HKUST's Xu Jialong: The agent moat has not yet solidified; model differences are more reflected in efficiency rather than disruptive breakthroughs

CryptoWorld News reports that on April 21, during the roundtable discussion titled “Decoding Web 4.0: When AI Agents Take Over On-Chain Permissions,” Xu Jialong, Associate Vice President of Hong Kong University of Science and Technology, discussed the “Agent moat” and said that there are differences in the model training paths and technical systems that different AI agents rely on, which leads to a clear split in the real-world user experience. Recently, some new models and tools have shown better performance in generation quality and execution efficiency, and even demonstrate higher ceilings in development output. However, he pointed out that, from the current stage, these differences have not yet formed a decisive generation gap; they are closer to “efficiency improvements” rather than a “paradigm breakthrough.” In other words, competition among agents is still in a phase of rapid evolution, and no stable or insurmountable technical barriers have appeared. The iteration pace of today’s AI agents and large models is extremely fast—new products or new capabilities appear almost every week—driving the industry to keep moving forward. But from the perspective of actual use and business decision-making, whether it’s necessary to continuously and frequently track these changes still requires careful evaluation.