Arbitrum Triple Dip (Delegate Incentive Program)

Proposals
,
21

it’s a requirement to get paid. currently, delegates spend time to dispute the results as well when that same energy could be put into evaluating what everybody else is contributing which would after the side effect of increasing the context of knowing what’s going on in the DAO.

this is also a concern for the current model, right?

actually probably even more so, because the bias would be bigger in a centralized evaluation model since there would be way less people evaluating. the advantage of a decentralized evaluation model is that the biases get faded in the end result because different human beings have different biases so they compensate for each other.

this is an interesting suggestion but I think we would learn way more if we would experiment with an approach that is more distant from the current status quo. we will have 1 year of centralized subjective evaluation to learn from already so we might as well try to learn something different for a bit, and only then go for a middle ground approach if we find that better.

2 Likes
22

This is probably the biggest flaw here. There is a lot of to be said on what we are paying delegates for.

The impact, today, has to go through voting. It can also happen through initiatives but the low hanging fruit (and fundamental need) is to have a strong base of voters.

Paying all delegates to watch what they are all doing in the ecosystem, is probably a collective waste of time, energy and money

  • we are likely pushing 15 to 30 delegates to read all threads of all others. Thinking about a 2 hours time to do it (and is quite conservative), we are putting on 30 to 60 hours per delegate
  • we are pushing delegates to evaluate contribution that might be impactful but, most likely, are not: the impact of single contributors 99% of time happens through structured programs
  • we are rewarding this work with 840,000 ARB in a year.

This is the wrong value proposition.

Note, it doesn’t mean we shouldn’t find a way to pay contributors. We definitely should. The system above is in my opinion a very wrong way of doing it.

1 Like
23

@JoJo as usual, you are twisting my words quite a bit.

To clarify, I said that for delegates that want to get Second Dip rewards, they need to:

  1. report their contributions in the forum, up to [5], in a short form and ideally with verification links
  2. read everybody else’s contribution reports
  3. distribute their voting power to the other delegates that they consider contributed the most that month, by voting in the offchain shielded weighted vote

So it’s not fair to say that we are paying delegates to review other delegates work. We are paying delegates for their contributions, but they only get paid if they self-report their contributions and vote on the monthly evaluation vote (which would required them to review other people’s work). It’s different.

Also, the benefit of “forcing” delegates to review everybody else’s work is that it is the only way to ensure everybody has the most amount of context of what everybody else is doing, and avoid duplicated work and even better, serendipitously provoke collaboration between delegates, which we desperately need to both onboard new talent and get better outcomes from those contributions.

Also, delegates don’t need to read all the forum posts of all the delegates every month, but only the ones, up to [5], that each delegate considered most impactful in their self report. This has the advantage of both being mindful of the attention span of evaluating delegates and opening the scope of contributions to be much bigger than just forum posts, which is currently the only thing awarded with Delegate Feedback in DIP 1.5/6/7 and makes up for 40% of the monthly score.

With this Double Dip design, delegates are able to report many more types of valuable contributions and get evaluated for those in a decentralized, transparent, and token weighted way.

The next step would be to create a template forum post, in a specific category of this forum, for delegates to create this monthly self-report, and that template should limit the description of each contribution to one paragraph only and offer an option to include a link to verify the contribution.
So, every month evaluating delegates would only have to read up to [5] paragraphs per each delegate that qualifies for Second Dip rewards.

1 Like
24

how can i vote on the contribution of 30 delegates, each putting out 5 different things they did, if I don’t evaluate what these things are? We are paying delegates for their contribution the moment in which they are evaluated by other delegates about these very contributions. This likely takes more time than anything else because of the staggering amount of different things that can pop up (from being in an event, to run a podcast, to spin up a pool for two new assets, to have a BD call with a team to migrate in our chain, to a report about the status of the eco, to whatever else comes to mind).

Again, will rephrase my previous post. You are jumping the gun here by thinking that what you just wrote here is something that brings value, or at least meaningful value, for the program, the DAO, and the delegates.
In my opinion, is not. A lot of delegates care only about a portion of the DAO and ecosystem and for good reasons: they work in a specific department, and don’t have the time, expertise, energy or even interest to evaluate other things. I think this is both fine and in line with normal social norms of people gathered in a place (real or virtual) to achieve a goal: they don’t necessary need the full context of everybody else. In the place I work, I don’t need to know every month the 5 best achievement of my other 30 colleagues to put it differently. So, why should we force delegates through this path? The answer is not “because we are paying them”: there are things more valuable then money here, like time and bandwidth to contribute to your protocol/program/initiative, to the ecosystem, and so on and so forth.

This is what I was trying to communicate to you, which in this discussion represent, effectively, a proposer of something we need to vote upon. And, for this reason,

I won’t comment neither on this statement nor the tone of it.

