[Non-Constitutional] DePolis, collective sensemaking for Arbitrum DAO

Proposals
1

TL;DR:

  • we, the team behind building InnoPolis during the ETHGlobal Bangkok Hackathon, propose creating a decentralized version of Polis for the Arbitrum DAO;
  • the original Polis has proven its effectiveness in large-scale irl-discussions but is not web3-compatible;
  • we would love to hear comments and suggestions from the Arbitrum community, either here on the forum or in the Polis conversation we’ve organized: https://pol.is/5jbzpjw2t8.

1. Abstract

DePolis (InnoPolis in ETHGlobal Bangkok Hackaton) is a decentralized version of Polis (https://pol.is/), an open-source technology for survey research that uses data science to gather insights.

In practical terms, DePolis is a tool for “smart” gathering of Arbitrum users’ and/or delegates’ opinions on specific topics. The term “smart” here refers to the process of collecting and displaying opinions in a clear, concise, and human-readable manner. This enables the Arbitrum DAO to better understand its users/delegates by identifying dominant opinions, points of consensus and disagreement, as well as major clusters of perspectives.

In its highest ambition, DePolis is a platform for enabling collective intelligence in human societies and fostering mutual understanding at scale.

2. Motivation

Original Polis:

  • is used as a deliberation tool on a national and/or regional scale in Taiwan, Greece, Germany, USA, and other countries (you can find case studies here);
  • was mentioned by Vitalik as a great social technology that could be used to enable more decentralized governance over contentious decisions (one, two, and three);
  • likely the best technology for large-scale conversations which can radically improve the structure of communication in DAOs;
  • is gradually starting to integrate into web3 governance frameworks (eg, Filecoin governance).

Our strong belief is that DePolis, if integrated wisely, could provide a way to create feedback loops between DAO/delegates/developers/protocols and users to truly grasp the “community’s mind” and channel those insights into DAOs and protocols decision-making.

3. Rationale

This proposal is aligned with Arbitrum’s community values:

  • DePolis is the tool that enables the collection of Arbitrum users’ opinions in their own words while providing clear and actionable insights into their collective views;
  • DePolis can serve as a tool for constructive governance participation, empowering Arbitrum users who currently lack the skills/time/knowledge to fully express their opinions in the governance forum.

4. Key Terms

Polis - a real-time system for gathering, analyzing, and understanding what large groups of people think in their own words, enabled by advanced statistics and machine learning, for more info check https://pol.is/.

DePolis - a web3-compatible version of Polis, which our team plans to develop for the Arbitrum community.

Topic - a question or subject to be discussed using DePolis (e.g., “How should the DAO allocate its funds?” or “What improvements are most needed?”).

Participants - entities/accounts eligible to participate in the conversation by voting and writing statements, which can be filtered using any arbitrary set of onchain data (e.g., “accounts holding/staking >100 ARB” or “accounts that participated in Tally voting > 5 times”).

Statements - opinions submitted by participants through DePolis (<140 characters) on a specific topic, which other participants can vote on.

Report - the summarized output generated after a conversation conducted on the DePolis platform. It provides a structured and visual representation of the collective insights and findings from the discussion (check example of Polis report from Gitcoin DAO conversation).

5. Specifications

I. How Polis for DAOs should look like

Conceptually, DePolis could serve as a tool for:

A. Gathering Arbitrum users’ opinions on contentious topics
Remember the spiciest conversations in the Arbitrum DAO (for example, the STIP proposal or the gaming ecosystem catalyzing discussions)? Imagine that, alongside forum debates, ordinary Arbitrum users were also asked to express their opinions in a simpler, more accessible way - by providing a short statement or upvoting/downvoting others’ statements. Now, what if these opinions were then transformed into a clear, human-readable report, summarizing key points of consensus and disagreement? Do you think this would change how decisions are shaped and communicated?

B. Gathering opinions from specific protocol/domain users
Imagine that before launching massive incentivization programs (like those for the GameFi or DeFi sectors), we could ask active users of relevant protocols for their thoughts on pain points and opportunities in the domains they use daily. Do you think such feedback could refine and improve the parameters of these programs?

C. Facilitating a “pre-forum” exchange of delegates’ opinions to map the existing opinion landscape
Do you think this could help structure and smooth discussions? Some Arbitrum delegates seem to believe so (eg, check Oversight Committee thread).

D. Some other ideas:

  • an automatically created incentivized poll for all Arbitrum users and/or ARB holders/stakers in case of highly contentious proposals appearing on the Arbitrum governance forum (the level of contentiousness could be determined using AI tools like x23.ai);
  • a tool that operates autonomously on its own website and can be initiated by any delegate on any issue;
  • a required procedure to be launched before/after any major grant program, enabling the identification of the opinions of the ultimate beneficiaries of such grant programs.

II. Why can’t we just use original Polis?

We can! Moreover, Polis has been used in numerous conversations among DAO members:

  • some conversations in Gitcoin DAO (one, two, three), check report example here;
  • a few conversations in Arbitrum forum (one, two).

