 move on to our next speaker, Axel Cortes-Cubaro, who is a research scientist at CryptoEcon Lab with a background in statistical physics. And he'll be speaking to us today about a gas model for a post-scarcity era. Take it away, Axel. Right. So, okay, so I want to make it. Okay. Right. So, my goal here is to introduce you a problem that maybe you don't know you have and then not give you the solution to that problem. But I just want to, like, introduce the possible space of solutions for this kind of problem, which is, in short, maybe there's a lot of people focusing on how do we scale the blockchain, make it more powerful, more powerful. But basically, to understand what could be the problems associated with that or if we want to think, maybe not just scale as much as we can, but scale as much as we should. That's the question. So, I'll just talk a bit about gas in general. So, yeah, just very briefly. So, gas is what I mean by gas is just like a unit of computational effort in the blockchain. So, all these transactions, all the nodes need to validate that it takes effort, everything that is included in the blockchain. So, it is the amount of gas that you could have in every block that you could use in every block is precious. It's scarce. So, you cannot put all you want in there because everyone needs to validate this. So, the point is, it's scarce, so it should be optimally used, utilized in a sense. So, the idea is that there should be an optimal block size. So, this is like the amount of gas that we can all fit in a block and we can handle it. Anything more than that and it's too much for the whole network to process. Anything less than that and the network is being utilized and that's not efficient. So, it should be used better. So, the idea is that you should price. So, it should be used optimally and the way that it's done is basically it should be priced at a market clearing price. So, the idea is that the amount of what you should pay for a unit of gas should be the optimal value such that only the amount of transactions that fit in the block size that you're aiming for can afford this price. So, anything else would be priced out. So, the idea is that you want to select only the most valuable transactions to be included on the block. So, you move the price until you reach that price that only this exact block size makes it through the market clearing price. So, this is in a normal like limited blockchain where you have scarce gas like only this amount of block size available. So, the next question. So, how do we kind of reach this market clearing price? There's different approaches to this. So, there's like the original Bitcoin approach is just a first price auction which basically amounts to like everyone just makes a guess about what is the market clearing price. So, you submit your transaction and you make a guess. This is what I think this is worth and everyone out there makes a guess and those with the higher guesses get through. That's a mechanism that has worked. Okay, but the idea is that this tends to be very volatile and people aren't great at estimating or so you might end up overpaying and the prices can fluctuate more. So, the approach we're using in Falcon and Ethereum as well is called the EIP 1559 where the idea is that as you price gas not just by the amount everyone voluntarily guesses but you have a base fee which is set by the network. So, this is posted, this is the base amount of fee that you need to pay for your gas and you can put something on top which is meant to be just like to cover the marginal costs for the miner to include your transaction. And then this base fee the idea is that this is now not paid to the miner but this is burned and this evolves in some like definite algorithmic way with the point being that this is adjusted based on demand on the network. So, in this case you allow the block size to be a little bit flexible but you set a target block size so this is the optimal block size that I want and if in a given block I spent too much gas then I raise the price such that the next block will be closer to where I want it to be. So, this adjust the base fee such that the idea is that this mechanism should equilibrate close enough to the appropriate market clear place because these are different options. So, this is what we are using. So, one point is that this price doesn't come from this mechanism. So, one of these shouldn't this EIP 1559 will not make your transactions cheaper or anything where this price comes from is just from supply and demand. So, these are two mechanisms just to try to find out what the price is but they are not what is used to determine the price. So, the price is simply just supply. So, in a fixed blockchain like we are talking about with a fixed with a target block size supply is fixed. So, this is the amount of gas I have for block then if more people want to use this if there's more demand that will lead to a change in the price. So, this is basically the kind of a symbolic equation showing the dynamics of if there's a change in demand if demand increases then I should increase the market clearing price such that I remain at my target block size and if it's demand reduces I should reduce the price. So, one question that we think about right as script economists that maybe many engineers think less about is like do we actually want to decrease the market clearing price. So, this is a big complaint okay gas fees are too expensive and so on. So, we could tell the engineers okay make it scale everything make it better make the blocks larger so we can feed more stuff and that should give us a lower market clearing price right and and that's something like CX describe is kind of what happened with this hyperdrive that suddenly we have this boom hour now we can process 25x times the amount of gas in the block and the question is this something we should do is this a good deal right and then we need to think about what is a good deal right and the reason why you you might think I rest so that makes sense but let's make it everything more scalable we can include more stuff but but there's something you're losing by by this collaborative which is you are reducing your your market clearing price right so miners were