Challenges and Approaches to Ethereum Public Goods Funding

Transcript

Introduction: Increasing Funding for Ethereum

So yeah, this is um it's a new talk. Uh it's something new I'm working on which is in the bare market. How do we actually increase all the funding which is there for Ethereum growth and development? A lot of earlier talks which I've given have focused on how we divide money or how we distribute money. But this is the first one on how we actually increase the amount of money that is there to distribute in the first place. Super relevant. How many people here think that's like relevant to what we are in right now? Yep. More than half the room. That's great. Um so like what is the status quo that we are walking into right now? Um I think we are walking into a Linux Foundation type of situation. So, Linux Foundation has about 300 million in annual revenue

which has been built up over 20 years and four revenue streams. Um, their bulk comes from so if you look at it, they started with a training program and certification and that eventually then led to um events. They make us they make some money from events but the bulk is what they call the platinum membership surge which uh basically you pay $500,000 and you get a board seat at the Linux Foundation and you can influence development. Um I think the key part is that kernel development is driven by corporate contributors and their Linux foundation is just a coordinator or an umbrella and if we don't want to be in a situation where the financial firms are the core contributors and we are just like a coordinating layer we really need

Linux Foundation Comparison & Risks for Ethereum

to change the trajectory that we're headed into now because that's what I think we're walking into unless we make some sort of intervention. Um so like what is the broad areas that we can look at? So there are three layers in the Ethereum ecosystem. There's the application layer, there's the smart contract layer or the protocol layer and then there's the consensus layer. So we should try to have different strategies to target where money can come from for each of these three layers. So we'll look at each of these strategies. the AI for new frontends, cyber resilient tax, some new software licenses we can come up with, dependency graphs and how people can fund dependencies and in protocol coordination. So before I get into each, I'll show you the overview of

The Three Layers of Funding (App, Contract, Protocol)

how it all works. The rule for storytelling is tell them what you're going to say, say it, and then tell them what you said. So this is the part where I tell you what I'm going to say. Um, so imagine that a user submits a $100 transaction on a swap in the world where we actually managed to implement some of these ideas or all of these ideas. Um, they use an RPC front end and that captures some ME fees. So they redirect a portion to a PG fund. We already see this happening with like some protocols like arbitum time boost where they get ME fees routed back to the DAO and the DAO can reinvest in Ethereum in like arbitum growth. Um so that's one level which is the application layer which you

can keep changing applications that's not fixed. What you can't easily change is the smart contract layer because that goes through audits. It's like a fixed back end and it's hard to change. So the smart contract the advantage is you can encode logic into it. So here you could encode logic into the smart contract that x amount should go towards supporting the ecosystem that we ourselves depend on. Um so that's the second layer of like how smart contracts can allocate and the third layer which is the deepest layer is coordination between validators who can coordinate with each other to actually distribute money into a central fund. Um so yeah this is the high level overview it seems simplistic now but I'll get into why each one is challenging and how we can

Overview Example: A $100 Transaction Funding Public Goods

overcome them. So on the protocol layer I'll start from the last which is the deepest layer and the protocol layer. So it makes sense business-wise for validators to actually fund Ethereum growth because if they grow the ecosystem that creates demand for blockspace and with demand for block space ETH value increases because of the burn and the flywheel effects that happen and um and then they can use some of their staking rewards to fund the ecosystem back. So this is like a very clear example of like a flywheel but we haven't actually seen it happening except with Octint like Nico you can put your hand up. Yeah, go Octint. Octant is maybe one of the few people who actually use staking rewards to fund the ecosystem, but we should be seeing a lot more of it because of this flywheel that

Validator Incentives & the Flywheel Effect

makes economic sense for them. A lot of validators I speak to don't care as much about the staking rewards as the principle which they stake. They want the ETH price to go up and they're okay with the rewards being used for um funding Ethereum growth and increasing the value of their underlying principle. But why don't we actually see this flywheel happening? Um my argument or my reasoning reason reasoning for it is the prisoners's dilemma situation where you have a strategy to either fund or to not fund and if you fund but the others don't fund the others still benefit from your funding so there's like an opportunity to defect like octant is great in funding but everyone else is getting all the other validators are also enjoying the benefits which octant

The Free-Rider / Prisoner’s Dilemma Problem

is incurring the cost on so an in protocol solution can be done to solve this free rider problem where if a majority of players agree that we should be funding the ecosystem then all of them have to fund all the people from the staking rewards will go towards funding it so that's kind of how we might solve this prisoners dilemma at the protocol level um so yeah this is like how it would work an implementation of it where all the stakers can like similar to how they put a gas flag saying I'm willing to increase the gas limits by from 45 million to 60 million, they can put a flag up saying, "I'm willing to put 5% of my staking rewards to funding the ecosystem." And if a majority of stakers then put the flag up

to 5%, then all the stakers have to now put 5% into the fund. So that's called signaling amount. So you're using the exact same thing. You're just allowing them to signal how much of their rewards they want to fund. And you're preventing free riding because the remaining 49% also have to fund if 51% say yes. So it's voluntary taxation, but there's no free riding. that's possible in it. Um how do we distribute it? We won't get too much because I said this talk is not about distribution but there is a mechanism we've thought of which relies on condor set voting where you have a splitter contract and then that has different allocations and people can propose a new splitter contract but I won't but like we have a session tomorrow I think 1 to 3 yeah at the turtle room where we'll be discussing