2 Likes
25

what if it was just [1] contribution a month per delegate?

the idea of this proposal is that these parameters are decided collaboratively to start and adjusted at any time (but not retroactively).

do you think it would be feasible if it was limited to just 1 major contribution per month? and the contribution description had a character limit and strict format?

I know and I agree with your description of the situation but the problem is that, in Arbitrum DAO (and most token weighted, single house DAOs) all delegates need to vote on all the things. And in Arbitrum in specifically right now, because of the quorum issues, more and more delegates that have very little context about the DAO, will need to be activated to vote on proposals. The goal of the First Dip is to do exactly that and I think we all agree that is required at this point. Therefore, we risk paying low-context delegates to vote, which we know will lead to bad outcomes and decisions. there’s no escaping that in my opinion.

that’s why i think that actually “forcing” delegates that want to get paid for their contributions to review everybody else’s contributions every month is a value add activity for the DAO, because as I said before, it will lead to more shared context, collaboration and serendipity between delegates.

but I think you and me will probably always disagree on this point. and that’s ok. that’s why we need diverse and more, way more, perspectives in this DAO.

26

Hey all!

As suggested initially here:

…me and the fine folks at @Curia transformed this Double Dip proposal into a Triple Dip Incentive Program.

I’ve edited the proposal to reflect the new program design with 3 separate “Dips”, and taking into account the feedback offered by several delegates above in this thread.

27

Thanks to @paulofonseca for incorporating the Peer Recognition Score (PRS) framework into the updated Triple Dip model. This concept grew from our earlier feedback on the need for a peer-driven evaluation layer to enhance the program’s transparency and fairness.

The parameters detailed in the PRS Notion doc are a starting point, and we are actively refining them to best suit the ArbitrumDAO context. We invite all interested community members to contribute their thoughts and feedback to help improve this system.

Based on the community’s initial sentiment in a Snapshot vote, we are prepared to move forward with a more detailed phase of refinement and modeling, including simulations to show how the system would function.

1 Like
28

Hey there!

Just posted an offchain vote for a temperature check on this proposal here.

To highlight the fact that this is not a binding vote, I’ve included an introductory text on the Snapshot proposal that reads:

Hello all!

This is not a binding vote. It’s just a temperature check vote to sense the appetite of the DAO towards experimenting with an objective, rules-based, and algorithmic based Delegate Incentive Program, that doesn’t require a dedicated program manager to run it, and rewards delegates in a predictable, transparent, automatic, meritocratic, and faster way.

Please read the proposal below to understand the proposed program process, general rules, and specific parameters. All specific [values] are just a indicative example at this stage and will be decided and put for a binding vote later, but only if this vote has a favourable result.

Thank you for voting!

Thank you for your attention!

29

Voting against this proposal in the temp check.

First, most of the points i posted above were not properly addressed. Just from an operational standpoint, thinking that a program that rewards delegates has none or minimal costs (in term of time, people and discussion) is quite naive. It doesn’t matter if, on paper, is totally algo based: delegates will complain, will ask for clarifications, will look for authority to address perceived issues. One of the thing I said, more than 1 year ago, to SeedGov, was that they were asking for too low of a pay considering the amount of trouble they were getting into, and looking at the amount of work they did per month for a year I was probably right.
All this to say: there is the need of a program manager, just because we are here about rewarding people for contributions here. Even just the simple action of voting might need tweaking: if we go back to a situation stip like, it might just not matter if a delegate is not able to vote on all proposals for example. TLDR: all algo only exists on paper.

I am also a bit surprised about the fact that the proposal got more complex over time, just now going to vote I saw that a third mechanism was added. I do understand that the third one is effectively a simplification of the previous second, going for a quarterly type of review.
But the second one, based on likes, IMHO doesn’t work.

While there are weights for “likes” based on who puts these out (author vs foundation vs dip delegate vs unknown user) the system does not take in account that there is an heavy skew of users activity in the forum itself.
As napkin rule, we can see that the biggest a delegate, the less the participation in the forum is, with a few exception. And this is still tied to what I posted above.

These users, these power entities of arbitrum, are just not here in the forum. Evaluating forum contribution based on forum likes is, in the end, the creation of an echo chamber, and doesn’t solve the problem (is it even a problem?) of big accounts/entities interacting with this very board.
It doesn’t matter currently if there are different weights: with the proposed one, 0.2 for unknown vs 1 in dip, you need 5 unknown/unregistered to make basically a valid vote, and we do have several forum members (thousands) that are registered, that do interact from time to time but that never were strong part of this community.

Even just cutting out this latter portion and keeping the DIP delegates as interaction, is fairly easy to see how the participants to this activity will only be a small subset of the current DIP enrolled people, which still lean toward the echo chamber approach.

