Written by: Haotian
After looking at the optimization points of ELIZA V1 and V2 frameworks, I feel that there has been great progress, especially in unified Agent wallet management, unified message management, and improved scalability. But to be honest, at this stage, the ELIZA framework is still at the optimization level of the basic development architecture, and has not significantly expanded its "imagination space". I personally look forward to the possibility of its "commercialization" and "chainization" in the future:
1) A complete set of agent interoperability protocols, including agent basic message protocol (format standardization, routing mechanism, state synchronization, etc.), agent dialogue interaction protocol (multi-agent dialogue management, context information storage and state synchronization, etc.), agent resource sharing mechanism (computing, storage and other resource allocation), agent task allocation mechanism (intention understanding, task splitting, progress synchronization, aggregation rules, etc.), agent identity and permission management layer (identity authentication, absorption scoring system, permission management mechanism, etc.);
2) A set of Tokenomics protocol, including governance token definition (ai16z or ELIZA?), Agent participant incentive mechanisms (Agent developer incentives, data contribution incentives, computing resource supply rewards, verification node rewards, etc.), Gas economic system (Agent call fees, Memory storage resource fees, cross-chain operation fees, deflation burning rules, fee distribution mechanisms, etc.);
Even if the goal of the ELIZA framework is not to be chain-based, how to modularly integrate into the consensus layer of various chains, and participate in governance and verification collaboration, etc., is also worth exploring.
In addition to these two core essential components for 'chaining', how to design a set of Agent asset circulation management standards similar to the ERC20 standard, incorporating a decentralized storage system, decentralized verification mechanisms (memory systems, behavior systems), decentralized resource allocation, and incentive mechanisms, etc., are all issues that ELIZA's long-term direction of 'chaining' reform needs to consider.
Perhaps only with a clearly defined 'chainable' roadmap can EliZA firmly establish itself as the leading AI Agent at the EVM level. But opportunities and variables also arise in this process.
What will ai16z iterate step by step? Can a new framework like Arc surpass others? Which framework can first have a clear route for chain-based advanced development? Who can first find a feasible commercial closed-loop path like Virtual? These are all directions to observe. Friends interested in the subsequent evolution of AI Agent framework standards can engage in more discussions.