Shen Ran (shenrancaijing) original
Organization | Jin Yufan
What topics will be the hottest in 2023? What is the most discussed topic among professionals during this period?
There is no doubt that it is the emergence of ChatGPT and the technological progress of AI. Of course, the words that have the most impact on people in the workplace are, "AI replaces humans" and "One billion workers are revolutionized"...
ChatGPT has brought us into the AI 2.0 era. AI is indeed affecting many industries in a way beyond our expectations. The two major topics derived from this are: Who can get on this AI express train? Whose jobs will be “killed” by AI?
Recently, Shenran Live Broadcast Room invited Bao Aile, co-founder and CMO of Lagou Recruitment, and Liang Gongjun, founder of Haina AI and former Microsoft technical consultant, to talk about "AI robbing war, who is it?" The most popular?" This topic launched an in-depth conversation.
Lagou Recruitment, where Bao Aile works, recently released the "AIGC Talent Supply and Demand Report for the First Quarter of 2023", which intuitively reflects the most popular positions and salary increases in this round of AI recruitment war from the data level. Specifically, the demand for AIGC talent positions increased by 42% month-on-month in March this year, of which 25% of AIGC algorithm engineer positions have a monthly salary of more than 68K.
Liang Gongjun used to be a technical consultant for Microsoft and invested in many AI companies. Currently, Haina AI led by him has provided AI interview services to more than 5,000 large and medium-sized companies.
The two guests provided some directional suggestions from the perspectives of talent market and technology development.
Quick overview of key points:
1. The city with the greatest demand for AI-related occupations is Beijing, and the most sought-after occupations are algorithm engineers, AI product managers, etc.
2. Within six months to one year after a company announces a general large-scale model, the demand for employees may surge; companies that develop applications in vertical fields currently have a large demand for employees.
3. In general large model companies, compound talents with both academic research and engineering delivery are the most scarce.
4. The so-called 35-year-old crisis in the Internet industry will collapse on its own in the AI 2.0 industry, especially in the To B field.
5. In the next five to ten years, AI 2.0 will "support" tens of millions of people just like the mobile Internet in the past ten years.
6. Now is the dividend period for AI technology. If you want to get on board, the next six months to one year is the window period.
7. The trend of “replacing people with AI” is penetrating from top companies to mid-level and tail companies at a faster rate.
The following is the live broadcast record, edited by Shenran:
Shenran: Does AI robbing wars really exist? To what extent does the AI megatrend impact the employment market?
Bao Aile: Judging from the recruitment statistics of Lagou, the AI recruitment war is more focused on the recruitment of technical talents, such as CV algorithm engineers, recommendation algorithm engineers, and NLP algorithm engineers. .
Let’s look at industry demand first. The five industries with the greatest demand for AIGC positions are firstly "artificial intelligence", secondly the content short video field, and the remaining ones are financial industry, enterprise services and data services, and intelligent manufacturing.
Look at the city again. The cities with the greatest demand for AIGC positions are Beijing first, followed by Shenzhen, Hangzhou, Shanghai, and Guangzhou.
There are indeed many high-paying positions emerging in the AIGC field. For example, the monthly salary of CV algorithm engineer, recommendation algorithm engineer, and NLP algorithm engineer positions, the 75% percentile is close to 60k; 25% of AIGC algorithm engineer positions have a monthly salary of more than 68K; AIGC The monthly salary of a vertical product manager reaches 59k at the 75th percentile, which is much higher than the salary level of other product manager positions.
Of course, the recruitment demand for these technical talents is not a short-term surge, but a long-term stable existence. Moreover, the requirements for senior talents and the salary levels provided are very high. It can be said that they are a small part of People's work.
What we have observed on the front lines of the industry is that some projects that investors are optimistic about and companies that train large Chinese local models are indeed very popular. They feel now like the mobile Internet in 2012, using one week as a year. On the whole, companies currently recruiting AI-related talents are mainly core companies, with some leading vertical companies also included.
Shenran: Have companies that make big models started to grab people? What are the differences in the talents that companies of different sizes compete for?
Liang Gongjun:There are currently about 30 companies in China that make general large models. The first 10 or 50 people in these companies' team formation may all be selected from old colleagues or internal recommendations. After 50 people, product managers and application layer engineers may be recruited from outside.
