Kai-fu Lee: AI scientists must do useful innovation
NetEase smart news April 27 news, the annual GMIC conference is held today at the Beijing National Convention Center. At the Leaders Summit in the morning, Li Kaifu, founder and CEO of the Innovation Factory, made a speech entitled "Scientist Entrepreneurship in the Age of Artificial Intelligence."
Li Kaifu said that scientists and entrepreneurs have great differences between their two roles. Scientists are pursuing scientific breakthroughs and stressing rigorous work and slow work. Entrepreneurs seek commercial returns, speed and rapid iteration. Kaifai Li believes that scientists' transition in entrepreneurship will be great. Innovation is important, but it is not the most important thing. The most important thing is to make useful innovations.
Kai-fu Lee said that the Internet age is destined to be started by sea turtles because turtles have seen the rise of the Internet in China abroad. The mobile internet application is destined to be a product manager's entrepreneurial business. Because we need to iterate products quickly in this era, people who understand technology and understand users and markets become proud of the times. In the O2O era, integrating ground sales and back-office technology is something we need to measure.
In the age of artificial intelligence, Li Kaifu believed that the most essential and most needed must be AI scientists, because today's AI technology has not yet entered the mainstream, and the AI ​​platform has not yet been produced. Therefore, AI applications cannot be blown out, and only a handful of people hold AI applications. The rising scientists can start a business.
For the development of artificial intelligence, Li Kaifu said that AI expansion will go through three phases. The first phase is to use existing big data; the second phase is to collect and upload the missing data; the third is the advent of unmanned robot era, from industrial to commercial home robots. From L2 to L3, L4, to a full range of driverless driving, these three steps are probably blueprints for the next five, ten, and fifteen years. (Small 羿)
The following is a record of Kai-Fu Lee's speech at GMIC 2017:
Kai-fu Lee: Professor Hawking made a very brilliant speech. I think it is very important for human beings to have visionary scientists like Professor Hawking to help our top geniuses and the most important politicians in the world to help us plan a better future. But his "super-intelligence" and "future artificial intelligence" have crushed humanity and want to avoid this situation. I personally think that this is not an inevitable result that can be inferred from today's science. Of course, non-necessary events do not mean that we don't pay attention to it, or that there are smart people thinking about it, but I think that the most important significance of artificial intelligence for everyone here today should be the following four things:
The first artificial intelligence will create great wealth, giving humans the chance to escape from poverty for the first time. Second, we have to worry about companies that have huge artificial intelligence power and data in their hands. They use data to do evil. The third is to see that artificial intelligence will replace the work of 50% of the people (in the next 10 to 15 years), what these people do, and what more important is education. The fourth is a very important topic. What are the missions of scientists, especially artificial intelligence scientists? What opportunities do they have? Will they have to start their own business or follow Hawking to find the future of mankind?
I myself was a scientist who started a business. I used artificial intelligence very early. I remember when I applied for Dr. Carnegie Mellon, the composition was to apply for artificial intelligence + the future of our future transformation of the world, hoping to participate in this industry. When AlphaGo and Deep Blue were not available in 1986, I wrote the first article about defeating the World Othello champion. Doctoral dissertation in 1988 was the first non-speech recognition at the time and was reported in the New York Times. After I entered Apple, I could see a very cute photo here. I talked about how speech recognition will enter the mainstream on Good Morning America. It was also reported by The Wall Street Journal that the stock rose 2.5 cents on the day and later Fall back. So I am full of confidence, because I am doing all kinds of top artificial intelligence work, that artificial intelligence and other technologies, such as the virtual world is about to enter the mainstream, so from the SGA company to do an internal business. What we did at the time was that we couldn't make every page full of 3D. 3D games and animations made the web page even more exciting. Let people browse not a web page but a room, which is very similar to today's VR. The venture was very unsuccessful. The investment of 20 million U.S. dollars and 100 employees were almost completely lost. I learned some lessons from this failure and I want to share it with you.
