How many people are needed to develop the ChatGPT that shocked the world? Hundreds, or thousands?
The answer is: 87 people.
To be honest, I was really shocked when I first saw this number. In my impression, when I read the media reports on the artificial intelligence talent reserves of major giants, there are often hundreds or thousands of people.
Like Baidu in China, there are thousands of people in the field of artificial intelligence, and Google and Facebook abroad are similar. But it is not these giants that set off this round of AI wave, but OpenAI, which was only established in 2015. At the beginning, there were only a dozen people, and until the beginning of this year, there were only about 300 people.
This makes me very interested in this team - what is the reason for such a small team to detonate this AI revolution? What kind of people are they?
01 The power of top talents
Physicist Landau once said: A first-rate physicist is 10 times as good as a second-rate physicist.
Looking back at the history of science and technology, we must admit that on the way to explore the unknown world, the most critical breakthroughs are often made by a few top talents, and the same is true in the field of artificial intelligence.
Although this company is not big, it is full of great gods-this is my biggest feeling after reading a report about OpenAI employees.
Let's take a look at the core personnel of OpenAI.
First up is CEO Sam Altman.
This buddy is young, born in Chicago in 1985, he is a legend: he can program at the age of 8, he started to create a company in high school, started a business after dropping out of Stanford University, and sold his company for 43 million US dollars at the age of 27.
In 2015, Altman held a dinner on artificial intelligence, discussing the current situation in the field of AI, how far away from human-level AI and other topics, and Musk, who has always been interested in AI, was also invited to participate.
Altman, like Musk, is an idealistic person. For AI technology, the two have the same idea and original intention: to build safe artificial intelligence to benefit mankind.
Shortly after that dinner, OpenAI was born. In order to ensure the "beneficiality" of AI research, several founders believe that: first, it must be a non-profit organization, so as not to have conflicts of interest affecting the AI mission; second, it must remain at the forefront of research.
It can be said that Sam Altman is the key to the OpenAI company from 0 to 1. It is this man who has pulled up a team and gathered a group of like-minded talents to set the tone for OpenAI's development direction and strategy.
However, Sam Altman's work focuses on the management level. From a technical point of view, the core figure in the development of ChatGPT is another technical master-Ilya Sutskever, the chief scientist of OpenAI, and one of the co-founders of the company.
Ilya Sutskever was born in Russia in 1985, before the disintegration of the Soviet Union, and his family lived in St. Petersburg. During the turbulent years following the collapse of the Soviet Union, the family immigrated to Israel and then to Toronto, Canada.
Sutskever had a strong interest in computers and mathematics since he was a child, and later studied computer science at the University of Toronto. During his college years, he participated in several international machine learning competitions. It was also at that time that he met his good friend Alex Krizhevsky (who is also Russian by name), another great man in the field of artificial intelligence.
After finishing his Ph.D., Sutskever went to work at Google, participated in the famous "Google Brain" project, and developed a neural machine system that drives Google Translate. He has also been selected as a technological elite under the age of 35 by the "MIT Technology Review", and he is a popular fried chicken in the field of artificial intelligence.
Such a great god is obviously not willing to work for others for the rest of his life.
In 2015, he co-founded OpenAI with Sam Altman, Greg Brockman and others. At that time, there were only about 10 employees. According to our standards, it was a veritable "small and micro company".
Several other founders of OpenAI once commented on Ilya Sutskever:
"Ilya was the source of the technical fundamentals, a clear-headed technologist with a wide range of knowledge and vision, and always able to get down to the nuts and bolts of system limitations and functionality."
Third, introduce OpenAI's current Chief Technology Officer (CTO), Mira Murati.
This is a very young woman, born in 1988, which is relatively rare in the artificial intelligence circle.
Like Ilya Sutskever, Mira Murati is not a native American.
She was born in Albania (the Eastern European country that often produces criminal groups in Hollywood movies), immigrated to Canada at the age of 16, and later went to Dartmouth College, one of the Ivy League schools in the United States, where she studied mechanical engineering. major.
Mira Murati worked at Tesla for three years, then joined OpenAI, and quickly became the chief technology officer. She is in charge of overseeing the development of ChatGPT. Although she is not a computer professional, she seems to be a very capable character.
In the core management, there is also a co-founder Greg Brockman, who is also a guy who has been passionate about entrepreneurship since college. He was not originally an expert in the field of AI, but he had a similar taste with Sam Altman and was pulled into the founding team.
It seems that the roles of the two of them in OpenAI are similar, mainly at the organizational management level, responsible for recruiting and expanding the team, formulating company strategy, corporate culture and so on.
