Lu Mingzhou composed himself and decided to answer his boss's question first, quickly getting into work mode.
"I don't have in-depth knowledge of the gaming industry, but I've had some contact with small and medium-sized companies before. The mobile game market is currently more popular than the PC game market. If you don't consider small, low-cost indie games, you can make them for tens of millions. As for large-scale games, including design, programming, and art, it's possible to invest hundreds of millions over several years."
After listening, Han Luyi had a general idea of what was going on.
Given the current revenue and profits of Origin Group, they can afford to support even a game development team that doesn't make money for the time being.
However, considering that every penny spent on this means that it cannot be spent on Computing Power and research into new models, such an investment must first prove its value.
Either it has intrinsic value that can generate profit.
Either it improves the Model Training effect through the data it provides, and thus the value of the corresponding model.
Zhao Wenyuan, who hadn't spoken until now, suddenly said, "Boss Han, why are we suddenly getting into games? It's true that games can make money, but... what does that have to do with our work on Artificial Intelligence? It's neither an application nor a model!"
Lu Mingzhou felt a surge of gratitude towards Zhao Wenyuan.
He certainly wanted to say that too, but it would be more appropriate for Zhao Wenyuan to speak from a research and development perspective.
Shifting focus to investing in other industries now would not only impact the company's core business and the ongoing IPO process, but also be highly unlikely to yield significant results, making it a losing proposition. Such a major strategic shift is a serious mistake.
Han Luyi smiled and said, "Old Lu, Wenyuan, you two go first. I understand that my words came out a bit suddenly and you can't accept them right away. Here's what I have to say: you can listen to my thoughts first."
"I ran into Zhang Wenduo before you came back and asked him about his project, that Social World Model. I found it quite interesting."
Zhao Wenyuan's heart skipped a beat upon hearing this, and he said, "Boss Han, isn't his project lacking data? Can you get your hands on the data?"
Han Luyi glanced at him and said, "You've hit the nail on the head. I can't get the data, but I can get the game."
"What?" Zhao Wenyuan was taken aback.
Han Luyi then shared his conversation with Zhang Wenduo that morning about the Data Bureau and how to solve them. He also recalled how, reminded by real-life events, he realized that games, especially massively multiplayer online games, are essentially miniature societies. Furthermore, because online games are played on computers, the in-game information players encounter is relatively closed off, making it easier to establish causal relationships based on data obtained in this environment.
After listening, Zhao Wenyuan fell into deep thought, and only after a long while did he speak:
"But if the sole purpose is to obtain data on human-to-human interactions, setting up a game company specifically for this, and developing one or even several multiplayer games, seems a bit too extravagant, doesn't it? Ultimately, what is the use of this Social World Model?"
Han Luyi smiled again.
What exactly can the Social World Model be used for? Zhang Wenduo himself had never thought about this question, but Han Luyi had been thinking about it since he first came into contact with the concept, and now he has gained some insights.
"Wenyuan, what do you think is the essence of the model proposed by Zhang Wenduo? Is it a multi-step logical deduction? Glutinous Rice Ball 1.0 is already very good. I believe that with your abilities, the 2.0 you lead will perform even better. We don't need to invest in this direction. Or is it predicting individual behavior? This doesn't reflect the influence of interactions between people." Han Luyi asked and answered himself, negating the two possibilities he had proposed, before saying the last one, "The most powerful aspect of this model is that it is designed for reflexivity. How does a decision, after interacting with the consequences of other decisions, ultimately influence the next decision? The greatest significance of this Social World Model lies in its ability to explain all these influences and form a logically rigorous chain."
After saying that, he glanced at Lu Mingzhou again and continued, "Aren't there many things in this world that exhibit reflexivity? Every time we study a specific problem, the first step is to abstract the real world, assuming an ideal environment and that external conditions remain unchanged. But this World Model itself can help us calculate the environment in which we will make the next decision and how much the accuracy of the decision will be improved."
"Think about it: financial investment decisions are a typical reflexive scenario. What do quantitative firms typically use Large Model for? Large Model are largely used to structure unstructured data, turning financial reports, news, conference calls, and research reports into signal data that decision-making models can directly access. However, the real quantitative logic still uses a combination of statistical models and manual rules. Why? Why can't we use Large Language Model directly for financial trading? Because current Large Language Model cannot solve the problem of reflexivity!"
"What if Social World Model could work?"
"And take our earlier example. What if the new model could predict the trajectory of unexpected events? What if it could make reasonable inferences and provide explanations for the subsequent impact of large-scale events? How valuable would such a model be?"
After Han Luyi finished speaking, Zhao Wenyuan and Lu Mingzhou both fell silent.
Zhao Wenyuan suddenly turned to Lu Mingzhou and said, "Old Lu, I don't understand the specifics of financial operations, but it seems we really have to play this game now."
Lu Mingzhou looked at Zhao Wenyuan with a complicated expression.
Old Zhao, you switched sides so quickly!
But it must be said that even he himself has been convinced.
Going public is important, but the process of going public should not slow down the company's current development; otherwise, it would be putting the cart before the horse.
Now, after Han Luyi Boss Han explained the two projects he proposed— Brain-Computer Interface and large-scale online games— Lu Mingzhou is convinced. It's clear that these two projects will play a crucial role in the future development of the Origin Group. Under these circumstances, it would be unreasonable to resist the right investment direction simply to reduce the workload of going public.
After thinking for a moment, Lu Mingzhou said, "Boss Han, since you think this direction is valuable, I suggest that we don't rush to set up a game company or recruit staff to create a game department. Companies like Google and Amazon have already proven that even for top global giants, even with plenty of money, the probability of successfully developing a game from scratch on the first try is still too low."
Han Luyi nodded and asked, "Then what suggestions do you have?"