Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Also talked about their respective life and work principles

From the editorial department of MEET2022 Qubit Report | Public Account QbitAI

“China has the smartest people to build China’s AI Whampoa Military Academy together to prove to the world that Chinese people can also do top-notch scientific research?”

You may not have imagined that the best computer research institute in Asia, which has trained countless talents for China’s AI industry, was born 23 years ago from such a “pictured cake” call.

The two parties on the call were Li Kaifu and Zhang Yaqin.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Today, one of them is the most influential technical VC representative. The innovation workshop he founded has become an accurate catcher of many technological innovation and cutting-edge technology companies. His scientific and technological books are sold worldwide.

The other is the founding dean of Tsinghua Intelligent Industry Research Institute (AIR) and the Tsinghua Intelligent Science Chair Professor. Not long ago, he was elected as a foreign academician of the Chinese Academy of Engineering. Together with the previous academicians of the American Academy of Arts and Sciences and the Australian National Academy of Engineering, he became “Academician of the Third Academy”.

They are similar to 1998, it can be said to open up a new era of China AI, and in 2021 the final days, the old friend on the same stage again after so many years, nearly two million visitors in front, Secret acquaintance past, but also to share Developed their respective views on the development trend of science and technology.

More exciting details, we will share it in full today~

About the MEET Smart Future Conference: The MEET Conference is a top business summit in the field of smart technology hosted by Qubit, dedicated to discussing the landing and industry applications of cutting-edge technology. MEET2022 takes the form of online participation. More than 20 live broadcast platforms and channels broadcast live broadcasts of the conference, and 2 million industry users participate online. In addition to the full media matrix of Qubit WeChat, Toutiao, Zhihu, Weibo, etc., dozens of mainstream media such as Xinhua News Agency, Phoenix Technology, Tencent Technology, and The Paper have reported on the conference and related content, with a cumulative online exposure of more than 2000 ten thousand.

Key points

Scientists need entrepreneur partners to start a business.

NLP technology has developed to a suitable venture/investment node.

This year, I am most optimistic about the intersection of AI, biological computing and life sciences.

The Metaverse will definitely come, but there will be no huge companies or applications within 5 years, which will take longer.

The question I most want to ask entrepreneurs is whether they have thought about the difference between starting a business and what they have done in the past, and whether there is resilience.

Stay with people you like and do things you like.

(The dialogue session was hosted by Qubit Editor-in-Chief Li Gen. In order to fully display the latest thoughts and thoughts of Li Kaifu and Zhang Yaqin, on the basis of not changing the original intention, Qubit edited the content of this dialogue. I hope I can give it to you. Bring more inspiration and thinking.)

Li Kaifu and Zhang Yaqin’s meeting in 1998

Li Gen : According to the legend of Jianghu, the first contact between the two was similar to today’s scene in the air. It was Zhang Yaqin who Li Kaifu actively contacted.

Li Kaifu : The first time I made a phone call.

Zhang Yaqin : Yes, the first time should be September 1998. Kaifu and I exchanged an email, and then Kaifu called me.

I remember quite clearly. At that time, Kaifu said that Microsoft was planning to build a basic research institution in computer science in China, and asked me if I would be interested in joining and starting a business together.

Kai-Fu Li : What I remember most clearly was that he “lied” Yaqin back to China (laughs).

Zhang Yaqin : Actually, I haven’t returned to China for many years. After returning home, Kaifu spent a whole day taking me to the best places in Beijing.

After that, there was actually another episode, which should be January 1999. At that time, the Microsoft China Research Institute was established. Shen Xiangyang and I returned to Beijing on the same day. As a result, Kaifu was immediately “fixed”—it was too busy at that time and my hair was extremely long. He was sent to the barbershop for a haircut.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Li Gen : When you met in 1998, what kind of trends were you optimistic about? Does Kaifu draw cakes for Yaqin?

Kai-Fu Li : It must be a big cake.

“Bill Gates gave 80 million US dollars, come and help me spend it together.”

“China has the smartest people, together to build China’s AI Whampoa Military Academy, it proves to the world that Chinese can also do top-notch scientific research.”

These are actually the dreams of a few of us. Later it came true very quickly.

Zhang Yaqin : At that time, we hoped to build the best computer laboratory in Asia and a world-class computer laboratory. I think this is achieved.

Looking back at Microsoft Research Asia in the past 23 years, it has indeed made a lot of remarkable achievements, with important achievements in the fields of multimedia, speech, new artificial intelligence algorithms, and graphics.

But I think it is more important to cultivate a lot of talents. Now whether in academia or industry, many of the CTOs of large companies or the founders of AI startups come from the Microsoft Asian Research Institute.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Li Gen : So 23 years have passed, are there any unexpected things?