But the original Polis has serious limitations:

  1. It is excellent for conversations within relatively small groups that are highly interested in quality discussions (e.g., among delegates on a specific topic), but it is not well-suited for managing large-scale discussions with thousands of participants.
  2. In the aforementioned low-scale discussions, manipulation of Polis is not a concern. However, if we imagine that opinions expressed through Polis truly matter and are embedded in governance decision-making processes, there will quickly be those looking to manipulate the conversation. The original Polis does not provide tools to counteract this.
  3. The original Polis is not compatible with web3 environments. It does not allow for defining the list of participants based on onchain data, does not enable incentivization for participating in conversation, and does not support integration with numerous platforms like Farcaster and Lens, among others.

III. So, what we wanna build for Arbitrum DAO?

We want to start with a relatively simple implementation of DePolis that, on one hand, will validate key hypotheses (regarding technical solutions, UX, organizational aspects, etc.) and include all the minimally necessary technical features, and on the other hand, will enable the Arbitrum DAO and its delegates to independently launch conversations without our involvement (including defining the participant pool and providing incentives for participants).

Core features of the first iteration of DePolis:

A) Conceptual:

  1. The ability for anyone to create and configure a conversation in the same way as in the original Polis.
  2. The ability to provide statements, with a limit of total statements per participant (unlike the unlimited submissions in the original Polis), and to vote on others’ statements.
  3. The ability to define a timeline for submitting statements and voting on them (e.g., 3 days for submitting statements, then revealing submitted statements, and then 3 days for voting).
  4. Configuration, moderation, and monitoring tools for administrators to effectively manage conversations (similar to the original Polis).
  5. The ability to incentivize the conversation by sending USDC/ARB to a smart contract associated with the conversation, which will distribute rewards among participants based on the conversation outcomes, with customizable reward distribution logic.
  6. Report generation accessible to all participants, summarizing the conversation and its outcomes.

B) Technical:

  1. A UX/frontend closely resembling the original Polis (examples linked below) but with a more modern and streamlined design.
  2. Wallet-based authentication with access restrictions based on onchain data (e.g., “accounts holding/staking >100 ARB”, “accounts that participated in Tally voting >5 times”, “accounts with >15 transactions on the Arbitrum chain in the previous month”, “holders of X, Y, and Z tokens”, etc.).
  3. A no-coding or minimal-coding method to set such access restrictions.
  4. Storing all important data (votes and statements) onchain/ipfs, so anyone can verify the accuracy of the reports generated through the frontend.
  5. The ability for administrators to moderate statements - removing statements from the frontend while keeping them stored onchain.
  6. An off-chain script to generate reports (onchain report generation would be too expensive and unnecessary).

The specification outlined above, including the technical solutions, may be adjusted during the development process.

Links:

As part of this proposal, we aim to conduct and moderate four conversations: two among delegates (without incentivization) and two among Arbitrum users (one without incentivization and one with incentivization).

The two conversations among delegates can be considered as test conversations to validate technical solutions. We do not expect these to yield any meaningful insights, but we would be pleased if at least 20 delegates or active forum participants take part. Accordingly, the topic of these conversations is not critical - it could, for example, involve a contentious issue relevant to governance procedures in the DAO that concerns delegates and forum members. Following the two test conversations, we aim to gather feedback from participants and use it to improve DePolis.

Additionally, we would like to conduct two larger and more engaging conversations involving active Arbitrum users and/or ARB holders/stakers. Examples of questions we propose for discussion include:

  • Does the Arbitrum community feel the DAO could improve transparency in how treasury funds are used? Would clearer reporting and progress milestones help build trust and accountability?
  • What impact do you think Arbitrum’s incentive programs (STIP, LTIPP, Gaming Catalyst Program, etc.) have had on the ecosystem? Have they driven meaningful growth, or were they just a short-term boost?
  • Should ARB emissions prioritize rewarding current stakers or subsidizing developers to grow the ecosystem?
  • Delegates and active forum members are encouraged to suggest more relevant and pressing questions as well.

For one of the conversations, we would like to incentivize participants (see below for incentivization methods).

IV. Controversial issues

There are certain architectural questions relevant to DePolis. Below is a list of these questions and our ideas on how to address them appropriately. The approach may evolve during development.

List of controversial questions

1. Sybils

When the results of conversations can genuinely influence decisions made by delegates and/or participants are incentivized, it is inevitable that some will attempt to game the system. We believe that carefully defining participation conditions and prohibiting accounts flagged as sybils from participating in the conversation will solve this issue. Additionally, this problem is not particularly relevant for the first iteration of DePolis.

2. Timings

In the original Polis, a conversation is ongoing - participants can submit statements and view others’ statements throughout the conversation. The downside of this is that early statements have a higher chance of receiving more votes. In our view, the default solution should be to provide a specific period (e.g., 3 days) during which participants can submit statements without being able to see those of others.

