Deconstructing AI Framework: From Intelligent Agents to Decentralization Exploration

Introduction

The development speed of the AI Agent track is astonishing, with frequent narrative changes. Recently, the market has focused on "framework-type" projects, and this subfield has seen multiple projects with market values exceeding hundreds of millions or even billions in a short period. These projects have pioneered a new asset issuance model: issuing tokens through Github code repositories, and Agents built on the framework can also issue tokens again. This model, which is based on frameworks and built on Agents, resembles an asset issuance platform but is actually a unique infrastructure model of the AI era. This article will start from the concept of frameworks and explore the significance of AI frameworks in the cryptocurrency field.

I. Framework Overview

The AI framework is a foundational development platform that integrates pre-built modules, libraries, and tools, simplifying the complex process of building AI models. It can be understood as the operating system of the AI era, similar to desktop systems like Windows and Linux, or mobile systems like iOS and Android. Each framework has its own strengths and weaknesses, and developers can choose based on their needs.

Although the "AI framework" is a new concept in the cryptocurrency field, its development has a history of 14 years. There are mature frameworks available in the traditional AI field, such as Google's TensorFlow and Meta's Pytorch. The framework projects emerging in the cryptocurrency space are designed to meet the demand for agents under the AI boom and extend to other tracks, forming AI frameworks in different subfields.

1.1 Eliza

Eliza is a multi-Agent simulation framework launched by a16z, specifically designed for creating, deploying, and managing autonomous AI Agents. Developed in TypeScript, it has good compatibility and is easy to integrate with APIs.

Eliza mainly targets social media scenarios, supporting multi-platform integration, including Discord, X/Twitter, Telegram, etc. In terms of media content processing, it supports functions such as PDF reading analysis, link content extraction, audio and video processing, etc.

The use cases supported by Eliza mainly include: AI assistant applications, social media characters, knowledge workers, and interactive roles, etc. Supported models include local inference with open-source models and cloud inference using the OpenAI API, etc.

1.2 G.A.M.E

G.A.M.E is a multi-modal AI framework for automatic generation and management launched by Virtual, mainly designed for intelligent NPCs in games. The feature is that it can be used without any coding background; users only need to modify parameters to participate in Agent design.

G.A.M.E adopts a modular design, with a core architecture that includes: Agent prompt interface, perception subsystem, strategic planning engine, world context, dialogue processing module, on-chain wallet operator, learning module, working memory, long-term memory processor, Agent repository, action planner, and plan executor.

This framework focuses on the decision-making, feedback, perception, and personality of Agents in virtual environments, and is applicable to metaverse scenarios as well as games.

1.3 Rig

Rig is an open-source tool written in Rust, designed to simplify the development of applications for large language models (LLM). It provides a unified operating interface, making it easy to interact with multiple LLM service providers and vector databases.

The core features of Rig include: unified interface, modular architecture, type safety, and efficient performance. Its workflow includes steps such as request handling, information retrieval, and response generation.

Rig is suitable for building question-and-answer systems, document search tools, context-aware chatbots, and even supports content creation.

1.4 ZerePy

ZerePy is an open-source framework based on Python that simplifies the process of deploying and managing AI Agents on the X platform. It inherits the core functionalities of the Zerebro project but adopts a more modular and extensible design.

ZerePy provides a command-line interface, supports OpenAI and Anthropic's LLMs, integrates X platform API, and features a modular connection system. Future plans include integrating a memory system, allowing the Agent to remember previous interactions and contextual information.

Compared to Eliza, ZerePy focuses more on simplifying the process of deploying AI Agents on specific social platforms, leaning towards practical applications.

2. A Copy of the BTC Ecosystem?

The development path of AI Agents has similarities with the recent BTC ecosystem. The BTC ecosystem has gone through stages such as BRC20, multi-protocol competition, BTC L2, and BTCFi. AI Agents are developing faster based on a mature technology stack, which can be summarized as: GOAT/ACT - Social-type Agents/Analytical AI Agents - Framework competition. In the future, infrastructure projects focusing on Agent Decentralization and security may become the theme of the next stage.

The AI Agent track is unlikely to replicate the homogenization and bubble process of the BTC ecosystem. AI framework projects offer new infrastructure development ideas, more akin to future public chains, while Agents are similar to Dapps. The future may shift from debates over EVM and heterogeneous chains to disputes over frameworks, with key issues revolving around how to achieve Decentralization or chaining, and the significance of blockchain in this domain.

3. The Significance of On-Chain

The combination of blockchain and AI needs to address its significance issues. Referring to the successful experiences of DeFi, AI Agent chainization may find value in the following aspects:

1. Reduce usage costs, improve accessibility and choice, allowing ordinary users to participate in AI "rental rights".

2. Provide blockchain-based security solutions to meet the security needs of Agents interacting with the real or virtual world.

3. Create unique blockchain financial models, such as Agent-related computing power, data tagging investments, etc.

4. Achieve a transparent and traceable reasoning process, surpassing the agent browsers provided by traditional internet giants in terms of interoperability.

4. New Opportunities in the Creative Economy

AI framework projects may provide entrepreneurial opportunities similar to GPT Store in the future. A framework that simplifies the agent construction process and offers complex function combinations is expected to gain an advantage, forming a more interesting Web3 creative economy than GPT Store.

Compared to the GPT Store, the AI Agent creative economy of Web3 may be fairer and introduce a community economic model. This will provide ordinary people with opportunities to participate, and future AI Memes may be smarter and more interesting than the Agents on existing platforms.