November 30-December 2, 2017 Entrepreneurial state 100 future leaders summit and entrepreneurial state annual meeting held in the National Convention Center, Peking University, director of artificial intelligence innovation center, founder of one of seven swordsman Baidu, coolest music founder Thunder delivered a speech titled "Artificial Intelligence and Innovation and Entrepreneurship" at the Entrepreneurial State Interstellar Camp Open Class and the 2018 Innovative China Spring Summit Project Joint Trial. The sharp points are as follows:
1, social progress, new technology not only in the preexisting industry to make it more efficient, but also gives you a greater chance. A little closer to nothing, turning some of the pent-up needs into a generally accepted and huge new opportunity.
2, artificial intelligence will be how to enter the community into the public? Very important three points, first, low skills to high skills; second, high data to low data; third, high economic efficiency to low economic benefits.
3, is a million dollars a good annual salary, or take a 10% stake in the company? AI is likely to create billions, billions, trillions of businesses, so dreams are more important.
My identity is more, but I have always insisted on the identity of engineers, I like to study technology, technological progress, and the combination of industry. In addition, I am also an entrepreneur, after graduating from Beijing University, Baidu, cool me, are constantly starting a business.
Now is the time of advent of artificial intelligence. Here I will share with you my understanding of the age of artificial intelligence, how much impact it can have on society, why it is now and in the future, and how we start a business.
Artificial intelligence on human impact in the end how much?
Let's look at some progress, first, object recognition or classification. In 2011, with the more traditional pattern recognition method, the accuracy rate of 74.8% can be achieved, that is, the error rate of 25%. The human is basically a 5% error rate, which is 5 times worse at this time. Now that deep learning has surpassed humanity, face recognition exceeds humanity, speech has recently reached or approached, and I think it will soon surpass humanity.
What is the time for business opportunities to mature? It is that this technology can be used in a certain way, and artificial intelligence is its core to replace a person in a field, or a professional skills. When it is close to the level of people, give it a little time, it will shine.
AlphoGO We all know that we sometimes watch the fun to watch, but we still have to look at the door. What does AlphoGO give us? Here we can explore two points, the first point, it evolved very fast, in less than two years, it can beat a professional two-stage player level to overcome Ke Jie, its ability to very Strong, basically you can get three wins, nine sections, if you insist on giving it a paragraph, you can think of it, this is a human can not understand the paragraph.
Second, once AI has a lot of data in one area, it quickly surpasses humanity when it finds the doorway. Recently, face recognition is rapidly landing, artificial intelligence once the data, the progress will be very fast, very strong.
The impact of artificial intelligence on human beings, Hawking believes that may be the most significant thing in human history. We look at the industrial revolution. I am a science and engineering student. So I am more respectful of the data. A professor at the University of California uses data to study economic history. Let's take a look at how much artificial intelligence has an impact on mankind.
In the past 3,000 years, the world's per capita income was basically equivalent to labor productivity or output per capita, per capita income, per capita output, which was about equal to labor productivity. The productivity of mankind has remained almost constant over the past 3,000 years, depending on the climate and watching the weather, and will be catastrophic in the event of a disaster or war. Therefore, it is a huge turning point until the industrial revolution. The mankind has seen a 10-fold improvement in labor productivity per capita in the past 200 years and has been practicing progress in the past 3,000 years. This is a huge impetus to the advancement of mankind by the Industrial Revolution.
Artificial intelligence in the end can do?
Artificial intelligence changed what human beings? From a social point of view, artificial intelligence, we say it smarter, then what has changed in the end? Agricultural society to industrial society, the most important change is what? Because energy and machinery work together to replace mankind's manual labor, thus liberating mankind from the most inefficient manual labor and turning it into skilled labor. So we start to create new things with our own wisdom and learn to master this stuff. Therefore, education now goes hand in hand with the industrial revolution. Before the Industrial Revolution, there was virtually no modern education. Previous education was a private school in China and mainly aimed at training officials. In the United States and Europe like the past, Oxford, Cambridge University, is a circle of interest in the rich community.
Let's see what intelligence does? Intelligence, coupled with intelligent computing through data, will replace human skilled labor, which today accounts for more than 90% of our total human workforce. When we think seriously and go out to look at the mortal beings in our society, most of them are skilled workers. Whether it is a driver driving a car, a waiter, a chef, a doctor, a lawyer, most people do skilled labor and only a handful of people Including scientific research. In human labor, most of the labor up to now is still repetitive mental work. We call skilled labor and become more and more skilled through constant training and accumulation, making things more and more efficient.
But unfortunately, this thing will become worthless after the advent of artificial intelligence. We watch auto-driving cars. In early 2017, I talked to many people and thought it would take three years to get there. Baidu should mass-produce a passenger car that drives under certain conditions on a small scale next year. Europe and the United States have also gone very fast in this regard. This opportunity is an opportunity between China and the United States.
