an LLM focused on Web3 and AI agents

Fetch.ai has just introduced ASI-1 Mini, the first large language model (LLM) native to Web3, which offers innovative access to decentralized artificial intelligence. ASI-1 Mini allows the Web3 community to invest in, train, and own AI models, reducing infrastructure costs and bringing new opportunities for businesses and users worldwide.

Fetch.ai and the innovative ASI-1 Mini: a scalable LLM designed for the Web3 world

Fetch.ai has announced the launch of ASI-1 Mini, a new large language model (LLM) designed for web3. Compared to other LLMs already existing in the decentralized field, ASI-1 Mini has been developed to optimize integration with blockchain-based applications, offering a more accessible alternative in terms of costs and AI infrastructure.

ASI-1 Mini offers high performance with only two GPUs, significantly reducing hardware costs, which are traditionally an obstacle for small and medium-sized enterprises looking to implement AI-based solutions. Furthermore, thanks to its innovative architecture based on the Mixture of Experts (MoE) framework, ASI-1 Mini is extremely scalable. It can adapt to different business needs and grow flexibly, handling increasingly complex tasks.

This makes it an ideal solution for a wide range of applications, from small businesses to more complex corporate realities. The modular approach also allows enterprise-level AI to reach a broader audience at affordable costs. In particular, ASI-1 Mini can be used for activities such as data analysis, business process automation, virtual assistance, and decision optimization, offering advanced tools for various sectors.

Fetch.ai and the “Black-box problem”: how ASI-1 Mini improves transparency in AI decisions

A common problem in traditional AI models, notably in deep learning systems, is the so-called “black-box problem“. This phenomenon occurs when AI models make complex decisions without providing explanations on how they arrived at such conclusions. The lack of transparency in an AI model can generate concerns, especially in sensitive areas such as health, finance, and justice, where it is crucial to understand the reason behind a decision.

A classic example concerns an AI algorithm used in the sanitary sector to predict the probability that a patient will develop a disease. If the model suggests a high risk but does not explain which factors were considered to reach that conclusion, doctors and patients might find themselves in difficulty making informed decisions. Another example concerns the use of AI in banking loans, where a model might deny a loan without clearly explaining which variables were evaluated, increasing the risk of biases or arbitrary decisions.

Fetch.ai has tackled this problem with ASI-1 Mini, which does not merely perform superficial reasoning but is designed to operate with continuous multi-stage reasoning. This approach allows for real-time monitoring and correction of decisions made by the system, increasing transparency and trust in AI. Although the model does not completely eliminate opacity, it provides more detailed and accessible explanations compared to other competitor solutions.

The Web3 freemium experience for FET holders

“`html

One of the features that distinguishes ASI-1 Mini is its freemium integration offered to FET token holders, the native cryptocurrency of Fetch.ai. FET holders can access ASI-1 Mini for free and its basic AI features, instead participating in a model that rewards investment in tokens.

“`

This system allows users to leverage cutting-edge artificial intelligence without the high initial costs typical of other traditional AI solutions. Essentially, ASI-1 Mini becomes an accessible resource for anyone who has participated in the Web3 ecosystem of Fetch.ai, reinforcing the concept of a decentralized community where users can earn tangible benefits from the use and development of AI.

The freemium system offered by Fetch.ai is not limited to a simple possibility of using the model. FET holders can also actively participate in the training and optimization of AI models through the ASI: platform. This allows users to contribute to the growth of the AI ecosystem by training specific models for different applications and sharing the results with the community. In return, users can earn rewards in tokens, thus creating a virtuous circle that stimulates adoption and innovation in Web3.

The connection with AI agents: the future of collaboration in the Fetch.ai ecosystem

Another crucial aspect of ASI-1 Mini is its ability to interact with autonomous agents in the Fetch.ai ecosystem. ASI-1 Mini is integrated with AgentVerse, the AI agents marketplace, which allows users to create and deploy autonomous agents that perform tasks in the real world through simple language commands. This connection between advanced language models and autonomous agents opens new opportunities for the automation of complex processes, from managing business operations to solving real-time problems.

Once fully integrated, ASI-1 Mini will allow users to request various activities, such as hotel or restaurant bookings, simply by making a textual request. These requests will be converted into micro-agents, hosted within AgentVerse, which will perform the operations autonomously. Additionally, developers will be able to monetize these micro-agents, contributing to the creation of an open-source ecosystem where customized AI functionalities will be easily accessible, fostering the development of an economy based on autonomous agents. With the evolution of AI and autonomous agents, these synergies could create new economic opportunities and enhance real-time decision-making capabilities.