complaining oh why did you make it so scalable now I'm making less money is the idea so yes I want to think a bit about so I want you to think about this question of what is what is a good deal for the network should I always scale or what should I do so one one strictly economic way to to think about this question is just to think about the the total network revenue right so the idea is that if I increase the supply the the market clearing price will be reduced so this is bad right people are making less money but the idea is that now more transactions can get processed such that sure each one will cost less but but there's more being processed so that's a net good right so so you can think of it so what is the there's this total network revenue it's just like what is the the current price times the amount of transactions that that you could process right so so if I scale and this gets lowered but this gets increased enough that overall this is this increases and that's what I would call a good deal economically okay so for instance this is what didn't happen right away in this hyperdrive that CX talks about so you scaled everything the base fees went down and the demand to reach that level that we needed the 25x is still working to reach that level so that's what could be considered a bad deal in this sense you might you might question like is this are we being superficial here that we're just like a good deal is just make more money but for now we'll leave it okay so this this is something to think about but that's that's one goal we could use to to work on but I think it's not that superficial in the sense that where so what we want is we want to have a like a healthy and very useful and active network and that involves miners being motivated right so so if the rewards if the revenue is good and rewards are good that the miners will be there will be more more miners or storage providers in our case in in in falcon joining and they need to be properly motivating to bring more miners and make it a good network so so so the general principle that I guide myself with is so the supply that we have that we provide by all the scaling solutions should be just appropriate for the level of demand of the network not like much higher than the demand record yes so just a little summary here of kind of a typical scaling solution that that we can see in the blockchain space but it's just like making a a roll up or a side net on the on the so so the idea is that as if you have your main net your falcon or your Ethereum or whatever and this is too expensive to use you can like make your own little side net where you run run your your your sub transactions in there so this is what happens like at an exchange or something so okay so so this exchange is going to make all these transactions and then only at the end report the the final result of that to the main net right and this allows you to do like more transactions maybe at a faster scale you can do more things without reporting each one of those without using gas on the main net for each one of those so you only spend it here and then report report only the result after after that and so that has a result so that that so having many side nets like this could take some some demand away from the main net right because these are all transactions that could have been performed on the main net but now they don't need to be performed there so that's something to to think about but one point of this is that right so the side net that you're running your sub transactions in should have a lower market hearing price that the main net if not like users would just go back to the original net but the tradeoff is that the main net is expected to be like the with the higher security and a better reputation right so that's why you want to report back your your final results to the main net and make them more permanent there because this is the idea so the subnets should be cheaper but less security less reputation and so on so that brings me to this is a this is a project that we're working with is mainly consensus labs of a project that is a solution we are working in in file point that this could be what scalability looks like for Falcon which is what they call hierarchical consensus and the idea is similar to to to this stuff about about side side nets that I talk about but just more here I could go I guess so so the idea so here you have your main net which I guess this represents time or something but anyway so here you have Filecoin let's say and then I have more demand that I can handle here so I spawn a subnet here so you can spawn a subnet and here I process transactions fast and then I am checkpointing total results back to the main net but not everything that happens here so similar to the to the side net roll up kind of thing and so here this one has two subnets but then the idea is that when there's too much demand here also these ones can spawn their own subnets and then these ones again could spawn their own subnets if they want and so on so so this this idea of of hierarchy of of nets that how they are connected is basically so this this one runs their own the sub transactions and then periodically they post their checkpoints back to the to their parent net right so so this interaction subnets checkpoint back with their parents so the subnets remove some demand from the parent but then they also ask for something because they need to spend some gas on checkpointing with the so there are checkpointing gas fees when they checkpoint with the parents and the right so the idea is also that the the parent chain should provide some level of security and reputation through to the subnets by keeping track of their checkpoints right so they're saying oh this this there is some more permanence to these transactions and also there will be collaterals that can be slashed to start running this net so so so you guys okay so I want to start a subnet if I want to start an official subnet with this parent I need to freeze some some amount of collateral that may be slashed by the parent if I do something bad that's the idea of hierarchy sorry right right so the idea is so now the the hierarchy is is limitless in principle so I can give spawning more subnets whenever