Article source: ChainFeeds
Author: 0XNATALIE
Since the second half of this year, the topic of AI Agents has continued to rise in popularity. Initially, the AI chatbot terminal of truths gained wide attention for its humorous posts and replies on X (similar to 'Robert' on Weibo) and received $50,000 funding from a16z founder Marc Andreessen. Inspired by its published content, someone created the GOAT token, which surged over 10,000% within just 24 hours. The topic of AI Agents quickly drew attention from the Web3 community. Subsequently, the first decentralized AI trading fund based on Solana, ai16z, was launched, introducing the AI Agent development framework Eliza, which sparked a token war. However, the community remains unclear about the concept of AI Agents: What is the core of AI Agents? How do they differ from Telegram trading bots?
Working principle: Perception, reasoning, and autonomous decision-making
AI Agent is an intelligent agent system based on large language models (LLM), capable of perceiving the environment, reasoning, and making decisions while completing complex tasks by invoking tools or executing operations. Workflow: Perception module (input acquisition) → LLM (understanding, reasoning, and planning) → Tool invocation (task execution) → Feedback and optimization (verification and adjustment).
Specifically, AI Agents first acquire data from the external environment (such as text, audio, images, etc.) through perception modules and convert it into structured information that can be processed. LLM serves as the core component, providing powerful natural language understanding and generation capabilities, acting as the system's 'brain'. Based on the input data and existing knowledge, the LLM performs logical reasoning to generate possible solutions or formulate action plans. Subsequently, the AI Agent completes specific tasks by invoking external tools, plugins, or APIs, and verifies and adjusts results based on feedback, forming a closed-loop optimization.
In the application scenarios of Web3, how do AI Agents differ from Telegram trading bots or automated scripts? Taking arbitrage as an example, users wish to conduct arbitrage trades under the condition that profits exceed 1%. In a Telegram trading bot that supports arbitrage, users set the trading strategy for profits greater than 1%, and the bot begins executing. However, when the market fluctuates frequently and arbitrage opportunities change continuously, these bots lack risk assessment capabilities and execute arbitrage as long as the profit condition is met. In contrast, AI Agents can automatically adjust their strategies. For example, when the profit of a certain trade exceeds 1%, but data analysis assesses the risk as too high, potentially leading to losses due to sudden market changes, it will decide not to execute that arbitrage.
Therefore, AI Agents have self-adaptive capabilities, with their core advantage being the ability to learn and make autonomous decisions. By interacting with the environment (such as market and user behavior), they adjust behavioral strategies based on feedback signals, continuously improving task execution effectiveness. They can also make real-time decisions based on external data and continuously optimize decision strategies through reinforcement learning.
Does this sound a bit like a solver under the intent framework? The AI Agent itself is also a product based on intent, with the biggest difference from solvers under the intent framework being that solvers rely on precise algorithms, which have mathematical rigor, while AI Agent decision-making relies on data training and often requires continuous trial and error during the training process to approach the optimal solution.
Mainstream frameworks for AI Agents
The AI Agent framework is the infrastructure used to create and manage intelligent agents. Currently, popular frameworks in Web3 include ai16z's Eliza, zerebro's ZerePy, and Virtuals' GAME.
Eliza is a multifunctional AI Agent framework built using TypeScript, supporting operation on multiple platforms (such as Discord, Twitter, Telegram, etc.), and through complex memory management, it can remember previous conversations and context, maintaining stable and consistent personality traits and knowledge responses. Eliza uses a RAG (Retrieval Augmented Generation) system, which can access external databases or resources to generate more accurate responses. Additionally, Eliza integrates TEE plugins, allowing deployment in TEE to ensure data security and privacy.
GAME is a framework that empowers and drives AI Agents to make autonomous decisions and actions. Developers can customize the behavior of the agents according to their needs, extend their functionalities, and provide customized operations (such as social media posting, replying, etc.). Different functionalities in the framework, such as the agent's environmental location and tasks, are divided into multiple modules for easy configuration and management by developers. The GAME framework divides the decision-making process of AI Agents into two levels: High-Level Planning (HLP) and Low-Level Planning (LLP), each responsible for different levels of tasks and decisions. High-Level Planning is responsible for setting the overall goals and task planning for the agent, making decisions based on goals, personality, background information, and environmental states, and determining the priority of tasks. Low-Level Planning focuses on execution, translating the decisions of High-Level Planning into specific operational steps and selecting appropriate functionalities and methods.
ZerePy is an open-source Python framework for deploying AI Agents on X. This framework integrates LLMs provided by OpenAI and Anthropic, enabling developers to build and manage social media agents that automate tasks such as posting tweets, replying to tweets, liking, etc. 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. Additionally, the framework offers Replit (an online code editing and execution platform) templates, allowing developers to quickly get started with ZerePy without complex local environment configurations.
Why does the AI Agent face FUD?
AI Agents seem intelligent, capable of lowering entry barriers and enhancing user experience, so why is there FUD in the community? The reason is that AI Agents are essentially still just tools; they cannot currently complete the entire workflow but can only enhance efficiency and save time at certain nodes. Moreover, in the current development stage, the role of AI Agents is largely concentrated on helping users issue MeMe with one click and manage social media accounts. The community jokingly remarks, 'assets belong to Dev, liabilities belong to AI.'
However, this week, aiPool released the AI Agent as a token pre-sale, achieving trustlessness through TEE technology. The wallet private key of this AI Agent is dynamically generated in the TEE environment, ensuring security. Users can send funds (e.g., SOL) to the wallet controlled by the AI Agent, which then creates tokens according to set rules and launches a liquidity pool on DEX, while distributing tokens to qualifying investors. The entire process does not rely on any third-party intermediaries and is completely autonomously completed by the AI Agent in the TEE environment, avoiding the common rug pull risks in DeFi. It is evident that AI Agents are gradually evolving. I believe that AI Agents can help users lower barriers and enhance experiences, even if it is merely simplifying part of the asset issuance process, it is meaningful. However, from a macro Web3 perspective, AI Agents, as off-chain products, currently only serve as auxiliary tools for smart contracts, so there is no need to overhype their capabilities. Due to the lack of significant wealth effect narratives aside from MeMe in the second half of this year, it is normal that the hype around AI Agents revolves around MeMe. Relying solely on MeMe cannot sustain long-term value, so if AI Agents can bring more innovative gameplay to the transaction process and provide tangible landing value, they may develop into a universal infra tool.