In-Protocol Funding Signaling (Voluntary Taxation Idea)

all of these ideas in more detail. So this is like how we might solve the free rider problem at the protocol level. We can imagine how L2s might be able to coordinate funding with each other and prevent a free rider and they can invest in L2 and blob research with validators. We can increase we can imagine how they can coordinate at the protocol layer and fund uh new Ethereum development to increase demand for block space. So we can think of a lot of futures where these stakeholders can collaborate at the protocol level. The disadvantage is it requires an EIP and that part can be a very uh rigorous process. So it has to be well thought out and you have to get it right on the first try. There's no fail fast or like you know make it wrong. You have to get it right on the

first try with the IPs. Um so this is slightly easier but still diff like still difficult. We're coming down the like difficulty stack. The hardest is protocol changes which we showed one example. The second is smart contract and force dependency funding. So on the left you see um on uh you see like Helios is an example of a repo which Ethereum depends on with A16Z and MEV boost relay by flashbots creates a lot of blocks. So all of them have a lot of dependencies. So here you can see that CPU ID is a dependency of me boost open telemetry is and this creates different weights like you can bet on the value of dependencies and then uh you actually get like you know all of the

Funding Smart-Contract Dependencies

dependencies of a smart contract and you also know the weights of each smart contract like of each dependency. So any interaction with a smart contract a portion of it can automatically automatically go to all its dependencies. So we have this already figured out and sorted out. It's a project called deep funding. You can build AI models to predict the values of this. Uh that market is live on deep.ca.pm. And if you're an expert in as if you're an expert developer, you can also give uh jury data. You can actually help resolve these markets by giving spot checks of each of these and determine these values and what profit and loss they make. I won't get into it now, but this is an allocation mechanism. But I think the larger point is we have a

method of giving you a smart contract, getting all of the dependencies a smart contract relies on and then figuring out how important each dependency is so that we just don't give equal value which I think is like communist and not right. I think like based on importance you should be funding not on equality. Um so yeah this is like a part of how you could do it at the smart contract level. Um I'll come to this slide a bit later and first talk about why would they do it in the first place? Like why would they voluntarily give up some of their revenue in order to fund their dependencies? Why wouldn't why wouldn't they just free ride? So this is something which Ava Black who's working on the EU cyber resilience act she came up with this great diagram where um so

Deep Funding & Dependency Graph Markets

the cyber resilience act forces every product owner to report bugs in their dependencies to the maintainers. they are basically liable for any vulnerabilities and dependencies that they use as a product owner. So that actually creates an opportunity for this sort of funding loop to happen where I have to comply with the cyber resilience act and that I anyways have to spend money with like for that and that can go by funding the dependencies to make sure that they stay up to date and they don't have vulnerability and then uh that increases security of my product. So I think there's this loop we can create which is opened up by the cyber resilience act and especially if we see a lot of financial firms start using

Cyber Resilience Act & Security-Driven Funding

Ethereum and if they are in Europe they would have to comply with this act starting in 2027. So we have some time to get the leg work up for this. The other part of compliance which Raymond Cheng is working on and we'll have a session with him tomorrow at 1 from 1 to 3 is open-source licenses that require a compliance is like product like is like some form of revenue sharing that if you're earning revenue by using our dependency you have to also give a portion of that revenue to us and like what can such a license look like while still maintaining the ethos of open source software in its principle. So these are some examples of a loop of why when people would want to fund them. Um this I'm going to go on a slight sidet

Revenue-Sharing Open-Source Licenses

track because we spent a lot of money to get this insight but it's a really good insight which is that uh if you have any funding round let's imagine you have a committee of five people that are deciding between 30 projects to fund. So the usual way you do it is those five members of the committee sit they deliberate they maybe take a month in deliberations and they come out with recommendations saying yes like you know we are giving this much money to project A this much to B this much to C and you as the funer have the option of either accepting the committee recommendations sending it back to them for revision or rejecting them and doing a new committee. Those are basically the three options you have. But we figured out that you can actually do it in like you can have a more efficient way of doing

Predictive Funding Committees & AI Markets

it where you have five people. You're not changing any of the processes. You have five members of a committee. Let them go through their process over a month deciding which project gets how much money but simultaneously start a competition where people can predict how much money each project is going to get. So you have like an AI competition like on Kaggle where they can say that I think project A is going to get $20, project B is going to get $15. Then after a month the committee actually comes out with its recommendation saying project A gets $22. So now you can actually determine who the winning model is. The winning model is the one which is closest to the committee recommendation. So the best part is I don't even need the permission of the

committee. I can just like uh like folk their uh like fork the eligibility list, create like a market where you can bet on how much money a project will get based on the committee recommendations and I can start a competition where AI developers can submit models giving breakdowns between projects. And the cool part of this like you know it's like you're not changing the process. You're not asking funders to to use a new process but what you are doing is giving funders three lists instead of one. So normally like jury votes is in the center. That is all we normally do in a committee process. But if we use this approach, you get the juror votes.