I needed there's no supply problem anymore all scalability is solved right so that's the idea of this so I grow the the hierarchy grows as needed and then right so if I'm a parent chain that have my own subnets I can kind of pull on them if I feel so I have some controls over my children nets by their checkpointing fees that I charge them and by the collaterals so I say if you want to start a subnet under me you have to pay a higher collateral so if these fees are high it could discourage the creation and survival of such subnets right so so I make it things too expensive for the subnets to to survive or if I want to spawn a lot of new subnets to take care of all this demand then I make these things cheap and they and they can join and then the idea is that while the whole hierarchy is limitless there's no there's no scarcity anymore each of the nets in the hierarchy is a finite net that each each of them should be utilized optimally and then kind of the guiding principle we want to use is that the overall total network revenue of the whole hierarchy should be maximum so growing this hierarchy should be in the name of it being better for the whole hierarchy so that's the idea right so so we can think of the model here what what we can do about this so this is the traditional model I I told you about about like a single chain that more demand brings higher prices and this is kind of how how this works in 1559 that this is the market clearing price is adjusted by depending if the block size was bigger than the target block size and I increase the price and so on so this is so this is this dynamic for the this simple case and what we're looking at now looks more something like this so this is for a given chain for a given net in the hierarchy so these are kind of the the knobs that you have here you have your your base fee your normal gas fees that you charge for transactions you have your checkpoint in fees and you have your collateral and you should adjust this this fees based on what's happening with the hierarchy so this is the data that you read to to make your choice here how to adjust this right so this is kind of the solution space I was talking about right so right so so each of these nets must keep their they have an individual goal of like keeping the optimal block size for myself for that one given net as there can be a eip 1559 like mechanism but then also they should have in mind like what's happening with the whole network so it's a collective goal so okay the total network revenue is increasing and if that happened because I see that more so this is the total amount of gas being spent in the network if more gas was was burned in the network and that was good for the revenue then I want to incentivize that more so I would make this fees checkpointing and collateral lower to invite more growth and if growth was bad then I want to pull back so this is kind of the idea and so you can so this is the main picture here and and this is not the main the solution because we still I'm not showing you like an exact formula like this how to play but there's many options of what we could do that will look like this but this is should come from something like this so so this is my final answer so how large should the hierarchy grow so it should be large enough and maybe a bit larger to maybe induce a bit more demand but it shouldn't be larger than it needs to be that's about all great thanks for that Axel and I think we have time for a couple of questions actually to the speaker um so we'll do this Oprah style if you raise your hand I will come at you with the mic or I will lead with my own question uh you mentioned that subnets inherent reputation and security features from the main net are is the reputation and security features of the sudden subnets are they strictly dependent on the main net is there any way for subnets to compete on features other than price if you start a subnet without a reason so but what I mean by this borrowing of reputation is just that you check points so you have a level of permanence that you know okay so the users of this stuff that knows there's something more permanent happening in the parent net and also that that you that you know this subnet paid some collateral that can be slashed if they misbehave so you know they're motivated to behave uh it could also become the subnet could become more valuable than that by providing a valuable service so you're also that subnet becomes a good subnet that is doing something good people want to use it so this is kind of a minimum value that you know that the user knows the subnet will have but the subnet can grow even above these values there is the idea fantastic thanks again Axel one more question over here I'm thinking about like the collective goal of maximizing the revenue of the network and wondering if there's any ambiguity and like how to measure that either what the elements of the network are not just miners or computers but humans and like how you measure revenue is it only in filecoin or are there other currencies right so this this this I guess this comes back to the question of the superficial goal or not so uh so the the idea here is kind of I'm assuming so right so the different subnets can actually do different tasks right so let's say we have filecoin that is about the storage and then right so you you you make a subnet that is focused on retrieval market and you're incentivizing people who retrieve files and so on they're doing different things but they have kind of what joined them what what we can measure them equally with is the amount of gas that they spent versus uh so it's just like on the computational power required for for each of these things so we are assuming that the computational power of all of them can be measured and compared and uh yeah so this is one simple goal which is uh okay so this is my definition but now kind of so now in the hierarchy version this would be a sum of disguise over all the subnets in the hierarchy so what's the market clearing price of each of them and the block size of each of them uh yeah so so this is again like a very cold economic target but it's one to to work with it could be yeah it's it's it's also open to maybe this is not the right target