I want to add a final remark. I was thinking about voting abstain, to also have more time and reflection on the proposal from the foundation as well that should follow this one next week, and give space for more discussion and alternatives.
But the more I read this proposal (that got more complex over time), the more I understand how it fails to address the fact that we have basically people doing activities and being rewarded for such activities. Thinking that this setup doesn’t require a central entity to which the community as a whole can fall back for issues, questions, problems and discussion, and thinking that this role is implicitly for foundation to take at almost 0 cost in term of time and energy, shows how imho there is little understanding for these social dynamics.
As a second note, the added complexity of this version of the program is not calibrated to solve for properly evaluating impact of activity of delegates, since it falls into an echo chamber approach.

2 Likes
30

Entropy will be voting against this proposal. In the past, we have seen formulaic scoring systems for forum comments reward/encourage low value behavior. The proposed program is overly complex and it is our view that such an initiative requires the subjective human input from a program manager, if we must have delegate incentives at all.

31

Thank you to @paulofonseca for driving this discussion and putting forward a concrete design.

We voted AGAINST in this temp check. While separating voter vs. contributor tracks makes sense in principle, as implemented here the design isn’t workable.

The 100% participation requirement is too rigid and will penalize good-faith delegates for travel/emergencies/scheduling. A realistic target would be 95–98% with an excused-absence path—but since that’s not what’s on the table now, we can’t support it.

Requiring everyone to post and review five “top contributions” every month creates heavy operational overhead and reviewer fatigue, likely discouraging smaller delegates—and, as several delegates have already pointed out, this would be overly burdensome in practice. And as JoJo noted, mixing voter and contributor tracks can penalize builder/protocol-focused delegates (who spend time on-chain/product work rather than the forum) and drift into an echo chamber.

If the proposal returns with softer participation thresholds and a lighter, less frequent, more objective contribution process, we’d be open to reconsidering. For this iteration, we are AGAINST.

1 Like
32

hey there @ITUblockchain

this vote is not to approve the specific parameters of [100%] voting participation for example. it’s just to do a temperature check if a design like this for a Delegate Incentive Program, would be something the DAO would be interested in, at all.

the proposal was changed to make this process for the now called Third Dip, a quarterly process, not monthly, precisely because of the feedback received.

33

Thanks, Paulo. We know this is a temp check and the review moved to quarterly. however; We’re still against. Even quarterly peer reviews can add extra workload (especially for smaller delegates), and also we think that a participation signal without a clear excused-absence path is too strict.

And also, Our concern on this issue still stands.

For this temp check, we voted against, but we’re open to re-evaluating a revised version in the future

34

We will have a DAO call in a little less than 1 hour to go through this proposal, get feedback on it, and discuss it at lenght. If you would like to join us, this is the call link: https://meet.google.com/yad-hqxo-kam

35

Today’s call recording is here https://drive.google.com/file/d/1oAIrkY-dr08V8dEPvf_1fgQ7PVr3y8RR/view?usp=drive_web

And the chat transcript is here https://drive.google.com/file/d/1EmFGoemrwX6CA3eao3LEWNwaCwAF0m4n/view?usp=drive_web

1 Like
36

Vote: FOR

I love this proposal, and I support exploring alternative mechanisms for rewarding delegates and contributors. I particularly appreciate the distinction made between these two groups, it’s something I’ve advocated for previously, and seeing it reflected here is definitely a step in the right direction.

I love the auto peer monitoring concept, which aligns well with Ostrom’s principles (I’m an Ostrom Maxi) and avoids the need for a lot of oversight using more of a coordinape approach.

I don’t like the idea of NO admin… We’ll need some form of program management, and I’d love to see a budget to expand AI assistance capabilities for this. While I recognize we’re not yet at the point where AI can fully automate this work, I think it can help a lot… if this passes I’d like to see OpCo involved in the program management in a more active way, supporting iterative improvement.

I think the model is simple and clear, we will need to tweak the numbers of course, but thats ok… It’s so great that this is coming from the community and not an AAE, we should encourage more grassroots efforts from outside our inner circles, this is probably 30% of the reason I want to support this proposal, it’s good enough to work, and we should have bias in favor of proposals that come from the community IMO.

Thank you Paulo for the great proposal, I doubt it passes but it really should, and has my support.

5 Likes
37

We thank @paulofonseca for his proactiveness in bringing a new DIP proposal. We believe, as @Griff states, more proposals should come from the community; however, that should not be the driving force for support.

