Zhu Xiaohu and Fu Sheng talked about model entrepreneurship and investment again, and this time the differences have decreased

Zhu Xiaohu and Fu Sheng stood together again to talk about AI.
On May 8th, at the 18th China Investment Annual Conference · Annual Summit, Zhu Xiaohu, Managing Partner of Jinshajiang Venture Capital, and Fu Sheng, Chairman and CEO of Leopard Static and Chairman of Orion Starry Sky, once again stopped their confrontation on the business and investment of the AI big model category.
A topic that has been directly thrown away is whether the future growth limit of the domestic large mold industry is in the format of the “AI Four Dragons”? The “AI Four Dragons” usually refer to four companies: Shangtang Technology, Kuangwang Technology, Yitu Technology, and Yuncong Technology.

Fu Sheng and Zhu Xiaohu do not agree with this. Fu Sheng pointed out that the industry format represented by the “AI Four Dragons” belongs to AI 1.0, and the AI skills and business structure of this era have not lost the rapid follow-up of technology giants. The era of AI big molds belongs to AI 2.0, which quickly resonates in the field of global technology.
In contrast, the time gap that domestic large model companies have is significantly extended, and the accompanying costs, technical cooperation, and trade forms will be provocations that a company cannot surpass.
Zhu Xiaohu directly stated that the lower limit of the domestic big mold category is the “AI Four Little Dragons”. He believed that the industry where the “AI Four Little Dragons” were located was not technically challenging, and they were at the forefront of each other for at most six months. One of the important reasons for their growth was that they encountered a huge industry surplus of security repair.
In his view, the form of large mold trade is “worse”: firstly, there is no difference in skills between companies; Secondly, from GPT-3.5, GPT-4 to GPT-5, each generation of skills has to be reinvested, and the investment scope has exponentially declined; Furthermore, the poor form of trade results in a monetization cycle of only two to three years for each generation of skills.
“I feel like Sam Altman is blowing the conch, GPT-5 is definitely not so close,” Zhu Xiaohu said politely. GPT-4 has already met the vast majority of trade needs, and he is no longer waiting for GPT-5. Fu Sheng also predicted, “If GPT-5 is not released for such a long time, it will either be due to little improvement in functionality or too much improvement in cost.”
This is not the first time the two have stopped talking about AI topics. In June last year, Fu Sheng and Zhu Xiaohu had a dispute over ChatGPT’s business in WeChat’s enemy circle, which they were preparing to attend to.
At that time, Zhu Xiaohu mentioned in an article that “ChatGPT is very unfriendly to conservative companies, so please abandon your financing ideals in the next two to three years.”. Fu Sheng quickly forwarded and criticized, “Half of the companies in Silicon Valley that have been operating have started (operating) around ChatGPT, and our investors can still be so ignorant and fearful.”
Not only that, the two sides have stopped multiple rounds of dialogue on whether the big model’s business can come at a cost, resulting in a distinct attitude between investors and business owners.
Regarding this debate, Fu Sheng expressed that one of the reasons why he shouted at Zhu Xiaohu from afar at the time was that action entrepreneurs did not like investors to “stand and talk without hurting their backs.”. He believed that the reason why Shouye was prosperous was because there were various non resonance factors. Secondly, at that time, big models had just become popular, but the boundaries of their talents were still unknown and there should be no rush to determine.
Fu Sheng felt that no matter how good a technique is, it must be packaged into a product, and this was the timing for the practitioner, but at that time, everyone was still unclear. “I think in the era of great skill revolution, there must be more opportunities for practitioners.”
Zhu Xiaohu, on the other hand, expressed that part of the reason why the dispute triggered the attachment was due to the excessive interpretation of the media, and the fundamental differences between the two sides were not too deep. “Today, to be honest, I personally am very optimistic about the use of AI, especially its outbreak next year.”
Beyond the big model, the topics of “patient resources” and “dangerous investment sharing” for venture capital itself are also important topics in this competition.
Fu Sheng believes that technology investment should focus more on long-term returns rather than short-term dividends. “In fact, the risks and benefits of technology investment are inversely proportional, and (excessively) focusing on dividends will miss out on companies like Toutiao.”
Zhu Xiaohu, on the other hand, should have pointed out that the previously mentioned “profit sharing” refers to the consumption track, which has become a common practice in the consumption industry. He also exaggerates that the condition for “patient resources” is that the invested enterprise has stable cash flow. “How can you be patient when you invest one billion yuan and haven’t seen a penny return for over a decade?” he said.