A few days ago, I had dinner with a friend who specializes in policy consulting. He sighed and said, "In the current AI industry, 99% of its value has been locked up by a few giants. ” The model, computing power, and data are all theirs, No matter how lively the open source community is, it only helps them test free features. After hearing these words, the siege lion's heart tightened. That's also why @ SentientAGI makes me feel different. What it needs to do is not the 'next ChatGPT', but an open, verifiable, and distributable AGI network. One sentence: AI should not belong to giants, but should return to the community. 1/ These days, the siege lion has re focused on SentientAGI. In just a few days, it played two moves in a row: One is the upgrade of GRID intelligent network, and the other is the opening of token snapshot and role system in Sentient Chat. Seemingly a functional update, in fact, its AGI network has officially entered the "economic layer experimental period". 2/ Briefly explain the logic behind these two points: ① GRID upgrade: It is not just an agent network, but integrates different models, data sources, and agents into the same autonomous framework. After the upgrade, the call of the agent, data flow, and profit settlement can all be traced on the chain, This means that AI networks are no longer "black boxes", but transparent systems with "accounting and profit sharing". ② Token snapshot+Spaces role mechanism: This mechanism allows users' contributions, interactions, and model usage history to be recorded as incentive data, Laying the groundwork for the next step of the 'open economic layer'. Simply put, it means that every participant, whether they are model trainers, data providers, or users, Everyone can be seen, measured, and motivated. 3/ What concerns me the most is the recognition Sentient has gained in the industry in recent days—— It has just been named the "AI Annual Emerging Project". This means that its philosophy is no longer just a community fantasy, And it has begun to be seen as a possible alternative path by the mainstream AI industry: From closed source models to open source protocols, from centralized platforms to community driven networks. 4/ The siege lion believes that Sentient's true ambition is not to compete with OpenAI and Anthropic for parameters, But rather, it is necessary to establish the underlying protocol layer of an 'AI supply chain'. OML fingerprint → model attribution and call verification; GRID → Autonomous Collaborative Network of Multi Agent Systems; Enclaves → Trusted Execution Environment ensures privacy; Chat+Spaces → User layer interaction and incentive interface. This chain is complete, systematic, and has inherent logic, It is completely different from the traditional path of "big model centrism". 5/ Of course, risks still exist: The robustness of the fingerprint mechanism under multiple retraining needs to be verified; The cost and deployment complexity of TEE confidential computing remain bottlenecks; Whether community governance can avoid being controlled by "technological oligarchs" is also a difficult problem. But what the siege lion is more concerned about is that it is at least trying to solve the "survival problem of open AI" through engineering methods. This is already rare and valuable in the current industry environment. 6/ Summary in one sentence: Others are still discussing whether AI can think, Sentient is already discussing who AI should belong to. Siege Lion believes that this may be one of the most worthwhile decentralized AI projects to track in the long term by 2025. It represents not only a protocol, but also a social experiment to reshape the ownership of AI.