The most realistic industry anxiety in the AI era, starting from “crossing the threshold”

This year, after going through a lot of industry dinners, everyone is talking about AI.

Everyone’s anxiety is real. The founder of a large model company told me that with the iteration of the base model and application engineering, the beginning of 2026 is an inflection point for almost all AI capabilities to become available – so overnight, the ‘virtual fire’ was transformed into visible productivity. A numerical planner friend used a large model to dismantle a successful product and push back, and some programmers raised a group of agents to help themselves run 24 hours a day. Missions, bug fixes… The game circle has its own technical genes, and soon a small number of AI masters will come up with some rather subversive solutions. But this demonstration effect inevitably amplifies the FOMO sentiment of more people.

After all, this is still a rapidly developing field – whether application consensus and best practice are still ‘changing every day’. More than a friend once told the teahouse gentleman that he put more energy into AI, but he couldn’t smooth the anxiety too much. Not long after OpenClaw became popular, the community began to discuss ‘new lobsters’ such as Hermes and Claude Cowork. harness New words such as engineering, MCP, and world models appeared one after another. The MPT spell, now in retrospect, seems to be a topic of a few centuries ago… ‘I didn’t need to learn AI knowledge this week, I won’t have to learn it next week.’ It is a common joke.

These phenomena are also signs, and to a large extent, they reveal a general emotion in the current game industry: if you want to grow in this wave of AI, you are afraid that the old knowledge system will not be defeated, and what is more afraid of ‘can’t learn’— – This round of AI wave, the trend of subverting the traditional game development model is very clear; but how to digest and absorb the massive new knowledge updated with the sky, and not be overturned by the turbulent huge waves, this is a rigorous challenge to the overloaded brain.

But I don’t think this anxiety is inexplicable – just a few days ago, the teahouse gentleman came to the scene of the Tencent game creation competition. In the eleventh year of the competition, they made a series of heavyweight adjustments around AI this year. After reading the new things they came up with, I think their efforts to help the new generation of gamers catch up with the AI bus, not only far exceeded my previous expectations, but also hit the essential problem that caused the current industry anxiety.

01

New paths, new solutions, new opportunities

Before introducing the changes of this year’s Tencent Game Creation Competition, first simply add the background, Demis Hassabis, CEO of DeepMind, mentioned a concise and comprehensive statement last year: the most important skill in the AI era is ‘learning learning (learning)’ how to learn)’. OpenAI founder Sam Altman once said that the AI era should be iteratively learning in AI tools; ‘AI godmother’ Li Feifei said that the value of knowledge in the AI era is not about what he has learned, but what tools can be used and how quickly he can get started.

The views of these technological leaders who are at the forefront of AI can be summed up in two points: First of all, it is not the learner who has a problem, but the ‘learning path has changed’ – the emergence of the AI era has too much impact on the human knowledge structure, which leads to the reconstructing of the old learning method. Secondly, the core of this new learning paradigm is to enter the iterative cycle; you must actively get started, not wait until you have learned to do it, but to go to the actual project, learn and correct while doing it.

These judgments are actually more realistic in the game industry. The game itself is a comprehensive proposition with the combination of technology, creativity, interaction and aesthetics. Many experiences have no ready-made answers, and can only grow out in the practice of the project. But there are still two landing card points – where does the good project that can be used to practice hands come from, and who will ensure that the trial and error direction will not go astray?

Focusing on the above two card points, this year’s Tencent Game Creation Competition has tried to give an understanding.

The former is their ‘traditional strengths’ – ten years after the competition, the Tencent Game Creation Competition has verified a set of mature talent incubation logic: not only simply run the competition, but also integrates knowledge sharing, expert counseling, and resource links. Promote student contestants to participate in the development of actual game projects, and promote creative talents from ‘with ideas’ step by step to ‘can make results’.

For example, this year, the main competition area of the Tencent Game Creation Competition will continue to open three major tracks: AI game track, Xiaohonghua game track, and creative game track. In previous years’ competitions, they received more than 13,000 entries, and hundreds of them successfully broke through. Among these contestants, some of them received millions of dollars in investment from top capital, and some projects were selected for the UNESCO Awards; some developers won with an AI native game. He won the gameplay innovation award, and then joined Tencent smoothly, and became the team backbone in three months… Many people have gone further on the career path of the game through the window opened by the competition.

The latter especially attracted the attention of the teahouse gentleman: the competition will continue to increase the size of this year on the basis of this verified training method. It not only emphasizes AI, but also brings out the top technical resources and expert resources that only a few big head manufacturers have, and builds for new developers. A new generation of ‘learning scenes’ that truly conforms to the current development characteristics of AI and can allow young developers to complete transformation in project practice. In the words of the director of Tencent Game School, the AI has now significantly changed the threshold of creation.

