Superfluid Reactor Summit: Managing Subscriptions in Web3

[Applause] [Music] so without further ado let's uh welcome our very first official talk which is tony from diagonal and tony's going to be talking about how do we actually successfully manage subscriptions in web3 so uh tony please welcome uh well welcome page and i'll let you take it from here hey carter how you doing today i am doing great all righty then let me kick you off um okay share my screen please just confirm we're all live okay uh already guys so my name is tony rosler i'm a co-founder of diagonal finance and today i'm going to talk to you guys about web free subscriptions so the agenda of the talk today is three-fold first we're going to look at why on-chain subscriptions are a hard problem and why largely we haven't seen much in the space kind of largely up until now secondly we're going to talk about the benefits of streaming subscriptions and lastly we will talk about why kryptonite payment rails are important alrighty so let's just kind of set the stage so first we're all familiar with webster payments and house experiences work with webto payments the way they work are essentially the bank pulls let's say ten dollars a month out of your account once a month now this works okay because you trust the bank to put money out of your account and if the bank acts maliciously then your account is ftic insured and you can always claim the money back now the issue is a little bit that if you try and move this model to blockchains you kind of run into some problems so the first is there is no native way to pull more tokens out of users accounts uh simply on the execution layer uh in blockchains now this is largely by because by construction blockchains are push-based payment systems where every change in state requires a transaction and every transaction requires an elliptical curve digital signature from the user's point of view so what that means in a very naive way if a user wants to submit a 10 a month payment they'll have to manually sign a transaction and kind of submit this payment however that's not really the full story because we can build kind of certain approval based applications on top of the execution layer um however these payment systems have certain flaws so all of these kind of pull-based payment systems largely boil down to this mechanism where a user will infinitely approve a contract and then a bot will on a kind of schedule schedule transaction execute transaction pulling money out of the user's account and then paying the merchant at the end of the month now the problems with these payment systems are firstly that they have trust assumptions so when you are infinitely approving a smart contract you are giving that smart contract the opportunity on the right to essentially empty your entire account for that token now we have seen time and time again smart contracts uh have these bugs when it comes to uh infinite approval and there have been plenty of exploits around this area and now the kind of important thing from a user's point of view is my wallet is not fdic insured so if the smart contract or the bot x unit transaction acts maliciously there's no way for me to get my funds back now this is a real problem the second point is counter counterparty risk so because payments are essentially enforced by external actors i.e the bots has to make a transaction once a month to execute that payment if the bot fails to execute our transaction you will not be paying the merchant and essentially the merchant will cancel your subscription and you will not have access to that service now this is an issue because suppose the bot just disappears then yeah there's no way for you to essentially contact our boat or sue that bot because there are in kind of ethereum and smart contracts agreements are enforced by code not by courts lastly an interesting point is kind of the lack of scalability so having kind of these scheduled transactions requires end payments to make end transactions so the question kind of becomes so who first of all pays for the gas and then what happens when the cost of execution of the transaction is greater than the fees earned from the bot for executing a transaction so should the fees kind of should the cost be pushed to the merchant side as in cannabis should be cut into their profit margin should the user pay the fees so it becomes more expensive than it would otherwise or should the service provider the the payments friday pay the fees now okay this is kind of there are many different ways you can go here but if we perhaps look at some numbers so this is some this is based on analysis i did in last february granted things may have changed but i looked at ustc and kind of the average uh on a 30 day kind of average what was the average erc20 transfer on the cost in some gst and arbitrary we have a dollar same for optimism and main house with v10 actually this is way more expensive than it would be in traditional fiat payment rails so if we kind of summarize the problem one thing it's the beach problem what would happen to your subscription if your payments friday went to the beach and essentially your subscription would cancel so this is not a decentralized famous rider and this is not um every kind of ecosystem we want to be building so right then so how does streaming solve this so first of all streaming has zero trust assumptions so the user pushes money continuously from their accounts to the merchant's accounts there is there are not they are not approving anyone money is pushed continuously secondly there is zero counterparty risk because payments are settled continuously so this