In the current market environment, these companies have obtained the most valuable funds, ranging from tens of millions of yuan to hundreds of millions of dollars. Therefore, you can pay close attention to these companies. They may have a large demand for employees within six months to one year after they announce the launch of general-purpose large models. However, making low-level large models is more of an opportunity for scientists and top engineering experts, these few people.
Companies that develop applications in vertical fields already have a large demand for employees. Because these companies were all established in the past three to five years, they have gone through the process from 0 to 1, but they have also encountered some problems. Now in the 2.0 era, they will usher in explosive development. I believe that in the next five or ten years, just like the mobile Internet in the past ten years, it will be able to "support" 10 million or 20 million people.
Shen Ran: In this war for talent, will more skilled people go to big companies? Will it be detrimental to startups?
Liang Gongjun: First of all, the environment in recent years has made many people feel that "big factories may not be stable, and small factories may not be unstable." Many people do not have such strong motivation to take risks. In addition, the current situation of AI is still unclear. These two factors have not yet caused a large-scale flow of talents to large factories.
But after half a year or a year, when the situation becomes clearer, I believe that more and more top talents will flock to representative companies of AI 2.0.
A major variable behind this is whether China’s own large model can catch up with GPT-3.5. Even if it catches up to 80% of the progress, it can promote large-scale commercial applications in many scenarios.
Shenran: In terms of employment, what kind of AI talents do domestic companies mainly lack?
Liang Gongjun: Looking through the resumes of the first 100 people on the OpenAI team, we will find that the people who really played a huge role are composite experts, and they share a common trait: It can not only do research and output papers, but also deliver projects on the ground.
The popular translation is that one can not only understand the principles of technology, but also consider various application scenarios and solve engineering problems. In the general large model company, this kind of talent is the most scarce and has the highest price.
Shenran: In the new wave of AI, are there any new jobs created?
Bao Aile:Many domestic companies are doing pre-trained large models. I believe there are some basic positions that still require people to participate, such asdata cleaning related to the AI industry, Data are marked . But it's definitely not mainstream.
The more mainstream and competitive positions are algorithm engineers, AIGC product managers, AI product managers, etc. For example, the product manager must choose a segmented field and have a deep understanding of the industry's usage scenarios.
Shenran: In the AI 2.0 era, will people with more seniority become more and more popular? Does the so-called 35-year-old crisis no longer exist?
Bao Aile: This is indeed a new change in the AIGC field and past technical positions. In the past, we often said that technical positions have a 35-year-old threshold, but judging from the age distribution of AIGC technical talents, talents over 30 or even 40 years old account for a large proportion. I believe that there are now a large number of experienced practitioners who are leading the current technological changes.
Liang Gongjun: This should be viewed from two aspects: B-side and C-side. My judgment is that there is a very big difference between the AI 2.0 era and the past 20 years. Before , the To C field was developed first and then penetrated into the To B field. Now, the To C and To B fields are developing simultaneously.
In the To C field, it is still the group of people who have the most keen sense of consumers who do it.
To B field will undergo huge changes. I think, To B field will have better cash flow. Because in the context of cost reduction and efficiency improvement, larger companies are more willing to spend money on AI. In order for B-side customers to pay the bill, AI talents need to have a good understanding of AI base and have rich industry experience.
One change brought about by this is that in the past, there was a 35-year-old phenomenon in the To C field. After 35 years old, it was no longer popular. But in the To B field, there is a greater need for mature practitioners over 35 years old. , they have a deeper understanding of the scene, a deeper knowledge of the industry, and are more patient with customers. So even at the age of 35, you can find your "second spring".
Shenran: From a low-level perspective of technology, how is this wave of AI different from the past?
Liang Gongjun:Technologies that can truly have a significant impact on human society have a very long maturity cycle, which can easily take 30 or even 40 or 50 years. AI is such a technology.
Microsoft has been conducting AI research since the early 1990s, and it has been 30 years since then. These few decades can be called the 1.0 era of AI.
On November 30, 2022, OpenAI trained the GPT-3.5 large model and released ChatGPT, which represented that AI has entered the 2.0 era and ushered in a true blooming of a hundred flowers with a single-point application that benefits the public.
