Proposal: Not missing the AI train

Proposals
1

Proposal: Not missing the AI train

Non-Constitutional

Context

This proposal is put forward by RnDAO as a critical need we have identified and not as a service provider bid. We don’t have a direct financial interest. As such, we invite the DAO to rally around this cause and offer not just feedback but also solutions and leadership.

Abstract

This proposal aims to make Arbitrum an ideal home for onchain AI Agents. We’ll set up an Agile team to identify, build, and communicate key enablers (developer tooling, documentation, and other resources as needed). This team can act as a key point of contact for AI builders, continuously identifying and rapidly delivering on what’s needed, thanks to continuous prioritisation and agile development (fast sprints).

Motivation

AI Agents are reshaping the internet as we know it. All kinds of interfaces, workflows, and processes are being disintermediated and a significant portion of the tools and ways of working we know will soon be replaced by swarms of AI agents coordinating. Said coordination can best be enabled onchain, and the race for where said coordination will happen is on. Blockchain ecosystems compete not only among themselves but also with Microsoft, OpenAI, etc.

If Arbitrum is going to have a chance, we need to act fast.

Approach

Speed should not be an excuse for doing things haphazardly and wasting money building the wrong thing. Onchain AI is a rapidly evolving space, and what’s needed will change frequently. A large program with many fixed milestones and detailed deliverables is likely to take a long time to negotiate and ultimately miss the mark as the landscape changes. So instead of committing upfront to a large build (waterfall approach), we’re proposing to set up a team that can work on fast sprints, experiment, and iterate (Agile approach).

To compensate for the lack of upfront visibility, we’re also proposing an oversight structure to ensure the team performs and remains Arbitrum-aligned.

We already have fund deployment systems like Questbook track for dev tooling. However, this system relies on builders proposing to create their own infra instead of building a project. And although service providers cover some gaps, there’s no systematic way to gather insights about what’s most needed. Also, the Questbook system is not currently creating RFPs and as such has to wait for proposals by 3rd parties. As such, this program is designed to work together with Questbook track managers (especially Dev Tooling but also making warm introductions to other tracks and Arbitrum entities as needed).

In essence, we’re proposing a public interface for AI builders where we’ll have the team at hand to deliver what’s needed and coordinate with other entities to avoid duplication of efforts and still mobilise all the resources at hand effectively.

Specifications

The base structure is that of a delivery team, a multisig for small expenses, and the MSS directly paying contributors and replenishing the expenses multisig as needed.

Team composition:

The exact mix of team members is not fixed, but rather used to provide a starting point for the initiative. The Program Manager will make hiring decisions and can restructure the team as needed.

  • Program Manager: in charge of managing the initiative, ensuring the team is functioning and has the right people involved and processes in place. In charge of alignment with other Arbitrum programs and organisations and reporting to the DAO. Also, get their hands dirty as needed.
  • Comms and Community Manager (DevRel): the first point of contact for AI builders. In charge of writing documentation, engaging with AI builders, promoting the program in coordinating with the AF, and making introductions for builders to other Arbitrum programs/stakeholders.
  • Technical Lead: scopes technical tasks and RFPs (to be funded internally or in coordination with Questbook track), oversees the delivery of dev tooling and technical infra needed. Also gets their hands dirty.
  • Additional team members: selected by the Program Manager (Expected 1-2 devs to support technical lead).

The program manager is elected via a DAO vote (snapshot vote with majority) and can be replaced via the same mechanism.

Expenses Multisig

With the Program manager, technical lead and DevRel as signers (⅔). The multisig is meant to reimburse small expenses such as SaaS subscriptions or conference travel if needed but can also be used for additional costs such as design and other unplanned needs.

Reporting and KPIs

Reporting requirement for the program manager to the DAO:

  • Disclose a detailed expense breakdown monthly
  • Maintain an up-to-date (updated at least bi-weekly) public kanban with work items undertaken by the team.
  • Monthly progress report, including achievements, challenges identifies, and priorities.
  • Make themselves available to answer delegates’ questions

KPIs:

  • TBD: which ones do you think it should have?

What will the team work on?

There’s A LOT of work to do, but as a starting point this is a good inspiration: https://www.coinbase.com/en-gb/developer-platform/discover/launches/introducing-agentkit

Budget

500k USD

Funds Management

  • The Arbiturm Foundation will receive the funds, convert to stables and send to the MSS. Any remaining Arb will be returned to the DAO.
  • Monthly payments to contributors are paid directly by the MSS.
  • Expenses multisig top up with up to $10k per month.

Steps to Implement

Rough timeline: with parallel election and onchain vote to accelerate the launch and avoid missing the AI train

  1. Proposal to gather sentiment and complete the team Feb 7 - 21

  2. Snapshot for sentiment check Feb 21-28

  3. Election: candidate submissions for program manager: March 3rd to 9th
    Program Manager candidates are encouraged but not required to disclose their team (DevRel and Tech Lead)

  4. Election: candidate election vote: March 10 -17th

  5. Onchain Vote: March 3rd to 25th

  6. Contracting: March 26th to April 12th

  7. Program kick-off: April 13th

4 Likes
3

Hey Daniel, thanks for your enthusiasm in keeping Arbitrum at the forefront of emerging trends. I have a few questions to help clarify the vision behind this proposal and ensure we’re aligning our strategic efforts with proven demand.

