GuShiio.com Article source: Shenwang Tencent News
Image source: Generated by AI
Behind Deepseek’s full swing throughout the Spring Festival is a group of “enigmatic wizards”(commented Jack Clark, former policy director at OpenAI).
Liang Wenfeng, founder of DeepSeek, once gave a rough portrait of this group of employees in an interview with 36Krypton:”They are all recent graduates of some Top universities, doctoral students and doctoral students who have not graduated, and some have only graduated. Young people of a few years. Nowadays, famous schools and high academic qualifications have become standard for practitioners in the AI circle.
The “2024 China Artificial Intelligence Entrepreneurs Portrait Report” shows that AI entrepreneurs are characterized by high academic qualifications and youth, with master’s degrees or above accounting for more than 70%. They come from more than 400 universities around the world, among which Tsinghua University and Peking University lead.
In addition, the post-95s generation in the AI crowd stood center stage. The “Global Artificial Intelligence Talent Tracking Survey Report” released by the McRoboluo think tank in the United States in 2024 pointed out that among the more than 30 high-profile Chinese start-up AI companies in the United States, half are entrepreneurs born in the 1995 era, including “talented girls” Guo Wenjing, Alexander Wang, who is worth more than US$1 billion, and others.
In China, Luo Fuli, key developer of DeepSeek-V2, Zeng Guoyang, CTO of Wall Intelligence, and Jiang Yuchen, founder of Wave Intelligence, are all born in the 1995s.
“China’s working-age post-95s population is about 200 million (aged 17-29). Young talents born in the 1995s grew up in an era of rapid economic growth. They live a prosperous life and are more inclined to balance work development and enjoy life.” The report “Talent Attraction Ranking of Cities in China: 2024” pointed out.
It can be seen that it is no accident that the post-95s generation can become the backbone and move to the center of the stage in the wave of big models. This is closely related to the tide of the times and personal choices.
Talented teenagers in the spotlight
Data provided by iMedia Research shows that the size of China’s AI model market will reach approximately 29.416 billion yuan in 2024, and it is expected to exceed 70 billion yuan by 2026, with strong development momentum.
The 2024 “Government Work Report” proposes to carry out the “Artificial Intelligence +” action, emphasizing the importance of AI in the digital economy and accelerating the realization of high-level technological self-reliance and self-reliance. Big models and AI are the star tracks of the moment, and it is the right time for the players in the past two years.
As we all know, AI is an industry with a high threshold, and entrants need absolute “hard power.” Throughout recent years, most of the young people who have approached the spotlight of the AI model circle have outstanding academic qualifications and strong technical strength.
Some time ago, it was rumored online that Luo Fuli, who Xiaomi paid tens of millions of yuan to hire talents, was a star in the big model circle. He graduated from Beijing Normal University with a bachelor’s degree majoring in computer science, and later graduated from the Institute of Computer Languages at Peking University.
After graduating from her master’s degree, Luo Fuli joined the Machine Intelligence Laboratory of the Aridamo Institute through the “AliStar Project”, was responsible for the AliceMind open source project, and led the development of a multi-language pre-training model VECO. The “Alibaba Star Project” is the highest level of Alibaba’s school recruitment. The average number of students admitted every year is less than 20, and the admission rate is less than 1%. In Ali’s propaganda,”every Alibaba Star is one in a hundred talents.”
As early as 2019, Luo Fuli was on the hot search list for publishing 8 top-issue papers (two of which were published) within a year. Her proposed word sense disambiguation method and text style conversion framework promoted the development of NLP technology.
In 2022, Luo Fuli left Ali and joined magic square quantization to engage in deep learning-related strategy modeling and algorithm research. Later, she moved to DeepSeek as a deep learning researcher and participated in the development of DeepSeek V2.
Luo Fuli’s fame is due to both the times and her strong professional abilities. Like her, she is also known as a “talented girl”, as is the founder of Pika Guo Wenjing. She studied in a competition class at Hangzhou No. 2 Middle School in high school. Her mother Guo Xiaoli was a top student at the Massachusetts Institute of Technology, and her father Guo Huaqiang was the actual controlling shareholder of Zhejiang Xindaya Company.
In her first year of high school, Guo Wenjing won the first prize in the National Youth Informatics Olympic League (Zhejiang Province Division). Later, she participated in the U.S. National Informatics Olympic Open for two consecutive years and won the championship.
In 2016, Guo Wenjing became the first student in Zhejiang Province to be admitted early to Harvard undergraduate courses. While studying for a PhD at Stanford University, Guo Wenjing found that various AI products on the market today were too cumbersome to use, so she resolutely decided to drop out of Stanford and founded Pika Labs, aiming to create an AI tool that makes video creation easier.
Investors in Silicon Valley were moved by the news. Within a year, Pika completed a Series B financing, with a total financing amount of US$135 million and the company’s valuation doubled to US$470 million. Half a year after the company was founded, Guo Wenjing launched Pika 1.0, which allows users to generate and edit different styles of videos such as 3D animation, animation, cartoons and movies with “zero threshold”.