So we can see that there is another 10 to 20 years of career as a driver, which will completely disappear from human history. Another example is the parking garage toll collector, before the door issued a card, brush out about the charges, and now license plate recognition has basically been automated, especially in places where the community started, the future of this position will disappear. Security, with ubiquitous cameras and artificial intelligence screening, the industry will disappear from human history.
We are rapidly changing because science and technology, because of the data, because of the superposition of computing power and other causes social progress in the rapid progress of the intellectual age will have a very far-reaching impact on society, it will humankind from the now well-known skills and labor forced innovation . Innovation, as a whole, is a thing that does not exist in the world. This kind of thing can be a product, a service, a business model, or a social structure that will have a bearing on social development To play a catalytic role, this is an effective innovation.
Let us take a look at the distribution and creation of wealth that is most relevant to us. In the agricultural society, the main wealth of this society is agricultural wealth and agricultural society, which is generally the number of animals and how much land it is. This is the most important part of wealth form. To the industrial age, agriculture as the most important industry in agricultural society, the main form of society accounted for 90% of the labor force, what happened in industrial society? Is it really agricultural death that the proportion of U.S. agriculture in less than 3% of GDP has changed from 90% in the past to 3%? No, these industries are still growing, but you are running, other industries are sitting rocket forward.
With the change of the form of labor, skilled labor becomes the main form, so the wealth that is in contrast with the skilled labor has become the dominant form of society. Both industry and service belong to skilled labor, including agriculture into skilled labor. So most of the wealth in this society is out of the ordinary, and today we see artificial intelligence, including we think back to the Internet age, as Alibaba this company has done a very large, it is a product of industrial upgrading, he is retail, but You think about it, what is Tencent industry upgrade? What is Google and Baidu industry upgrade? Can you find the benchmark? Not easy to find. why? Social progress, new technology not only makes it more efficient in the industries that existed before, but also gives you a greater chance of turning some of the previously pent-up needs into a generally accepted one that will then lead to a huge new birth opportunity.
This is also what we need to think in the business. A leap from the industrial age to the intellectual age will reshuffle the wealth. The social wealth you see today will add up to the total market capitalization of these listed companies. In the next 50 to 100 years, the corresponding forms of these wealth will be reduced to not To 10%, the other 90% of the wealth, will be a new form, this is a great opportunity. Artificial intelligence will open a whole new era for mankind and will create countless opportunities for us. If we can grasp these opportunities, we will be able to create great companies.
Which industries will be replaced by artificial intelligence?
Artificial intelligence is not something new. Deeper neural networks existed decades ago, but can shine today, with the core of the rise of deep neural networks or the rise of a series of algorithms. Why is this? Recall that we learned the most famous formulas in physics, the formula of mass energy. When this formula was first proposed, it is theoretically deduced that there is no other use. When the first atomic bomb exploded, we did not really know how powerful it was. Deep learning is the same, it requires data, requires more computing power.
Let's take a look at past developments. Human data is increasing at 50%. The computing power of human beings is almost matched by improvement, doubling every 18 months to two years. These things are very scary, and our minds are seldom indexed, up 50% a year. Some people say that sometimes I invest 100% a year or so. The problem is that you can not always be one hundred percent, and you can earn 50% a year. You are more cattle than Buffett.
What if it doubles every 18 months? How many times will it grow after 15 years? Is 1000 times, it is very scary, 100 million times in 30 years, 1 billion times in 45 years, and this is the speed. Today's computing power is not enough today's data is not enough, as long as the line is on the line, one day is up to achieve.
Therefore, we see that artificial intelligence will have the most profound impact on all industries in the future. Why? If it replaces skilled labor, the main cost of almost all industries, except for the energy industry, is the human cost. Why so many people concerned about self-driving car? Have you ever wondered how much money you pay when it comes to a taxi? About driver-related can account for half. You say that I bought this car. In the production process of this car, what is the value of the heap of scrap iron and the cost of manpower behind it? When these costs are gradually realized by artificial intelligence, the cost will be greatly reduced, so you will use Lower prices get better service. This is an irreversible process of development, but there are also far-reaching social problems behind it.
We see the robot, this is a big area, the entire industry will be reshaped. We may have seen Shandong dumplings assembly line, no one from start to finish, done quite good. Another example is the autopilot, it will change the traffic, transportation, logistics three industries, these three major industries accounted for about 20% of GDP, if this is changed, what does it mean? As the operating system of the three industries, how much is the company worth? This contains the possibility of birth than Alibaba also larger company opportunities, which is why everyone is investing, entrepreneurship is a very big hot spot.
