AI Agent (Artificial Intelligence Agent) refers to a system or program that can autonomously perform tasks on behalf of users or other systems by designing its workflow and utilizing available tools. It can interact with the environment, collect data, and use the data to perform automated tasks to achieve predetermined goals.
Humans set goals, but AI Agents independently choose the best actions to achieve those goals. For example, when you are sick and ask what medicine to take, general AI will give you the specific name of the medicine, but AI Agents may ask you about your current body temperature, body sensation, drug allergies, home address, etc. It will automatically query the information, place an order directly, and deliver the medicine to you, and tell you how to use it and what the backup plan is. The information provided by LLM helps you make decisions, and all the links in the middle are handed over to AI Agents, and you only need to click to confirm.
How AI Agents Work
At the core of AI Agents are Large Language Models (LLMs). For this reason, AI Agents are often referred to as LLM Agents. Traditional LLMs (chatGPT, etc.) generate responses based on the data used to train them and are constrained by knowledge and reasoning limitations. In contrast, agent technology uses backend tool calls to obtain the latest information, optimize workflows, and autonomously create subtasks to achieve complex goals. In the process, autonomous agents learn to adapt to user expectations over time. Agents are able to store past interactions in memory and plan future actions, thereby facilitating personalized experiences and comprehensive responses.
Application of AI Agent in Blockchain
From the emergence of GOAT in early October 2024 to today, the AI sector has flourished, with a market value of $17.65 billion in just two months. The explosion of AI agents is very fast.
However, from the initial pure AI meme form of GOAT, it has now become more substantial. This is undoubtedly a positive growth for the ecosystem.
Inventory of key projects
1) Virtuals
As the AI agent track has developed to this day, the Virtuals ecosystem has firmly established itself as the leader. The Virtuals protocol is mainly focused on the entertainment field. As an AI agent distribution platform, Virtuals allows all builders to easily create and distribute AI agents, and can obtain protocol income from AI agents through tokenization and decentralization.
The biggest advantage of Virtuals is that it solves the pain point that non-professional AI people cannot participate in the AI boom. In addition, as the leader of the sector, it has a certain user base and is relatively stable in terms of market value.
2) ai16z
First of all, ai16z has no connection with A16Z. The only connection may be that the Twitter account has been followed by the founder of A16Z.
ai16z is different from Virtuals. In the beginning, ai16z started as an AI meme. Later, the Token incentive mechanism was incorporated into the ai16z plan. ai16 has a strong technical background and developed the Eliza framework. The Eliza code ranked first in the December global trend list of GitHub.
In the future, ai16z will launch several games based on the Eliza framework, and will focus more on creating a practical Al Agent investment tool, DeFi Al Agent. As its founder said, the goal of ai16z is not to create an AI robot that imitates a16z, but to beat it in the investment field where it is best.
3) Pure AI meme
In the last round of meme craze, AI meme projects such as GOAT and ACT were the first to ignite the market and set off a wave of enthusiasm for the AI track. Among them, after ACT was launched on the top platform, its market value quickly soared from about US$20 million to more than US$600 million, further increasing the market's attention to and investment enthusiasm for AI meme.
However, after only two months, market preferences have clearly shifted from pure AI Meme to the field of applied AI Agent. Although the market value of AI Meme projects such as Fartcoin and Turbo has also experienced rapid growth during this period, even breaking through the high point of one billion US dollars, overall, the popularity of AI Meme has declined significantly.
Prospects of AI Agent in the blockchain field
There are many pain points of AI big models that can be solved in the blockchain. For example, computing power problems can be solved through depin, funding problems can be solved through Token + incentive mechanism, and data security can be solved through TEE or MPC solutions. However, there are also many problems faced, such as incompatibility between public chains.
