Deconstructing AI Framework: From Intelligent Agents to Decentralization Exploration

Introduction

Recently, the narrative of the combination of AI and cryptocurrency has evolved rapidly. Market attention has shifted to technology-driven "framework" projects, and this niche has given birth to several dark horse projects with a market value of over one billion dollars in just a few weeks. These types of projects have derived new asset issuance models: issuing tokens through GitHub repositories and reissuing tokens based on frameworks developed by Agents. Based on frameworks, Agents serve as applications, forming a unique infrastructure model for the AI era. This article will start with an introduction to frameworks and explore the significance of AI frameworks for the cryptocurrency industry.

1. What is a framework?

AI frameworks are underlying development tools or platforms that integrate pre-built modules, libraries, and tools, simplifying the process of building complex AI models. The framework can be understood as the operating system of the AI era, similar to Windows and Linux in desktop systems or iOS and Android in mobile devices. Each framework has its own advantages and disadvantages, allowing developers to choose based on their needs.

Although "AI frameworks" are a new concept in the cryptocurrency field, the development of AI frameworks has a history of nearly 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 cryptocurrency are designed to meet the large demand for Agents under the AI boom, and are derived into other fields, forming AI frameworks in various sub-sectors.

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 and supports multi-platform integration. Features include comprehensive support for Discord, automated accounts for X/Twitter, Telegram integration, and API access. It supports the processing of multimedia content such as PDF documents, link content, audio transcription, video processing, and image analysis.

Eliza currently supports four types of use cases:

1. AI Assistant Applications: Customer Support, Community Management, Personal Assistant
2. Social Media Roles: Content Creator, Interactive Bot, Brand Representative
3. Knowledge Workers: Research Assistants, Content Analysts, Document Processing
4. Interactive Roles: Role-playing, Educational Tutoring, Entertainment Robots

The models supported by Eliza include:

• Open source model local inference: such as Llama3, Qwen1.5, BERT
• OpenAI API Cloud Inference
• The default configuration is Nous Hermes Llama 3.1B
• Integrate Claude to handle complex queries

1.2 G.A.M.E

G.A.M.E(Generative Autonomous Multimodal Entities Framework) is an auto-generating and managing multimodal AI framework launched by Virtual, primarily aimed at intelligent NPC design in games. The feature is that even low-code or no-code users can use it, and they only need to modify parameters to participate in Agent design.

The core design of G.A.M.E is a modular architecture where multiple subsystems work in coordination, including:

1. Agent prompt interface: Developer and framework interaction interface
2. Perception Subsystem: Receives input information and synthesizes it.
3. Strategic Planning Engine: Establish long-term goals and specific action plans
4. Global Context: Environmental Information and Game State Data
5. Dialogue Processing Module: Handling Messages and Responses
6. On-chain wallet operator: may involve blockchain applications
7. Learning Module: Update Knowledge Base from Feedback
8. Working Memory: Storing Short-Term Information
9. Long-term memory processor: Extract important information and sort it.
10. Agent Warehouse: Save Agent Attributes
11. Action Planner: Generate specific action plans
12. Planner: Execute action plan

G.A.M.E is suitable for gaming and metaverse scenarios, with multiple projects already adopting this framework.

1.3 Rig

Rig is an open-source tool written in Rust that simplifies the development of large language model applications. It provides a unified interface for easy interaction with multiple LLM service providers and vector databases.

Core features:

• Unified Interface: Simplify Integration Complexity
• Modular Architecture: Flexible and Scalable
• Type safety: Ensures code quality and runtime safety
• High performance: Supports asynchronous concurrent processing

Rig is suitable for building question and answer systems, document search tools, intelligent chatbots, and content automation generation scenarios.

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( before the Twitter) platform. It inherits the core functionalities of the Zerebro project but is designed to be more modular and easier to extend.

ZerePy provides a command line interface for managing AI Agents. The core architecture is based on a modular design, including:

• LLM Integration: Supports OpenAI and Anthropic models
• X platform integration: directly call X platform API
• Modular Connection System: Easy to add support for other platforms
• Memory system ( planning ): achieving contextual memory

ZerePy focuses on simplifying the deployment of AI Agents on the X platform, while Eliza places more emphasis on multi-agent simulation and extensive AI research.

2. The Replica of the BTC Ecosystem

The development path of AI Agent is similar to the recent BTC ecosystem: BTC Ecosystem: BRC20 - Multi-Protocol Competition - BTC L2 - BTCFi AI Agent: GOAT/ACT - Social/Analytical Agent - Framework Competition

The AI Agent track may not replicate the historical smart contract chain. Existing AI framework projects provide new infrastructure ideas, more akin to future public chains, while Agents are similar to future Dapps.

Future debates may shift from the EVM versus heterogeneous chains to framework disputes. The key issues are how to achieve Decentralization or chainification, and the significance of development on the blockchain.

3. What is the significance of going on-chain?

The combination of blockchain and AI needs to consider its significance. Referring to the successful experiences of DeFi, the reasons supporting the chainization of agents may include:

1. Reduce usage costs, improve accessibility, and allow ordinary users to participate in AI "rental rights".
2. Provide blockchain-based Agent security solutions
3. Create unique financial gameplay for blockchain, such as investment opportunities related to Agents.
4. Achieve transparent and traceable reasoning to improve interoperability.

4. Creative Economy

Framework projects may offer entrepreneurial opportunities similar to the GPT Store in the future. Frameworks that simplify the agent building process may have an advantage, creating a more interesting Web3 creative economy than the GPT Store.

There are many unmet needs in the Web3 space. Introducing community economies can make Agents more refined. Future AI Memes may be smarter and more interesting than the Agents on existing platforms.

The Agent creative economy will provide opportunities for ordinary people to participate, and in the future, AI Memes may far exceed the current level.