Medical, US GDP18%, China 6%, is also a huge industry. In the field of medical imaging, we can say with certainty that in the next five years, radiologists will certainly be no more than computer vision without incurable diseases. Specialist supplementary diagnosis and treatment to automatic diagnosis and treatment, this path should go in the next 10 years, these things are in rapid development, it certainly hidden huge business opportunities, of course, also contains a huge human progress. Once the virtual doctor can be done, we think that if a virtual doctor can reach the rank of associate director, what happens to society? What happens to the business model? It is crucial that the director of Kyowah can see the number of patients a day. An old expert can see up to dozens of patients a day, but how many patients can an AI physician see per day? Have you thought about it? It will be an unlimited amount of patients, not enough machines on the line, just as you search today is the same reason, with you on Alibaba to buy things is the same, afraid of many? Just add the machine on it.
So the final arrival of artificial intelligence medical treatment and the real transformation into a product for public use will change the status of human medical care for the first time and it will make a great progress. We call medical treatment fair. Since then, both Beijing and impoverished mountain children , At least in this case of medical treatment, is likely to be more fair.
So any great company in it actually makes a great contribution to mankind. If you think carefully, you can get a similar conclusion, Google, Baidu, a search engine company has a very great contribution to mankind.
Financial is also, looking to the future, the top 5 companies in the market are now all Internet-related, including Apple and Microsoft, which are also rapidly becoming Internet-based. Therefore, we infer that in the next 20 years, the five largest global market capitalization companies may not only have 5 companies, but will all be artificial intelligence companies. Some modern ones may have survived, but more are still new ones. The chances are huge.
Artificial intelligence entrepreneurship what kind of opportunity?
How artificial intelligence will enter society? Into the public? Very important three points, first, low skills to high skills. Second, high data to low data. You put a genius there, you do not let him read a medical book, you let him see the patient does not make sense. Intelligence is achieved, but there is no data it really no way. Automated driving is also due to lack of data. In the past, many industries were connected to data consumption. Therefore, it is a natural accumulation of data, but many industries do not accumulate data, so we need to accumulate from scratch.
Third, high economic efficiency to low economic efficiency, where money more natural where to go. Like autopilot, 19% of China's GDP is well worth it, and a single technology can win this market. Therefore, these three articles will determine the basic path of artificial intelligence landing. Whether there is any money or not, is technology not difficult to solve with relatively simple technology?
This is a golden age of entrepreneurship, a great opportunity, but this is related to start-ups? Better today, that there are still opportunities for start-ups. If put into a year ago, are concerns, take a look at the above we talk about, the first talent here there are people who want to recruit artificial intelligence, do you know recruit a Peking University Tsinghua graduation, with the AI-related PhD, How much do you have to pay? No 500,000 can not get it. If this person sent a better essay, R & D ability is very good, done practical things, basically have to run millions, this is the status quo, really expensive.
The second thing data, BAT is the data, our start-ups where the data? No data how to do? No one can recruit people yet.
So are these challenges ahead of us, are entrepreneurs still having a show? How successful start-up company?
I used to go to the New Oriental class before. I think the remark from Teacher Yu Minhong though it feels like a little chicken soup, is right: looking for hope from despair, life will be brilliant. In fact, all the challenges are opportunities, a hero in troubled times, behind every advantage, there is a huge disadvantage, are double-edged sword. Talents are expensive, yes, but do people really want to work? Today you see the leader of these start-up companies are a few years ago in Baidu, Google artificial intelligence research, to a certain time he came out of business, he pulled a few small partners, to give enough shares on the line. Is a million dollars a good annual salary, or take a 10% stake in the company? AI is likely to create billions, billions, trillions of businesses, so dreams are more important.
Big companies do have a lot of data, but do not forget, they have data that's all about application scenarios. Baidu search data, Ali sales data, does have, but it also has a lot of places no data, he has automatic driving data? Medical data? Agricultural data? There are a lot of places that are not there yet, which is at least fair. Therefore, there are still many opportunities for starting a business. In fact, today, the wave of mobile Internet, especially related to O2O, tell the truth, big companies burn but start-up companies. We also have VCs. Some donate money to you. Big companies have to consider him. Of the profits, can not be from a for-profit enterprises into loss-making enterprises.
Artificial intelligence plus transportation, artificial intelligence and medical care are all added, but be sure to focus on industries that did not exist before and solve them today. It will be a tremendous opportunity that is as good a chance as alternative. From the agricultural age to the industrial age, 90% is not the same as before, and the same thing from now to the future.
Here I give a whole new opportunity, such as smart assistant, in fact, everyone wants an assistant, your assistant will want an assistant. Why can not everyone do this because they can not afford it. So when the cost of virtual assistant becomes zero, this thing is not become a matter of flowering everywhere.
Therefore, we are in the face of a great historic opportunity. The opportunities are actually very good. I think that in the next 20 years, because of the opportunities given by artificial intelligence, society will surpass the opportunities in the Internet age and will be far-reaching for the human age than the Internet. We are in a golden age more than gold, this time if we can seize the opportunity to find the right direction, we can do one thing seriously, I think the result will be very brilliant and brilliant.