When we discuss the impact of AI models on workers, we can't help worrying that some occupations will be replaced by artificial intelligence in the future. However, it is undeniable that AI is "destroying" some jobs, but also "creating" some jobs.

In April, a company in Hangzhou for 3.2 million annual salary recruitment of visual algorithm experts on the microblogging hot search, become the best footnote of the shortage of AI talent.

The reporter of "Science and Technology Board Daily" interviewed a number of large model companies, as well as industry headhunters and job seekers, and a personal feeling is that the market demand for AI talents has been rising in waves since this year.

Some job seekers believe that this is a rare era of dividends, firm "All in" AI; there are also people who worry that this time will be replicated when the "AI four little dragons" of the high start and low walk, so choose to wait and see.

It is necessary to point out that, between entering the large model industry and the rising salary, there is not an inevitable relationship. The technical characteristics of large models determine the strongest demand for senior algorithmic technical talent, the highest salary can be more than a million, but the middle and lower levels of algorithmic engineers can often only be flat salary or even pay cuts to jump ship.

One side is lively, the other side is cold, this is the current situation of the AI talent market.

Some people All in, some people wait and see

At the end of April this year, Song Goose (fancy name) left Tencent, where he had worked for two and a half years. Two days later, he immediately joined a Beijing AI big model star startup.

The rapidity of the action, the decision to choose, unusual. Moreover, he moved his family from Shenzhen to Beijing, which increased the shock and curiosity of his friends. One friend commented, "I've never seen you so sharp in my life".

As for the reason for jumping ship, SongGoose wrote in WeChat public number "SongGoose has something to say": "Aspire to realize AGI (General Artificial Intelligence)".

"I think AI will really affect the next ten years at least, this is not a hanging in the air completely did not land the technology, this is a thing that can immediately change the world. In the next ten years, all industries will bring no small changes because of AI." Song goose said.

Pine Goose's choice of career choice, but also just a microcosm of the current development of the large model industry.

Pulse high employment talent think tank data show that this year's spring recruitment period is the peak of large model startups, the demand for talent has risen sharply. in February, the volume of new jobs issued by AIGC increased by 75.6%, and continued to rise in March, with the ringgit increase rising by 12.56%. With the end of spring recruiting and the end of the wave of entrepreneurship, the demand for talent has fallen back. in June, the big model companies wind up again, the demand for talent rose again by 4.16%.

As a large model of this "entrepreneurial pioneer", Wang Huiwen from February this year will be in its social account widely issued "hero post", many users in the comment area to leave a message to seek cooperation, ask the requirements of the job application, urging the progress of the delivery of resumes, etc., we do not want to miss this entry! The opportunity.

The big model startup wave is visible to the naked eye in attracting talent to enter, but not everyone is as determined as SongGoose All in AI.

Another big model star startup in Beijing has been recruiting technology and algorithm related talents in the first half of this year. But the company's executives told the "Science and Technology Board Daily" reporter, even if the company's founding team background hard, and in the first half of the year also got the head of the investment institutions of the hundred million yuan of financing, but the progress of recruiting "is not smooth", the reason is that the job seekers are holding a wait-and-see attitude, they feel that the situation of this round of AI is uncertain, "worried about the big model companies live, but not all of them are so determined All in AI," the company said. "Worried that the big model enterprises will not live long".

Job seekers are worried, not without reason.

AI has been proposed since 1956, and has experienced several ups and downs since then, before ChatGPT, the latest domestic AI boom is represented by the development of the "Four Little Dragons of AI", but the degree of these companies burning money and the mismatch between the output of the landing, and the inversion of the stock price of the listed companies and the promotion of the hotness of the market is doubly disappointing.

Shen Jia is a senior recruitment consultant who has been serving the talent recruitment market for a long time, and has been providing recruitment, enterprise management, investment and financing docking services for large model industry companies since August last year, and he has basically dealt with the earliest batch of large model startups in the market.

Although AI is as hot as it is today, the industry cycle is constant. Shen Jia summarized a characteristic to the reporter of "Science and Technology Board Daily", this round of AI large model startups, and the AI four dragons have a very different characteristics, that is, the former in the beginning of the venture on the commercialization of the consideration of the business growth, and the pursuit of internationalization, they try to subvert some of the mistakes of the predecessors.