When speaking at MIT one day, I asked every speaker to leave a sentence after the speech. The sentence I left is -- "Innovation is important, but not the most important thing. The most important thing is to be useful. Innovation." Scientists are often impressed by their own research, their own cool things, and also think that the cool things they see are what the world needs. But the fact may not be the case. We think about how cool 3D is, but we didn't think of whether 3D hardware is ready or not. What are the users' needs for the 3D world? Why do VCs invest in us? How do we make entrepreneurial businesses that make money and have economic value? I basically missed all of them. The path that entrepreneurs should take. What I did was top research. I entered the top companies. The products I made still suffered from Waterloo. What do artificial intelligence scientists think? In essence, scientists and entrepreneurs are very different. Scientists pursue scientific breakthroughs. Entrepreneurs pursue commercial returns. Scientists pay attention to rigor. Scientists pay attention to speed. Scientists have to work slowly, and entrepreneurs have to iterate quickly. These six things often run in the opposite direction. The most important of these six things was that during my reading of a doctoral or professorship, I deeply understood what was the first question I asked each time in the field of scientific research. Did anyone else do this? It is not brand new. If someone has done something to see others and see if they have added value, the added value is not as great as the value of the breakthrough. Therefore, every scientist constantly asks for innovation, and innovation is done by others who have never done before. Work, this sentence represents a sense of science, the pursuit of scientific spirit. But what does an entrepreneur, or a VC, pay more attention to? How to build a product, how to generate business value. Even in VC's investment process today, think about every team we invest in risking talents, business risks, competition risks, and risks of execution. We don’t take technology risks anymore, so we would rather look at a team. Say: This technology has already been proved, just apply it in the scene. Just two paragraphs ago, we all saw that the nature of scientists is very different from the nature of entrepreneurs and VCs. A job that is not pursued by others before it is not considered whether it has any use, but only need to make money does not want to take risks in technology. Both Usually do not come together. And when it comes to the time, scientists may develop into a less pragmatic, or less rapid, iterative, or not pursued, or do not concentrate, scientists are very clever, because of the years and years of development under such cultures and genes. Everyone has a lot of ideas. A start-up company makes a point every day and the company will die because everything is done. So STEVE 燘 LANK, the father of lean entrepreneurship, helps scientists to start a business. The conclusion is that scientists must be careful, because the topics are often out-of-the-world and there is not much market; second, the choice of topics is very different from the tune; Third, scientists are reluctant to admit It is very likely that you do not have the insight and execution to convert technical cream into value. Every scientist who wants to start a business must be sincere about whether or not he or she will face these problems. I think you want to be confident when you face everyone, but you should know for yourself whether it is a problem for you.
Of course, today, science entrepreneurs have had a best time in history and can see that in addition to today's artificial intelligence, in the blockchain, life science, high-energy television, cell expansion, gene programming, almost every field All of them are opportunities for entrepreneurship. I have absolutely nothing to suppress the meaning of scientists participating in entrepreneurship. It is only necessary to think clearly when participating in entrepreneurship. In the past, so many scientists included how to avoid the dead spots they encountered, and there are successful cases of my friends. Dr. Lee has created valuable companies, and domestic companies have seen many similar companies. But to see more of the professor's obedience to his position, the United States is a typical example of the United States, Stanford and Google founder professors did not participate, but the technology is authorized 340 million US dollars, which is a model.
When it comes to AI entrepreneurship, why do AI need AI scientists? Every time science companies have done such entrepreneurship, the Internet age is destined to be started by sea turtles because the turtles have seen the rise of the Internet abroad in China. The mobile internet application is destined to be a product manager's entrepreneurial business. Because we need to iterate products quickly in this era, people who understand technology and understand users and markets become proud of the times. In the O2O era, integrating ground sales and back-office technologies is what we need to measure. The United States Mission and Dripping are such collocations.
The age of artificial intelligence came. The AI ​​scientist must be the most important and most needed. Because today's AI technology has not yet entered the mainstream, and the AI ​​platform has not yet been produced, AI applications cannot be blown out. Only a few hands hold how to apply AI. Scientists can start a business. But these scientists have the characteristics of all the scientists just mentioned: pursuing innovation, writing papers, and not knowing the market. Most of the scientists’ entrepreneurship is doing face recognition and computer recognition in China. Dr. Zhang Hongjiang, who is here, is the originator of this area. His disciples and grandchildren are doing face recognition work. This to some extent we see that our scientists are lack of imagination, artificial intelligence has so many applications, face recognition is definitely not the best one, and scientists need a person who understands the sales of companies to match with them. Of course, it is chaotic. The two unrelated people, but it is like people like Feifei Fei, can collide with the real super unicorn AI idea. AI itself is not a consumer application, of course, BAT is very fortunate to apply to consumers, but we AI startup can not bring their own traffic, there is no use of traffic and data? So the AI ​​that was made was still applied to the enterprise, and corporate finance and medical data were available. As a result, his plan required company sales in his company and he needed to understand the AI ​​solution. This is a golden collocation to solve AI entrepreneurship. So the participation of AI scientists is very important.
The biggest breakthrough in AI was deep learning seven years ago. We can understand that deep learning is a super EXCEL table. After a lot of data is thrown in and lost a data, it can be used to make predictions, judgments, or classifications. Many faces lose to know who is who, many chessboards, and once go tournaments are lost. When you go in, you will know which move you are going to use. Many Taobao products and users are thrown in to know what you want to buy. So future AIs will surely know what you want to eat tonight, know better than you; know where you go for vacation, know better than you do; and even what type of spouse you might like to be unmarried may know better than you. This is a prophetic AI that can make very strong predictions about the future. AI is a single-domain big data-driven engine at this stage. It can be thought of as a black core and can enter various fields. AI expansion will definitely go through the following three phases.
