Search
On August 6, the 13th Internet Security Conference (ISC.AI 2025) convened in Beijing. During the event, Zhou Hongyi, the founder of 360, highlighted that despite the ever-growing capabilities of large models, they alone are insufficient. He identified two significant challenges large models face in enterprise applications: limited reasoning abilities and an inability to autonomously execute tasks. While reasoning abilities have seen notable enhancements over the past year, the issue of autonomous task execution remains unresolved. Zhou Hongyi underscored that the evolution of large models into agents represents an inevitable trend. Agents possess the ability to comprehend goals, plan tasks, invoke tools, and maintain memory, facilitating a seamless transition from demand to result. The crux of this evolution lies in their capacity to utilize tools and plan intricate tasks.
-
C114 Communication Network -
Communication Home