Specifically, this year, they first faced the contestants of the three major tracks, and fully opened up a large number of AI tools and capabilities within Tencent, including CodeBuddy, Visvise, MagicDawn, etc., covering AI. Agent, 3D generation, 3D rendering, game testing, game voice and other comprehensive capabilities – in terms of AI tools, let the contestants and Tencent’s own employees stand on a starting line.

More importantly, this year’s competition has newly established an AI game expert committee, from various Tencent game departments such as Tianmei and Photons, and more than a dozen AI-related experts are invited to sit in the town – that is,The ‘traders’ who created big DAU products such as ‘Peace Elite’ stepped out of the scenes, showed their valuable experience, and served as the expert think tank of the contestants, and participated in the whole process of propositions, counseling, review, knowledge sharing, etc.

The teahouse gentleman said that for AI games that are still in the fog, the industry significance of this matter may be more profound than it seems.

02

When the ‘Tutor Tiantuan’ of the big factory is incarnated as a leader

After the unveiling of the competition, the teahouse gentleman also contacted several mentors who participated in the AI game expert committee this time. The exchange with them has made the teahouse gentleman further convinced that the real scarcity of this Tencent game creation competition is not only the award or the bonus itself, but that it has systematically gathered a group of people who are working on the front line of AI game practice into the same field.

2026 Tencent Game Creation Competition AI Game Track Topics

When the overall AI exploration paradigm in the industry has not yet been clarified, these developers represent the people who have gone the furthest in the field of AI games in the entire industry. Therefore, for the contestants, through in-depth contact with these ‘leaders’, what they can come into contact with is not the abstract concept and paper trend of ‘sitting and talking about the Tao’, but the method, judgment and working methods that have been tempered in the mature projects of large factories. This is an opportunity for exercise.

Especially these developer mentors, they have both a solid technical perspective of traditional developers – they have many years of project development experience – they are also explorers and seekers in the AI era. It is meaningful enough to share the lessons learned.

For example, I met Yang Mengzhou, IEG Technical Art. Before the opening of the competition, he was already a well-known ‘one-person team’ in IEG – he could use AI capabilities to complete the development of art, programs, music and other full-stacks by himself, and land an entire project.

And he himself is not only a very distinctive sample of personal ability, but also previewed how the ‘super individual’ in the AI era can break the ability boundary between various functions in advance, and what is more valuable is the process of his step by step to today. Mengzhou told me that he is a technical art background and has more than ten years of Unity development experience. He has been comprehensively researching AI development since 2023, and has grown from an art with a ‘slight understanding of programming’ to a full-stack developer.

This unique experience of gradually running through the workflow makes Mengzhou attach great importance to the control of the process in AI development. As the official lecturer of Unreal Engine, he also has a lot of lectures and shared experience. I have come into contact with a large number of newcomer developers and are well aware of their common misunderstandings. He told me that he hoped that through this competition, the players would realize that if they want to achieve good AI development results, one is to avoid ‘what you want to do’, and the other is not to superstitious the ability of the model to ‘generate in one sentence’. It is necessary to formulate enough detailed planning before design, and design the demo iteration process in layers. In the end, the product will return to the first nature of ‘fun’.

Instructors have their own expertise. For example, Mony, who is the person in charge of the Tianmei J1GBT algorithm group, is the representative of the game AI actual combat school. Last year, the team and the project team worked together to overcome various technical problems such as high concurrency and consistency of people, so that the AI ‘Parrot Butler’, who can listen to and talk, officially entered the ‘QQ Speed’ mobile game.

‘QQ Speed’ mobile game ‘Parot Butler’

MONY, a Ph.D., has the rigorous attitude of a technician, and he has provided many technical insights to the contestants of the competition. For example, he encouraged contestants to ‘think the model’, and the small team’s engineering strength is not enough, so you might as well use the technical aesthetics – only use the large model where necessary, and avoid repeating the wheels in fine-tuning models. These experiences come from a technical iteration of the ‘Parrot Butler’, which has been optimized into a series mode of ‘one function and one large model’ at first, and has been optimized into a large model + multiple specialized small models.

For small teams, time and trial and error costs need to be more carefully calculated, so you might as well keep the project more certainty and stability, and save a lot of development time. Mony also suggested that all small teams participating in the competition should do a good job in the complete context management of the project, so that the AI can better understand the project intention and provide the most accurate technical suggestions.

These first-line experiences from super products and super teams are the best creative points for creators. Like a realistic scenario that IG’s AI researcher Shiwei restored to me: an inexperienced team faces technical difficulties, how to determine Is it a personal ability card that can be overcome with hard work, or is it an industry problem that the entire industry does not solve?

