Collected KyberDAO Q&A

(1) What is your opinion on the recent sub 1% participation rates (% of holders) in dao votings on Aragon, maker & 0x respective daos. What are the main reasons for that?

First, we need to consider what we classify as a “DAO,” and neither Aragon nor 0x really fit the elegant description of a DAO as initially proposed by Vitalik:

Automation at the center, humans at the edges.

Why is this the case?

Foremost, both 0x and Aragon both have centralized proposal curation. In the case of 0x, the only rights token holders have are a veto authority. For Aragon, proposals undergo curation first by “AGP Editors” (two handpicked individuals by the Aragon team) and then a 3-man association. In both of these cases, the relationship Vitalik describes is explicitly inverted: humans are at the center, and automation (the actual nuts and bolts execution of proposals by the blockchain) exists at the edges.

MakerDAO is different, as large portions of their governance exists on and is automated by a blockchain, yet has the same issue with participation. What gives?

In MakerDAO’s governance article, they explicitly hint at the problem:

Combined with the majority wins tenet, it should be clear that the main responsibility of MKR token holders is to be cognizant of all proposals that enter the voting construct. In fact, the lack of quorum implicitly incentivizes all MKR token holders to be involved.

… but, they fail to articulate a solution to this issue: rather, they anticipate that participation will be an incentive-in-itself. However, we understand this isn’t usually the case, through a concept called Rational Ignorance:

Rational ignorance is refraining from acquiring knowledge when the cost of educating oneself on an issue exceeds the potential benefit that the knowledge would provide.

The cost/benefit analysis of voting can be applied to both the Aragon and 0x cases as well in a similar way: why would anyone feel incentivized to thoroughly research proposals and vote when a handful of highly curated proposals are the only proposals brought to the table? The most extreme case is 0x, where their last proposal was a technical upgrade presented by their dev team.

Generally, your question is really getting at costs of voting, and what motivates people to pay these costs by participating. The answer to your Q is: the costs are too damn high! When we consider that there’s no cost to not voting – since the experts already figured out the best proposals to vote on anyway – and that the results are largely certain (to date I believe only 1 AGP has been vetoed) – it becomes clear why there’s no real participation in their system.

(2) how would you tackle the issue of most people not voting, especially due to the prevalent technical risks associated with using unproven smart contracts / removing funds from cold storage

There’s three Qs here:
i. How do you tackle the issue of most people not voting
ii. How do you address technical risks
iii. How do you address UX friction in voting (removing funds from cold storage)

I’ll answer them 1-by-1:

i.) In the DAOstack framework, we actively consider the costs of voting to the collective attention of the voters, and have designed a consensus protocol around it. The protocol is designed, actually, not to increase participation in the voting for every proposal submitted, but to ensure that there is alignment with the entire voter pool even when participation is low. The DAOstack Whitepaper (pg 22) discusses the mechanics of aptly named “holographic consensus” in detail:

In holographic consensus we want to eliminate possible mismatches between decisions made by small groups on behalf of the greater majority and the decision that would have been made by the greater majority itself, if enough attention was available. To eliminate possible mismatches, we introduce a novel prediction game, in which anyone — not necessarily a DAO reputation holder — can place predictions, backed by stake, on the fate of proposals… Correct predictors make profits, and incorrect ones lose their stake. By playing the prediction game predictors contribute to the governance process by signaling to, filtering and maintaining the governance process.

Finding out these social mismatches, or inefficiencies, is the most important part of making a collective attention more efficient and a governance system more scalable. By providing stake-backed predictions, the predictors provide valuable signals to the governing network about proposals that are under or over evaluated… once a proposal is being boosted, there is enough predictors’ skin in the game to make sure the process is effective and that whoever needs to be participating indeed participated. We use crypto-economics incentives to “outsource” the navigation of the collectiveattention of DAOs to a professional network of predictors.

The main takeaway from all of this is that with our solution we are not trying to boost participation for every decision taken by a DAO; in fact, we would argue that this is a bad thing as it risks tiring out the collective attention of the voters, especially if there is a large frequency of decisions to take. Holographic consensus tactically aims to boost participation when there is ambiguity as to whether a proposal’s expected outcome is aligned or not with the presumed outcome of all DAO members participating.

Additionally, there are some main technical differences between DAOstack and the aforementioned token voting systems discussed, the foremost being that Alchemy Earth is optimized as a Reputation voting system. Two recent posts describe this in detail:

A Guide to DAOstack’s Initial Reputation Protocol
Reputation vs. Tokens

Generally, with our Reputation voting system, proposals tend to have a greater level of participation than the original token voting examples presented.

ii.) Our opinion is that technical risks are best approached through community bug bounties and extensive auditing. One example is the dxDAO bug bounty, which successfully resulted in a bug being reported. We’ve worked together with ChainSecurity on DAOstack’s smart contracts as auditor – for reference, they were the ones who caught the Ethereum Constantinople launch bug prior to the fork.

But – more than this – it’s important to start by managing small sums of funds, and then scaling up to larger fund management. Good risk management is the best way to start transitioning to a decentralized treasury, something the original DAO didn’t practice.

iii.) Moving funds out of cold storage for the purposes of voting does not exist with DAOstack, because it uses a Reputation based voting system. So you can keep your tokens tidily tucked away without the hassle of moving them back and forth.

