[ARFC] Cancel Llama Service Provider Stream

@Llamaxyz has done some fair work in the past. I understand the discussion is for the remaining time. The payment amount for the remaining 2 months (and initial engagement) do appear to be bloated for the services being provided. AAVE DAO should be concerned about this.

I am in agreement for a pricing negotitation/discount on services left for 8 weeks.

A point was mentioned earlier that DAO has allocated this budget and service contract, however, the same DAO can also cancel it before expiry. Thats the beauty of DAOs.

Moving forward a framework is essential to ensure the service provider, community, DAO and token holders are getting a fair agreement in place in terms of $ value as well as tasks (for bear and bull markets).

Finally, are there any other service providers that need to be reviewed in terms of pricing and contribution value?

I am sure that If we get more conscious about the DAO spending, we will have some conclusions on a short term that will probably end in some proposals. However let’s keep that on standby for now until this vote happens and then we can discuss about it on a forum thread.


Thanks everyone for your feedback. It seems like many community members are open to a reduced pricing for the last two months, which we are willing to do. We request that a Snapshot on this subject has three options for the community to choose from: cancel, reduce pricing, and no change.

Here are a few high priority tasks we have for the last two months:

  • Curator: dedicated contract for managing the exchange of assets with MEV and price-impact protection. It will help Aave swap assets, acquire LSTs, and transfer assets
  • StrategicAssetManager: a dedicated contract for managing Aave DAO’s strategic assets

These contracts have taken months of technical research, development, and testing, along with coordination with external technical teams like Yearn, Balancer, and Cowswap. It is difficult to hand over contracts like this so late in the development process. They will be useful open-source tooling for Aave that can be used by many contributors for the years to come.

We are also working on the following tasks:

  • Data analytics dashboards
    • Shortfall events analysis
    • Investment performance reporting
    • Risk analysis for DAO holdings
  • Aave financial reports for July, August, and September
  • Treasury management payload

We hope to arrive at a reduced pricing that is favorable for the DAO.


Let’s strive to maintain the quality of our discussions by avoiding one-sentence or low-value replies and staying focused on the main topic. This approach will enhance the clarity of the proposal and empower the wider community to make an informed decision when this ARFC is escalated to the snapshot stage.

Our community is known for its clear and accessible debates. Divergence in opinions is not only welcome but also a strong indicator of decentralization.

We believe that all parties involved, as well as the wider community, have had ample opportunity to express their views on this topic. Let’s continue to uphold these standards as we move forward.

It’s important to note that a stream cannot be modified once it’s started - it can only be cut and then a new one re-initiated. It’s quite revealing that an entity, which has cost millions for its “engineering” efforts, lacks a basic understanding of how Aave DAO’s smart contracts operate.

Those who are in favor of a “reduced” stream are encouraged to vote YAE on the upcoming snapshot to cut the current stream. Subsequently, they can create another ARFC to establish a new, ad-hoc stream that better aligns with their preferences.

This proposal will be escalated to the snapshot stage on Sunday, July 30th, and voting will commence on Monday, July 31st. We look forward to your active participation in this important decision-making process.


The Snapshot temp check is an offchain vote for the DAO to voice its opinion on whether it wants to cancel the stream, reduce the stream, or keep the same stream. All three are distinct options. If the Snapshot is passed to reduce the stream, a single proposal payload can be created to both cancel the current stream and initiate a new, reduced stream.


If the option is to reduce the stream, how much will it be reduced to? This will help us make a more informed decision.

We would propose doing 50% of our original contract price for the next two months. While this is a significant and unexpected reduction for us, we would like the opportunity to complete the work we have outstanding for Aave.


For the sake of absolute clarity, we have updated the Motivation section of this ARFC.

The primary aim of this ARFC is to terminate the stream of a service provider that has consistently underperformed and overcharged. There have been numerous instances where the actions of this service provider have justified termination.

These actions form the core motivation for the publication of this termination ARFC.

Discussions regarding the future scope of this service provider role, the associated budget, and potential candidates fall outside the purview of this ARFC and should be addressed separately in an appropriate forum post.


Marc continues to state factual errors that we have politely corrected but he’s yet to respond to a single correction. Now, he has completely updated the original forum post and people cannot see it. Making false accusations, being corrected, and then deleting the original accusations is against the spirit of a good faith discourse.

The original forum post read that the reason for ending Llama’s stream was:

As most members of TokenLogic are no longer working at Llama and were responsible for most, if not all, contributions to Aave

As we pointed out, this is factually incorrect as only one member has left and everyone else is still an important part of Llama; our 35+ proposal payloads have been written by engineers who are still full-time at Llama. Marc has not responded to this and has since changed the narrative to something else.