is incredibly important for businesses where essentially they provide a service to the user throughout the month and then expect to be paid by end of month a good example of this is let's say an rpc provider like infuria alchemy where you are allowed to make rpc calls and at the end of the month they send you a bill and they expect to charge for that bill now these merchants have a massive problem which is essentially that users can rack up huge bills that they are not able to pay for and there's really no clean way to mitigate this because there's you cannot settle continuously you cannot settle every time you basically give value so with streaming you can do that which is quite beautiful where essentially every every single second the payment is settled and money's leaving your accountant in your account so this reduces the risk for the merchant dramatically lastly scalability is an incredibly important way so you only have to submit a single on-chain transaction to open a subscription and a single on-chain transaction to close a subscription and lastly something which is very cool is uh and also important is capital efficiency so merchants are paid every second which essentially means they can always put their money to work always be earning yield um and just dramatically increases their capital efficiency all right then well let's also kind of think a little bit more broadly so why is streaming not only better on those verticals we touch but why is it logically better in general so let's consider netflix so as we know netflix has been in use a lot recently for essentially they've been struggling in terms of their revenue in that business model so if we look and if we and this is not all that surprising really if we look at netflix's business model so netflix charges 10 a month now this is a bad basis model if we because essentially you have a cohort of people who have a higher willingness to pay and ten dollars a month and these people are getting a bargain where they are paying less than they're willing to pay you also have a code of people who are not willing to pay 10 a month they're willing to pay less than that but they're being pressed out of a netflix subscription so this means that netflix is just capturing a very small area under the demand curve and leaving a lot of money on the table now if you introduce streaming uh something very very interesting happens so first of all if you just make netflix a ten dollar subscription a streaming subscription you're essentially now creating an infinite number of price points below the demand curve there is now a price point for a user who wants to subscribe for just one second a month there is now a price point for user one subscribe for two seconds a month or maybe you just watch 10 minutes on netflix a month and now the total revenue is just integral into this area and as you can see the total revenue captured by netflix is much much larger now that's just with one package we can consider a multi-package setup where you have a team family whatever and what you essentially see is that you can now capture all area under the demand curve okay now another important thing is that uh when we are dealing with streaming payments we are dealing with on-chain verifiable payments payments which are by construction crypto native and this really allows us to reimagine and rethink uh from first principles how we think about payments so i know if we look at the history of crypto and d5 one thing that you should always keep in mind is that you should maximize for composability at every vertical of the stack so there's a reason unit swap tokens and v2 tokens we are c20 and v3 tokens are 721 same similar with a token to c tokens or erc20 so by moving state into a tokenized state you become composable and in the context of subscriptions that can mean quite a lot but we will discuss two things the first is that subscriptions themselves can be composable they can be represented in the form of nfps all right so let's just imagine for a second you owned an nft saying you were the first subscriber to the joe rogan podcast now clearly this nft would have economic and social value in today's society um however sort of along the way joe rogan would have experienced a lot of benefits as well because the moment you bought nft you now acquired a real stake in his success so the better joe rogan does the better like kind of the more valuable nft and the better you do similarly this provides joe rogan with organic distribution because essentially now joe rogan's subscribers are now his greatest brand ambassadors because incentives are aligned and lastly this can be very interesting from joe rogan from a business point of view because you can now essentially identify your subscribers you can identify who was early and you can reward them appropriately you can assign them property rights for example early access to podcasting episodes or special meet and greets as stand-up gigs and he doesn't even have to provide all these services themselves people can build services based fundamentally around these uh kind of building blocks because again the subscription is fundamentally composable now another vertical which is very interesting to kind of investigate is your service or your business itself being uh on chain composable so what what i'm kind of what you see on the screen here is something which is a little specific to superfluid and also what we're doing at diagonal but nevertheless i think it was interesting to share so the way diagonal represents your business is that we represent your business as an nft in the form of a specifically it's an nft and a