For this particular proposal, we are concerned about a few points, in particular:

  • We are not aligned with the scoring methodology around forum activity and believe that it can encourage low-value or spammy behaviour, and doesn’t necessarily translate into meaningful contributions.
  • The peer-to-peer review process adds burden on delegates, as mentioned by others; this will inevitably take precious time away from actually contributing. In our opinion, efforts should be focused on reducing delegates’ workload, not increasing it.
  • While the current DIP is not perfect, the role of a program manager with contextual knowledge about previous programs and Arbitrum’s direction is an added value not to be overlooked. We should strive to harness, develop, and foster contextualised knowledge that can make programs more effective. We believe this can be achieved without resulting in an overly-centralised program.

For these reasons, we will be voting AGAINST this proposal.

3 Likes
38

Hey @paulofonseca, as you mentioned on yesterday’s call, the main question in your Snapshot vote is whether we want a predictable, algorithmic DIP.

Your proposal has more hard-coded rules, while DIP 2.0 is more flexible (meaning the PM or OpCo can adjust the rules as needed).

At ETH Warsaw, Mateusz from Arbitrum Foundation asked me the same question: do I prefer a predictable, hard-coded system or a flexible one? I told him I prefer flexibility, and I still feel the same way.

The main reason is that every system eventually gets gamed. That’s why we need flexibility, so that rules can be quickly adjusted when that happens. This naturally requires some level of centralization, in the form of a program manager.

And it’s not just about preventing the system from being gamed. If we notice that a rule (or set of rules) within the DIP isn’t working as intended, it should be possible to change it and try something else, without having to go through the governance vote each time.

That’s why I voted against your proposal on Snapshot. That said, I think your proposal includes some interesting ideas that could be incorporated into future iterations of DIP 2.0.

39

Thank you for the proposal!

I have carefully followed all discussions regarding the DIP.

During the October 21st discussion, @paulofonseca , you mentioned that, in essence, the proposal is about whether we want an algorithm to determine the distribution of rewards among delegates. Personally, I find this approach efficient. I agree that an algorithm, not a person, should make that judgment. However, as presented, the proposal focuses mainly on showcasing the triple DIP rather than exploring the differences between a system governed by an algorithm and one led by a human group. Therefore, inevitably, what is being evaluated is not only the concept but also the way it is presented. At least according to my understanding.

There are parts of the proposal that, given the current state of the DAO, I disagree with, for example, using likes as a form of evaluation. About the rest, I would happily discuss if we had all the maths done.

In general, prompted by the broader discussion around the DIP, I’d like to say the following:

As I mentioned during the call, there’s no need for tension. I understand the need for incentives for larger delegates, since they play a key role in helping reach quorum.
However, smaller delegates need clarity on two important points:

  1. Does participation matter in terms of pluralism and genuine decentralization?
    If yes, then, plain and simple, there should be some form of reward, even if symbolic, for smaller delegates as well, within a proof-of-work logic.

  2. Is there an intention within the DAO to create space for smaller contributors?
    If so, we need to understand what kind of contributions the DAO expects from us, so we can focus our energy productively instead of on irrelevant issues that lead to frustration and friction.

I believe that by answering the two former questions, the noise over the DIP will stop and we will be able to focus on how we can attract more capital and increase revenues, since this is the sector that demands the most focus and energy.

Summing up, even if I do not totally agree with the proposal, I will vote Abstain, because I do not want to oppose @paulofonseca ’s proposal. On the contrary, I recognize the effort and commitment he’s put in. He clearly respects the ecosystem and is an active participant. His proposal reflects one core value: fairness!!

2 Likes
40

Hey @TempeTechie just to clarify that, as it says in the intro of the offchain vote proposal text, the main thing this vote is trying to do is:

…to sense the appetite of the DAO towards experimenting with an objective, rules-based, and algorithmic based Delegate Incentive Program, that doesn’t require a dedicated program manager to run it, and rewards delegates in a predictable, transparent, automatic, meritocratic, and faster way.

also, regarding this argument of yours to justify a centralized program manager instead:

do you think the reason for @SEEDGov to have changed (sometimes retroactively) the rules of the current DIP (going from DIP 1.5, to DIP 1.6, and now to DIP 1.7), was because the program was being gamed? can you produce evidence of past participating delegates “gaming” the current system to substantiate your argument?

because… on the other hand… I can produce evidence of a centralized program manager gaming the current DIP system, last month:

and clearly abusing it’s power to actively penalize contributors based on forum comments, in May:

and also abusing it’s power to straight up suspend contributors for tweeting an image, in June:

so I would ask, aren’t we really capable of figuring out a way, collaboratively, to have a Delegate Incentive Program that rewards delegates for the value they provide, with a collection of rules that can’t be gamed?

I mean… I personally believe that the whole point of smart contracts running in public decentralized blockchains, is indeed to design incentive systems that can’t be gamed. That’s why I’m here. And that’s precisely what I’m trying to do with this proposal.

And I would appreciate feedback regarding exactly that, how do you think the specific rules and parameters of this program can be gamed?

1 Like