Meitu's Wu Xinhong responds to competition from large models: vertical applications are like professional tools, Meitu's application data is still growing rapidly.

The debate surrounding large models consuming applications has sparked market concerns, and the AI application sector as a whole is facing setbacks. At a recent company annual meeting, Meitu's CEO, Wu Xinhong, discussed the conflict between models and applications. He revealed that even after the release of Nano Banana, Meitu's application data continues to grow rapidly, showing a synergistic effect between general large models and applications. Wu Xinhong admitted that general large models are "unstoppable," leaving little space for applications, but at the same time, the efficiency of general large models in specific vertical scenarios is not very high. He likened large models to a "Swiss army knife," capable of handling general needs and daily tasks, while vertical applications are like scissors, craft knives, fruit knives, nail clippers, and other specialized tools, meeting specific needs in different scenarios. In Wu Xinhong's view, there is always space for application developers in each era, the key lies in deep exploration of high-value vertical scenarios. These scenarios often have very rigid demands, high costs, and low efficiency issues, with customers willing to pay for them, and services can also create highly elastic growth space. The key to the competition between applications and large models lies in whether one can gain the mindset of "I am the most professional in this vertical scenario," solving the last mile and long-tail demands.