Author: 0XNATALIE
Since the second half of this year, the topic of AI Agent has continued to rise. At first, the AI chatbot terminal of truths attracted widespread attention for its humorous posts and replies on X (similar to "Robert" on Weibo), and received a $50,000 grant from a16z founder Marc Andreessen. Inspired by its posts, someone created the GOAT token, which rose by more than 10,000% in just 24 hours. The topic of AI Agent immediately attracted the attention of the Web3 community. Later, the first decentralized AI trading fund based on Solana, ai16z, was launched, and the AI Agent development framework Eliza was launched, which triggered a dispute over uppercase and lowercase tokens. However, the community still has a vague concept of AI Agent: What is the core of AI Agent? How is it different from the Telegram trading robot?
Working principle: perception, reasoning and autonomous decision-making
AI Agent is an intelligent agent system based on a large language model (LLM) that can perceive the environment, make reasoning decisions, and complete complex tasks by calling tools or performing operations. Workflow: Perception module (obtaining input) → LLM (understanding, reasoning and planning) → tool call (executing tasks) → feedback and optimization (verification and adjustment).
Specifically, AI Agent first obtains data (such as text, audio, images, etc.) from the external environment through the perception module and converts it into structured information that can be processed. As a core component, LLM provides powerful natural language understanding and generation capabilities, acting as the "brain" of the system. Based on the input data and existing knowledge, LLM performs logical reasoning to generate possible solutions or formulate action plans. Subsequently, AI Agent completes specific tasks by calling external tools, plug-ins or APIs, and verifies and adjusts the results based on feedback to form a closed-loop optimization.
In the application scenario of Web3, what is the difference between AI Agent and Telegram trading robots or automated scripts? Take arbitrage as an example. Users want to conduct arbitrage transactions under the condition that the profit is greater than 1%. In the Telegram trading robot that supports arbitrage, users set a trading strategy with a profit greater than 1%, and the bot will start to execute. However, when the market fluctuates frequently and arbitrage opportunities are constantly changing, these bots lack the ability to assess risks and will execute arbitrage as long as the profit is greater than 1%. In contrast, AI Agent can automatically adjust its strategy. For example, when the profit of a transaction exceeds 1%, but the risk is too high through data analysis, and the market may suddenly change and cause losses, it will decide not to execute the arbitrage.
Therefore, AI Agent is self-adaptive, and its core advantage lies in its ability to self-learn and make decisions. Through interaction with the environment (such as the market, user behavior, etc.), it adjusts its behavior strategy according to feedback signals and continuously improves the performance of task execution. It can also make decisions in real time based on external data and continuously optimize decision-making strategies through reinforcement learning.
Isn’t this a bit like a solver (slover) in the intent framework? AI Agent itself is also a product based on intent. The biggest difference between AI Agent and solver under intent framework is that solver relies on precise algorithm and has mathematical rigor, while AI Agent decision-making relies on data training, and often needs to approach the optimal solution through continuous trial and error during training.
Mainstream framework of AI Agent
AI Agent framework is the infrastructure for creating and managing intelligent agents. Currently in Web3, the more popular frameworks include Eliza from ai16z, ZerePy from zer, and GAME from Virtuals.
Eliza is a versatile AI Agent framework built with TypeScript. It supports running on multiple platforms (such as Discord, Twitter, Telegram, etc.), and through complex memory management, it can remember previous conversations and contexts, maintain stable and consistent personality traits and knowledge answers. Eliza uses the RAG (Retrieval Augmented Generation) system, which can access external databases or resources to generate more accurate answers. In addition, Eliza integrates TEE plug-ins, allowing deployment in TEE to ensure data security and privacy.
GAME is a framework that enables and drives AI Agents to make autonomous decisions and actions. Developers can customize the behavior of agents according to their needs, expand their functions, and provide customized operations (such as social media publishing, replying, etc.). Different functions in the framework, such as the agent's environmental location and tasks, are divided into multiple modules for developers to configure and manage easily. The GAME framework divides the decision-making process of AI Agent into two levels: high-level planning (HLP) and low-level planning (LLP), which are responsible for tasks and decisions at different levels. High-level planning is responsible for setting the overall goal and task planning of the agent, making decisions based on the goal, personality, background information and environmental status, and determining the priority of tasks. Low-level planning focuses on the execution level, converting the decisions of high-level planning into specific operation steps, and selecting appropriate functions and operation methods.
ZerePy is an open source Python framework for deploying AI Agents on X. The framework integrates the LLM provided by OpenAI and Anthropic, enabling developers to build and manage social media agents and automate operations such as posting tweets, replying to tweets, and liking. Each task can be assigned different weights based on its importance. ZerePy provides a simple command line interface (CLI) for developers to quickly start and manage agents. At the same time, the framework also provides Replit (an online code editing and execution platform) templates, through which developers can quickly start using ZerePy without complex local environment configuration.
Why does AI Agent face FUD?
AI Agent seems to be smart, can lower the entry threshold and improve the user experience, why is there still FUD in the community? The reason is that AI Agent is essentially still just a tool. It cannot complete the entire workflow at present, and can only improve efficiency and save time at certain nodes. Moreover, at the current stage of development, the role of AI Agent is mostly focused on helping users to issue MeMe and operate social media accounts with one click. The community jokingly calls it "assests belong to Dev, liabilities belong to AI."
However, just this week, aiPool was released as an AI Agent for token pre-sale, using TEE technology to achieve trustlessness. The wallet private key of this AI Agent is dynamically generated in the TEE environment to ensure security. Users can send funds (such as SOL) to the wallet controlled by the AI Agent, and the AI Agent then creates tokens according to the set rules, starts a liquidity pool on the DEX, and distributes tokens to qualified investors. The entire process does not need to rely on any third-party intermediary, and is completely completed by AI Agent in the TEE environment, avoiding the common rug pull risk in DeFi. It can be seen that AI Agent is gradually developing. I think that AI Agent can help users lower the threshold and improve the experience, even if it only simplifies some asset issuance processes, it is meaningful. But from the macro perspective of Web3, AI Agent, as an off-chain product, currently only serves as a tool to assist smart contracts, so there is no need to over-exaggerate its capabilities. Since there is a lack of significant wealth effect narratives other than MeMe in the second half of this year, it is normal for AI Agent's hype to revolve around MeMe and become popular. MeMe alone cannot maintain long-term value, so if AI Agent can bring more innovative gameplay to the transaction process and provide tangible landing value, it may develop into a common infra tool.
Catherine
Kikyo
Alex