ChainCatcher "From Cryptocurrency to Smart Economy" Roundtable: AI Agents are shifting from conversational to execution-oriented, with trust and verifiability becoming key bottlenecks for scalability

At the "Crypto 2026: From Cryptocurrency to Smart Economy" themed forum held in Hong Kong, guests including KiteAI's Asia Pacific head Laughing, Sentient's Asia Pacific head Anita, Brevis content director Nic Tang, Mentis product head Jerry, and Predict.fun's Asia Pacific BD head Ah Huang Ricardo engaged in a roundtable discussion on the theme "AI × Crypto --- The Next Generation Digital Economy Foundation."

Regarding whether the outbreak of open-source AI Agents (such as "raising lobsters") is a short-term sentiment or a long-term trend, Laughing believes this marks a paradigm shift of AI from "conversational" to "executive," signaling the beginning of a long-term trend. However, Agents lack a physical identity, making it difficult for merchants to trust their trading behavior, and issues of data leakage and uncontrollability remain obstacles that must be overcome for scaling. Jerry pointed out, based on his own usage experience, that current Agents face security risks such as memory confusion, infinite loops consuming tokens, and accidental file deletions. Although newly emerged Agents have improvements, they still have shortcomings. Ah Huang Ricardo stated that behind the short-term sentiment is real user demand, but truly reproducible profitable trading Agents are still very rare, and there is a long way to go.

Addressing core bottlenecks such as trust, security, and verifiability, Nic Tang emphasized that the current execution process of Agents is like a black box, where users cannot verify whether their decisions are executed as promised. Zero-knowledge proof technology can mathematically prove the source of the Agent's output and compliance of behavior while not exposing privacy. Anita added that Sentient is committed to building an open and collaborative AI agent network, promoting the deep integration of AI and blockchain in identity, data, and incentive layers, providing foundational support for the next generation digital economy.

Regarding the application scenarios that may scale first, the guests generally believe that the programming field is relatively mature, and Agents in prediction markets perform better than humans in specific tracks due to their rationality and lack of emotions. In the short term, they are optimistic about API Marketplace and cross-platform shopping Agents, while in the medium term, they are optimistic about content creation AI (such as AI influencers). Agent-to-Agent interaction and the Agent/Skill Marketplace are seen as important future trends.