Original Title: “2023 Game Industry Opening Volume: Start with AIGC Game Talents”
Source: Jinghe Author: Gui Zhiwei
Wen Jie game company spring recruitment.
Recently, the author has observed that with the explosion of the AIGC concept and the official opening of APIs by ChatGPT, more and more game companies are recruiting/cultivating relevant talents.
In terms of salary and treatment for AIGC-type talents, game companies have swept away the haze of “cost reduction and efficiency increase” last year, and the monthly salary of regular positions is basically around 30k. For top companies like NetEase, the basic salary for AIGC positions in key projects is even 60K+.
Could it be that the game industry will start trading in the AIGC field in 2023?
Game companies “compete” for AIGC talents
Judging from the recruitment website information, companies currently recruiting AIGC-type positions include: Netease, 37 Interactive Entertainment, Giant Network, Mini Ideas, IGG, Kunlun Wanwei, etc. (Search scope “Shanghai”, keyword “AIGC”, industry “game/Internet”)
In terms of job categories, the above-mentioned companies are currently recruiting jobs that can be divided into two categories: AIGC art, AIGC algorithm research. Overall, the salary level of AIGC algorithm engineers is relatively high.
In terms of job responsibilities, AIGC art is mainly responsible for production resources, as well as tuning keywords and parameters. For example, the position of “original painting (AI drawing European and American cartoons)” on Giant Network requires workers to use AIGC tools such as Midjourney/StableDiffusion to make 2D original paintings, and then use PS to correct the AI generated pictures.
Tencent (Beijing) “AI Original Artist – Far Future Science Fiction Oka SLG” position requires workers to study different AIGC tools to help the art department produce art materials; write and enrich the company’s AI keyword vocabulary for other artists to use.
AIGC algorithm research focuses on the creation and improvement of production tools. For example, in the recruitment job descriptions of 37 Interactive Entertainment, Giant Network, IGG, etc., the main purpose of AIGC algorithm engineers is to “improve the production efficiency of project content”.
These include: in-depth understanding of the art design and production process, and improving the production efficiency of art content through AI technology/tools; solving the specific needs of collaborative projects in terms of visual, audio, and text generation through AIGC, proposing solutions, and assisting in project implementation, etc. .
It is worth pointing out that, in addition to AIGC’s strategic layout (referring to AIGC’s general technology for improving content production efficiency, expanding intelligent dialogue and interaction, etc. The project needs to recruit AIGC talents.
As a blockbuster product under NetEase, “Egg Party” can be said to be a leader in the casual competitive game track in the past two years. At NetEase’s 2022 financial report analysis meeting not long ago, Ding Lei revealed that the DAU of the game exceeded 30 million, making it the game with the highest daily activity in NetEase’s history.
In addition to the attributes of casual competitive gameplay and cute visual style, the important reason that promotes “Egg Boy Party” to become a big DAU mobile game is that it incorporates the UGC model – Egg Boy Workshop. Jinghe has mentioned in many articles that the greatest significance of AI to UGC at this stage is to lower the threshold of creation and accelerate the implementation of ideas.
Recently, NetEase has also launched a number of AIGC-related positions, including AIGC tool development master, AIGC platform product manager, AIGC level generation researcher, etc. Among them, it is clearly stated in the job responsibilities of the AIGC level generation researcher to promote the implementation of AIGC technology in “Egg Party”. In terms of job salary, Netease AIGC level generation researcher is currently the highest among the three, at 60-85K.
Some industry observers said that AIGC is an important round of technological innovation for the game industry, which will comprehensively benefit the game industry chain from multiple dimensions such as production efficiency, product innovation, and production capacity. But the question is, what is the real talent pool in this relatively new field? Right now, major game companies are full of recruitment benefits, which can also be regarded as a new round of talent competition.
From internal learning to external recruitment of professionals
Compared with game companies actively recruiting AIGC talents from outside this year, according to the author’s understanding, everyone’s attitude was still conservative last year.
When AI painting and ChatGPT emerged in 2022, Jinghe had exchanges with the heads of many domestic game companies. At that time, everyone chose to set up an “innovative research group” internally to learn how to use AI tools or pay attention to AIGC technology application fields.
The person in charge of a game research and development company in Beijing said that AI technology research requires a lot of resources and costs. For small and medium-sized teams, it is more likely to choose general-purpose AI products in the industry for further research.