Back to the candidate's perspective, Wang Fang (a pseudonym), who previously provided high-end executive search consulting services for the tech and Internet industries, has also begun to identify candidates for the big-model businesses of these companies this year. He told Sci-Tech Board Daily that now when job seekers look at opportunities in the big model industry, there is one thing they must take into consideration, and that is the big model business of this company, which must have a landing scenario, as this is the only way to ensure revenue and maintain the normal operation of the company.

On the other hand, in recent years, the macroeconomic environment and the Internet's "optimization wave", most people choose to stabilize, get up and jump ship, many people do not have so strong risk-taking power.

Wang Fang said, some job seekers think, from the current relatively stable line of business to a less familiar, the future is not clear on the large model business, "most people do not dare to go, they are afraid that the project does not land, or landing less than the boss expected, will be laid off.

Higher salaries of millions of dollars a year, lower salaries stay put

Loose Goose joined the current big model company, or the continuation of the Tencent's old profession - product manager, a highly mature and important role in the Internet industry, reference to the "father of WeChat" Zhang Xiaolong.

But Song Goose said, to big model industry "this is the sacrifice of originally clear career growth path", because he is not sure two years later, big model itself still need product managers, not to mention the career development path.

What Songoose says is true. At least for the time being, product managers are not the most in-demand type of talent in the big model industry.

Shen Jia told "Science and Technology Board Daily" reporter, from the industry development trend, from the middle of last year to the present, whether it is a large factory or startup team, are still focused on the big model to lay the foundation. Therefore, at this stage, the underlying technical talents, such as NLP algorithm engineers, deep learning and other positions, are the most urgently needed group of people in the big model talent market.

The data from the Pulse High Employment Talent Intelligence Center show that the supply and demand ratio of analog chip design, integrated circuit IC design, wind control algorithms, AIGC algorithm engineers, natural language processing and other positions is less than one.

And this kind of the tip of the pyramid of that batch of people - mainly refers to both research, but also able to output Paper (thesis), but also able to take into account a variety of application scenarios, landing the delivery of the project, is the first echelon of the first echelon of the big model vendors of the first to compete for the object.

To what extent is this talent being competed for? Wang Fang gave an example to a reporter from Sci-Tech Board Daily.

A head Internet company located in Shenzhen, now recruiting needs are mainly the world's top AI talent, its internal recruitment team, the external also hired headhunters, "the current situation is that these two teams are grabbing the same set of people, grabbing to the end of the day, we all feel tired, it's not interesting, because the supply of talent is only these, and can be persuaded to come into the job even fewer ".

Salary, from its recent services to recruiting companies and candidates in the process, the industry's top talent, the value of this year is once again get a boost, jumping the salary of the people are almost always a substantial increase.

Shen Jia also told reporters that the public is familiar with the companies in the big model business, is recruiting AI talent globally, for the top personnel to open the salary is "regardless of the cost", but Shen Jia did not tell reporters this "regardless of the cost" of the number of how big! The company's main goal is to provide the best possible service to its customers.

Recruitment platform data may give some guidance. Pulse high employment talent think tank statistics show that in the pulse platform, such as machine learning, speech recognition, AIGC algorithm engineers, natural language processing and other positions, the highest salary of more than 900,000; more core recommendation algorithms, algorithm engineers, computer vision positions, the highest are more than a million dollars a year.

But if you think, as long as into the big model of this line, the salary will naturally rise, then it is very wrong.

Wang Fang also said another industry status quo. According to their recent offer situation, the middle or bottom of the algorithm engineers, or belong to the flat salary jump, or even pay cuts, even if there is a small part of the salary increase, it will not be more than 20%.

Pulse high employment talent think tank statistics, test engineers, product managers, etc., these can be a large number of people from the Internet industry leveling over, talent supply and demand ratio are greater than 2, these people are also relatively good recruitment.

"The big model industry needs or high technology and top talent, most of the middle and lower level personnel, the salary level is not much higher, will not be because of the big model hot, their salary also due to the industry dividend and follow the rise." Wang Fang said.

"Adjust your expectations, don't dream big."

The heat of the big model is growing at an exponential rate in the whole society, and nowadays, people have been used to compare this technology application to the Internet era 20 years ago, which contains unlimited opportunities, and people expect the new BATs to grow from it, and then give birth to one after another myth of ordinary people's wealth creation.

Now a very obvious social symptom is that everyone is looking for big model opportunities - even if only a little bit, people do not want to miss the technology to bring a huge era of dividends.