3. Incentivization

An intuitive solution seems to be incentivizing two categories of statement authors:

  • those whose statements received the highest number of upvotes (participants who helped identify points of consensus);
  • those whose statements received the most upvotes within their opinion groups (participants who helped identify points of disagreement).

The question of whether voting itself should be incentivized - and if so, under what principles - is something we propose to leave open at this stage.

4. Seed Comments

The original Polis includes the creation of seed comments - comments submitted by the conversation owner (more details here). In our opinion, dividing the conversation into two phases (see ‘Timings’ above) eliminates the need for seed comments. Those with something to say will do so in the first phase, while those who prefer to understand the landscape of opinions before contributing can do so during the second phase.

5. Moderation

The original Polis allows for various moderation approaches (details here). In our view, a permissive moderation approach should be chosen, allowing statements that violate predefined rules or are repetitive to be removed from the interface (but retained onchain).

6. Anonymity

In some cases, linking statements to specific accounts might discourage participants from expressing their genuine opinions. However, integrating onchain logic with anonymity poses a complex technical challenge and is a subject for future research.

7. Interpretation

Is the standard Polis report sufficient for consideration in decision-making processes? In our view, it is - currently, the report is clear, accessible, and informative enough for delegates to use. However, proper interpretation of conversation results remains a topic for future research.

6. Steps to Implement, timeline

1. Development phase 1, core features and UX:

  • frontend with web3 authentication and popular wallets support
  • functionality for checking customizing participation criterias
  • decentralized statements and votes storing via blockchain/ipfs
  • public backend for heavy math operations and verifiable report generation

Estimated timeline: 1-2 months

2. Conducting two test conversations among Arbitrum delegates, collecting feedbacks

Estimated timeline: 2-4 weeks

3. Development phase 2, improvements based on feedback received from delegates

Estimated timeline: 2-4 weeks

4. Conducting two test conversations among Arbitrum users

Estimated timeline: 2-4 weeks

Overall Cost

According to our calculations, developing the technical components will require the work of three technical specialists (two backend and one frontend) over a period of two months. Based on this, we propose setting the compensation at $60k (in ARB tokens) as compensation for developers + $5k (in ARB tokens) to cover the incentivized conversation with Arbitrum users. If needed, the compensation can be distributed in installments, for example:

  • $20k after the completion of Development Phase 1
  • $20k after the completion of Development Phase 2
  • $20k after conducting two DePolis conversations among Arbitrum users

Team

Alexey, Web2/Web3 Security Specialist in Decurity, Blockchain Researcher, GitHub

Nikolai, Web2/Web3 Appsec, Blockchain Researcher

Bogdan, Web2/Web3 Appsec, Blockchain Researcher

Daniil, Security Researcher

Roman, x.com

Will be happy to answer any questions! We also invite everyone to participate in the conversation on Polis - https://pol.is/5jbzpjw2t8 (conversation topic: “DePolis as a deliberation tool for Arbitrum DAO”).

1 Like
2

Thank you for this proposal. I find it very interesting, and I have a few questions that I hope can be clarified further.

1.Purpose of Polis
Is the goal of Polis to replace the current Forum? If so, will the data accumulated on the Forum over the past two years be migrated to Polis? If not, how will Polis avoid requiring Delegates to repeat discussions that have already taken place on the Forum? Additionally, what mechanisms will ensure effective collaboration between the two platforms?

2.Long-term Maintenance
I noticed that the costs outlined in the proposal focus only on development. Will you provide long-term maintenance and support for DePolis after its launch? If so, what is the budget for this? How will you handle updates, improvements, and issue resolution moving forward?

Looking forward to your response and further discussions!

1 Like
3

hey there @DePolis

was this proposal written and posted by Alexey Posikera? if so, the linkedin profile in the github account that was shared does not exist.

so, could you provide more background info and social profiles for the team members and their past achievements?

4

Interesting proposal. Our team has a few questions:

Is it possible to provide results showing the successful implementation of dePolis in other similar DAOs within the Arbitrum?

Is there any specific reason for setting this character limit?
Perhaps it would be better to remove or increase it so that the community is not restricted when sharing their opinions.

How can dePolis prevent this issue?

It seems like a good way to encourage the community to participate in discussions, but how do you plan to determine who will receive the rewards? There is a risk that users might post low-value comments just to claim the rewards.
How do you plan to secure more funds for these incentives?
We currently have the DIP that incentivizes the community to generate valuable comments, and this seems to have a similar purpose.

1 Like
5

Thanks for sharing the DePolis proposal. This could be a great tool to help the DAO better grasp user and delegate views, saving us a lot of time instead of going through hundreds of replies.

However, I have some reservations about moving the discussion to a new platform. Introducing a separate space for conversations would fragment our community interactions. Instead, I recommend focusing on extracting and analyzing the discussions already happening in our existing forums. Similar to what @Maets23 is already doing with SimScore, using our current infra could make things run more smoothly. Perhaps there’s even potential to merge both projects, combining the strengths of DePolis and SimScore to create a better tool.