In the proposal, you mention the concept of a “public interface for AI builders”. What might that look like in practice? Is it a hub where AI agents could feed information from? ie. somewhere developers can easily access the tools, documentation, and resources they need to build and deploy AI agents on-chain or do you have another model in mind?

It would also be helpful to hear about some examples of AI-related projects that could benefit today from something like this being built. This could provide a clearer picture of how a dedicated team might add value/help these AI builders/projects make their work easier.

Additionally, the proposal requests a $500K budget, with $372K allocated to team salaries, but there are no specific KPIs outlined yet. I understand that an agile approach is valuable in fast-evolving fields, but given the size of the proposal, it would be nice to see more details on how progress will be measured and how the proposal will deliver tangible results, before it moves to snapshot/temp-check.

Finally, I recall that the foundation previously had a $1M grant fund aimed at “powering AI innovation on Arbitrum”. Do we have any updates on that program? Is it still active, and are there any exciting projects emerging from it that might relate to or benefit this proposal?

1 Like
4

@danielo Will builders also be able to apply with their AI project ideas and get funding from this program?

1 Like
5

There’s a huge opportunity for Arbitrum here. Moving fast with an agile approach makes sense.

Potential projects I’d be really exited about:

  • AI agents with the end goal of buying ARB, similar to how a16z trades memes with its own token—but instead, focusing on real investments to ultimately buy back ARB.
  • Start using AI to assess proposals and run a pilot—I want to see the potential of unbiased governance.

A budget of 500K or any amount can’t be properly evaluated at this stage, as there are still many unknowns. But I’m excited about this one!

3 Likes
6

I imagine that primarily as a point of contact. Having a human to talk to is critical to build trust. That could look like a telegram chat with the DevRel person there doing Q&A (supported by an AI to avoid repetitive questions like TogetherCrew has built for example).

There should likely be some sort of docs/wiki, and where that is built and how it looks like would be decided by the team in collaboration with AF and others. The key here is the manpower to do it. I’m specifically not suggesting we fund some documentation hub because there are already interfaces and maybe the easiest is adding a new page in the existing website or so. But all that requires a lot of user journey work and is the type of things the team would be doing anyway.

I’m thinking any team building something like this. In RnDAO we have 2-3 (Harmonica and TogetherCrew) that could have benefited form this and I’m sure there are many others. Defining the user persona is part of the work the proposed team would need to do. And as mentioned, RnDAO is not benefiting from this so we don’t have the bandwidth to to this work, we’re relying on delegates to help source other examples and co-build the proposal.

agreed, please provide suggestions :slight_smile: as mentioned, this proposal needs to be cocreated

the program is still active as fas as I know. The difference is those grants are about funding AI builders but we’re lacking the DevRel (infra, tools, documentation, customer support) component of supporting those teams. Deploying cash alone rarely works, we need to make Arbitrum hospitable to them

7

no. There are multiple grants programs already for that.

2 Likes
8

what do you mean by this? not sure I understand

1 Like
9

Hello, we appreciate the proposal.

We are not sure we understand, will the team only be responsible for providing information, or will they assist the projects in other ways?

Do you know if there are any interested candidates for the team positions you mentioned?

We see that the budget calculation is based on six months of work. Was this timeframe chosen for any specific reason?

1 Like
10

the team will do everything they can to support builders. That includes creating, documentation, developer tooling, infra, etc. Hence why contemplating up to 3 devs focused here

feels like enough to give it a try and then check if renewal/continuation makes sense. Maybe 12 is better with a snapshot vote at 6 months to check for continuation. This way the team can focus on delivering and less on politics during their second quarter…, but it’s very hard to pass stuff… so not sure 1mn instead of 500k will make this viable to start

11

imagen
imagen435×270 5.96 KB

I mean this, how am i suppose to say thats a good investment or not with that information?

Edit: And TBD KPI

12

Thanks for putting this up, Danielo.

This proposal feels vague. AI agents are making waves; this we are fully aware of. However, we are unsure of how this proposal benefits the DAO in its current form.

Asides the budget, is it really okay to put together a team before knowing what exactly they would be working on/coordinating? We suggest we start discussions on possible tangents we could take with AI agents before considering teams and funding.

2 Likes
13

Hi @danielo thank you for preparing the proposal. I’m fully on board with the idea of positioning Arbitrum to be ready for the upcoming AI innovations.

I do have one question regarding the $500k budget. Could you clarify the period for which this budget is planned? I’m assuming it’s for 6 months, but I’d appreciate some insight into why this specific timeframe was chosen.

14

In addition, I’d also appreciate more clarity on the success criteria for this initiative. It would be helpful to understand the vision for the team and as you already said KPIs (which I expect will be defined).

1 Like
15

Please see here

as mentioned, this needs to be co-created. Which success criteria do you think it should have?