It is understood that at that time, the company directly asked the art VP to lead the team to research and learn to use AI tools. Of course, this model is still applicable to many medium-sized game manufacturers in Beijing, Shanghai, Guangzhou, and Shenzhen.
Some game manufacturers are relatively ahead. Jinghe mentioned in the article “How to integrate AI painting into the workflow, this is the answer of Tianmei Rubik’s Cube Longyuan Network”, that AI painting tools have been integrated into the daily workflow of the project team last year.
Talk about the value of AI painting in game development and operation. They believe that AI painting is more of a tool that can improve work efficiency. However, it has no way to replace the designer itself, or it can expand the designer’s imagination. For example, help plan quick and concrete ideas, have a more intuitive communication with artists or leaders; or do some quick material testing to help the project verify the direction of the attempt, etc.
But not everyone has a strong tolerance and acceptance of new things. It is understood that last year, some companies required employees to passively learn AI tools or integrate AI tools into their workflow, and some employees even chose to leave their jobs for this reason.
In the latter’s view, when producers and planners think that concept maps are collected and recombined to generate a large amount of image information, they will equate AI painting tools with concept artists. The result of this is that the initiative and recognition of concept artists are reduced.
To put it bluntly, artists believe that art is the creation of advanced thinking, and they have talent, value and irreplaceability. When AI quietly captures these fields, it is difficult for them to admit that what they do is just work, just workers on the assembly line.
“High-powered talents often need to learn from the old and learn from the new, but few people can do this.” A Beijing game manufacturer revealed, “They will choose high-potential talents who have worked for 2-3 years and have a good understanding of the art workflow to learn and research AI. and other innovative tools.”
In general, in 2022, domestic game manufacturers will mostly set up innovative research groups internally for AIGC technologies such as AI painting to try to integrate AI tools into the workflow. However, the attitude towards implementing AI painting into the development of specific projects is relatively conservative.
However, after entering 2023, the attitude of game companies towards AIGC has obviously changed, or some of them can no longer hold it.
A few days ago, Xindong CEO Huang Yimeng tweeted that he recently chatted with two game teams, one of the team cut down the original painting outsourcing company, and the second team cut down the translation outsourcing company. AI has actually begun to affect many people’s jobs.
He believes: “AI may not replace you, but people who make good use of AI as a role model can replace you in the future. Everyone needs to be prepared to meet changes and embrace changes, and use their original professional skills and holdings. AI as a tool for its own use.”
It is worth mentioning that with the rise of the AIGC wave and the gradual maturity of AI drawing tools such as Midjourney, some training institutions have smelled a new wave of “knowledge payment” business opportunities. As far as the author knows, there are already some design/art training institutions on the market that have launched paid teaching prices for AI smart painting.
On the one hand, in the eyes of these people, AI painting will have an impact on UI and daily operation jobs, and there is a learning market demand; on the other hand, they believe that AIGC may be a new round of “industrial revolution” that accelerates the industrialization of the Internet. Future Trends.
Right now, game companies are actively recruiting AIGC professionals from outside through recruitment. Compared with passive learning and acceptance last year, they are obviously more active this year. To some extent, this also reflects from the side that AIGC is of great value to game development and operation.
Tengwangmi layout AIGC 3D asset generation
Looking back at the changes in the game industry in recent years, from cloud games in 2019, metaverse in 2021, to AIGC in 2023, there will be “technical changes” almost every two years.
Facing the so-called tuyere, manufacturers may have different entry angles, time points, and optimistic attitudes, but they will basically wait and see or explore.
At present, the layout of domestic game companies in the field of AIGC can be roughly divided into three types: internal self-construction, external investment or cooperation, and some choose to find another way.
Jinghe once analyzed in the report “71 Pages of Depth: Revolution of Game Technology Power, Tengwangmi Embracing the AIGC Era”, at this stage, AIGC presents characteristics such as dots, discontinuity, and weak connections in the R&D and operation links. In fields such as battles, numerical values, and plots, AIGC’s landing games are relatively mature.
Inside the game manufacturer, there are already corresponding platforms and templates. Typical examples include Tencent (CDG AI Lab, IEG Cros, etc.), NetEase (NetEase Fuxi, Mutual Entertainment AI Lab, etc.) and Mihayou (Anti-Entropy Research Institute). Medium-sized game companies have also bound high-quality entrepreneurial AI targets in the market through investment and other forms. For example, Lilith invested in AI technology developer Qiyuan World.