Kai-Fu Li : What we could expect at the time was that the talents we trained would become deans, deans, and vice presidents of various top universities in China, and they would produce many good papers. Some may also join top companies and bring up an entire AI team. For example, Baidu, Tencent, and Alibaba now have many talents with a background in the Asian Research Institute.

But what was unpredictable at the time was the arrival of the wave of AI entrepreneurship. For example, the four little dragons of computer vision today all have the background of Microsoft Research Asia.

Zhang Yaqin : What surprised me was the rapid application of deep learning.

Specifically, it is AlphaGo. Because I also play Go. Before AlphaGo and Li Sedol played, someone asked me who would win. I said it must be Lee Sedol. It will take ten or twenty years before machines can beat humans.

Finally, we saw that the current AlphaMaster and AlphaZero play with people, which can make the top three players, which is a big shock.

In addition, machine learning is used in various fields, including life sciences, which I did not expect at the time.

Scientists need entrepreneur partners to start a business

Li Gen : This year, the topic of industry-university-research transformation and scientist entrepreneurship is very hot. It seems to be similar to the situation when Kaifu persuaded Yaqin to return to China to join Microsoft China Research Institute 23 years ago. I don’t know what Kaifu thinks about it?

Kai-Fu Lee : I think now is really the best era of hard technology entrepreneurship.

In addition to the field of artificial intelligence, there are particularly good entrepreneurial opportunities in the fields of chips, enterprise-level software, biological sciences, and medical devices. Moreover, the established company really has hard technology, differentiated and sustainable advantages.

In such an environment, these startups will inevitably need the participation of top scientists, and we also feel that some scientists are very interested in participating.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

This way of participation is different from traditional startups. In many cases, a scientist first takes a few years of students from his own project to do some exploration, and then slowly separates it out. He/she himself still stays in the university and allows the students to become CEOs.

But in all these scenarios where scientists start a business, an essential element is that someone can understand the product, understand the market, and know how to do marketing.

If this is not done well, scientists use top-notch scientific research methods to build a company. It is actually good at first, but it will be difficult for the company to enter a state of operation, product, revenue, and profit afterwards.

Therefore, a scientist needs an entrepreneur partner to start a business . This is what we have seen after investing in many scientists. A few scientists are interested and able to grow into entrepreneurs and become all-rounders. This is of course the best, but it is hard to find.

Li Gen : Does Yaqin agree with it? It seems that Tsinghua AIR is such a model.

Zhang Yaqin : I totally agree with this point. Scientists and entrepreneurs have different abilities and interests. A good scientist may not be a good entrepreneur. To make a technology company successful, scientists also need entrepreneurs.

A person is a good scientist and entrepreneur at the same time, there are examples of this, but this is a kind of accident.

For example, Intel should be the most successful. At that time, Gordon Moore and Robert Noyce were both scientists in the laboratory. They relied on invention and technology to create Intel. There is also a professor at MIT, Irwin Jacobs, who invented some very important parts of CDMA and later founded Qualcomm. Kaifu’s friend Rodney Brooks and I were also professors at MIT, and later founded the sweeping robot iRobot.

There are good examples, but in most cases, doing scientific research is different from doing business. In addition, entrepreneurship must be focused. Scientists with ideas can actually work for a few years in the company before returning to academia.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Li Gen : So who does Kaifu think is the more successful industrial and entrepreneurial scientist?

Kai-Fu Li : Actually, I think the most successful is not a scientist. The most successful should be Steve Jobs. He subverted seven industries, including PC, mobile phone, music, animation and so on.

His characteristic is to be able to identify particularly good technology, and then use it to solve real problems, or to create a future he can imagine, and then use the technology according to the future shape of the product in his mind.

He may not have invented any technology himself, but he has the ability to recognize heroes with insight.

So I think if there are more people like Jobs in this society who can match up with scientists, not only can they think of what users want today, but also what users want in the future, I think this example is even better. good. As a scientist, Yaqin gave many examples just now, and I agree.

Bottom logic of entrepreneurship: technological breakthroughs will bring industry disruption

Li Gen : Kaifu has worked with many people and experienced several waves of entrepreneurship and innovation. Is there any underlying logic or fundamental law in these wave cycles? Do you think you can follow?

Li Kaifu : Actually, there are. First, if you study an industry, you can basically draw a line and see the trend of technological maturity.

For example, computer vision, its recognition rate has improved very slowly over the years, but there was a leap in improvement from 2011 to 2012. At this time, it gives us a powerful hint: the time to invest/start a business has come.

