The New DeFi Meta with Paradigm’s Head of Research, Dan Robinson

welcome bankless Nation to state of the nation where we dive down into a topic that is currently floating around the world of crypto this week on state of the nation we are talking with Dan Robinson head of research GP at Paradigm and recently across my Twitter feed I saw a tweet thread from Dan Robinson about how there are five big ways that uniswap X changes the game for swapping in defy swapping in crypto and I learned a ton in this tweet thread and I thought it would be super useful if Dan Robinson came onto the show and explained each

point himself here and so that is what you are about to get here today because uniswab X doesn't just stop at uniswap X it also brings with it a bunch of new changes for the Dex landscape that impacts so many other things across defy across ethereum it impacts his Bridges it impacts Roll-Ups it impacts Mev it impacts market makers and so there's about to be perhaps if you believe in this new intent based Paradigm which is something that we will to find in the show it flips a lot of what it means to

be a DEX or to be a deck swapper on its head so you're going to learn about all these details in this coming new meta for defy as well as what it means for you the user you the swapper and also perhaps you the lp if you are an LP in this world of D5 as well then we also open up some other doors is what does the future of elping yields in uniswap look like what can uniswap V4 do for LPS well uniswab X is focusing on swappers and why swappers and LPS are the two

parts of a Decks that need to be held in the highest regard and why everything else is secondary and then also we open the door to what flashboss is working on with Suave which is at the very end of this episode overall you're going to learn a ton about the Dex landscape that is changing quickly both with unison V4 and with uniswap X as well as some other things in the intent base World which is a Hot Topic in the Dex and Dev landscape quickly before we get into our interview with Dan uniswab is a main focus of this episode in uniswap is also a sponsor of

bankless bankless also holds a supply of uni tokens which you can see at bankless.eth and also bankless.com disclosures so let's go ahead and get right into our episode with Dan Robinson from Paradigm but first a moment to talk about some of these fantastic sponsors that make this show possible especially Kraken our preferred exchange for crypto in 2023 before you get to uniswap on chain you first have to bridge your money from off chain from the Trad World consider using Kraken our preferred exchange for crypto in 2023 if you do not have an account with Kraken consider clicking the link in the show notes to

get started with Kraken today Kraken Pro has easily become the best crypto trading platform in the industry the place I use to check the charts and the crypto prices even when I'm not looking to place a trade on Kraken Pro you'll have access to Advanced charting tools real-time Market data and lightning fast trade execution all inside their spiffy new modular interface kraken's new customizable modular layout lets you tailor your trading experience to suit your needs pick and choose your favorite mod modules and place them anywhere you want in your screen with Kraken Pro you have that power whether you are a Season Pro or just starting out join thousands

of traders who trust Kraken Pro for their crypto trading needs visit pro.cracken.com to get started today introducing polygon 2.0 the value layer for the internet for too long the limitations of blockchains have held back app development and stifled user adoption the internet allows anyone to create and exchange information what's missing is a value layer that lets anyone Exchange store and program value that's where polygon 2.0 comes in polygon Labs has unveiled a series of innovations that will radically alter the polygon ecosystem and web3 as a

whole by leveraging groundbreaking ZK Innovations such as polygon ZK evm the next iteration of the Best in Class plonky 2 proving system and a first of its kind ZK powered interoperability layer polygon 2.0 will give users and devs unlimited scalability and unified liquidity right now there is a polygon Improvement proposal regarding a potential ZK powered upgrade of polygon proof of stake if approved polygon proof mistake would become a layer 2 zkevm validium so make your voice heard on this proposal by joining the polygon Discord today you have a chance to help

the polygon Community give the internet the value layer it deserves are you planning to launch a token is your token already live and are you granting your employees and contractors vesting token Awards and are you trying to figure out how to take care of taxable events for your team toku makes implementing a global token incentive award simple with toku you will get unmatched legal and tax support to Grant and administer your Global teams tokens Toca will help you navigate across the life cycle of your token from easy to use pre-launch token Grant award templates to managing post-cliff taxable events with payroll

for legal finance and HR teams is a huge complex task to have to comply with labor laws payroll and tax obligations tax reporting and crypto regulations in every country that you employ someone it's difficult time consuming manual and costly and is drawing more attention from Global regulators and governments toku makes it simple for leading companies in the space protocol Labs Heder era get coin and many more so if you want some help in navigating the complex world of token compliance go to toku.com bank list or click the link in the description below Bank of this nation I would love to introduce you to

