Recently, Zhu Xiaohu, Managing Partner of Jinsha River Venture Capital, stated that they have been paying close attention to the application of AI, focusing only on AIGC applications that can be commercialized and achieve Product-Market Fit (PMF). He also mentioned that AI applications will definitely explode at this time next year. It is true that at the beginning of each cycle, hardware and infrastructure make more money, but in the end, it is the applications that make more than ten times the money. Open-source small models at the edge are definitely the future direction, offering numerous business opportunities. Previously, Zhu Xiaohu talked about high dividend returns in the secondary market this year, and also emphasized the importance of dividend returns in consumer projects. He explained that dividend returns have become a common practice in the consumer industry. Additionally, he expressed optimism towards AI applications, especially the upcoming explosion next year. He believes that small models with hundreds of billions of parameters may be able to run directly at the edge in the future. Zhu Xiaohu has been focusing on AI applications that can be commercialized and achieve PMF, and he is confident that AI applications will explode next year. He mentioned that they are currently closely monitoring developments in this area. Regarding Fu Sheng's belief that the business model of the AI "Four Little Dragons" is poor, Zhu Xiaohu is even more radical in his views. He thinks that the technology behind the "Four Little Dragons" has no threshold, and can at most lead others by half a year. He believes that there is not much difference in technology, and even large models are not much better. Zhu Xiaohu's opinion is that GPT-5 may not be as impressive, as GPT-4 already meets the majority of commercial demands. He compared this development to the X86 evolution, where upgrades after the 586 version were marginal. He does not have high expectations for GPT-5. He also believes that open-source small models at the edge will be the future direction, offering numerous business opportunities. He sees great potential in this area and predicts a burst of activity in the next one or two years. Many people are already exploring this direction, and he suggests focusing on opportunities in this field such as robots and emotional machines, which require small models for specific applications at the edge. According to Zhu Xiaohu, from PC internet to mobile internet, applications have been the most profitable. At the start of each cycle, hardware and infrastructure make more money, but in the end, it is the applications that bring in the most revenue. This time with large models, the same pattern will likely occur, with applications earning more money. He believes that applications will definitely explode next year, as open-source models have reached a sufficient level of intelligence.