Natural language processing is also the same. There has been little progress over the years, but after deep learning, especially the recent self-supervised learning, you can see a particularly big improvement in 2018 and 2019.

Second, we can see the successful scientific research results. AlphaFold2 is a good example: Now that such a difficult problem has been solved, can’t we find a way to target and find other solutions for good small molecules or large molecules?

The third method is to project from one industry to another. For example, in the PC era, there are tools, then games, then social, then e-commerce, and then multimedia.

According to this set of logic, when the mobile Internet is up, project it over, in fact, similar conclusions can be obtained.This is why the innovation workshop invested in tools in the early stage, such as Meitu, and later in social networking, such as Zhihu and so on.

I think these logics are not difficult. Everyone can use these three methods. I think there are more other ways to calculate the maturity of future technology and how quickly it can become a usable product.

Li Gen : I don’t know if Yaqin shared anything about the underlying logic. For the three events just mentioned, are there any fundamental laws or underlying logic?

Zhang Yaqin : Let me give an example to summarize the logic.

Just now Kaifu gave examples of computer vision and natural language. In fact, the same goes for the industries I work in, such as music and video.

I don’t know if you remember that Sony has a product called Walkman, which everyone uses. Suddenly, music became digital, MP3 appeared, digital music encoding and compression technology matured, and hard disk storage technology was also mature at that time, the three things together formed a cycle of innovation. You can see that the entire music market industry was overturned at that time.

Video is the same. In the past, the technology and industrial form of analog video has been subverted after the emergence of digital video.

At the same time, we have seen the progress of deep learning, forming a perfect storm. The combination of breakthroughs in data, computing power, and algorithms has disrupted many industries, so I think this is a deep logic.

Li Gen : So if you choose a new field/new track that you are most optimistic about this year, how would you choose?

Zhang Yaqin : One is the intersection of AI, biological computing and life sciences. The other is driverless and smart transportation. There is also IoT, especially the recent application of IoT technology in the new areas of dual carbon.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Autonomous driving is an interesting industry whether it is L4 or L5. It has been more than 100 years since Ford invented the Model T 100, and it has basically followed the same underlying technology: fuel engine, gearbox, clutch, and wire-controlled chassis, but now it has really been subverted.

When talking about automobiles, we are talking about battery technology, electronic control technology, intelligent decision-making, artificial intelligence, software, and chips. The elements of technology are undergoing major changes.

So this industry is a new industry, not mechanical engineering, but smart industry. I am very optimistic about it.

Li Kaifu : I will talk about 3, and I will talk about 1 in depth.

First, I agree with Zhang Yaqin, in the fields of AI, pharmaceuticals and general health.

The second is the field of robotics, especially in industrial manufacturing. As the world’s factory, China is in great need of technical solutions that can reduce costs to continue to maintain our leading position in the world.

The third is natural language.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Throwing an article and a book today allows AI to answer questions, and the effect can already surpass humans, just like letting AI recognize things and animals in photos. The reason is very simple. Once AI surpasses humans, it can do many things that assist humans, and it can also do many tasks that replace humans.

One characteristic of natural language is that its understanding is different from that of human beings, but the answer is that it surpasses human beings. This kind of understanding can be projected in various application fields, so speech recognition will become more accurate, and machine translation will be more accurate. It is even conceivable that a few years later, when you go abroad and put a headset in your ears, you can complete the communication in real time. Translation work.

Further down, you can look forward to answering questions. Answering questions is often what the boss wants employees to do for help.

If intelligence surpasses employees, they will first become their assistants, and then they will replace them.

This will also be reflected in search engines, from a keyword that brings a lot of websites to entering a question to get the answer immediately. In the field of advertising, there are thousands of people, and you can make a targeted advertisement for everyone, so that you can’t stop wanting to buy this product. Even for everyone, there are thousands of people to create an article, a poem, a short video…

These of course need 5-10 years to mature. During this period, I feel that whether it is starting a business or continuing to advance technology, there are great opportunities, such as how to not make the model so large-not every startup company can buy a very expensive machine. There are also opportunities in architecture.

Quick question and answer

Li Gen : What is the biggest technological progress this year?

Li Kaifu : AlphaFold2 decodes 98% of the protein.

Zhang Yaqin : I agree. In addition, there have been many advances in quantum computing. For example, the USTC team successfully developed a 62-bit programmable superconducting quantum computing processor (Zuchongzhi) this year.

Li Gen : Is Metaverse a technological breakthrough?

Kaifu Li : Metaverse is a new device in the future, just like computers and mobile phones, a huge APP, a huge scene. It will definitely come, but there will not be a particularly large number of companies or applications within 5 years, which will take longer.