Dan Robinson he is GP and head of research at Paradigm he over the years has done a lot of deep thinking about markets auctions and liquidity and D5 with a particular Focus I'd say on uniswap new Swap and uniswap V3 recently uniswap Labs released uniswap x a new Dutch auction mechanism that Dan thinks changes The Meta for swapping assets in defy and also Mev chain interoperability and many more things the last time we had Dan on Bank list was September of 2020 almost three years ago if you've ever heard Ryan or I say the phrase

ethereum is a dark Forest it came from that episode Dan welcome back to bankless thanks for having me Dan you recently wrote a thread which triggered my imagination and was the impetus for bringing you onto this podcast that was titled five reasons I think uniswab X changes the game for decentralized exchange Mev and interoperability and this already kind of gave me um a model an agenda for this episode which we want to get into but just I really want to start at the highest of levels can you maybe Define the landscape for why something like uniswap

X is needed in the defy sphere uh what it what are the current variables or the current things that are producing this need for uniswap x yeah so I can talk about that and and I think it'll also talk a little about unisoft V4 and why I said where I see that fitting in so from my perspective decentral exchange research which is where I've spent a lot of of um my career and my research Focus has been um at Paradigm and with uniswap do you Subs exchange research in my view has to be about reducing the amount of

value that leaks out of the system um The decentralized Exchange system and so with uh uh when you were in a decentral exchange was particularly with automated market makers you have on one side liquidity providers and on the other side swappers and swappers are coming in because they can get the most liquidity from or the best prices from your decentralized exchange and the query providers are hoping to earn the best uh uh returns from um from these trading fees being paid and a lot of what I would say uh in research in

decentralized exchange design um that I've seen happen elsewhere has been focused on trying to tip the balance one way or the other between take money out of the swapper's pockets and put it in LP's Pockets or vice versa um and I think when you when you see a lot of reactions to some to the centralized exchange resources often saying oh you know this hurts LPS uh to benefit swab versus swab versus benefit LPS and in my view I think it's just sustainable Dex research is going to be about reducing the the value that leaks out of this system entirely it's not

going to be about you you can't just like benefit swappers by making lpus make less money you can't just benefit LPS by having um swappers get worse prices you actually have to find where where is the parts of the system where you can benefit both or reduce the leaks that come out of the system and almost all those leaks go into Mev and so I think the types of decks research the valuable lines of Dex research correspond to different types of Mev that come out of the uh of the system now um I think there's three kinds of Mev there one is just eip559

um uh uh burn just East it's been that's being burned gas cost from the transaction um and that's being paid by both swappers and LPS effectively um that's I think a really important one and unusual V4 and unusual X both address that one but for the other two one is uh loss versus rebalancing and that's um uh losses that liquidity providers suffer to Arbitrage or someone's coming in and they're trading on this pool at a better at a better price than the liquidity Rider baby should be giving them so liquid dividers is losing money to arbs and that all goes basically to

Mev um in the modern system um and so you just saw before and I can talk later about a little about how we think about that is about uh addressing those losses to um uh uh to Arbitrage I mean opening up new ways for developers uh for amm designers to reduce that but uniform X in my view is largely about uh the other side is trying to help the swappers and one way that a couple ways that swappers lose money um they can be sandwiched uh you can have your transaction actually executed

the worst price because somebody traded ahead of you you could have just ordinary um slippage you're trading at a price um and you know the prices moved away from you on the decks and so you're getting worse price than you expected and you could just be getting a worse price that is available somewhere else and in my view if if uniswap the system is not providing swappers with the best possible execution they could be getting anywhere um then what are we even doing here and so that's the um that's the motivation for me with unison X is how can we actually protect swappers the swapper side of this from from getting worse prices than they could otherwise I think

to reword I thought what I what I heard from you research in amm and and just you know dexes on on ethereum and encrypted broadly has two players that you think need to be held as first class citizens of the highest regard and everyone else should actually be minimized and these two players are swappers people that want one token and have one token and want a different token and then liquidity providers who are providing liquidity and I think what you're saying is that research in the

realm of dexes in in crypto needs to figure out how to optimize for retaining value by the the net between these two players and anyone else Mev arbitrages the ethereum protocol with eip1559 needs to be men while these two players are maximized that's how I would reword what you were saying is that fair yeah I think that's that's basically how I think about it and then uniswap X is specifically a technology that benefits

swappers the most it's yes largely focused on benefiting swappers but I think again in a way where you're benefiting primarily retail swappers right it's not about it's not about helping Arbor Treasures um trade against you know Universal LPS more effectively it's it's the people who we care most about actually helping here okay so I'd actually like to take this opportunity to start with your your Tweet thread and number one because I think this is where this goes your first tweet in your thread goes the architecture of uniswap X opens up a vast design space for dexes and so maybe we can talk about like I think the the

