Threshold
I haven't been able to test the tool yet. I'll abstain until I have a solid opinion about what's being built.
Dear @1THD...9Eoe,
Thank you for sharing your thoughtful perspective, we want to address the concerns directly and with clarity.
First, we completely agree that AI should never become a centralized “go-to” tool for DAOs. That is precisely why this project is being built with decentralization at its core. We are committed to ensuring that the AI agent is open-source and designed to be progressively decentralized, so that it is not “owned” by a single team, but instead by the Polkadot DAO community as a whole. This way, anyone in the ecosystem can audit, adapt or build upon it.
Second, it’s important to highlight that this isn’t about creating a closed tool funded by Treasury that could risk centralization. On the contrary, the goal is to provide foundational, open infrastructure that the DAO and its members can expand upon. Just as DAOs already fund common-good infrastructure, this initiative ensures that the tools serving our collective decision-making are trustworthy, transparent and community-controlled.
Third, while it’s true that DAOs could eventually build their own versions of AI tools organically, the reality is that someone has to take the first step. Whether that’s Parity, another team, or us, a DAO product must start with a committed builder. Our team is in a strong position to do this responsibly, with both the technical expertise and the philosophical alignment to serve the community’s needs. We are already leveraging decentralized LLMs and will open source the code that is directly relevant to the Polkadot DAO, ensuring this product isn’t just another proprietary black box, but a living public good.
Finally, we believe this approach reflects the true purpose of a DAO: empowering the community with tools that they own, govern and improve collectively. Treasury funding here is not a subsidy for centralization, it is an investment in decentralization, in open innovation and in the tools that help every voter and member participate more meaningfully.
We look forward to your thoughts on our reply and your support of our proposal.
Dear Proposer,
Thank you for your proposal. Our first vote on this proposal is NAY.
The Medium Spender track requires 50% quorum (at least 4 aye votes) and simple majority of non-abstain votes according to our voting policy v0.2, and any referendum in which the majority of members vote abstain receives an abstain vote. This proposal has received two aye and two nay votes from eight available members, with two members abstaining. Below is a summary of our members' comments:
The feedback noted a promising start with a clear intent to streamline OpenGov processes and improve due diligence, thanks to the tool’s open source elements and phased approach. One group appreciated the project’s potential to consolidate scattered governance data and boost productivity, while others expressed reservations about its ability to detect duplicate proposals and maintain full transparency, especially given that an earlier version remained closed. Concerns were also raised about the elevated funding request relative to the project's initial scope and alignment with free software principles. In the end, these mixed assessments contributed to a decisive NAY vote on the proposal.
The full discussion can be found in our internal voting.
Please feel free to contact us through the links below for further discussion.
Kind regards,
Permanence DAO
Decentralized Voices Cohort IV Delegate
📅 Book Office Hours
💬 Public Telegram
🌐️ Web
🐦 Twitter
🗳️ Delegate
While this tool has the potential to be a valuable asset for the community, I don’t currently see clear advantages over what’s already achievable with existing AI solutions. Given how rapidly commercial and open-source AI systems, both free and paid, are evolving, I wonder whether it will be sustainable for the team to maintain a specialized AI focused exclusively on governance.
Additionally, since there isn’t a defined revenue model to support long-term development, I’m concerned this could eventually lead to recurring funding requests from Polkadot’s treasury. It might be worth carefully assessing the long-term viability and differentiation of this project to ensure it remains both effective and financially sustainable.
TruthDAO voted NAY ❌
Members felt that DVs already represent a stronger governance practice, and that the community does not need an AI system to broadly influence or control collective judgment. We do, however, see AI as potentially useful for tasks such as summarizing proposals, though at this stage it does not yet seem capable of supporting sound decision-making.
In addition, the proposal states that this tool could:
• Increase user productivity by up to 300%
• Eliminate over 120,000 hours of manual governance work
• Improve workload efficiency by 50%
• Save the Treasury billions of dollars in unnecessary spending
Some members viewed these claims as somewhat overstated. In particular, it’s not clear how figures such as 300%, 120,000 hours, and 50% were calculated. And given that the Treasury has been running for just over three years without yet spending a total of $1B, it’s difficult to see how the claimed billions in savings could be substantiated.
Finally, as mentioned in feedback on the first proposal, if the team has strong conviction in this project and believes it can play an essential role in OpenGov, then the best path forward would be to first deliver a usable version.
📖Truth DAO Governance Statement
💭 Email: open@truthdao.cn, Telegram
🗳️ Delegate
Edited
Security concerns were raised about the API and the responsiveness to mitigate the security issues was not meet at the time.
Our members raised questions about the utility (no other supported language than English) and the lack of global integration in the existing governance tools.
-> The community already supported different tools in the past, and the lack of integration is a main factor that would make a real discoverability problem from users, they won't know about it and won't use it.
Some other members raised that the cost is obviously too important, anyone wit hbasic AI skills could develop an even better tool for way less or for free.
This is Johan, representing the PBA Alumni Voting DAO.
Our community has cast its vote on this referendum as Yay.
Our community felt that:
PBA Alumni encourages continued work on these ideas and proposals. Refinement and further implementation are key—what is a Nay today may become a strong Aye tomorrow.
Threshold
Saxemberg has voted AYE on the Polkadot referendum 1717 Resubmission: Clarys.AI - OpenGov Data Backbone [...] The previous observation is kept, OpenGov is a great candidate for AI efforts like this should be welcome just like other chains are doing it.
This referendum is eligible for vote overrule:
https://voting.opensquare.io/space/the-sax-guild/proposal/QmXR3xyFJdK47DtW3BxqP8d83mZ32atkrjAHyLcNdYhrK6