Additionally, I’m curious about the possibilities of integrating DePolis with our existing governance tools like Snapshot and Tally. For instance, incorporating a feature within the voting dashboard that highlights the main ideas and sentiments of a particular proposal could provide voters with valuable context, helping the decision-making process without requiring them to go through all the discussion of the particular proposal.

Regarding the budget, I’d like to know if the initial $60k will cover all expenses or if there will be additional costs later, such as maintaining the DePolis after it’s built.

5 Likes
6

Thank you for mentioning me and SimScore in your reply.

I agree wholeheartedly that devs job when envisioning improvements to governance is to make delegate / voters life easier, ideally seamless.

SimScore is a clustering and prioritization tool. It accepts “replies”, author and ID#. It returns clustered priority list, relationship graph and pairwise similarity table.

During SimScore’s development it became clear that we should not develop a front end. Their are too many options. So instead we commissioned our dev team to change our original POC to a SimScore.API. As you probably know, SimScore results can easily be token weighted. The APIs dev is nearly complete…documentation etc is forthcoming.

I like Pol.is. It is a great tool to determine sentiment. I believe a depolis API would be another great addition. Arbitrum would then know, clusters, priority and sentiment of replies within the Discussion Forum. Making proposers and delegates life easier.

I great project depolis and SimScore could work on is “pulling” all the replies out of the forum, snapshot and tally so the APIs can connect to Forum Discussions. In fact, Proposal.APP may have already done this work @paulofonseca

Delegates would have forum replies available in chronological, priority and sentiment order with a click.

Let you guys get some sleep at night, instead of pouring over complicated strings of replies.

4 Likes
7

The main question is: Do we really need this?

Migrating from the existing forum to a new platform would be a significant challenge, so the new platform would really need to make a difference.

I think the value of DePolis could be shown (or disproved) before we make a decision on this proposal. I suggest that the authors of DePolis use the content of this forum to show how it would look like on DePolis and what insights would it bring to DAO members.

Even better would be to programmatically fetch new forum content to DePolis, and show what benefits DePolis brings over this forum.

I understand this would require additional work (coding), but I think it’s not realistic to expect funding in advance, before showing the real benefits. Let alone to persuade people to ditch the existing platform and move to a new one.

1 Like
8

Interesting proposal. I looked at how it works on the site - I liked it.

There are a couple of questions about the application:

  • What questions exactly should be used for your platform?
  • Why can’t I use the usual built-in forum poll?
  • Why is Snapshot voting not enough?
1 Like
9

Note that Snapshot already built a Pol.is sort of solution (cc @cp0x) . Actually months and months ago, maybe even 2 years!

Unfortunately, Pol.is is a very time-expensive mechanism so it never got traction. That’s why SimScore has been designed with that criteria in mind: making sure that delegates don’t have to spend A LOT of time.

Another issue with Pol.is is the lack of deliberation, and here tools like Harmonica can help.

4 Likes
10

Thanks for the questions!

1. Purpose of Polis

The purpose of Polis (I’ll call it DePolis so we can distinguish the existing Polis from DePolis, the system we’d wanna create based on the original Polis) is not to replace existing governance mechanisms but to facilitate large-scale discussions involving many actors.

Ideas about how DePolis could be integrated into decision-making processes are described in “5. Specifications. I. How Polis for DAOs should look”. Even when these ideas focus on delegates (e.g., “Facilitating a ‘pre-forum’ exchange of delegates’ opinions to map the existing opinion landscape”), they’re not meant to replace the current forum.

Anyway, in my opinion, the most canonical, straightforward, and practically useful use case is “Gathering Arbitrum users’ opinions on contentious topics” and/or “Gathering opinions from specific protocol/domain users”. In this case, overlapping topics aren’t a problem (moreover, they must overlap!), and effective collaboration between the forum and DePolis will be organically facilitated by the delegates, who are likely interested in considering DePolis participants’ (= Arbitrum users’) opinions.

In other words, forum - for skilled participants (delegates, service providers, large stakeholders), DePolis - for ordinary users.

2. Long-term maintenance

As stated in the proposal, “we want to start with a relatively simple implementation of DePolis that … will enable the Arbitrum DAO and its delegates to independently launch conversations without our involvement (including defining the participant pool and providing incentives for participants)”. So:

  • yes, the costs focus only on development;
  • maintenance for DePolis after its launch does not require an additional budget, updates/improvements/issue resolutions will require it.

The point we must achieve, no matter what, is:

  • the development phase (with the functionality described in the proposal) is successfully accomplished, DePolis code is open-sourced and publicly available;
  • the Arbitrum community has a working system that can be used without our involvement.

In my opinion, achieving these goals ensures that the grant’s money is not wasted, and even in the worst-case scenario, the Arbitrum community has a working system and is able to maintain it on its own.