secret sauce that unit swap is uniswap X is bringing to the table uh we might actually need to Define what unit swap X is for people that didn't listen to our episode with Hayden um but I definitely recommend that as well so what what is the new thing that uniswap X is bringing to the table what's the new Secret Sauce here yeah so first I'll say I I don't like to use the term new because I think a lot of as people as many people have pointed out a lot of the ideas in unit swap acts have been around in ethereum deck space actually some since before uniswap um and I think some of the concepts especially around sign signed limit

orders I think are um uh you can trace back through through Xerox or even before um and so I think part of what Ubisoft X is about is are we actually at a time when some of these Concepts that have been around for a while make really make sense and start to um uh be worth integrating into how into how Dex is actually done on ethereum which right now again is dominated by Design chain flows on um on amms um and I I'm I think actually that there's a lot to be gained from from going to into the into these digging up some of these uh ritual ideas so yeah so so the key

idea for me of you know swap X is that we're moving from a system where all transactions created by retail order uh users are expressed as transactions with a specific um uh path you know imagine just like a an itinerary of here's I'm going to take exactly this this route on the subway um to get from 0.8 about B to an intent um uh or a limit order and an intent if you think about it is just it's just a statement like I want to get to the point I'm at Point a I want to get to point B but I don't necessarily care

um how I get there and um you know there's been a lot of of discourse on intents and I think um and and you know a lot uh prior even just on the on the more General concept and decks of of limit orders but I think it's very important that we start actually operating um with user intents I think they're they're incredibly flexible and powerful and you're gonna swap backs the main thing to me that it achieves is um starting to move toward uh toward users expressing their trades as intense as limited or saying I'm willing to I want to trade at the best price possible

no worse than this particular price rather than um expressing I want to trade on this exact amm so people who have swapped on universal before do you you open up with the asset that you have you open up the asset that you want and then you click confirm and then you like approve that transaction your metamask your Ledger or whatever and then you like go to etherscan to watch that transaction get verified in that what you're saying is that is a specific the user has chosen their point a and their point B and then the app the web app actually codifies that into a specific

transaction and then the user commits to that transaction and then that is committed to on chain and then a bunch of Mev happens in the background that you the user it doesn't really see but definitely does happen and so I think that that's the current meta of swapping and what you're saying is there's this new alternative called intense which is can you explain like how that actually gets expressed and then how that actually turns into the outcome that the user wants how is that different from an actual codified transaction and how does it actually achieve the same result

right so the way it works is the user you know the user still sees a similar interface on the on the web interface and but when they go to sign um to sign their trade instead of signing um a an ethereum transaction that is again constructed by the by the interface um they sign just an off-chain message they sign they sign um you know a message in a particular format um but that isn't an ethereum transaction and then it's the job of the entire Mev superstructure

um out there to get this turn this trans this intent somehow package it in a transaction that somebody creates um and get it included on chain but this can this could you know in intense in general could include anything it's it's a more General format than a than an ethereum transaction and um this the unit X uses a particular format where the user commits to a particular um uh uh price the price that they're that they're willing to trade at um supports the ability to do a a Dutch order which is

um a a kind of off-chain Dutch auction where they sign this order like uh they say here's my Decay rating basically it's going to change the price to make it more basically get a worse worse or Worse price the longer goes uh goes on until you hit something until you get some limit and that's in order to do it to facilitate some off-chain price Discovery for it and then it also allows the ability to pick a particular filler um uh who gets exclusivity on the trade for for a couple blocks and that's in order to to enable them to um potentially through through a request

for quote like system off chain find someone to fill this this quote better but all those all these are details about how this particular transaction format works and it's I think how we see um uh this is sort of the initial way that usobex helps find users the best price but again to me the key is you're moving into now I'm signing a message that just says uh expresses my my trading intent and it's up to the system to try to find the best price for it so in the example that say just for round numbers ether's two thousand dollars uh and in the old way is you would have uh

