How AI Agents are Reshaping the Future of Web3?

Introduction

The AI Agent has stirred up a wave in the crypto space, and Shaw, the founder of ai16z, is undoubtedly at the forefront. ai16z is the first AI Meme-themed on-chain fund, which has quickly grown to become an AI DAO with a market capitalization of over $2.5 billion on Solana within just a few months. Its core product, ElizaOS, is a multi-agent simulation framework that occupies about 60% of the market share for Web3 AI Agent development.

Recently, Shaw accepted an in-depth interview with Tom Shaughnessy, co-founder of Delphi Digital, sharing insights on AI Agent use cases in the Web3 industry and predictions for the future. The discussion covered important topics such as Agent development frameworks, token economics, and the future of open-source AGI platforms. Below is a summary of the main content of the interview.

Entrepreneurial Experience and Journey in Asia

Shaw reflected on his journey from early setbacks in open source projects to focusing on AI Agent research, and then to founding ai16z and Eliza. He mentioned that he recently had intensive exchanges with local developer communities in Asia, where he saw many interesting projects, especially optimistic about integrating AI Agents into DeFi protocols to lower the entry barrier for users.

AI Agent + DeFi Use Cases and Practicality Discussion

Shaw believes that the three most promising types of Agents currently are: those that can help users make money, those that can bring products to the right customers, and those that can save users time. He is particularly optimistic about applications in the DeFi space, such as bots that automatically identify arbitrage opportunities.

Shaw pointed out that LLMs are not suitable for direct trading, but they can convert unstructured data into actionable structured data. They are studying a "trust market" to analyze the recommendation information of top traders.

Shaw believes that, compared to simple buying and selling operations, we need a more flexible and reliable approach to utilize AI Agents for DeFi operations.

The Value of the Agent Framework and the Development Advantages of Eliza

Shaw explained the advantages of the Eliza framework: it is an end-to-end solution where developers only need to focus on the product itself. Eliza provides a simple abstraction, greatly lowering the development threshold.

The choice of TypeScript as the development language is due to the large number of JavaScript developers, and many Ethereum developers are already familiar with this ecosystem.

Eliza has comprehensive features, including voice-to-text, PDF processing, etc., which is also why many people choose to use it.

The Vision of Open Source AGI and the Future of AI Agents

Shaw believes that AGI is essentially the form of the internet, a system composed of numerous agents. The open-source and public development model can enhance the efficiency of capital allocation in the market and accelerate the development in this field.

In the future, agents will trade tokens with each other, and almost everything can be tokenized. Cryptocurrencies provide a new way for capital allocation, accelerating innovation and the realization of future visions.

Discussion on Token Economics and Value Capture Mechanisms

Shaw stated that the tokenomics model of ai16z needs to continuously evolve. They plan to open-source the Launchpad, but the long-term strategy is to invest in technologies that can expand the overall ecological value.

In the future, a positive cycle may form: the increase in token value drives the development of the ecosystem, which in turn enhances the token value. Shaw hopes to ultimately achieve a self-operating state.

Exploring Cross-Chain Development and Blockchain Selection

Shaw believes that AI agents should cover all blockchain platforms and provide services based on user distribution. He stated that the choice of Solana is mainly because users and some key projects are on this chain, but there is no strong preference for the platform.

The Transformation from "Garbage Bot" to Utility

Shaw expects that Agents capable of performing surprising things will emerge soon. Although some useless Agents may always exist, more Agents of practical value will appear in the future.

He also mentioned that there may be an Agent verification system in the future, similar to the trust scores of decentralized exchanges. Open source will become an important incentive means, and a new type of "programmer influencers" may emerge to assess the legitimacy of Agents.