At the same time, the most preferable scenario for us is when DePolis finds its PMF (within and potentially outside Arbitrum DAO), and we continue to maintain and upgrade it as a tool for Arbitrum DAO, as well as an instrument/product for other communities. It’s really hard to predict future scenarios in this case, but they include requesting Arbitrum DAO for additional funding (the budget will depend on the scope of our future work), requesting grants from other grant providers (likely interested in governance-related products), receiving funds from investors, etc.

11

Hello, @paulofonseca !

This proposal was primarily written by me, Roman (you can find my X and Paragraph/Mirror accounts above), and is based on ideas and input from the entire team. Regarding Alexey’s profile and the team’s profiles:

  • Alexey: the link GitHub profile was broken, fixed now.
  • Nikolay: an application security specialist currently working at a web2 security company. Fields of expertise include reverse engineering, penetration testing, web audits, mobile app audits, web3 audits, and software development. Holds several security certifications (OSCP, OSWE, OSED, KLCP, OSWP, BSCP) and has over 5 years of experience as a pentester and auditor.
  • Bogdan, Daniil, Alexey: experienced professionals in web2 and web3 security audits, currently working at a web3 security audit firm. Along with Nikolay, they have participated in and organized numerous CTF (security) competitions and have formed a team of web3 researchers.
1 Like
12

Thanks for your feedback and comments!

1. Statements <140 characters

Most conceptual features of DePolis mirror those of the original Polis, including a character limit for participants’ statements. To understand this limitation, it’s important to highlight the key differences between governance forums and DePolis (“Gathering Arbitrum users’ opinions usecase”):

Purpose:

  • forum: a decision-making hub - platform for in-depth discussions among skilled participants. Topics are debated in a sophisticated and often technical manner, with the goal of producing proposals, resolving disputes, and shaping policy.
  • DePolis: a signal-gathering tool - platform for wide-scale participation, allowing anyone to express their opinion quickly and easily. Contributions are simplified, typically through short statements or by agreeing/disagreeing with others (upvoting/downvoting).

Irl-comparison:

  • forum: parliament or council where elected representatives deliberate and negotiate policies.
  • DePolis: “smart” non-binding (advisory) referendum, where the general population can vote or express opinions directly on key topics.

Accessibility

  • forum: naturally exclusive due to the knowledge and skills required to participate effectively - primarily accessible to skilled participants, such as delegates, service providers, and large stakeholders with specialized knowledge.
  • DePolis: radically inclusive by design, inviting participation from anyone regardless of expertise - open to everyone, including ordinary users.

The character limit in DePolis ensures that the platform remains accessible and manageable for broad participation. In other words, nobody is likely to vote on 100+ statements if each is over 500 characters long. This limit encourages concise, focused statements and avoids the complexity of long, forum-style debates. However, the exact character limit is flexible and can be adjusted based on user feedback.

You can also explore the vTaiwan governance framework, where Polis is used alongside other deliberation platforms like Discourse (forum).

2. The issue of manipulation

The most obvious manipulation scenario is a sybil attack. For example, a malicious actor might try to manipulate grant-related DePolis conversations to push a specific grant policy. The naive solution to this issue would be: “we must prevent sybils from participating”. In the original Polis use cases, sybil attacks weren’t a major concern. Some mitigation measures included using Facebook and Twitter authentication in combination with moderation policies. However, in a “highly-sybilled web3 environment” these approaches won’t work.

The core idea here is that preventing sybils from participating essentially boils down to properly defining the list of conversation participants. This depends on the topic of the conversation. Our approach is to define these participant lists ad hoc based on onchain data, minimizing the risk of sybil attacks. For example:

  • in GameFi-related conversations, participants could be accounts that have interacted with a predefined list of smart contracts deployed by GameFi projects;
  • in DeFi-related conversations, participants might include the top-500 users of the top-10 Arbitrum DeFi protocols;
  • in general conversations, participants could be active users of the Arbitrum chain (e.g., those with 100+ monthly transactions and holding >10 ARB).
  • and so on.

These ad hoc approaches can also be combined with other techniques, such as excluding accounts flagged as sybils or applying general conditions like “holding >0.01 ETH during the month”.

3. Incentives

As stated in the proposal, an intuitive solution seems to be incentivizing two categories of statement authors:

  • those whose statements received the highest number of upvotes (participants who helped identify points of consensus);
  • those whose statements received the most upvotes within their opinion groups (participants who helped identify points of disagreement).

Hence, posting low-value comments will not result in any rewards.

Whether people would want to participate in conversations by upvoting/downvoting without incentivization is a more complex question. In real-life Polis conversations, people engage because the discussed topic is important to them for some reason (e.g., it could be related to Airbnb or Uber city policies, climate change issues, etc. — you can find more case studies here. If we adopt the much-loved analogy of “blockchains = states”, we can assume that people will be willing to participate in discussions simply because the topics matter to them. However, whether significant participation in conversations can be achieved without any incentivization (but with minimal informational support from Arbitrum DAO voices) can only be determined in practice.