Zhang Yaqin : I think we need to look at Metaverse with an open mind.

If the Metaverse is a fusion of the real world and the virtual world, but the real world is the main one, we are already doing it.

Including Kaifu’s Cosmo, which was made 25 years ago, when VRML was made, which used AR and VR technologies to connect the real world and the virtual world. These mainstream companies, such as Facebook, Microsoft, Baidu, and Tencent, are all talking about Metaverse. I think the current product form is an expansion of technology, which makes sense in itself.

If the metaverse is to create a completely virtual world and has nothing to do with the real world, I think it may have gone a little too far. In addition, there is also a lot of hype now about buying and selling real estate in Metaverse.

Really speaking, we must use an open mind to see the direction of future technological development, but many technologies are not mature and need to be continuously developed.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Li Gen : When some young entrepreneurs and scientists come to you, what questions do you like to ask?

Kai-Fu Li : I would most like to know if he has thought about starting a business, and how different is it from what he has done in the past.

I would prefer the scientist-type entrepreneurs, the diversified and the type that has been very thorough, but most people are unlikely to do it. But at least I want to hear that he will understand a lot of things he doesn’t know yet, and he has the confidence to overcome, instead of saying “I can do everything”, because science is done well, but it is only a small part of entrepreneurship.

Zhang Yaqin : Let me add another toughness. Since starting a business will encounter many setbacks, it is crucial to have the resilience to survive this period of time.

Li Gen : Do you have a core principle that guides your work and life?

Li Kaifu : Follow your own heart and do what you love to do. Talk about innovation, but don’t just think about innovation, but do useful innovation.

Zhang Yaqin : Stay with people you like and do things you like.

Li Gen : What is the next small goal in life?

Zhang Yaqin : Build Tsinghua AIR (Intelligent Industry Research Institute of Tsinghua University).

Kai-Fu Lee : I want to prove that high-tech investment returns are higher than the Internet.

Kaifu Li talks with Yaqin Zhang: Scientists need entrepreneur partners to start a business, and they look at Metaverse with an open mind | MEET2022

Li Gen : We are in an air conversation today. Yaqin is doing video and multimedia. How can we be sure that the Kaifu teacher on the opposite side is not forged by AI?

Zhang Yaqin : Look at your eyes.

Li Gen : How did Mr. Kaifu determine that the opposite of the video is the real Zhang Yaqin?

Li Kaifu : I think you two didn’t have a good haircut, especially Li Gen. With AI, the hairstyle should be better.

Li Gen : The next question is, if we have superpowers and can replicate one of the other’s abilities, which ability will the two choose to replicate the other’s?

Kai-Fu Li : I will copy Zhang Yaqin’s 12-year-old university genius, and there are a lot of beautiful sisters around me. I really hope to have such special abilities. (laugh)

Zhang Yaqin : You still don’t get fat by eating delicious food every day (laughs). Seriously, I think Kaifu has a kind of insight into the future, and it can also use very simple language to summarize very complicated things. This is an ability that I particularly admire.

Li Gen : The last question. The two are friends for many years. What makes the other person most happy during this process?

Kai-Fu Li : It must be a great experiment to find AIR, which is very difficult to do. At that time, he proposed to do this, and we all sweated for him a little bit. But now he has done this very well. Not only has he recruited a group of superb people, he has done a lot of good research, and he is now also connecting with incubation and entrepreneurship. I am especially happy for him about this matter.

Zhang Yaqin : First of all, I would like to thank Kaifu for your support and encouragement. I am now working on AIR, and I often go back to when we founded Microsoft Research Asia 23 years ago. This experience at the time has been especially helpful for me now to start a business.

Kaifu is now engaged in investment and incubation. I hope that the companies he invests in will have more IPOs. He is very happy every time I see his company’s IPO.

Li Gen : Due to time, the peak dialogue between the two has come to an end temporarily. It has always been our Qubit’s wish to invite Kaifu to talk with Yaqin, not only because of their achievements and status today, but also It is because of their choice 23 years ago that to some extent created the foundation for us to talk about the intelligent future. More importantly, they are still tirelessly cultivating talents and bringing forward advances in intelligence. Thanks again to Yaqin, and again Thanks for the reopening at the end of the video~


Posted by:CoinYuppie,Reprinted with attribution to:https://coinyuppie.com/kaifu-li-talks-with-yaqin-zhang-scientists-need-entrepreneur-partners-to-start-a-business-and-they-look-at-metaverse-with-an-open-mind-meet2022/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.

Like (0)
Donate Buy me a coffee Buy me a coffee
Previous 2022-01-03 08:10
Next 2022-01-03 22:27

Related articles