3) is it more efficient for a project like kyber to remove its centralized governance now that ensures speed of decision making or would you recommend to do it later after a certain traction has been proven and decision making can be conducted slower and require more decentralization?

A lot of this depends on what aspects of Kyber you’re trying to decentralize. With most orgs, and especially blockchain projects managing a decentralized application, there’s a wide space to play in.

Does Kyber want to decentralize…

  • A microgrants process, to develop and promote the project?
  • The management of its application’s smart contracts, so the DAO becomes the owner and effective controller?
  • The curation of a registry of tokens listed on KyberNetwork?


My take on it is to run a pilot decentralizing the simplest items first, such as a microgrants process, before looking at bigger picture items, such as the application itself, or a token registry. So I would say: after a certain traction has formed, in a low-risk, no stakes environment, with a small sum of funds (perhaps $10k/month worth of KNC, ETH, DAI…)

One thing to consider: there’s no reason a business with a specific set of KPIs to hit cannot exist as a sub-agency within a larger DAO. In fact, it might be valuable for a big “Alliance DAO” to have one or multiple businesses sitting within it, acting in the interests of the crowd, receiving grants as part of the ecosystem.

4) what metrics would you propose to measure the successful implementation of a dao?

Some questions to ask:

  • Is the DAO being effectively managed by a small association of managerial operators, or through a widely decentralized crowd of greater than Dunbar’s Number?
  • Does the DAO have a source of sustainable revenue, so it can continue to grow and support itself?
  • Is the DAO receiving many proposals for different types of jobs – such as community growth, product development, and content creation – or just a few proposals focusing on a very narrow perspective?
  • Is the DAO’s activity growing (in terms of participating voters) or shrinking?

There are different success metrics for a DAO, but I think these are the most important from the get-go.


Thanks a lot for the very informative answers @papa_raw.

At the end of the day what it boils down to is the mechanism design of the respective DAOs, basically which desired objectives they try optimizing for.

If I understood correctly, DAOstack’s point of view is that DAOs should not optimize for high participation, but rather for coherence and efficiency in decision making on a large set of proposals. What is still unclear to me is how exactly you would measure that. You proposed the following to metrics among others:

  • Is the DAO’s activity growing (in terms of participating voters) or shrinking?
  • Is the DAO being effectively managed by a small association of managerial operators, or through a widely decentralized crowd of greater than Dunbar’s Number?
  1. Based on the DAOstack framework, wouldn’t you optimize the system to have a) shrinking participating voters and b) a small association of managerial operators rather than a large set of voters on average for the sake of coherence and efficiency of the system?

  2. Building on top of 1), does DAOstack aim to reduce the participation rate by regular voters in order to focus their collective attention solely on the extremely important, ambiguous decisions that occur fairly infrequent (e.g. when a large malicious voter votes on something that clearly goes against the collective interest), while leaving the regular votes to a kind of “full-time” participating minority that gain extra value from actively participating in prediction markets which enables them to overcome the ‘rational ignorance’ dilemma?

  3. Side question beyond the scope of discussing DAOstack: Do you think a technocratic system would serve the interest of the people more effective than a pure, direct democratic one?


If I understood correctly, DAOstack’s point of view is that DAOs should not optimize for high participation, but rather for coherence and efficiency in decision making on a large set of proposals.

This is for the most part true. We are aiming to optimize for expanding participation when the outcome of a proposal is ambiguous (increasing the sample size will better and better reflect the majority outcome).

(1) For the most part, this is theoretical, but:

  • Generally, we expect to see a small group of highly active managerial operators. This is consistent with the history of decision-making: most communities are steered by a minority of decision-makers.
  • Holographic consensus aims to involve the large set when there’s ambiguity in the outcome of the decision, as previously mentioned. This is why it possesses the characteristic of a quiet ending period, that is, the time by which a proposal is passed through relative majority is extended if the outcome changes within a parameter period as defined by the DAO.
  • I’m not sure what you mean by shrinking participating voters, but yes, as a DAO scales we should expect a decay of Reputation held by any historical bloc as the DAO mints new Rep, inflating the general pool, unless that bloc increases its member count or designates additional Rep to itself.

(2) We don’t aim to reduce the participation rate, but attention is a scarce resource. When the collective attention of a crowd is activated, it has a “cooldown” period; thus it does not make sense to focus this attention all the time on many proposals.

The full-time participating minority can participate in the prediction market aspect (by boosting proposals and generally steering the DAO), but they don’t have to. We imagine a network of Predictors working across all DAOs to turn social arbitrage into economic profit (that is, predicting correctly; our plan is to go one step further and develop tools to automate prediction).

Our expectation is the rational ignorance dilemma can be elegantly resolved when there’s an incentive structure to “pay for” the cost of attention, namely the prediction game.

(3) Since this is outside the scope of DAOstack, happy to speak for myself: I think that technocrats have a particular type of way of making decisions that isn’t always optimized towards great governance, especially if the shared ideals of a community have nothing to do with technology. Can you imagine, for instance, an association of self-employed mothers being run by a technocratic elite? Absolutely not.

I think what you’re getting at is more particular to the specific use case of the KyberDAO, and I would say you need to have a diversity of opinion specific to the DAO’s community to govern it; that is, traders, token holders, developers, researchers, and I imagine even those organizations that choose to list their tokens within the Kyber Network. If you could slice up the Reputation between these segments in an effective, logical way, I’d expect an alignment and fair representation of the community’s myriad values. This to me should be the goal.

1 Like