In the first phase, the existing big data was used, and BAT was in use. Today's headlines, quick hands, drops, and US missions are all in use. In addition, the financial sector can be used. For example, the Qianbao Group of the Zhirong Group, which we invest in, releases a loan of 3 billion yuan a month because it can activate existing data and user data to make microfinance loans, as well as medical treatment. The second stage is to collect and upload the missing data, collect the faces with various cameras, collect 50 billion faces, and recognize 3 million faces at any time. This is not a human function. It is a superhuman function. The third is the arrival of the era of unmanned robots, from industrial to commercial to home robots, from 2 to 3, 4 to full-scale unmanned driving. This is our potential investment in technology. This trilogy is probably the blueprint for the next five, ten, or fifteen years. The future described by Hawking is true. It is not certain that AI will have conscious awareness, human emotions, control of humanity, our tools, whether it will re-iterate itself, re-rewrite itself, etc. These are unknown, but Known can launch these applications. After the application is launched, huge results will be produced and huge value will be generated. The state will levy taxes and replace a lot of work. These jobs can be used to retrain the laid-off people with a large amount of tax subsidies to reform education.
So the era of artificial intelligence has tremendous changes to the economy. What about 50 percent of laid-off workers? What should we do in the future? (PPT chart) Turning counterclockwise from 12 o'clock this point, red is a job that is bound to be replaced, green is a job that is being modified, yellow is a job that cannot be replaced for the time being, and shifting to the top is the hardest job to replace and can replace And not being able to replace it is easy to explain. What can replace is that big data can make a decision for an objective function and make better decisions than people. Then you can replace it. Most of the work is like this. Green means that when one day the machine can do a better diagnosis than the doctor, but the machine is cold, the doctor can package a human interface, so that the patient can get comfort, let the placebo effect start, so that the patient survives and feels better. On the right these are what artificial intelligence can't do today, including arts, anthropology, managers, decision makers, and even the largest inventors. Therefore, in the talent structure of the AI ​​era, we see that there are a large number of service-oriented talents, including Luo Ji's thoughts of stacking divisions and loving companionship. Human love cannot be replaced by machines. Upwards are people who will use artificial intelligence as a tool. For example, physicians become transitionians of AI tools and people. Then they are inventors of new technologies in each field. Then they are cross-domain workers. field. Of course, the most advanced is to invent new AI and control AI. They write people not only the smartest and the most technologically aware, but also have the greatest scientific and technological revolution.
How does AI participate? There are four ways to speak from the bottom of the downturn: First, to raise your sleeve and start your own business, this is the most difficult, because scientists are essentially innovative, not business value. Go down to find a business partner. On the top right, stay in school, continue to innovate, and delegate technology out to students or others. This is something I greatly encourage. The last one is to provide open source, release data and content. Today's AI scientists suffer from the lack of data in the hands of BAT, hoping to use more open-source methods to do their work. Therefore, we really hope to encourage scientists to see clearly that the researcher is the pure land for scientific research. Do not join the company or start a business because you can make money. Top scientists are our scarce resources. We hope that you will abide by your position and push technology to a higher level. Deep learning is only the first step of artificial intelligence. There are many opportunities in the future. Observing your own scientific research positions will also be of commercial value, such as being authorized by technology. If you decide to start a business, you also hope that you can understand that if you start your business, you need to know that customers are God. You need to know how to get the most valuable VC money to help you make up for shortcomings, such as innovation workshops, to be able to be disciplined and efficient. The solution to the problem, rather than the one-by-one question, is not to ask the question, but to solve the problem. When solving, we should pay attention to efficiency, do what you are good at, find partners, and make up for your shortcomings.
As a society, there are many things that must be done to enable professors to get benefits, reputations, and interesting startups. In terms of remuneration, let the entire research community increase the compensation of scientists. In terms of childbirth and grand prizes, such as the Turing Award and the Science Award are all very good support. Resources should give more data, not only to BAT have the largest data, professors also have the largest data. Time to think about how to solve the problem of not wasting time. Stanford uses technology to authorize the way that CMU allows professors to start a business as a way to take shares. This is a very good way. The international community, especially in China, must think hard. How to make the soil so that entrepreneurs can do what they are good at? Let scientists do what they are good at, let the organic combination of the two, and do not force every scientist must go out to start a business.
Therefore, the conclusion is that scientists are very hard-working and very important. We should support the top scientists by cherishing the national treasure. Thank you.
燑br>
Vapes Maskking,Maskking High Pro Vape,Vape Maskking Cigarettes,Maskking Vape Pod,Vape Pen Kits
Shenzhen Axiswell Technology Co., Ltd , https://www.medhealthycare.com