2 000 usdc and then you would say hey give me one ether and then you would ultimately get like 0.987 either before a number of different reasons one of them would be gas fees one of them would be slippage one of them would be front running and a bunch of different participants both humans and protocol eat away a little bit at that and and I think the way that this intense system works is you actually since you don't broadcast a transaction you actually uh excuse me you do broadcast a message but you do not make a transaction and so I I would

guess maybe this like this package that you're talking about you're just saying hey I have 2 000 usdc the minimum price that I'll take is one ether for that but if anyone can give me more than that then you are allowed to fulfill this order and then Dutch auction is just to Define index auction is you set a high price and then over a period of time that price comes down little by little by little and I think that's kind of the magic can you can you elaborate on why that's so magical for how that actually reduces Mev

yeah yeah so the idea there is but when I want to set I if I want to discover the price suppose I don't know at all what the price is um and this by the way I use Dutch auctions all over in like every mechanism because I think they're just a fantastic um uh way to get decentralized price Discovery um uh but I think this is what it's you know it's one component of a piece of uh of your sub Expo but the way it works is um when you put if you publish this order to the world um with some particular starting some particular starting price if that price

is is uh if anyone in the world can actually fill that their incentivized to do it as fast as possible um and so you'll you'll actually get include they can fill that profitably so you'll get that included um as quickly as possible if nobody can fill that then okay maybe that price is too high I'm going to start lowering it um and as soon as if if the um uh system is sufficiently competitive and there's enough there's enough potential bidders out there and nobody um and people aren't censoring multiple blocks at a time and again I think these are assumptions that are worth worth uh

questioning um but in that case basically as soon as this order becomes profitable for anybody in the world to fill they'll fill it and we have the assumption that basically once you hit that point that's the right price to actually fill it at now again like it there's some there's some possibility you could lose you could lose a little the slippage from the system like if the price you know uh moves in the other direction at the same time that your price fails you maybe don't get the best um execution possible and I think there's that's why there are these other systems as part of it um and to me unit sub X again is not about just the Dutch order it's not about the RFQ it's about this General concept that we actually get a lot more

flexibility from intents um but Dr what that's that's one mechanism that unisubex uses to try to make a competitive market for filling orders I think another way the way that resonates with me in my brain is is that the Dutch auction mechanism say say ether is 2 000 price two thousand dollars give or take but like you said you actually don't know in the in one single moment of time you actually don't know what the fair price of ether is so you start at like 2010 and this Dutch auction mechanism lets that decay you know 10 cents every second until till someone fills it and I think the the

idea here is that the the the most profitable person who can make that trade profitable first at the highest amount at the highest price possible will fulfill that because they can squeeze a profit out of it and and it allows the margins of what you call the system like this Mev system the market maker system it allows the first reasonable rational person who can make a profit will fulfill that order first and that's how you're coming to the conclusion that well yes that was the

best possible price for this swapper right that's exactly right okay and so I think this is this is why you come into the second tweet saying it's a better foundation for order flow auctions order flow auctions versus what's the opposite of an order flow auction what's the other half yeah well so right well so so backing up a little about order flow auctions so I think we've seen some systems um and flashbots uh has one called Mev share and I think there's some others out there um with the ideas right now when users trade on decentrals exchanges often they

um they may get sandwiched they may get front run um and one way that you can potentially just make the user strictly better off than they otherwise would be assuming already that there's submitting a transaction is instead of just letting anybody in the world um uh fight around the transaction or even maybe instead of just letting this transaction get included and see what happens um we can auction off the right to background your transaction and a background happens because uh you know when if I trade on an amm I might push the price it might be such a large trade that it pushes the price to a level

where it's profitable to trade after me and so I could potentially get a rebate on my on my transaction um by if my transaction gets included by basically selling off the right to say you're you're allowed to trade immediately after me that's a right that I in some sense have although it it can be hard to facilitate this Market between me and somebody who might want to background me and so if I can get someone to pay me for the right to backward me I will get a rebate effectively on um on my transaction so this is how order flow actions have been implemented um by flashbots on others

um with the transaction model um now the thing is I it's sort of a hilariously inefficient way to actually pay somebody for the Mev that's being created by their transaction because what's actually happening there is I'm I'm trading at a worse on on the amm at a worse price than I actually want to trade at I you know I want to trade maybe like at the current price or at a or at a better price but um uh I'm trading at some at some bad price on the mm such a bad price and causing so much price impact that actually is profitable for somebody to pay me to be the trader afterward um to trade the the amen back in the

other direction and this is you know this involves at least two transactions on chain um it involves fees being paid to liquidity writers who haven't actually really done anything because what's happening in this transaction is economically is that I the user I'm actually just trading with some with some backgrounder I'm I'm buying a token that they are selling and I'm buying you know at a at a such a bad price for me that they can sell it back to the amm at a better at a price that is profitable for them and so we should just cut out the middleman here we should actually just have the user and the the

