People often blame the difficulty of decentralized AI on technology, but in reality, the biggest bottleneck lies in the economic and incentive mechanisms. How to enable computational resources, AI models, and data to circulate efficiently within the network, while ensuring that contributors can retain asset ownership and receive appropriate rewards? This is the key factor that determines whether a project can succeed. Recently, I came across a research plan from a project that addresses these two issues together—by designing a clear economic model that allows network participants to retain ownership during interactions involving computation, models, and data, while also earning corresponding incentives based on their contributions. This approach is quite worth paying attention to.