Kai-fu Lee: I will never upload the brain, I will cooperate with AI

(Original title: Kai-fu Lee: I will never upload the brain, I will cooperate with AI)

Reprinted from: INSIDE article: Li Baifeng

Kai-fu Lee recently became a “missionary” of artificial intelligence. He not only wanted to create the “innovation workshop” he founded as the most knowledgeable artificial intelligence venture, but recently released a new book called “artificial intelligence.” Why is Lee Kai-fu so optimistic about artificial intelligence? Is he optimistic or pessimistic about the impact of artificial intelligence development?

How to determine which AI startup team to invest in?

As the VC with the most knowledge of AI, how does the innovation workshop led by Kai-fu Lee decide which artificial intelligence team to invest in? Kai-fu Lee had said in the past that 50% of the investment in innovation workshops was focused on entrepreneurial teams in the field of artificial intelligence, and continued to invest US$100 million in face recognition, microfinance and other companies.

Li Kaifu does not deny that there are many "broad-talking" startups in the field of artificial intelligence that are not ideal investment targets, so they need to be carefully selected. Asked whether the innovation workshop is using artificial intelligence to select the artificial intelligence team worth investing in, and Kai-fu Lee said frankly that there is absolutely nothing because venture capital work is the least easily replaced by artificial intelligence, so in the process of finding investment targets, It does not use artificial intelligence, but instead relies on more valuable human intelligence.

So how did you pick the team?

1. 燑/strong> Innovation Workshop does not invest in any unproven technology.

2. 燑 strong strong strong / > 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 / > 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 / > 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 即使 / > 即使 即使 即使 即使 即使

3. Is the team's talent really great? Venture capital does not expect teams to invent new technology, but at least it is necessary to understand how to master the technology.

The above are the three basic principles for investing in a startup team. However, investment is not only responsible for picking a good team, but especially in the field of artificial intelligence, which is almost always a top scientist, investment is actually very much invested in it and keeps in touch with the team. Discuss how to apply, there is no opportunity for cooperation, create more business value, and become the team's most trusted advisor.

Therefore, some artificial intelligence teams currently investing, such as face recognition technology, can try to apply to unlocking mobile phones, e-commerce services, and security applications. The microfinance team, because many people in China do not have credit cards, but it is embarrassing to borrow money from acquaintances, this demand does exist, but the bad debt rate is high. Through the application of artificial intelligence, can be used to determine whether this person can borrow his money? The bad debt rate can be lowered down. These applications are becoming more and more mature. In terms of small loans, it is now possible to reach 1.5 million single bills per month, each amounting to 1,000 yuan, and accumulating 18 billion yuan a year. This has exceeded the size of banks. .

The technology, coupled with clear business needs, can have explosive growth.

Artificial intelligence technology is mature, but it has not yet been platformed

When it comes to the development of technology, artificial intelligence is still very much emphasizing the participation of scientists. It is very different from the past that the development of Internet entrepreneurship does not attach much importance to academic or academic qualifications. Will academic qualifications be re-emphasized in Silicon Valley?

Li Kaifu believes that scientists are indeed very important now that artificial intelligence is at an early stage. After the maturity of technology, the technology can be platformized. When an engineer can open a tool box and select the right tool to use that type of labor. Smart technology does not require so many scientists, and artificial intelligence is only universal. The current mainstream deep learning technology is very mature, but it has not evolved into a platform, so it requires a lot of scientists.

The characteristic of artificial intelligence is that it will evolve itself, even to such an extent that human beings cannot read or understand. So if artificial intelligence is used to deal with whether or not to lend to someone, because the review mechanism is artificial intelligence to learn from it, was rejected and do not know why, this will not cause problems? Kai-fu Lee is optimistic that he can actually use this advantage in reverse, just because people can't read it and cannot explain it. It is also a better explanation. Through the application of artificial intelligence, some commercial secrets have also been protected. However, the reason why the loan was rejected is unknown, but it is still determined by the computer. Is it really a problem?

The closure of the seven black holes will eventually be replaced by more open ecosystems

In the process of bringing artificial intelligence into technology, open source is actually very important. Kai-fu Lee, for example, originally closed Wintel ecosystem, with the trend of action, the open source code has now become the mainstream, the emergence of Github, the open ecosystem has laid a good foundation for the development of artificial intelligence. However, in addition to code, data is also open source. However, there are currently seven black holes, namely, Google, Facebook, Microsoft and Amazon in the United States, and BAT in China. These enterprises have a great deal of information, but they go in. If you can't get out, like a black hole, this is not a good phenomenon for the development of artificial intelligence. Of course, these seven black holes are relatively more advantageous.

These seven companies are of course very large on the Internet. However, in some traditional and physical industries, the influence of these companies is actually very small. Therefore, we can expect that in the future there will be more open information systems to replace these seven. Big black hole. In fact, for scientists, these seven black holes caused a lot of problems because the development of artificial intelligence technology requires this information, but this information has not been shared, so the development of scientists is limited, which is why many scientists later If we do not go to these seven companies, we cannot continue our research. However, after all, this is still an act that people criticized as an over-protection of self-interest. When scientists do research, they have to “begging” information. Why then?

Therefore, the practice of innovation workshops is not only open-source code, but also open source materials, and cooperation with students, because many of the more senior engineers actually do not understand artificial intelligence, but now many college students have learned, in the past Stanford University machine learning There are 80 people in the course and 1000 students this year. These college students who understand artificial intelligence will be talent pools for artificial intelligence development. They will be very hungry to find platforms to develop their own talents, and we will try to provide them with opportunities.