If the issue was truly about performance, why did Marc not inform us about this before in the first 10 months of our contract? This is the first we are hearing directly about a desire to cancel our stream. There was a lot of time to initiate a discussion about cancellation, especially after our detailed 6 month update. We post updates every month (1, 2, 3, 4, 5, 6, 7, 8, 9). We host Twitter Spaces every month for community members to join (e.g. 1, 2, 3, 4, etc.). Marc could have responded to one of the forum post updates (he hasn’t responded to a single one) or joined one of the Twitter spaces and let us know of this desire to cancel the stream. He has also never expressed this via private message. Why is it being done when there are 8 weeks left?

The post said that we don’t have much to do over the next two months, which is also untrue.

As we responded previously, we are developing the Curator and Strategic Asset Manager, which are complex contracts. These contracts have taken months of technical research, development, and testing, along with coordination with external technical teams like Yearn, Balancer, and Cowswap. It is difficult to hand over contracts like this so late in the development process. We are also working on a treasury management payload, three data analytics dashboards, and Aave financial reports for July, August, and September.

He said:

It’s important to note that a stream cannot be modified once it’s started - it can only be cut and then a new one re-initiated.

We pointed out that this is factually incorrect as a stream can both be stopped with a smaller, new stream initiated in a single transaction, i.e. using the same proposal payload. He did not respond to this.

He criticized Llama for its engineering efforts in this post, but he has privately complimented our engineering work as high quality and has said that he often re-uses variations of our payloads for his own AIPs given how good they are.

We have been receiving messages in private from current and former Aave contributors and community members expressing a similar sentiment, but who are scared to communicate it on the forum because Marc is the largest delegate. He currently has 247k AAVE delegated - nearly 80% of the quorum requirement - so it is challenging for people to speak against him when he almost entirely drives a vote. This is in addition to being a paid service provider to Aave. He was the largest voter for his own proposal in April for Aave DAO to fund ACI.

Given that ACI’s 6-month contract ends soon and partly overlaps with Llama’s work, there is a clear conflict of interest in this discussion. It would be as if Gauntlet put up a proposal to cancel Chaos Labs’ stream.

Llama’s contract with Aave ends in 2 months and we do not intend to ask for a renewal, so we’re one of the few groups that feels comfortable voicing what’s on a lot of people’s minds.

We have been willing to compromise and request to work at a reduced price though it is a significant and unexpected reduction. We hope the DAO should at least discuss a reduced price as several community members have brought it up as a viable option.


In the short time since we posted this, we have already received a number of messages of private support. People have told us that Marc is also a forum admin and Snapshot admin. Being the largest delegate by far, a paid service provider, a forum admin, and Snapshot admin is a lot of power for one person in a decentralized organization to have.


We would like to clarify that the Aave Chan Initiative votes with exactly zero of our own AAVE tokens - they are all delegated to third-parties. Furthermore, we receive zero delegations from companies, protocols, or entities. Our independence is paramount to us, and it is this independence that allows us to call out actors like Llama and ensure the DAO is free from excess.

Our voice represents hundreds in the community. When we vote, their voice speaks. We are ushering in a new era for the Aave DAO, an era of maturity, sustainability, and professionalism. The termination of your engagement is a significant step forward in our journey toward becoming a mature DAO.

We are committed to fostering diversity and professionalism within the delegate community. To this end, we have undertaken several initiatives:

  • We allocate a significant percentage of our own lean budget to fund the Orbit program. This allows third-party delegate platforms to receive funding and contribute to the DAO.
  • We actively support and encourage an independent DAO-funded Orbit program to reduce reliance on the ACI and promote diversity. reducing miss-allocation of DAO resources toward efficiency can free-up resources to fund this kind of programs
  • We have co-authored and supported numerous proposals with third parties to help them gain more influence within the DAO and provide them with the necessary support.

Can you name another entity that invests substantial resources to support its own “competitors”? We do this because we represent many voices within this community, and our primary duty is to the DAO.


We believe the discussion is getting a bit out of subject here arguing about powers. We do believe there is a conversation to be had on a different forum post about the ACI or ANY entity having so much admin power even if the delegations received are pretty well decentralised.
Correct us if we’re wrong but isn’t it possible to have multiple admins on Snapshot and the Forum? Good chance to create an “Admin Committee”.

We recommend that we focus for now on the actual ARFC which is the cancelling of the service provider stream for @Llamaxyz.


Agreed that the ARFC* has changed intention, but prob a good idea from @0xkeyrock.eth and others to just focus on the remaining Llama stream for now.

My 2c:
Unfortunately, the DAO probably agreed to a bad deal with Llama and overpaid, and maybe Llama under-delivered. There was a clause that after 6 months the DAO could revisit the deal and cancel it. No one said anything. I also am not aware of any feedback or communication with Llama that they were underperforming or not keeping up their end of the contract. So again that’s on us - Llama communicates well and gives plenty of avenues for feedback.

Imo, to rug them with 2 months left without any prior communication of poor performance or gross negligence is wrong. The DAO made a bad deal, then didn’t fix it when we had the chance. I think the right thing to do here is honor the contract and learn from our mistakes, not screw a partner bc we screwed up. Does it hurt the bottom line? Yes. Is it the high integrity, right thing to do? I think so.