On the day it was launched, the product became popular in the domestic and foreign technology circles. It also became popular, including Guo Wenjing’s father’s company, Xintaya, which had 6 daily limits in 8 days. The company’s total market value reached 8.3 billion yuan, setting a new high. To this end, the company also issued a special announcement clarifying that it had not invested in Pika, and the story of “the father depends on the daughter” is popular among people.
Not only Luo Fuli and Guo Wenjing, there are many talented teenagers in the AI circle. Zeng Guoyang, who taught himself to code at the age of 8, and who won the Olympic Games championship was sent to Tsinghua University. He was once praised as a “genius boy” by the outside world.
During college, Zeng Guoyang joined the NLP (Natural Language Processing) Laboratory of Tsinghua University and became one of the earliest large model researchers. Later, as a backbone, he joined the Wudao Wenyuan Chinese pre-domestication model team. In 2020, as a main member, he and his team developed the country’s first large Chinese model CPM-1, which aroused heated discussion in the industry.
After graduation, he and his mentor Liu Zhiyuan jointly established Wall Wall Intelligence and served as CTO. He was 24 years old this year. In May 2024, Zeng Guoyang led the team to develop the model “MiniCPM”, which was praised as the “Performance Small Steel Gun” by the industry.
They fit their times
Not only these already famous post-95s, AI-related entrepreneurial fields are also constantly pouring in new faces from the post-95s generation.
Jiang Yuchen, the founder of Waveform Intelligence, was born in 1995 and has a doctorate in artificial intelligence at the Federal University of Technology in Zurich. Less than a year after her company was founded, her company developed “Frog Writing”, an online text AI writing product driven by Weaver. One year after its launch, the product has registered more than 300,000 users, generated nearly 20 billion words of text, and paid users use it for an average of more than 6 hours a day.
There is also Song Yachen, who is working hard in the field of 3D large models. Born in 1998, he graduated from Johns Hopkins University and worked in Shangtang Technology and MiniMax. At the age of 25, he founded the 3D large model company VAST. Within a year and a half, VAST became the leading company in this field, accounting for more than 70% of the market share.
The optical standard, which focuses on photonic computing chips, was also jointly founded by two people born in the 1995s. One of the founders, Xiong Yinjiang, is a master’s degree in computing science from the University of Chicago, and another co-founder, Cheng Tangsheng, graduated from Oxford University with a Ph. D. The two of them chose AI computing chips for their entrepreneurial direction to serve AI reasoning and AI training scenarios in the era of large models.
Not only domestically, but from a global perspective, the post-95s academic talents have a tendency to become the mainstay. Scale AI, a once sensational artificial intelligence company, has become an AI start-up myth in the industry because it has received 7 rounds of financing and a valuation of 100 billion yuan. The leader behind it, Alexander Wang, was born in 1997. He began to interact with programming at the age of 5. He was admitted to the Massachusetts Institute of Technology at the age of 18. He dropped out of school during the summer vacation of his freshman year to create Scale AI.
Today, Scale AI has become the world’s leading AI data service provider. Alexander Wang himself has also become one of the world’s richest people under the age of 30, becoming one of the youngest entrepreneurial giants in Silicon Valley. At the age of 27, he is worth more than US$1 billion. Some people even describe him as the “next Musk.”
The post-95s group has been described as a generation that is confident, optimistic, creative, and courageous. They dare to express their ideas and pursuits, do not blindly follow, emphasize individual differences, and pursue a personalized attitude towards life. As “Internet aborigines”, the post-95s generation came into contact with the information society early, praised new technologies, and also had keen insight and creativity.
LinkedIn once conducted a survey of the professional attitudes of thousands of working people of different generations in first-line and “new first-line” cities. The survey found that the acceptance of emerging industries born in 1995 is higher than that born in 1980/90s, especially for high-end technology industries, including big data, cloud computing, AI, robotics and other fields.
A headhunter in the AI field told AI Lightyear that judging from the recruitment situation, corporate needs are concentrated in a small number of colleges and universities with relevant laboratories, and doctoral students in laboratories are more popular. The headhunter believes that there is currently an urgent need for algorithm pre-training talents in China, and most companies will recruit directly from core laboratories. At present, most of the top universities in China have opened artificial intelligence laboratories.
However,”academic qualifications are only the stepping stone, and ability is the iron rice bowl.” As an industry with high barriers, academic qualifications and intellectual knowledge may be only the first step. DeepSeek’s employment outlook is also in line with this. Founder Liang Wenfeng told 36Krypton that the main members of DeepSeek’s team are recent graduates of top universities and doctoral students who have not yet graduated. In addition to colleges and universities, DeepSeek also takes a fancy to candidates ‘competitive results and innovation.
Luo Fuli and Zeng Guoyang have also publicly stated that individuals like challenging things and do not like boring and repetitive work. Being in the comfort zone for a long time means that it is difficult to grow.
Zhang Peng, founder of Zhipu AI, told “AI Guangnian” that its recruitment mainly focuses on two points: learning ability and innovative thinking. “The starting point is very important, but the starting point only represents past accumulation, and the future growth space depends on learning ability.” Zhang Peng revealed that the interview will examine the speed and quality of interviewers ‘grasp of information in the technical and business fields.
The future is here. On the AI stage, the post-95s generation who has moved to center stage will create more new possibilities.
