren’t for the headway made by DJI in the field of automation control, we might still be behind by leagues in the future. Our second topic is, what kind of effects will AI have on commerce in the next couple of years, and how? Will I be changed by it, or will I influence the change? Dr. Li has had many practical experiences in the manufacturing of AI I would like to hear his views on AI and how future sectors of commerce will affect the pace, process, and model of its production.

Professor Zexiang Li: As of right now, AI is very important to us as a tool. It has infiltrated every industry, in particular, the manufacturing industry. However, remember, AI is but one tool amongst many. If you want to apply it well, you must first have a thorough understanding of the problems existent in our current manufacturing technologies and scenarios. If AI is manufactured solely for the sake of having AI, then it will not have any practical use besides wasting investors’ money and creating one app after another. It is very dangerous to view AI as a panacea. Especially now, as chip-purchasing has become an issue, there might be tendencies towards solving the problem by doing everything fast and expecting results immediately. This way of thinking is just incorrect.

Dr. Hongjiang Zhang: To add, Cao Yi from Source Code Capital has repeatedly stressed their investment concept of the “Big 3” fundamental drivers (the 3 by 9 matrix). Each section of the “nine verticals” represents a scenario that is useful to us. Traditional economies such as traveling, food, and lifestyle have always existed. The horizontal “three” investment drivers refers to Internet+, AI+, and Global+. Internet and AI should be used to benefit the traditional economies or send good models of our “3 by 9” thinking abroad. Our enterprises should develop technologies from scenarios and not the other way around. Nowadays, China’s economy, the manufacturing industry, and the Internet industry have become massive in scale, and many problems may be solved by Internet+ and AI+. This should be the perspective we undertake regarding the application of AI. Incidentally, this touches upon an earlier point I made. Scientist-entrepreneurs tend to develop technologies first and then try to fit them into scenarios, but a better model is to develop technologies for and from scenarios.

Peng Zhang from GeekPark: It is what we have been saying all this time. Should we hit a nail with a hammer or hit a hammer with a nail? I find Dr. Li’s establishment of Songshan Lake Xbot Park very meaningful. In the future, we should expand the industry of robot and intelligence manufacturing instead of trying to define problems with AI. There was a theme in the discussion this morning between entrepreneurs. They said that opportunities will always be there and are just waiting for the right technologies, people, and scenarios to cause a “big bang.” Technology continues to iterate, and so do people. It is important to think in what ways we should use sophisticated technologies to solve problems.

Last question: DJI has become a very successful company. As the chairman of DJI, you played an important role in the early stages of their development. I think that DJI is a company with industrial power, and instead of a certain technology, industrial power is found at its core. Perhaps my judgment is inaccurate. Thus I would like to hear your views on what the most prominent value of DJI is.

Dr. Zexiang Li: The industrial power you speak of is perhaps a company that combines software, hardware, product, scenario, and technologies. Shenzhen and Hong Kong have provided a very particular environment and set of conditions, which has enabled young people with clever ideas to combine elements such as hardware, software, product, and manufacture together. Furthermore, they have been able to define innovative things that were not there before. In this respect, DJI has probably opened a new door. You do not have to follow in others’ footsteps and make what others already have with less cost, or sell it at a lower price. You can use this great environment and platform to create something entirely different.

Peng Zhang from GeekPark: What you said was quite profound. What I see is the domineering presence of its competitiveness, supply chain, as well as product manufacture and design. You said that what DJI does has never been attempted by other technology and hardware enterprises, and this is in my opinion, what sets Wang Tao apart from the entrepreneurs born in the 60s and 70s – he is not afraid of being different, and this is quite respectable.

Now I would like to hear from Dr. Zhang since you have been observing large and rising companies in the technology industry, you have witnessed every wave of entrepreneurship out there. Which qualities should a company concentrate on in order for it to make thousands and billions in profit, or to become a truly great company?

Dr. Hongjiang Zhang: This is a big question, and my answer will probably not be as incisive as Professor Li’s. If we look at AI companies, many technology entrepreneurs graduated from Microsoft Research. I have gathered some thoughts on the patterns of their mistakes and successes. If we look at technology development, then the first thing we should think about is: how can we transform technology into platform technology? However, if we consider it from the point of AI development, then it is the question of how long it will take for our technology to turn technological barriers into data barriers, for data must be obtained. The data obtained will initiate a cycle of product development. Once you have a product, you will get feedback from users, which will enable technologies to be improved, and then the number of users and the volume of data will increase. This loop is a very traditional iteration cycle of the Internet.

The same cycle applies to AI as well. Technology drives products, products acquire users, and users generate more data. This cycle has to keep spinning. Eventually, the barriers before you, will no longer be technological barriers, but iteration barriers. These barriers will become higher as the iteration cycle becomes bigger and spins faster. Your technology will become increasingly easier to establish as you see more user cases, gather more data, and define more apps as the iterative loop runs. Today, DJI’s drones occupy 70% of the market. It will certainly see more possibilities for application and more problems in the process of manufacturing. It is important to know which areas can be improved in efficiency.

Entrepreneurs born in the 80s did indeed have fewer burdens than those born in the 60s and 70s. First of all, they did not establish enterprises as a means to escape destitution. Our generation had to overcome the challenges of survival, challenges like eating just 500g of meat a month. These were not problems that they had to think about. Hence they were able to look farther than us and have higher pursuits in technology.

Recently I went to Shenzhen and visited Professor Li’s Songshan Lake Xbot Park on my way. I saw that China is making giant strides in the fields of manufacturing, technological upgrades, and replacing traditional manufacturing with digital. We have pioneers like DJI to thank for, they have led progress in so many manufacturing industries. Let us think about the areas of our business, and which of them could be digitalized, software-ized, and systematized. This will be our key to overtake the “race track” and find an opportunity for establishing systems in the future.

Peng Zhang from GeekPark: I very much agree with you on the point of creating a closed loop and running it efficiently. Every company should go through this process, not just AI companies. I think that all technologies will eventually be integrated into commerce and scenarios; this is perhaps the matter that all entrepreneurs will reflect upon.

Our time today is limited, thank you very much for introducing these profound thoughts. You have given us a thorough understanding of how the exchange of values occurs and how positively closed loops form between technology and commerce. Thank you, everyone!