1 Like
16

That’s the essence of Agile, to be able to dynamically adapt. But to be clear, this came after @AlexLumley shared a whole proposal someone had drafted to build an AI framework. The issue is that it’s an emerging area and things will change and move fast. There’s a lot to work on, starting with creating something like Base’s https://www.coinbase.com/en-gb/developer-platform/discover/launches/introducing-agentkit

1 Like
17

Thanks a lot for your proposal Danielo.

I think it has its merit & I have seen the products in action that you and your team have built.

This being said, given some feedback is around trying to figure out whether there is some space for AI agents in Arbitrum/where this space should be: wouldn’t it make sense for the first SOS to be completed?

This would give this proposal a North Star to attach itself to.

4 Likes
18

If I understood the proposal correctly, the first six months are focused on:

Some ideas for KPIs and success criterias:

  1. Team set-up It would be important to have the right people in the team - as time horizon is quite limited, they have to start executing and delivering immediately.

• KPI: Full team formed and operational within X weeks.
• KPI: First deliverables available in the first X weeks.

  1. Backlog Creation & Prioritization – Clearly ideate and define the infra stack, tooling, and documentation needs. If this is an Agile initiative, then Epics and Stories should be created and prioritized (ideally on super early stage).

• KPI: X Epics/Stories defined and delivered within X timeframe.

  1. Marketing & Awareness – Spreading the word early is key to attracting AI builders and fostering engagement.

• KPI: Number of engagements (e.g., partnerships, community discussions, event participation).

  1. Early Adoption, Usability & Testing – Having initial users onboarded within the first few months to validate the framework.

• KPI: X AI builders actively testing the framework.

  1. Collaborations & Feedback Loop – Establishing partnerships with AI builders or projects that can provide real-time feedback and help refine the tooling.

• KPI: Onboarding strategic AI partner - who use the framework
• KPI: X collaborations initiated and feedback cycles completed.

3 Likes
19

I am not technical but have been following AI developments and it seems there is a lot of frameworks already available. The idea of AI Agents seems to be largely chain-agnostic, with the blockchain layer abstracted away. Have you considered this aspect? It might be worth exploring collaborations with other prominent infrastructure providers, such as Virtuals or the Aliza team, to see how Arbitrum can best position itself in this evolving landscape. Maybe what’s needed isn’t just developer tooling but also a plugin for Arbitrum—something that enables AI agents to have wallets and interact on-chain seamlessly.

I’d love to hear more about the core of this proposal and what kind of deliverables you have in mind. If there’s anyone in this community already working on AI agents, it would be great to get your perspective on this as well.

20

I’m curious to understand how this proposal fits in with existing initiatives by the Arbitrum Foundation, like the Trailblazer AI Grant Program and the broader efforts to onboard builders into the Arbitrum ecosystem.

The Trailblazer AI Grant Program is already allocating $1M to support AI-driven projects on Arbitrum. You’ve answered a question about differences with it saying:

However, there are already efforts in that direction as well. For example, Arbitrum is sponsoring ETHGlobal’s Agentic Ethereum hackathon, which is all about AI agents on Ethereum. This initiative seems to overlap with that—are we duplicating efforts here, or is there a clear distinction in approach?

Also, you mention:

It’s unclear how you plan on these resources be used to onboard AI builders to Arbitrum. Is it via a coordinated effort with the Arbitrum Foundation and other programs like Questbook?

Finally, given the emphasis on avoiding wasted resources and “building the wrong thing,” has there been any analysis of the projects already funded by the Trailblazer program? It seems like that would provide useful insights before launching another initiative.

21

I am a bit puzzled by this proposal. I understand is fishing for feedbacks to make it more concrete but it seems like it still lacks the business plan behind.

You present the idea: don’t miss the AI train. We could probably say that we already did miss the bulk of the first wave but more waves are to come, and advocate how this second wave is currently being build in the same place in which first wave was born (base, solana).
Ok, so the proposal is about “creating a team that can guide AI teams”.

So, the PM and the devrel are the key figures here. But i see an issue.

I don’t want to go over a dao vote to choose a pm for this. We need a person that is highly knowledgeable on the field. We won’t get it through an election we will have to externally source it.

Same for devrel. Especially in a moment in which we don’t have a devrel figure in the foundation, feels utopistic to find a good one for the dao.

I don’t get this, nor I get in general what funds are for. My guess would be: to pay for people in this team and attached costs. But seems a bit too generic.

This is where the proposal falls short imho. Is too vague as goals, and is too vague in the deliverables. I get that you want to crowdsource a big part of it but feels like a meaningful part of the vision should come from you, the proposer, here. I don’t see it yet.

All in all as of now i think this proposal at the current stage does not make sense, not really for the capital but for the mindshare we need to collectively spend on top of this.

Arbitrum is behind in the AI. To even have a remotly chance to catch up, a structure like the one in stylus program in which OCL was able to have the technical lead being highly opinionated on what they think the dev environment might need 1 year from now is probably a more constructive way.
We need someone that has hands in the dirt already, both in AI and in Arbitrum. This figure is the PM but I think we have very low chances to find a figure matching this skillset.

5 Likes