Although AI has been developing in the past, it has never ushered in a real explosion. For example, In the past, many customers purchased AI products out of a "sense of technology" gimmick. After three to five months, they discovered that they needed to hire specialized people to "serve" the AI. Because the reliability and validity are poor and the effect is not obvious, customers are unwilling to renew their subscriptions for a long time.
With the open source of many new underlying bases at home and abroad, various C-side and B-side application scenarios have emerged on a large scale. In the past six months, new applications have been launched almost every day, and AI has finally become everyone's Available technologies.
Those of us who are really engaged in technology are still very excited, because this may be a once-in-a-lifetime major opportunity. Like myself, since the release of ChatGPT, I have been watching various new technologies and developments every night for two or three months, and I was so excited that I couldn't sleep.
Shenran: Will the trend of To B companies competing for AI talents become more obvious?
Liang Gongjun:In the current environment, the larger the company, the more stable it is. Of course, companies are also worried about missing out on the big AI trend, so they are also making some arrangements, such as adding more people with AI backgrounds to the IT department and increasing internal budgets for purchasing AI systems.
Shenran: In this round of AI trends, which type of company are you more optimistic about?
Liang Gongjun:Whether it is a low-level general large model or a vertical large model, the competition is nothing more than the key elements of computing power, algorithms, data, and cloud infrastructure. The biggest threshold is data. Whoever has high-quality industry data will have a greater advantage. The advantage of Chinese companies is that in a large unified market like China, they have a huge advantage in data volume.
In terms of the underlying large model, there is still a huge gap between us and OpenAI in the United States and its competing companies. However, Tencent, Alibaba, Baidu, and Huawei are all struggling to catch up. They need more encouragement and applause.
As for the industry model of vertical scenarios, the industry should have more confidence. We believe that it is very likely to repeat the development of mobile Internet in the past ten years, that is, in three to five years, ranking in the top ten , many of the top twenty AI 2.0 vertical scenario companies will come from China.
Shenran: For those who have a certain technical background and want to get on the AI 2.0 train, do you recommend joining now or waiting?
Liang Gongjun: From the perspective of a bystander, if you have a firm judgment and you or your family have a relatively strong ability to resist risks, you might as well make a decision while most people are waiting and watching. Determined to "get in the car." Of course, for most people, the best choice is to first ensure that their location is stable.
Bao Aile: I think the cruel thing for students who are concerned about AI opportunities is that if you have the opportunity to participate, you may already be in it. This group of people They are technical talents in the industry; if you do not have the opportunity to participate, then you may have to wait a little longer until it reaches the application level to see if there are any employment needs related to your past work experience.
Shen Ran: Let’s talk about bosses and companies in detail. Under the general trend of AI, have bosses’ employment strategies been adjusted?
Bao Aile:The recruiter can be summarized into three states.
The first type of bosses, are also the first wave of people who eat crabs. They have already taken action to lay off unnecessary people and quickly apply AI. For example, some game companies have laid off their team of painters and instead use AI tools to create value for the company.
The second type of boss is willing to embrace new changes, but instead of laying off people first, they adjust the organizational structure and hope to lead the team to speed up learning how to apply AI. As I know, one company has formed an AI brain team. The view of this type of boss is that the future team will be like AI, like a jellyfish, first it needs a brain, and then it needs countless tentacles. The tentacles are actually not important, but the brain must be smart enough.
The third category is the majority at the bottom of the pyramid, which has not actually changed at all. Although the concept of AIGC is relatively popular in new technology and Internet circles, it has not penetrated into a wide range of traditional industries, and it is still too early to talk about the application of AI.
Shenran: Is there a trend for companies to replace people with AI?
Liang Gongjun:The idea of AI replacing people has been discussed for many years. This trend is indeed accelerating now and is penetrating from head companies to waist companies and tail companies.
The fastest responders are mostly leading companies, such as Wal-Mart and SF Express among the world's top 500 companies. The emergence of ChatGPT is equivalent to a market education for these leading companies, allowing them to quickly consider using AI to improve efficiency. Our customer conversion cycle has been greatly shortened. The vast majority of waist companies are not very urgent in their understanding and actions of AI.