As a side case of jumping from the last gold industry to the next potential gold industry, after going from the Internet to the big model industry, a very large number of friends ran to ask SongGoose if there are any reliable recommendations for good opportunities in the application layer? Should I join a big modeling company? How to evaluate whether a startup company is reliable? What if I get laid off when I enter a startup during the trial period?

After communicating with 43 friends, and in response to some questions, SongGoose wrote a 5,000-word article in WeChat public number about AGI's "Job-hopping Thoughts and Suggestions", which to a certain extent fits the current situation of the industry.

For example, to evaluate a startup company, from the team background, financing, geographical situation and other perspectives to analyze the company he is now working for, for example, Tsinghua Department, the first echelon of large models, venture capital circle star team.

Venture capital circle star team is often also a priority item for job seekers now. Wang Fang told reporters that at present, the star big model company is indeed the consideration of most job seekers, compared with other startups, the former has investment from institutions, abundant funds, the probability of problems in a short period of time is small, the team will be more stable.

Loose Goose also said that the domestic AI application layer is still a chaos, VCs seem to be lively, but not much out of the blue; seize the user, seize the scene, can be landed, there is always value; the future opportunities in the application layer, can form a very deep barriers, and now there is no point in time to explode the application layer ......

Although SongGoose is currently slightly worried about his role as a large model product manager, Shen Jia believes that the application layer of the large model will burst out with great market opportunities, and the consumption of talent here will be the largest, and as the industry moves forward, the future product manager with AI industry experience is the most valuable.

The wildly advancing wave of Chinese big model startups has recently braked sharply due to the illness of Wang Huiwen and the acquisition of Lightyear Beyond by Meituan, which has added some cold colors to the current reality of China's big models.

The industry is in its infancy, rugged and chaotic. In the face of the company's situation may occur temporary retraction and even entrepreneurial failure, loose goose only said, "at least to do a good job of a company did not pay you three months, and then closed down the minimum expected", "we recommend that we do not have too high expectations of financial returns, and do not do some big dreams ".

On the day he left Shenzhen for Beijing, Song Goose took a picture of a billboard at the Shenzhen airport, which he felt fit the current state of mind and industry environment.

It reads, "But wherever people can imagine something, someone will be able to realize it."

When we discuss the impact of AI models on workers, we can't help worrying that some occupations will be replaced by artificial intelligence in the future. However, it is undeniable that AI is "destroying" some jobs, but also "creating" some jobs.

In April, a company in Hangzhou for 3.2 million annual salary recruitment of visual algorithm experts on the microblogging hot search, become the best footnote of the shortage of AI talent.

The reporter of "Science and Technology Board Daily" interviewed a number of large model companies, as well as industry headhunters and job seekers, and a personal feeling is that the market demand for AI talents has been rising in waves since this year.

Some job seekers believe that this is a rare era of dividends, firm "All in" AI; there are also people who worry that this time will be replicated when the "AI four little dragons" of the high start and low walk, so choose to wait and see.

It is necessary to point out that, between entering the large model industry and the rising salary, there is not an inevitable relationship. The technical characteristics of large models determine the strongest demand for senior algorithmic technical talent, the highest salary can be more than a million, but the middle and lower levels of algorithmic engineers can often only be flat salary or even pay cuts to jump ship.

One side is lively, the other side is cold, this is the current situation of the AI talent market.

Some people All in, some people wait and see

At the end of April this year, Song Goose (fancy name) left Tencent, where he had worked for two and a half years. Two days later, he immediately joined a Beijing AI big model star startup.

The rapidity of the action, the decision to choose, unusual. Moreover, he moved his family from Shenzhen to Beijing, which increased the shock and curiosity of his friends. One friend commented, "I've never seen you so sharp in my life".

As for the reason for jumping ship, SongGoose wrote in WeChat public number "SongGoose has something to say": "Aspire to realize AGI (General Artificial Intelligence)".

"I think AI will really affect the next ten years at least, this is not a hanging in the air completely did not land the technology, this is a thing that can immediately change the world. In the next ten years, all industries will bring no small changes because of AI." Song goose said.

Pine Goose's choice of career choice, but also just a microcosm of the current development of the large model industry.

Pulse high employment talent think tank data show that this year's spring recruitment period is the peak of large model startups, the demand for talent has risen sharply. in February, the volume of new jobs issued by AIGC increased by 75.6%, and continued to rise in March, with the ringgit increase rising by 12.56%. With the end of spring recruiting and the end of the wave of entrepreneurship, the demand for talent has fallen back. in June, the big model companies wind up again, the demand for talent rose again by 4.16%.

