GuShiio.com AMMO defines an abstract space called MetaSpace, allowing all data surrounding the AI Agent to be deployable within the space in the form of Vector vectors.
It took some time to carefully read Ammo’s new white paper and felt a lot. Below, share some inspirations:
1) The market’s pursuit of AI Agents is essentially that it does not satisfy the fact that AI is just a query tool in the Copilot model. Users should answer what AI asks, but should be more like the companion growth model of Buddy’s model, capable of understanding, thinking, and proactively creating value. Push it to people. This is the key to AI Agents being lifted to a narrative level;
2) The traditional web2 AI single model started out as “instrumentalized pragmatism”. It is easy to form isolated islands of data sources in multimodal collaboration, and it is difficult to truly achieve intelligent breakthroughs. Although web3 proposes the individual autonomy of AI Agents. Form, but it is still far from achieving the goal, and AI’s autonomous decision-making is far more complex than imagined. Only by letting AI do assisted automated learning and path recommendation can people enhance the “symbiotic model” of AI autonomous learning through feedback can it truly become the dominant direction of AI Agents in the next sense;
3) AMMO defines an abstract space called MetaSpace, allowing all data surrounding the AI Agent to be deployable in the space in the form of a Vector vector. It is like how the blockchain initially defined Hash, which led to all subsequent links. The same protocol and application form. This original form starting with Vector can not only serve web3, but is also a framework standard suitable for web2 multimodal. Coupled with the MAS multimodal collaboration system on top of it, it can transform AI’s current academic direction. The “think tank” orientation becomes a “practical” orientation toward practical application scenarios such as work, games, and education;
4) How to understand it in a popular manner? We regard MetaSpace as a large shopping center. Each functional layer belongs to a SubSpace, and each area has a different knowledge base. The Buddies system is an intelligent shopping guide system. Goal Buddies serves as a professional shopping guide to select some high-quality products. Recommend for you; User Buddies is more like a personal assistant that can provide customized solutions based on your consumption habits and budget;AiPP collects feedback suggestions like a general service desk to improve service quality;
Overall, it is necessary to make AI Agents run through necessary components such as MetaSpace+Buddies+AiPP human-computer feedback system to truly accelerate the mass production and practical implementation of AI Agents;
5) The white paper shows more of an offline AI Agent multimodal collaboration framework and engineering implementation ideas. Some definition standards on the combination chain, including ID identity system, Memory memory system, Character feature system, Context context management, Oracle system and other component definitions still need to be further explored (the “chained” general standard framework I often mentioned before);
above.
It should be said that this is the most emotional and pragmatic project that has been seen in recent times with the implementation of macro architecture and applications, as well as engineering realization ideas, but maybe everyone has a sense of abstraction after reading the above. It is true that AI Agents are far from being truly popularized and applied on a large scale than imagined, but there are indeed more and more excellent teams coming in, and some innovative solutions and ideas are also being planned. The market is waiting for an innovative “singularity”. The birth of “singularity”.