Shenran: Which industries are already replacing people with AI or machines?
Liang Gongjun:As mentioned in the "14th Five-Year Plan" employment promotion plan, the more "simple and repetitive work links and dangerous, complicated and dirty jobs", the easier it is to be replaced or replaced by AI. AI greatly improves efficiency.
The operational data of our Haina AI also reflects this talent migration trend.
First of all, the manufacturing employment volume has indeed dropped significantly compared with the same period last year. Second, after the epidemic, the mid-to-high-end service industry is recovering rapidly. For example, the recruitment volume of Wal-Mart and SF Express has increased significantly compared with the same period last year. In other words, The demand for basic blue-collar workers like assembly line workers will become less and less in the future, the trend of machines replacing people is accelerating, and the released labor force will quickly migrate to the mid-to-high-end service industry.
Shenran: So far, the trend of soaring demand for AI-related jobs is still continuing or has it stabilized?
Bao Aile:The salaries of these positions did not rise overnight. However, because the topic of AI has attracted great attention, more people have noticed, "It turns out they make so much money."
However, technical talents are also divided into layers. Some are leading the trend, while others are working at the bottom. High-end AI R&D talents have always been the best technical talents in the industry.
I just saw a netizen saying in the comment area, "It turns out that the income gap between everyone started when the liberal arts and sciences were divided into classes." I think this is the case. When we look at the overall salary in the Internet industry, technical talents have always been higher than other positions.
Shenran: Do Internet people still have a chance to "get on the bus"?
Liang Gongjun:I think that AI 2.0 is a once-in-a-decades major opportunity for those who were originally engaged in the Internet and mobile Internet, as well as the fresh graduates who want to use AI. I suggest that we should pay attention to AI 2.0 with an open mind, especially the opportunities in vertical scenarios.
If you want to "get on the bus", take advantage of the current technology dividend period. Because the industry will use new AI 2.0 technology to redo almost every life scene and industry scene.
Shenran: As a liberal arts student with no technical background, I am very anxious about the AI trend. What suggestions do you have?
Bao Aile:This kind of anxiety may not be caused by AI, but more from too many external changes, which reduces your predictability of your career future.
The only way to relieve anxiety is to make yourself a relatively top talent no matter what industry or position you are in. If you can be in the top 10% or 15% compared with your peers and peers, then you don’t have to worry, because no matter how the industry changes, even if AI completely “destroys” the entire profession, then As a leader, you will definitely be the first to be noticed when you change careers.
Liang Gongjun:Just like the emergence of electricity, it is bound to allow humans to move from basic jobs to higher-end jobs. When AI comes, many people will think, "What are people going to do?" In fact, there are many positions that are warm, emotional, and require creativity, such as tutors in the high-end service industry, the elderly care industry, etc.
I have been using ChatGPT for about half a year. I asked it some questions and it helped me write a "draft", but it still required me to reorganize it. It was unable to deliver a truly high-quality article. .
Shenran: Finally, I would like to end today’s live broadcast with a question from the audience. What type of talents are the most sought-after now? How do non-technical talents become such people?
Liang Gongjun: Don’t be blown in different directions by different winds. You must mark the "coordinates" of your life. Choose a field and accumulate it in a ten-year cycle, I believe that if you can become an expert in your own field, big or small, then all new technologies are just tools for you. At the same time, keep an open mind, embrace new tools and technologies, and use them more in your work.
Bao Aile: First, no matter what industry you are in, you must be keen on learning new technologies, have some basic understanding of human-computer interaction and artificial intelligence, and be able to learn quickly. Second, your experience in your industry is invaluable. No matter how smart AI is, it will eventually be applied to the real world, where it needs to serve customers and create value, rather than creating a new world from scratch.
Don’t panic or be anxious. The final winner must have a relatively stable mentality.The AI wave is bound to come. When it comes to your industry, the test will be whether you can become a person who can apply this tool faster and better.
*The title picture and the accompanying pictures in the article are from pexels.The above is the detailed content of With a monthly salary of 60,000, which AI positions are in demand?. For more information, please follow other related articles on the PHP Chinese website!
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