As a large model of this "entrepreneurial pioneer", Wang Huiwen from February this year will be in its social account widely issued "hero post", many users in the comment area to leave a message to seek cooperation, ask the requirements of the job application, urging the progress of the delivery of resumes, etc., we do not want to miss this entry! The opportunity.

The big model startup wave is visible to the naked eye in attracting talent to enter, but not everyone is as determined as SongGoose All in AI.

Another big model star startup in Beijing has been recruiting technology and algorithm related talents in the first half of this year. But the company's executives told the "Science and Technology Board Daily" reporter, even if the company's founding team background hard, and in the first half of the year also got the head of the investment institutions of the hundred million yuan of financing, but the progress of recruiting "is not smooth", the reason is that the job seekers are holding a wait-and-see attitude, they feel that the situation of this round of AI is uncertain, "worried about the big model companies live, but not all of them are so determined All in AI," the company said. "Worried that the big model enterprises will not live long".

Job seekers are worried, not without reason.

AI has been proposed since 1956, and has experienced several ups and downs since then, before ChatGPT, the latest domestic AI boom is represented by the development of the "Four Little Dragons of AI", but the degree of these companies burning money and the mismatch between the output of the landing, and the inversion of the stock price of the listed companies and the promotion of the hotness of the market is doubly disappointing.

Shen Jia is a senior recruitment consultant who has been serving the talent recruitment market for a long time, and has been providing recruitment, enterprise management, investment and financing docking services for large model industry companies since August last year, and he has basically dealt with the earliest batch of large model startups in the market.

Although AI is as hot as it is today, the industry cycle is constant. Shen Jia summarized a characteristic to the reporter of "Science and Technology Board Daily", this round of AI large model startups, and the AI four dragons have a very different characteristics, that is, the former in the beginning of the venture on the commercialization of the consideration of the business growth, and the pursuit of internationalization, they try to subvert some of the mistakes of the predecessors.

Back to the candidate's perspective, Wang Fang (a pseudonym), who previously provided high-end executive search consulting services for the tech and Internet industries, has also begun to identify candidates for the big-model businesses of these companies this year. He told Sci-Tech Board Daily that now when job seekers look at opportunities in the big model industry, there is one thing they must take into consideration, and that is the big model business of this company, which must have a landing scenario, as this is the only way to ensure revenue and maintain the normal operation of the company.

On the other hand, in recent years, the macroeconomic environment and the Internet's "optimization wave", most people choose to stabilize, get up and jump ship, many people do not have so strong risk-taking power.

Wang Fang said, some job seekers think, from the current relatively stable line of business to a less familiar, the future is not clear on the large model business, "most people do not dare to go, they are afraid that the project does not land, or landing less than the boss expected, will be laid off.

Higher salaries of millions of dollars a year, lower salaries stay put

Loose Goose joined the current big model company, or the continuation of the Tencent's old profession - product manager, a highly mature and important role in the Internet industry, reference to the "father of WeChat" Zhang Xiaolong.

But Song Goose said, to big model industry "this is the sacrifice of originally clear career growth path", because he is not sure two years later, big model itself still need product managers, not to mention the career development path.

What Songoose says is true. At least for the time being, product managers are not the most in-demand type of talent in the big model industry.

Shen Jia told "Science and Technology Board Daily" reporter, from the industry development trend, from the middle of last year to the present, whether it is a large factory or startup team, are still focused on the big model to lay the foundation. Therefore, at this stage, the underlying technical talents, such as NLP algorithm engineers, deep learning and other positions, are the most urgently needed group of people in the big model talent market.

The data from the Pulse High Employment Talent Intelligence Center show that the supply and demand ratio of analog chip design, integrated circuit IC design, wind control algorithms, AIGC algorithm engineers, natural language processing and other positions is less than one.

And this kind of the tip of the pyramid of that batch of people - mainly refers to both research, but also able to output Paper (thesis), but also able to take into account a variety of application scenarios, landing the delivery of the project, is the first echelon of the first echelon of the big model vendors of the first to compete for the object.

To what extent is this talent being competed for? Wang Fang gave an example to a reporter from Sci-Tech Board Daily.