At this critical confusion point, the ‘old drivers’ of the industry can undoubtedly save a lot of trial and error costs, so that creators can devote more time to effective creation. The group of researchers represented by Shiwei in the expert group of the competition is born in a technical background and has been engaged in the most cutting-edge research in the field of AI for a long time. I have seen a large number of technical solutions around the world, with a broad technical perspective. Their participation can be like a searchlight, illuminating the blind spots that the contestant team themselves cannot realize.

To borrow a very popular speech from Jiaotong University, the founder of Mihayou, Daweigra, recently – ‘If you want to come out, the most important thing is to come out first.’ For many AI game developers, making it first, running, and receiving feedback in the real environment is an important part of growth in itself.

We can even see that in addition to the R&D mentor in the AI field, the members of the AI Game Expert Committee this time also include business experts from teams such as IEGG distribution and investment department. In last year’s Tencent Game Creation Competition, they helped the contestants of excellent works to connect with investment and distribution resources, linking the idealistic environment of ‘competition’ with the real commercial market. It not only brings the project online, so that developers have the opportunity to go further, but also make the ‘realistic feedback’ to the extreme.

To put it simply, what the Tencent Game Creation Competition has done this time is through their reserves throughout the competition and the full link ability as a giant. , really built such a platform that allows various projects including AI to have the opportunity to move forward and let learning not stop on paper.

Even one step forward, there are a large number of creative teams in previous years in the Tencent game creation competition, there are a large number of creative teams breaking through creatively with exquisite gameplay. As newcomers in the industry, the shortcomings are often lacking in development experience – but if more powerful AI tools are handed into their hands, they will provide them with timely guidance from senior developers when they are facing key development cards. Under such multiple escorts, can the AI native game creativity that has been looking forward to for many years in the industry emerge faster and better?

This matter is not unpredictable.

03

In the era of urgent need for ‘opening’, Tencent Game School has set a new benchmark

Going back to the question at the beginning of the article, all anxiety actually comes from the fear of the unknown. For everyone in the industry, the most real anxiety at the moment is actually not whether the AI will come, but that the industry has not fully precipitated a set of best practices that are clear and stable enough.

But from another angle, breaking this anxiety may not be as difficult as expected. AI technology is naturally dismantling many old barriers. The ridicule of the phrase ‘I didn’t have time to learn anything last week, I don’t need to learn this week’, to understand from another angle, in fact, it can be equivalent to another sentence: when you start learning AI, it is not too late. Because it is not the starting point for a person to really decide whether a person can keep up with AI, but whether he has entered the correct learning mechanism.

The same goes for the whole industry. The most effective way to alleviate this collective anxiety is not to continue to create a sense of mystery and to seal the experience within individual companies, but to gradually transform those AI possibilities that are still in the unknown into a method and path that can be implemented and reused.

From this perspective, Tencent’s actions in the past two years are actually very representative. To a certain extent, Tencent has been consolidating its ‘pragmatic big factory’ image. For example, we have previously reported that at this year’s GDC, Tencent has come up with the largest number of AI issues in all companies in the industry. In a public speech, the technical solutions, exploration paths and landing results behind the project are shared with the industry as candidly as possible, and the heart is drawn and the lungs are drawn. WorkBuddy and CodeBuddy, which were launched by Tencent in March, were originally self-developed by Tencent and used by internal employees. Agent tools, but they choose to open to all the public during the ‘lobster fever’ period, so that ordinary people are also benefited by the equality of technology.

Now, what Tencent Game School has done in the Tencent Game Creation Competition has also continued the same idea, and even went a step further. It does not just keep ‘opening’ at the conceptual level, but further implements opening up to a more specific and critical level such as talent, technology and business resources.

The combination of these three constitutes a really special part of this competition. It not only gives young developers a stage to show creativity, but tries to build a more complete growth environment: someone leads the way, there are tools available, there are projects to do, and there is also the opportunity to really push the work farther away.

This may also be the most important point of the industry’s attention in this Tencent Game Creation Competition. In an era of rapid changes and the half-life of knowledge is getting shorter and shorter, the Tencent game creation competition provides a paradigm, setting a precedent that can be imitated – instead of firmly guarding against persistence, it is better for enterprises to open up resources, share experience, and explore together. Not only to ’employ’, but also start ‘education’; not only to find people who know how to AI, but also to help more people learn to grow in the AI era.

Let more creators stand on the shoulders of giants, get in touch with the frontier of AI game development faster, and accelerate the talent reserve and cognitive upgrade of the entire industry through opening-up practice. In the end, it is not just a certain game, a certain team, but the innovation speed and talent density of the entire industry.