Other questions:

  • overlapping with DIP: as mentioned above, the forum and DePolis have quite different natures and purposes.
  • securing more funds for incentives: if the Arbitrum community finds DePolis and its outputs valuable for governance processes, it will likely be inclined to approve incentives for new conversations.
1 Like
13

Thanks for your ideas and comments!

I want to reiterate - we’re not talking about moving discussions to a new platform or creating any kind of “competition” between DePolis and the existing forum. DePolis (at least in its first iteration) is intended as a useful tool for delegates and the broader Arbitrum community to gain insights into the collective views of Arbitrum users, rather than as a separate space for in-depth conversations. I also suggest checking my earlier response to Argonaut regarding the differences between the forum and DePolis.

I’m aware that x23.ai is doing a great job analyzing discussions in existing venues, but have never heard of SimScore. Will dive deeper in the project - thanks!

I expect Snapshot and Tally to integrate Polis-like mechanics in one or another form in the coming years. However, at this stage, our proposal focuses on the very first iteration of a decentralized Polis. I also want to emphasize that the goal of DePolis is not to relieve voters from “going through all the discussions on a particular proposal”, but rather to provide a tool for gathering the opinions of Arbitrum users, which can be radically different from those expressed in forum discussions.

Regarding the maintenance of DePolis, I suggest checking the response I provided to Larva above.

1 Like
14

Thanks for your feedback!

From your response, I gather that there’s currently no publicly available GitHub or documentation for SimScore, correct? If there’s anything related to SimScore that can be explored, I’d appreciate it if you share it here or via direct message!

Regarding frontend vs. API

I don’t think the core entity of DePolis should exist in the form of an API. As I mentioned in my earlier replies, the primary use case of DePolis is not to analyze clusters or sentiment within the forum but rather to engage those who find the governance forum inaccessible (due to lack of time, skills, knowledge, or other reasons).

In the current governance framework, ordinary Arbitrum users lack both a way to express their opinions in a way that can be heard and the motivation to do so. I’d argue that the “opinion transmission system” operates almost identically to legacy governance systems: delegates are supposed to represent the interests of the network’s end beneficiaries (Arbitrum users), but in reality, no one truly knows what those users think.

In my view, SimScore and DePolis serve different purposes:

  • SimScore collects and automatically analyzes opinions from various levels of governance participants;
  • DePolis gathers opinions from ordinary users who, under other circumstances, would never express their thoughts in any form (except perhaps via a “Support request”).

Though there’s obviously synergy and potential for collaboration!

16

Thanks for your questions!

What questions should be used for your platform?

Technically, a conversation on any topic can be organized using DePolis. We can imagine various scenarios of how DePolis could be integrated into governance processes (check “5. Specifications. I. How Polis for DAOs should look” in the proposal), but as a starting point, it’s better to choose the most straighforward and solid use case. To identify this use case and understand what questions DePolis could be used for, it makes sense to explore the following:

1. Polis irl success stories:

Polis case studies can be found here. Polis has been used in numerous cases in Taiwan (and is integrated into the vTaiwan governance processes). Examples include debating technology regulation questions such as “Should Uber be allowed?” or “Should alcohol be available for purchase online?”. Other topics include:

  • “What do we have to do today to live in a climate-friendly future tomorrow?” (Austria)
  • “Airbnb in Athens and Thessaloniki” (Greece)
  • “Improving Bowling Green/Warren County” (Kentucky, USA).

You can find more case studies here.

2. Vitalik’s thoughts on using Polis in decentralized governance:

Vitalik (an active proponent of Polis) describes it as a social technology to help communities identify points of agreement between sub-tribes who might otherwise disagree on many issues (source). In another blog post, he highlights Polis as a tool for large-scale conversations to be used before governance participants solidify their opinions and input them into governance (source).

3. Polis blog and documentation:

Best practices for creating discussion topics are described in the official Polis documentation (source). Key takeaways include:

  • open-ended conversations are good;
  • Polis is most effective for controversial issues with a diversity of opinions;
  • keep it short (long topic descriptions are bad).

Combining these insights, we can identify what types of questions would suit DePolis:

  • open-ended questions that encourage participation from a large number of stakeholders with diverse opinions;
  • questions that address highly contentious topics, where opinions are polarized, or where there is a significant need to identify common ground;
  • questions tied to actionable outcomes, where the results of the discussion can directly inform governance decisions or policies.

Assessing how contentious a topic is requires further exploration, but we can find clues in the Arbitrum governance forum. For example, large grant programs are definitely contentious topics. Discussions related to defining key development directions, policies, or guiding principles for the DAO are also well-suited. Seasoned Arbitrum delegates likely have additional ideas for topics that meet the criteria described above.

Why can’t I use the usual built-in forum poll?