Looking back on Taiwan’s academic circles, according to Li Kaifu’s observations, researchers are in a state of almost giving up, because there is no more information than people, and it is not like the United States or China has seven big black holes, and Taiwan does not even have black holes. Therefore, if there are Taiwanese professors or students who are interested in the development of artificial intelligence, they can develop more or participate in the open source ecosystem.

The future development of artificial intelligence has three stages

Regarding the development of artificial intelligence, Li Kaifu mentioned that it will go through three phases:

1. First, apply existing data.

2. The next step is to develop new applications by collecting new data with more new sensors and hardware.

3. The final step is full automation.

The time for these three phases will probably be the next five years, ten years and fifteen years. The beginning of the application of artificial intelligence is the technology industry, such as Google's search has actually been applied for a long time, and now is the technology company if you do not use artificial intelligence technology, it is a bit behind the times, there is no data, isn't it strange?

In addition to Internet companies, who else has the data? The financial industry, and the financial industry is born with artificial intelligence. It is a data-based industry. No matter whether it is banking, insurance, loans, credit, or investment, it is all about data processing.

Kai-fu Lee is such a look, if an industry has a large amount of data, data with a label and is a single area, it is the most ideal data for artificial intelligence applications.

The financial industry is in line with these three standards. And the financial industry is very rich. Entrepreneurs will swarm into this industry and make good use of artificial intelligence.

The next step is medicine. Medicine also has a lot of data. It's like imagery is data. If you have friends in the radiology department, then you have to make a change. Looking at X-rays, nuclear magnetic resonance, and tomography, we needed to accumulate experience in the past. However, how can anyone's experience be inferior to artificial intelligence through the entire database? Facial recognition, the most powerful police in the tradition, knows who is not a fugitive at a glance, but now it is setting up a camera at the intersection. Which policeman can do better than artificial intelligence? 24 hours without rest, complete database, fast identification and low error rate. In addition to images, DNA is also a kind of data. The processing of these data has a breakthrough after the maturity of deep learning, and it will also be the focus of future development. Next, it can develop more sophisticated and customized through DNA. Medical.

Even doctors have to change their job content

So don’t look at the doctor. Now it’s a good job. Artificial intelligence can do better than doctors in some things. For example, people’s hands will shake, the machine will not cut, people’s experience will be limited, and the machine’s experience will only be used by the database. Size limitations. Next, the doctor is not a doctor, it is rather a bridge or "interface" between the machine and the patient. Actually, this is not a bad thing. Now that many doctors have to look at the doctor's office and have to do research while working hard. In the future, when artificial intelligence replaces most of the work of a doctor's office, doctors can do more research and medicine will have a faster breakthrough. Now.

When it comes to medical and artificial intelligence applications, it is reminiscent of IBM's Watson who also took part in medical research on cancer, but what's interesting is that Lee Kai-fu talked about the development of many artificial intelligence, but he seems to talk about IBM? Li Kaifu believes that IBM's artificial intelligence is not based on deep learning technology that has made significant progress. In the past, IBM did pick a few specific areas to provide their services and earn revenue. In fact, if you really want to think about what top scientists with artificial intelligence are serving at IBM? What breakthrough technology? What terrible data? It does not seem to have.

Uploading the brain is a manifestation of vanity

So does Kai-fu Lee want to upload his brain? Li Kaifu believes that people's lives will only make sense when they are finished. Uploading the brain is a kind of vanity. They feel that they are great and they must stay. Think of the world as physics, appearance, spirituality, and loving existence. Let's explore it. If it is a physical world, is every brain worth saving? We will soon be defeated by artificial intelligence! And if the world has only one side of physics and appearance, then there is no meaning to be alive. When it is dead, it is dead! And if it is a spiritual, loving world, can these upload? Therefore, Li Kaifu prefers to maintain a human being who is still self-aware and does not want to upload his own brain.

Don't want to upload your own brain. Does it work together with AI to create better performance than in the past? Kai-fu Lee believes that many people believe that the cooperation between humans and artificial intelligence is 1 + 1 = 3, but it should actually be 1 + 100 million = 100 million and 1.1. In most areas, human beings are very small and artificial intelligence is very large, so all people must know how to apply artificial intelligence in order to get things done.

Artificial intelligence came after

Of course, there are still many limitations to artificial intelligence today. For example, scientists are trying hard to make artificial intelligence not require a large amount of data, or artificial intelligence can only deal with problems in a single area, and it will be stuck in cross-domain. Therefore, there are many things that can be done and worth doing.

First of all, artificial intelligence replaces the work of human beings is certainly far more than the work of creation, but some of the work that was neglected in the past will now be gradually re-emphasized, like the so-called liberal arts students, fields of philosophy, sociology, anthropology, art, etc. AI can't do these things, it can change these areas. It is not that these areas will make good use of artificial intelligence, but that artificial intelligence will replace a lot of manpower and that more manpower will have to develop in these areas. It is not a bad thing.

Second, careers will change. In the past, doctors were the main force in the medical process, but then it may become a bridge between patients and artificial intelligence, or switch to do research. Like a teacher, it may no longer be teaching but student care. Therefore, the service industry and volunteer trade unions are the main tasks for humanity in the future. Kai-fu Lee believes that human beings are born in the world. Shouldn’t they produce more psychological collisions and exchanges? Now that we are all busy at work, artificial intelligence just released humans. It is actually a good thing.

Universal Vacuum Cleaner Dc Motor

Universal Vacuum Cleaner Dc Motor,Dc Motor Vacuum Cleaner,Dc Brushed Motor,Motor For Vacuum Cleaner

Zhoushan Chenguang Electric Appliance Co., Ltd. , https://www.vacuum-cleaner-motors.com