supra now a super app is a smart contract where users stream money in to smart contracts and the smart contract redirects that money automatically to the merchant and pays them every single second so what you now have is you have a business which has predictable on-chain verifiable revenue which means it can do things like take out under collateralized loans because now external actors can see this is a business this business makes this amount of money every month it's this reliable and we can actually offer you a loan based on your in based on your risk profile not the one of some risk agnostic pool okay another kind of interesting area about this is that you can also sell your business on exchanges like openc or even potentially fractionalize your business creating kind of shareholders and community and distributing value to your community now lastly kind of we want to touch on why are kryptonated payment rails important so we'll perhaps only touch on two of these verticals but let's start with dallas so as we know there's a dow there's a kind of a b2b dowel service industry that's really booming at the moment and has been for a long time and essentially all of these services are sas services which are providing services with ours in return for conversation now the one thing dials will have common is they want to settle in crypto on cryptonated payment rails because interact with fiat is a massive headache and more specifically if they can pay in their own token that is a huge benefit so clearly in this scenario having crypto native subscriptions encrypted native payment rails is very important secondly you can think more in the future perhaps uh gaming so if we can see the metaverse such as decentraland where you can create your metaverse to your own space you can even i remember reading snoop dogg was hosting concerts in decentraland and the kind of economic activity which you can do in these metaverses is becoming increasingly sophisticated and i really believe will kind of eventually will be just as sophisticated as the economies we see in the kind of the real world today society now this means there will be service providers and this means people will be demanding crypto needed payment rails because they will want to settle in token like uh i find it very hard to believe that one day let's say the central line of sandbox will redirect you to a fiat gated checkout link where you have to put your credit card down and dodge yourself to let's say buy an in-game item okay so that was kind of the talk uh and again yeah i'm tony i work at diagonal finance if any of this sounded interesting to you and you think streaming subscriptions uh are powerful please get in touch uh either through kind of twitter or you can even reach myself on twitter or our website and finally we're hiring for several engineering roles so yeah have a great day everyone tell me that was an awesome talk and a really great overview i think uh i mean i'm super excited and obviously good to see you and uh on the other side after being a hacker too my only kind of question is um you kind of talked about how all these things are working now and what you are able to do um curious like what you think is missing in this space and what would actually supercharge yeah um yeah well so what i think is missing is frankly a clean abstraction to in use this technology so streaming payments is like what i think we will discuss is very cool and this this is very powerful however if you want to bring this piece of technology to the masses to businesses or even other protocols every protocol is not going to write the solidity to integrators themselves they're not going to pay the money for the audit they're not going to build off chain monitoring infrastructure to make sense of the payments they're not going to build the relevant dining running flows and the sdks to integrate into their existing payment systems so if you really want to bring this to the masses you need a clear abstraction which truly makes it just as simple as integrating string payments as it is to integrate stripe or whatever it may be there really shouldn't be a trade-off from a developed point of view and i don't think there there will be in the future so this is what i really think is missing here um and also largely what motivates us at diagonal got it so i got just it's not like we we think we need a new opcode or a new framework for everybody to generalize deployment and manifest that that yeah that's interesting i don't necessarily think we do so there there have actually been people who've suggested op codes in the future uh previously but this was uh rejected uh tremendously and i think there's a nice blog post by vitalik why something like an alarm op code it doesn't make sense specifically and doing asynchronous actions uh it becomes very complicated um and yeah it's just this should you don't want to do this at the execution layer this is a protocol layer problem and then the question becomes do you want to have do you want to do stuff based on external actors and trust assumptions or do you want to have on chamber or reliable value transfer and i really think it's yeah we think it's this one i think uh i don't think many people disagree on you uh well that's awesome um well i am super excited to kind of see how diagonal finance progresses and uh yeah just generally the fact that you can now use this thing to make a new way to think about under collateralized loans is uh to me super exciting so hopefully that becomes more and more common and uh really appreciate you taking the time and uh giving this amazing talk sure no problem thank you very much