For several reasons:

  • most regular Arbitrum users do not visit the forum;
  • even if a specific forum poll gains significant attention and many people participate, the voting results will not be representative, as anyone can vote in forum polls (there is no protection against sybil attacks and no way to restrict the list of participants to a specific community or sub-community);
  • the forum only supports standard polls where the responses are predefined by the poll creator, whereas DePolis allows participants to express their opinions in their own words;
  • unlike standard polls, DePolis enables the identification of key opinion groups among participants and provides deeper insights into the community’s needs and preferences;
  • several other advantages of Polis compared to traditional surveys are described in this article: Examining Pol.is as a Survey Methodology.

Thus, DePolis is fundamentally different from built-in forum polls, and the output it produces provides decision-makers with far greater and more profound insights into user perspectives.

Why is Snapshot voting not enough?

For the same reasons as with built-in forum polls. Snapshot provides voters with a predefined list of options, whereas DePolis allows participants to express their opinions in their own words. Additionally, Snapshot is designed to be a decision-making tool, while DePolis is more suited for an earlier step - determining the landscape of opinions, “checking the temperature” of the community, etc. Furthermore, Snapshot is a tool for token holders, which makes it impossible to use Snapshot to gather opinions from, for example, active Arbitrum DeFi users who may not hold tokens.

1 Like
17

I’d like to emphasize again - we’re not talking about moving discussions to a new platform or creating any kind of “competition” between DePolis and the existing forum!

I’m afraid it’s impossible to use the content of this forum to test DePolis. DePolis doesn’t “work” with information that already exists in the forum but instead requires a different type of input - statements and votes from conversation participants. However, as a “proof of usefulness”, I suggest you check out these Polis reports (generated from conversations conducted through the original Polis):

DePolis and the forum serve different purposes - please refer to my reply to Argonaut above, where I’ve detailed their differences.

1 Like
18

Thanks for your feedback!

Never heard of Snapshot using any Polis-like mechanics, nor have I seen any mentions of Snapshot in the Awesome Polis repository. I’d be really grateful if you could share more info on how Snapshot uses Polis!

Can’t agree with this statement. While Polis can indeed be time-consuming when used irl - for example, there’s a great description of how a multi-layered discussion incorporating Polis was organized in Austria (source) - the generalization feels off.

I see from the last five Arbitrum Snapshot votes that the average time between the start of a forum discussion and the vote on Snapshot is 2–4 months. It seems that starting a DePolis conversation in parallel with the forum discussion, running for about a week, wouldn’t consume anyone’s time excessively.

I’d also like to reference Vitalik’s post (try not to overuse references to him, but it’s particularly relevant here), where he argues that tools like Polis “allow decisions to be made quickly, at large scale, and in a way that favors quality in a dynamic way that allows experts to quickly rise and fall with each individual topic or decision”.

There’s no doubt SimScore is a great tool for transforming existing data from forum/Snapshot/Tally into actionable insights! However, I’d like to point out that DePolis has a different goal. Unlike SimScore, DePolis’s primary purpose is gathering and analyzing opinions from those who don’t currently participate in existing discussions (the 99% of Arbitrum users). I’d also recommend checking my reply to Maets23, where I mentioned other differences between DePolis and SimScore.

Could you elaborate more on this? As for Harmonica, I see that they position themselves primarily as a tool for initial sense-making in small groups (e.g., core teams), which is very different from DePolis’s objectives.

19

Am I correct in understanding that the goal of DePolis is to enable the average user to express ideas, participate in discussions, and ultimately improve the efficiency of Arbitrum governance? If so this point is very valuable, I am concerned about several issues, 1, whether it can really be used:
This tool can’t just be a round of discussion in the DAO on the end, it has to become part of the governance process, for example, before each important proposal, you must use DePolis to collect community opinions, the final report and consensus directly into the voting reference, so that the tool can have a real value, rather than form and function. 2, user participation threshold:
Currently DAO user participation is limited, many people are just spectators, DePolis should consider lowering the threshold of participation, such as with a simple voting + praise mechanism, with clear rules of reward, so that users feel that participation is valuable, rather than a waste of time. But if it involves rewards here there will be a big unclear mechanism,3 the incentive mechanism needs to be clear and simple, and the amount of incentive should be dynamically distributed and adjusted according to the influence and importance of the discussion. How to realize this I feel is difficult. How do you think about it. 4, on the long term question, how do you aid in integrating DePolis with existing governance forums and voting systems, such as mandatory DePolis reports attached to each proposal, formal voting, etc. How about all the links articulated?

1 Like
20

While reading, I was particularly struck by the “Controversial Issues” section in point 6, where it mentions:

In some cases, linking statements to specific accounts might discourage participants from expressing their genuine opinions. However, integrating onchain logic with anonymity poses a complex technical challenge and is a subject for future research.

It would be interesting to evaluate the possibility of implementing a delayed publication approach on the blockchain. During the discussion period (for example, 10 days), participants’ comments and votes would be kept in a standby state and not immediately stored on the blockchain. This would allow participants to express their opinions freely without external influences or biases.