In addition, the top start-up AIGC company in the market has served some projects of game companies based on relatively complete solutions in some fields. For example, Mutong Technology’s “MLBB” uses hyperparameter AI Bot, Longyuan Network’s “Auto Chess” uses Xingzhe AI Bot, and Lingxi Interactive Entertainment’s “Three Kingdoms: War Chess Edition” uses Qiyuan World Game AI solution.
Specific to the three leading manufacturers of Tengwangmi, Jinghe learned that the common point of their AIGC landing in 2023 is the automatic production of 3D assets. The difference is that Tencent Games will devote part of its energy to animation generation; NetEase Games will focus on human-machine collaboration and radiate to the field of construction machinery; Miha Games will focus on Luming AI to create a strong intelligent Crypto human.
Among the leading game manufacturers, Mihayou is relatively special. Its strategy is self-development + investment to walk on two legs. It is reported that Mihayou is preparing to participate in the third round of financing of the domestic AI startup MiniMax, which is also the third time Mihayou has invested in the company.
A person in charge of a leading investment institution told Jinghe that Mihayou has formed a five-member investment team to search for AIGC manufacturers in the primary market. As early as before the Spring Festival in 2022, Mihayou had already invested in MiniMax, which was valued at US$200 million at the time. After the Spring Festival in 2022, MiHoYo made additional investment, and the valuation of MiniMax reached 500 million US dollars.
You may ask, why did Mihayou invest in MiniMax three times in a row? The author guesses that from the perspective of Minimax’s products and the company’s core technical capabilities, they and MiHoYo create strong intelligent Crypto humans, which are highly complementary in technical dimensions.
Public information shows that Minimax’s first product, Glow, is an AI chat software that focuses on social interaction. The basic principle is similar to ChatGPT, the difference is that Minimax is a multi-modal AI large model. Simply put, that is: input text – output text; input text – output vision; input text – output voice.
Compared with ChatGPT’s powerful information retrieval and knowledge service capabilities, Minimax focuses more on the emotional service layer. For example, users can set the visual image and voice of the AI chat object, and even develop character (design and train AI communication style).
It meets the capability stack needed by MiHoYo to create a strong intelligent Crypto human Luming AI. Last year, MiHoYo’s virtual idol, Luming, launched its first live broadcast. The official announced that it adopted the mode of “motion capture to generate motion + anti-entropy AI to generate sound” for live broadcast. Judging from the effect, Lu Ming’s voice was not smooth enough at the beginning, but the movements such as pulling hair, backing hands, and dancing were relatively smooth.
Jinghe analyzed in the article “Cooperating with game companies to divert water, this is the AIGC new business model that Xiaoice wants to create”, Luming Voice adopts real-time voice dubbing + voice convert solution, which is driven by a one-to-one conversion method , not artificial intelligence sound generation.
In view of this, we believe that MiHoYo has shortcomings in the field of AI real-time voice generation, and needs to rely on the strength of partners (the word “real-time” is added, mainly because the effect of using AI dubbing for the characters in “Undecided Event Book” is not bad) . MiniMax revolves around the three major AI generation core technologies of “text, drawing, and voice”, which is exactly what Mihayou urgently needs at this stage. From this point of view, it is reasonable for Mihayou to invest in MiniMax three times a year.
From a long-term perspective, MiHoYo’s vision is: “By 2030, create a virtual world where one billion people around the world would like to live in it.” To achieve this, Mihayou needs to rely on internal and external forces.
Around games and derivative virtual idols, game manufacturers are accelerating the use of AIGC tools and even developing new pipelines. We learned that there are already companies on the market that are moving towards TexT to Game, that is, exploring one-click game generation.
In their view, since ChatGPT can communicate with people fluently, and Midjourney can produce pictures by typing text in, it is reasonable to directly use text to generate games.
In general, both major and medium-sized manufacturers are actively recruiting AIGC talents such as Midjourney and ChatGPT, or top game companies such as Tengwangmi are actively developing their own AIGC tools. All this shows to a certain extent that the influence of AIGC on the game industry does exist and has a far-reaching impact.
If you ask me, what is the trading volume of the game industry in 2021? The answer is senior talent at the producer level; what will the game industry do in 2022? The answer is to see who cut off the project team, the most invalid manpower; as for the transaction volume of the game industry in 2023? The answer may be AIGC’s low-end and high-end talents.
What do you think?
Source of information: compiled from 8BTC by 0x Information.Copyright belongs to the author, without permission, may not be reproduced