daos

Nouns

This is a piece by @alexmurray, @ivolver, and me.

We are implementing a governance AI agent for Nouns. The agent will be purely transparent.

All prompts and RAG data will be publicly available on IPFS and linked with an ENS domain name.

Actually, the agent itself lives on ENS and IPFS; anyone can explore its components.

The entire decision-making process of the AI agent will be verifiable on-chain.

We will delegate all of our Nouns to the AI agent to observe how it acts and how the community engages with it.

Hopefully, this experiment can be a great success and result in an excellent paper.

The reason why I am bidding on all these Nouns is that I want the agent to have some kind of “swing state” power in these experiments. This way, people will want to engage with it and try to win its votes.

We are newcomers to the community and have a lot to learn about the DAO. Any feedback would be appreciated!

Many thanks!

https://substack.com/inbox/post/153783194

Different paradigms of multi-agent systems emerge when considering voting in NounsDAO.

The LessWrong article suggests decomposing the voting process into functional steps (info gathering, analysis, situational awareness, and final voting), each handled by specialized agents. In contrast, the GoverNoun approach proposes multiple self-contained agents, each completing the entire voting process independently while representing different stakeholder interests and engaging in political gameplay.

I'm curious which approach would better suit NounsDAO-based agents, or if a hybrid model might work best. Looking forward to exploring this further.

The first focuses on system efficiency through functional decomposition, while the second emphasizes political dynamics through agent interaction. Both offer interesting perspectives for DAO governance design.

Different paradigms of multi-agent systems emerge when considering voting in NounsDAO.

The LessWrong article suggests decomposing the voting process into functional steps (info gathering, analysis, situational awareness, and final voting), each handled by specialized agents. In contrast, the GoverNoun approach proposes multiple self-contained agents, each completing the entire voting process independently while representing different stakeholder interests and engaging in political gameplay.

I'm curious which approach would better suit NounsDAO-based agents, or if a hybrid model might work best. Looking forward to exploring this further.

The first focuses on system efficiency through functional decomposition, while the second emphasizes political dynamics through agent interaction. Both offer interesting perspectives for DAO governance design.
https://www.lesswrong.com/posts/jpGHShgevmmTqXHy5/decomposing-agency-capabilities-without-desires