Recently, I have been reading several updates from @ floor_io. It's not difficult to understand them separately, but when viewed together, the direction becomes much clearer. On one hand, we are expanding the assets and trading entry points in Crypto, while on the other hand, we are pushing AI infrastructure towards real-world scenarios such as government, agriculture, and public services. Let's first take a look at the recently noteworthy FOMO flock, whose FOMO has been integrated into Robinhood Chain. Currently, people are still familiar with tokenized stocks, memecoin, and DeFi, but FLock wants to incorporate a completely different asset into them: AI Model FOMO itself is a Fair Open Model Offering. Simply put, it's not just about issuing tokens to an AI model, but about trying to connect the Model Token with the real model inference and AI usage requirements on the FLock API Platform. In this way, the model is no longer just a piece of code in the background, but can gradually enter the same set of on chain economic relationships of "issuance, trading, and invocation". I think this direction is more interesting than just talking about AI tokens. In the past, Crypto has capitalized many things, such as stocks, gold, treasury bond and accounting power. Now FLOCK wants to move forward: If an AI model itself can be continuously called and generate usage requirements, can it also become an on chain asset? And FLock doesn't just leave the entrance on Robinhood Chain. FLOCK has previously adopted Chainlink CCT and connected Base, Ethereum, BNB Chain, HyperEVM, and Robinhood Chain through CCIP/Transporter. The distribution channels on the chain are being established, and the centralized market is not stopping either. On September 12th, OKX has launched FLOCK/USDT perpetual contracts, which can be traded on the web, app, and API, with a maximum support of 20x. So I prefer to look at these two things together. Robinhood Chain undertakes FOMO, Model Token, and subsequent on chain AI asset economy; OKX has added a trading portal for FLOCK in the centralized derivatives market. One is expanding the ecosystem and the other is expanding the market, but if FLock only has these actions recently, I think it is still easy to be classified as a "Crypto+AI" project. What really made me feel like there was a difference in the route behind it was actually another line. FLock and the UNDP Dominican Republic Country Office signed a new MoU on September 4th. This matter is almost completely different from the two directions of Crypto mentioned above. The two sides are discussing practical issues such as digital innovation in the public sector, data governance, trustworthy public systems, as well as social protection, climate/environmental sustainability. And it's not just after signing the MoU that we start looking for applications. In May of this year, the FLock team conducted a crop microinsurance pilot in the Dominican Republic, attempting to use AI for risk assessment, blockchain for payout/payment processing, and privacy preserving AI to avoid direct exposure of raw sensitive data for the purpose of using AI. This case actually makes it easier for me to understand why FLock has been working on decentralization and privacy AI. If it's just generating an image or writing a paragraph, ordinary users may not even care where the data behind the model is. But if it were government data, medical data, agricultural data, financial data, and personal privacy data, things would immediately be different. These scenarios are not about 'not wanting to use AI', but rather many raw data that cannot be easily handed over to external platforms. FLock's previous approach to advancing sovereign AI in Sarawak was actually similar, using federated learning to keep data as local as possible while completing model training and collaboration. So stringing together these recent actions, I think it's more interesting than watching any official announcement alone: Robinhood Chain → Expanding AI Model's On Chain Asset Scenarios OKX → Expand FLOCK's trading portal UNDP Dominican Republic → Advancing AI infrastructure towards real public service scenarios The first two address distribution and liquidity in the crypto world, while the latter answers another longer-term question: Is there really any place that needs this AI infrastructure after leaving Crypto? If FOMO's Model Token can generate more real inference demand in the future, and at the same time, government and public institutions continue to produce feasible pilots, then FLock's line will not just tell DeAI another new story. What it really wants to connect may be the AI asset economy and the data usage needs of the real world.