Sentient Trilogy 3. Yesterday, after discussing the topics of Sentient open source and open weighting, I found that some people still couldn't distinguish the relationship between Sentient technologies and products (some say there are too many terms) ）， So we decided to turn this tweet into a trilogy, where those who haven't read the first two can climb the stairs. Firstly, Sentient Foundation, equivalent to OpenAI, is the main developer of the entire protocol and also the issuer of the coins. Sentient @ SentientAGI is the protocol they developed (although officially referred to as "supervision") ）， SENT is a governance token Then Sentient protocol governs a platform called Grid. What does Grid do? The official definition is an intelligent network that represents open AI fighting against closed source AI giants. In human language, Grid is equivalent to a large local mutual aid group, where various AI models (especially priority open source AI and open weight AI), AI agents, training data, computing power, and various tools are available for group chat. According to their plan, queries sent by users to the Grid will be split and routed to the appropriate AI to receive orders, helping users get the best processing. Finally, Sentient Chat is a chatbot and a terminal that truly interacts with users. So Sentient ≠ ChatGPT, Sentient Chat is equivalent to ChatGPT. The difference between Sentient Chat and ChatGPT is that Sentient Chat is connected to the Grid network, and theoretically, the requests sent by users are processed by the most suitable AI model or agent routing found through Grid analysis. In other words, Sentient Chat is equivalent to the gateway of the Grid network. Next are their several core technologies. Firstly, ROMA. ROMA is not AI, it is a set of task splitting techniques. The parent node splits the complex target into subtasks, passes them as context to the child nodes, and finally aggregates the solutions of the child nodes to provide the final result. ROMA is not a worker, it is equivalent to a middle-level leader responsible for breaking down the tasks of the big leader (user) and handing them over to grassroots employees (Grid network AI) for processing. For example, you ask an AI based on ROMA to write a report on the climate differences between Shanghai and New York for you. Based on ROMA, the parent node can decompose this into specific subtasks: subtask 1 studies the climate in Shanghai (led by Chinese AI), and subtask 2 studies the climate in New York (led by American AI). Once both subtasks are completed, the parent node will find an AI skilled in comparative analysis to create the final comparative task, and finally summarize these results into a comprehensive report. However, this technology is still under development and Sentient Chat itself has not yet used ROMA. Next is OpenDeepSearch. OpenDeepSearch is also not AI, it is a technology developed by them for deep search, which has achieved good results in several previous question sets. And Dobby, this is finally AI, it's an open weight model. From the code analysis, it should be based on Meta's Llama 3.3, which is also a member of Grid's large-scale group chat. However, Sentient Chat is not using Dobby. Dobby is still under development, although it will be one of the many routing sources for Sentient Chat in the future. Finally, fingerprint recognition is one of Professor Pramod's core technologies. Because Grid is a large-scale order taking network, to use the familiar analogy of food delivery, on the one hand, it needs to attract customers, and on the other hand, it also needs to find delivery riders. Because Sentient advocates open AI to combat closed source, the more open source and open weighted AI naturally integrated, the better. One difficulty with open AI in the past was that it was difficult to make profits, so it advocated embedding fingerprints into these open AI and then adding them to the Grid network. After the addition of fingerprint function in the future, if someone uses these AI in the future, the payment amount and SENT rewards will be traced back to the debuggers and developers of these open AI, giving these developers the motivation to optimize these AI and have the possibility of profitability. Finally, regarding governance, Professor Pramod mentioned that AI should be loyal. Closed source AI is loyal to the development company and board of directors, while open AI must be loyal to the entire DAO, and the SENT coin they want to issue is the governance token of this DAO. Finally finished sorting it out, I couldn't help but feel uncomfortable seeing these concepts and nouns mixed up in other people's articles. Now it's refreshing. Finally, please attach a complete relationship diagram for the reference of future researchers