These four people are probably the most critical souls of OpenAI. They have laid the foundation for OpenAI's success from two levels of organizational culture and technology.
Looking at it, the United States deserves to be ridiculed as a "lighthouse country", and it is still far ahead in attracting top talents. Two of these key players are new immigrants from other countries.
Here, by the way, Russia.
Russians are quite talented in the fields of mathematics and computers, and many geniuses emerged. However, after the disintegration of the Soviet Union, many went to the United States to contribute to the construction of American empire. For example, Sergey Brin, one of the founders of Google, is a Russian immigrant. Ren Zhengfei The old man also often let Huawei go to Russia to snatch people.
In addition, Chinese descent accounts for about 10% of ChatGPT’s Chinese population. They are basically graduated from top universities such as Stanford, Tsingbei, Huake, and the University of Hong Kong.
02 The power of youth
Another thing that impressed me about the OpenAI team is that it is young.
The average age is only 32 years old, and even the core management including the CEO are born in the 1980s. In the team that developed ChatGPT, there are 28 people under the age of 30, and most of them are under the age of 40.
In our traditional impression, a major scientific and technological breakthrough is often "led by academicians, supported by special national projects, and coordinated by multiple departments", etc. It is hard to imagine that a group of youngsters in their early 30s can make a big move to change the world.
This kind of reminds me of the era of the explosion of physics in the first half of the 20th century. Einstein proposed the theory of relativity at the age of 26, Marie Curie discovered polonium and radium at the age of 30, and Yang Zhenning won the Nobel Prize in Physics at the age of 35... The building that changed the physical world , actually a group of young people.
In the field of engineering, such examples are not uncommon. For example, in Tesla, a few Stanford college students had a sudden whim in the garage, whether it was possible to connect thousands of lithium batteries in series to drive a car. They assembled prototype cars and tested them by themselves, and brought about the birth of Tesla step by step.
In fact, I have seen discussions on this aspect on the Internet in recent years: the age when a person is most creative and active in thinking is actually 30-40 years old, but it is a pity that in reality, people who hold the right to speak However, it is often a group of seniors who have already established their status in the world.
An ideal place like OpenAI, even in Silicon Valley, is often hard to come by.
03Values and sense of mission
Corporate culture and values sound like a very empty thing, especially in our place where "the higher the level of dedication, the narrower the ratio of treatment", everyone will have a headache when they hear the boss talk about corporate culture.
But value is really an intangible and powerful thing.
The main founders of OpenAI are all great gods who have achieved financial freedom at a young age. They decided to devote themselves to the development of artificial intelligence. Their original intention was really to develop AI technology that benefits mankind. To disaster, commercialization and making money is not the starting point.
In contrast, Li Yanhong happily declared to reporters that "80,000 customers are waiting in line to cooperate with Baidu" after Baidu released "Wen Xin Yi Yan", which is a bit low-level.
In the early stage, OpenAI has always been positioned as a non-profit organization, and it is regarded as a technological paradise by many young people who are interested in AI. Here, they can directly participate in the most cutting-edge and creative AI projects, and devote themselves to technological innovation without distraction.
Some people say that the victory of OpenAI is the victory of a group of technical idealists, which makes sense. Perhaps because of this, ChatGPT was not first born in giants such as Google, Microsoft, and Meta, but from a more "pure" small team.
However, the Silicon Valley giants are indeed the Whampoa Military Academy for cultivating talents. The top three employees of OpenAI are Google, Meta and Apple.
03 Population and talent
To be honest, only a few years ago, I was optimistic about the narrowing of the gap between China and the United States in the field of artificial intelligence.
After all, China's population base is there, and the number of science and engineering talents trained every year is several times that of the United States, and the user scale is much larger than that of the United States. It requires talents, data and data, not to mention money, no matter how you look at it. Everything is ready.
But now it seems that population and talent are not necessarily strongly related.
We have indeed cultivated a large number of science and engineering talents, and have the so-called "engineer bonus" that many countries envy. However, in the field of cutting-edge technology, the most difficult breakthroughs may require one or two top talents.
Without Jobs, there would be no Apple Empire today; without Musk, there would be no Tesla and SpaceX.
The hottest AI topics in recent years, whether it is AlphaGo, unmanned driving or ChatGPT, which defeated humans for the first time, are always the Americans who detonate it first, and then we follow up quickly. But blindly following up behind other people's buttocks, I am afraid that we will never be able to catch up.
This goes back to the famous "Qian Xuesen's question": "Why can't we always cultivate outstanding talents who are leaders?"