A head Internet company located in Shenzhen, now recruiting needs are mainly the world's top AI talent, its internal recruitment team, the external also hired headhunters, "the current situation is that these two teams are grabbing the same set of people, grabbing to the end of the day, we all feel tired, it's not interesting, because the supply of talent is only these, and can be persuaded to come into the job even fewer ".

Salary, from its recent services to recruiting companies and candidates in the process, the industry's top talent, the value of this year is once again get a boost, jumping the salary of the people are almost always a substantial increase.

Shen Jia also told reporters that the public is familiar with the companies in the big model business, is recruiting AI talent globally, for the top personnel to open the salary is "regardless of the cost", but Shen Jia did not tell reporters this "regardless of the cost" of the number of how big! The company's main goal is to provide the best possible service to its customers.

Recruitment platform data may give some guidance. Pulse high employment talent think tank statistics show that in the pulse platform, such as machine learning, speech recognition, AIGC algorithm engineers, natural language processing and other positions, the highest salary of more than 900,000; more core recommendation algorithms, algorithm engineers, computer vision positions, the highest are more than a million dollars a year.

But if you think, as long as into the big model of this line, the salary will naturally rise, then it is very wrong.

Wang Fang also said another industry status quo. According to their recent offer situation, the middle or bottom of the algorithm engineers, or belong to the flat salary jump, or even pay cuts, even if there is a small part of the salary increase, it will not be more than 20%.

Pulse high employment talent think tank statistics, test engineers, product managers, etc., these can be a large number of people from the Internet industry leveling over, talent supply and demand ratio are greater than 2, these people are also relatively good recruitment.

"The big model industry needs or high technology and top talent, most of the middle and lower level personnel, the salary level is not much higher, will not be because of the big model hot, their salary also due to the industry dividend and follow the rise." Wang Fang said.

"Adjust your expectations, don't dream big."

The heat of the big model is growing at an exponential rate in the whole society, and nowadays, people have been used to compare this technology application to the Internet era 20 years ago, which contains unlimited opportunities, and people expect the new BATs to grow from it, and then give birth to one after another myth of ordinary people's wealth creation.

Now a very obvious social symptom is that everyone is looking for big model opportunities - even if only a little bit, people do not want to miss the technology to bring a huge era of dividends.

As a side case of jumping from the last gold industry to the next potential gold industry, after going from the Internet to the big model industry, a very large number of friends ran to ask SongGoose if there are any reliable recommendations for good opportunities in the application layer? Should I join a big modeling company? How to evaluate whether a startup company is reliable? What if I get laid off when I enter a startup during the trial period?

After communicating with 43 friends, and in response to some questions, SongGoose wrote a 5,000-word article in WeChat public number about AGI's "Job-hopping Thoughts and Suggestions", which to a certain extent fits the current situation of the industry.

For example, to evaluate a startup company, from the team background, financing, geographical situation and other perspectives to analyze the company he is now working for, for example, Tsinghua Department, the first echelon of large models, venture capital circle star team.

Venture capital circle star team is often also a priority item for job seekers now. Wang Fang told reporters that at present, the star big model company is indeed the consideration of most job seekers, compared with other startups, the former has investment from institutions, abundant funds, the probability of problems in a short period of time is small, the team will be more stable.

Loose Goose also said that the domestic AI application layer is still a chaos, VCs seem to be lively, but not much out of the blue; seize the user, seize the scene, can be landed, there is always value; the future opportunities in the application layer, can form a very deep barriers, and now there is no point in time to explode the application layer ......

Although SongGoose is currently slightly worried about his role as a large model product manager, Shen Jia believes that the application layer of the large model will burst out with great market opportunities, and the consumption of talent here will be the largest, and as the industry moves forward, the future product manager with AI industry experience is the most valuable.

The wildly advancing wave of Chinese big model startups has recently braked sharply due to the illness of Wang Huiwen and the acquisition of Lightyear Beyond by Meituan, which has added some cold colors to the current reality of China's big models.

The industry is in its infancy, rugged and chaotic. In the face of the company's situation may occur temporary retraction and even entrepreneurial failure, loose goose only said, "at least to do a good job of a company did not pay you three months, and then closed down the minimum expected", "we recommend that we do not have too high expectations of financial returns, and do not do some big dreams ".

On the day he left Shenzhen for Beijing, Song Goose took a picture of a billboard at the Shenzhen airport, which he felt fit the current state of mind and industry environment.

It reads, "But wherever people can imagine something, someone will be able to realize it."