At the first capital market summit of Guolian Minsheng Securities, Kai-Fu Lee, CEO of Zero-100 Things and Chairman of Innovation Works, delivered a keynote speech on the future development of artificial intelligence. He predicted that 2025 will be the first year for the large-scale implementation of AI applications, marking a new stage of development for artificial intelligence technology.
Kai-Fu Lee pointed out that compared with cloud computing, the burst cycle of the AI2.0 application layer will be significantly shortened, and a comprehensive breakthrough is expected to be achieved within two years. He further analyzed that the current artificial intelligence ecosystem is not healthy, mainly because profits are concentrated in the hands of a few chip companies. A healthy ecosystem should be the application layer (App) to get the most profit, followed by the platform layer and model layer, while chip companies are at the end of the profit chain. He predicts that in the next two years, the artificial intelligence ecosystem will be gradually adjusted to form a more reasonable profit distribution pattern.
From a historical perspective, the IT industry has experienced two major revolutions: the PC era and the mobile Internet era. These two revolutions not only changed people's lifestyles, but also greatly promoted the growth of global GDP. Today, generative AI is leading the third IT revolution. Compared with the previous two revolutions, the AI era will bring more profound changes. The PC era introduces computers into the office, the mobile Internet era allows people to connect with information, applications and others anytime, anywhere, and the AI era allows every application to have the ability to surpass human intelligence.
Kai-Fu Lee also mentioned that the future development of AI models will enter a watershed, mainly divided into two types: one is a trillion- or even ten trillion-level super-large model, and the other is a trillion-level large model. The model of the 100 billion level is pragmatic, fast and low-cost, and is suitable for a wide range of application scenarios; while the super-large model has higher intelligence and broader development potential, but it also requires more resources and investment.
Overall, Kai-Fu Lee’s speech paints a clear blueprint for the future of artificial intelligence. With the continuous advancement of technology and the gradual improvement of ecosystems, AI will play an increasingly important role in various fields and promote social and economic development.