Using Markets to Audit Grant Decisions

You also get the market which is betting on what the juror votes are going to be. So the market makes money based on the difference between the market prices and the jury decision. So there's an arbitrage opportunity if the market prices are higher, but you think the jury votes are going to be more closer to what your model is predicting. And then you have the winning model which is anyone can submit a model predicting the breakdown between projects uh and then you can just take the output of the winning model. So the advantage is that there's actually competition here. The committee has to be on top of its game because they know that if they give bad recommendations the person who's giving them money can just like opt for one of these other two lists if they want which is a nice feature. Competition is always

good. And uh the other nice feature is that uh if a member of the committee is corrupt that he is giving the money to one of his friends or his project members or anything you can actually question them because you you can say that the market or the winning model is predicting that this project should only get this much but there's a huge discrepancy where the committee is giving them way more than what the market is saying that they should have got. So why is that discrepancy there? So you can ask questions and maybe they'll have a valid answer for it, but you at least have a basis to question the committee decision which normally you don't have the context or the time to do that. So you have alternate numbers. So this is like a general call to action. Anyone who's running funding, how many people run funding rounds here?

App Layer Opportunities & AI Front-Ends

Um, okay. Not bad. So there's like four or five people who have. So any funding round you're doing, just like put this as an add-on uh as an add-on feature. And the best part is you don't even like we don't need to ask you for permission. We just need to get the list of projects that are competing in the round and then we can do it all on our own where we can start the market and all of that. So we're not asking you to change anything at all. Um so yeah this is the part about the smart contract and force funding and how you can get weights between them. Um so maybe the last part now reaching the end of the talk um is the app layer which is the easiest in terms of actually going to market and actually trying something. Um so I think

the famous quote was software is eating the world and just recently Mark Andre said that uh it's in all probability software all software is going to be open- source and all value is going to be flowing to chips and to like physical bits and atoms. Um which is funny coming from the person who predicted software who predicted right at that time that software is eating the world but now he's saying that like bits and atoms are going to be the one which capture the value and all the software is going to be open source and a large part of that reason is AI and um like just yesterday there was someone who vipcoded an entire Ethereum execution client. Isn't that just like insane? Like they voded an

Why Blockchain Wins in the AI Era (Open Backends)

Ethereum execution client and it actually works. So I think we're going to see major boosts in terms of that and that creates an opportunity for the blockchain community in particular. When like anytime a lot of developers ask me why should I build on blockchain my usual answer is it allows for separation of the back end and the front end. So in a regular web 2 app you are the only the company is the only one with access to the back end. So in Twitter or X only they have access to the back end of posts coming in and so they have access they have a monopoly on the front end whereas with decentralized applications a might be the smart contract but anyone can build a front end which uses a so and I think that is actually going to

Public-Goods Fee-Redirecting Interfaces

give a competitive advantage to blockchain in the AI era because with AI anyone can easily spin up a new front end and then that gives an opportunity to the public goods world as well where we can create or spin up a front end which actually lets you um say that okay all like a percentage of fees are going to go towards a public goods fund or to support the ecosystem that that this app depends on um so I think we can spin up multiple frontends from back end so like let's spin up a front end for a which allows us to do it let's spin up a front end for unis swap which allows us to do it and I think there's a real advantage here that we should be taking care that we should be doing in the AI era and maybe the Last part is uh

Frontier Research Contractors Model

again like another colleague of my team. All of this a lot of these ideas have just been from colleagues in my team like these are like Rahul's idea which is around frontier research contractors where um there's a gap in the ecosystem between startups which want a strong product focus and large research organizations which think in decade timelines. Um so I think there's a gap in the ecosystem. So if you look at how we get concrete funding to organizations like say L2B today L2B is a good example where they are a research organization which researches decentralization in rollups and they are like they're not exactly a startup they return the money to investors which is amazing. They give back the money saying we want to just remain as a public good. So they are but

like so how do they get money today? So they get about like half of it through grants and half of it through consulting with companies which are deciding which rollup stack to build on. What level of decentralization do they require? So that's maybe an example of being a frontier research contractor where you're in the trenches with the startups that have a strong product focus and you're still a large research organization which ultimately does require empirical data. And this is maybe the last thing I'll say before questions but um it's like Emanuel Kant said that um concepts without percepts are empty while percepts without concepts are blind. So if you just have a concept but you have no experience of it on the ground that's an empty set and

Closing Thoughts & Philosophy of Research

if you just have experience but you have no larger idea then you're operating blind. So I think this is a good way for public goods to immediately access funding which is work in the trenches with startups, get money to deliver on their goals and you can use that to feed into your larger research as well. Um so yeah that's my talk.