Once the discussion period concludes, all comments and votes would be transparently published on the blockchain, ensuring the verifiability and traceability of interactions. This mechanism would balance the need for privacy during the debate with the transparency necessary for data-based decision-making.

Additionally, we could implement a system for evaluating the quality of comments, allowing the most valuable and genuine contributions to stand out. This way, we ensure that the opinions expressed during the process are authentic and relevant, preserving the integrity of the decision-making process.

This solution would not only protect participants’ privacy while fostering open and honest debate, but also provide transparency and accountability at the end of the process.

Does this seem viable to you? Certainly, it is necessary to assess how feasible and achievable what I propose is to ensure that it can be effectively implemented.

2 Likes
21

Thanks for your feedback and questions!

The short answer is yes, you’re correct. Two minor corrections:

  • “participation in discussions” for an average user in the context of DePolis in most cases means upvoting/downvoting others’ statements (but not coming to the forum);
  • “improving the efficiency of Arbitrum governance” mostly stems from the ability of governance participants (delegates, Foundation, large stakeholders) to understand the community sentiments from DePolis reports.

There are some other ideas on how DePolis could be integrated into governance processes (for example, it could be used by delegates and forum participants; check “5. Specifications. I. How Polis for DAOs should look” in the proposal), but the logic here is: “let’s start from the most straightforward and practically useful use case and see if it works out”.

In my opinion, for DePolis to be a working tool, we’ll need:

  • some informational support from Arbitrum voices in the form of invitation to conversations in tweets/posts/etc.;
  • a more or less stable framework of “when and how to use DePolis” — this is a topic for a separate discussion. If DePolis shows its usefulness (that users will come and vote, and that reports will provide some insights) after our test conversations, we’ll likely have this discussion.

I’d say the best initial use case for DePolis is to “check the temperature” on spicy discussions. Who and when should start the conversation, and where should the incentives come from - I don’t think we can answer these questions right now. Not because they’re too complicated but because imho the best approach here is to look at how the test conversations went, articulate possible options, and discuss them all together.

Regarding “becoming part of the governance process” - I think DePolis and its reports would be useful even without deep integration into governance processes. If we have a big enough conversation conducted through DePolis, delegates would likely use these results in forum debates (hence, community opinions will be taken into account, and Arbitrum governance will already be a little bit better and more effective). Then, if we’re at this stage, it’s time to think about deeper integrations into the governance framework.

DePolis is supposed to have the lowest possible participation threshold — users can either provide their opinion in the form of a short statement or upvote/downvote others’ short statements. Afaik in the original Polis the average ratio of statement providers to voters is around 1:10, so we can expect that the default option for most users will be simple voting.

Agree, the incentive mechanism is one of the most complicated and nuanced questions. I think setting up a proper incentivization framework for DePolis is an iterative process, where we start from the most intuitive solution - determine the size of incentives in advance and incentivize two categories of statement authors:

  • those whose statements received the highest number of upvotes (participants who helped identify points of consensus);
  • those whose statements received the most upvotes within their opinion groups (participants who helped identify points of disagreement).

This logic provides a clear and simple mechanism (both categories of statements and authors are visible in the DePolis report).

If initial conversations show DePolis’ usefulness and DePolis is further integrated into governance processes, we can imagine various approaches to the question of incentivization (the choice of a specific approach depends on many factors, including those mentioned in question 1 above), for example:

  • if the need to launch a DePolis conversation is determined/proposed by the Foundation or the proposal author, the Foundation or the proposal author could define the size and logic of reward distribution at their discretion;
  • if a DePolis conversation is launched automatically and includes something like an incentivization fund, the total size of incentives could dynamically increase depending on the number of participants in the conversation (which would serve as an indirect indicator of the discussion’s importance);
  • DePolis could work alongside AI tools like x23.app and adjust the size of incentives based on the level of importance of the discussion as determined by such tools.

Imho the question of DePolis integration into governance forums and voting systems is a matter of a fairly distant future and will only become relevant after:

  • DePolis has a battle-tested product and has demonstrated its usefulness in numerous discussions;
  • at least a few DAOs are frequently using it.

I do have some thoughts regarding long-term possibilities for DePolis, for example:

  • DePolis could be integrated into forums/Snapshot/Tally so that delegates could easily access reports - however, it’s not clear whether these integrations should be done at the DAO level (e.g., attaching reports as part of a proposal standard form) or at the level of the governance venue;
  • it’s not clear where DePolis will show its usefulness - only in large-scale, highly contentious discussions, or as an auxiliary tool for many proposals;
  • there’s a chance that DePolis will change direction - for example, by adding AI features, which would change its value proposition (for more info on the Polis/AI intersection check out this paper) or by being integrated into info finance tools;
  • DePolis could converge with other deliberation tools (some of them were mentioned by @danielo).

The bottom line here is that yes, DePolis has many positive long-term development possibilities (the most obvious being closer integration with existing governance platforms), but there are also many unknowns. Therefore, an iterative approach to DePolis development, starting with the parameters and goals described in the proposal, seems logical.