Just one dead philosopher’s opinion tho, I respect the views of others if they feel differently.

Disclaimer: These are personal views and do not represent those of my employer, yadda yadda etc etc.


My interests are exclusive with the Aave DAO on this topic, and the “bargaining” happening at the moment, replacement options (e.g. TokenLogic), and misc are pretty ridiculous in my opinion.

This should not be a matter of how “cheaper” the DAO can re-negotiate a deal, items “left to complete”, items that have been delivered (which they are a fact, but they were a duty, compensated), future strategy of the DAO, etc.

The scenario is pretty simple: there are reasons to open a vote for termination because an engaged treasury manager has not fulfilled its obligations of something as simple as managing runway.

Does everybody understand how serious is that and how bad to set a precedent that is acceptable? I think all contributors are flexible and will accept the situation, but why all other service providers should just not stop working, with the DAO not fulfilling its obligations to them?


Just my 2ct on this, as I think “rug” is definitely the wrong word for this. If you pay for a service but it’s not provided you probably should stop paying for it.

I can’t really rate past performance holistically, and I don’t agree with some criticism brought up here.

In regards to treasury management though to me feels a bit like neglectance to now be in a situation were providers cannot be payed - especially when you have a treasury monitoring website & pitch treasury management as part of your agreement.

If to much was expected due to big scope is another question, but seems like ppl expected them to do “these kind of things” but they didn’t.

Arguing that curator (as noted before a quite complex contract using cow to swap assets) is on the way is a bit weird imo.

While the curator might be used for swapping assets it’s also important to note that the DAO has more than enough USDC to fulfill the outstanding streams - just on wrong chains(so no curator/swapping or similar is needed, just bridging).


Our proposal to increase USDC on mainnet is up and will be voted on in the next few days. We have informed all relevant service providers about the progress of our proposal over the past few weeks and they were appreciative of the heads up. This temporary issue faced by service providers in the last few weeks will be resolved as soon as this proposal passes.

Appreciate you validating that the curator contract has complexity to execute correctly:

curator (as noted before a quite complex contract using cow to swap assets)

We are building the curator as a sustainable long-term solution for the Aave community. Just for others on the forum to understand: it is not possible to perform swaps directly from the Collector / Treasury contract. After much discussion, BGD and Llama have concluded that it is not viable to upgrade the Collector Contract to facilitate the efficient management of the DAO’s assets. As a result, a series of separate contracts are required. The Curator receives assets from the Collector / Treasury Contract, swaps (or deposits) assets and sends the newly received asset to another address. The Curator is intended for exchanging assets without the need to write a custom implementation every time the DAO needs to swap tokens as has been done previously. A custom implementation not only requires developer resources to code but also requires significant review time, along with additional security risks. Since this is a specific problem experienced by the DAO, we built a general solution to solve it.

the DAO has more than enough USDC to fulfill the outstanding streams - just on wrong chains(so no curator/swapping or similar is needed, just bridging).

As addressed above, the Curator is separate from the USDC payload but it is an important contract we will deliver over the next two months.

Re. bridging: Polygon and Avalanche are the main networks with significant USDC to bridge. Bridging here is done from EOA to EOA. Let’s look at Polygon as an example: a withdrawal of an ERC20 on Polygon happens from the same address across networks. We’d need a smart contract that has the same address across both chains to handle this flow, which might not be possible as the ERC20 implementations are different across networks (thus resulting in different bytecode). More realistically, we’d need to withdraw from the collector to a multisig address and then once the funds are received on mainnet, send from the multisig to the collector. We have always prioritized the safety of funds and keeping funds in the possession of the DAO at all times. This is the first time on Aave that such a bridging would be done. Here is what Polygon said in our conversation regarding bridging:


This proposal has been escalated to Snapshot stage voting starts tomorrow.


Following the governance Snapshot vote, an AIP cancelling @Llamaxyz streams and creating ad-hoc 45 days streams will be created today.

The AIP vote will start tomorrow.

This proposal was an example of an engaged community and decentralization with a very high participation rate from the community.

The ACI would like to thank everyone involved in this vote.


We thank the community for voting for the reduced stream option. It’s unfortunate to have our stream unexpectedly reduced for the last two months, but we look forward to completing our remaining deliverables for Aave.

The reduced stream option should be 50% of our original contract payment as stated above and as the community voted for. This is equivalent to $55k in USD-denominated stablecoins and 283 AAVE. Canceling the original stream and creating a new one can be done in a single payload.

There are 62 days between today (Aug 3) and the end of our contract stream (Oct 4). After accounting for the days involved in AIP 1-day delay, voting, queueing, and execution, there should be 5 days left. Therefore, the remaining stream should be for 57 days.

Here is the reduced stream calculation:

Stream details can be verified here:

  • aUSDC - call getStream with streamId 100003
  • AAVE - call getStream with streamId 100001

Please let us know if there are any questions.


This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.