background are basically trade with each other but the transaction format doesn't allow this kind of um order flow auction and so that's why when you move into an intent suddenly you know all I'm saying is I want to actually trade I'm not saying it has to be on this particular amm so that you have to extract this value from you know by back running and then making a side payment to me you can just come in and trade with me at a better price the the metaphor that I'm getting here maybe you'll you'll like this one is is like it sounds like we're trying to build a ship design a ship that doesn't leave a large wake and it sounds like in this um

if the current Paradigm of uniswap trading where the reasons why um the pools get balanced or who we're trading with is because in order to fulfill that order there needs to be enough wake for uh arbitragers to come settle the waters and right now uh with with this off-chain intent model we actually are just designing a ship that doesn't disturb the waters that much is this a fair metaphor yeah I like that I think I think that's the idea is you're ultimately getting more efficiency from from doing it this way because it's just

a more flexible format and so these things that people are are it's so needed that it actually getting hacked into the current model and I'll give you another example which is just in time liquidity or jit liquidity which I've been going on like a like a rampage about for about a year um partly in preparation for for anticipation of Unison X coming out I'm talking about how I think so jit liquidity the way it works is when somebody's trading and they're going to trade at a bad price um someone can come in and provide concentrated liquidity uh right around the current tick uh right before their trade and then withdraw it right afterward and what they're functionally

doing is they're improving the user's price impact they're actually helping the user a little bit um you know by trading against the user but this is this too very gas and efficient a lot is being paid in 1559 um and uh it's not in this case like not actually going to liquidity providers um who's it going to it's basically just being wasted or it's going to to gas costs um all the inefficiency of this of the system was really what we want is okay okay if I'm if I'm a user I'm about to trade at a bad price can someone just come in and save me and again the intent allows us to do this more efficiently your second tweet your second point in

your five point uh tweet thread is it's a better foundation for order flow auctions that return Mev to users this is what we're discussing correct that's right yes so if you want to if right now the way that those work and the way they fit into the unit swap model uh involves like this this multi-transaction expensive um and often inefficient uh protocol for for like back paying someone to background them when really again like when you've got an intent it's much easier to just batch these together and and match them against each other so maybe to summarize points one and point two to the best of my ability there's

two there's kind of two birds with one stone in this intent model we get the Elegance of the Dutch auction model which like we said earlier naturally emergently finds the most profitable actor who can give you the best trade and still be profitable themselves and because it's not on chain we don't have all of this wake in the transaction so there's not just in time liquidity there's not back running there's not two trades that help settle a first trade it is all done off chain which helps retain

value from going to eip1559 and has helped retained by swappers and also therefore liquidity providers so that's my summarization of your first two points would you say that's fair that's right although we're gonna I think I would separate out the Dutch auction Point um as being maybe maybe part of yeah it's part of Point yeah it's part of point one where it really is like this is the kind of thing you can do with intents that you actually can't do with um with ordinary transactions on amms beautiful okay let's go into number three which I love love is so simple

it's still early which is you say that we've opened up a new frontier of of exploration so what so elaborate on like what this new frontier looks like what's needed in this new frontier what kind of infrastructure do we need to settle on this new frontier like what's left to build here in this new design landscape to help this efficiency become even more efficient yeah so I think part of so part of the motivation for you to sub X is okay we see all this Behavior that's happening on chain with say like jet liquidity or um or people

like designing Mev uh order flow auctions of like you know MAV action background in actions right and thinking okay like this is a sign that actually there's some serious inefficiency here that should be that should be streamlined that should be improved and I think in my view the the design I think does this and and gets users better prices um and uh and say you know it is more more gas efficient in many cases um and I think that's that's you know why why I think it's I'm excited about it out right now but really once you're now we're up in intense space

um this guy is really the limit on just like how you can actually uh match these orders and um and uh process them so for example right now the way that usubex works every every order is um well there's there's there is there's a star of Q processors I alluded to earlier where um you can um use you can choose a particular filler to fill your order and give them exclusivity for a couple blocks in exchange for basically giving you this this good quote early on um but right now that's you know the trades are being matched primarily against these um these fillers what what happens when you actually start to match

uh user trades against other user trades um when you do something more like a batch auction um like like cow swap uh innovated I think what happens when you start to do um uh you can do some sort of more interesting math with that right like when you have a bunch of when suppose you have like a bunch of user orders you can execute them all at a common clear advice so there's algorithms for trying to do that um you can you can do like you know match ring trades where you have multiple um uh uh assets being traded and you try to like maximize the total amount of volume these can be you know you can

encrypt these orders right like you can um and I'm not sure how much other some of these other Solutions have have done this yet but you can actually say like all right we're gonna have everybody um do a sealed bid batch auction that's actually enforced by cryptography um where you encrypt your orders you you do a this homomorphic computation on it to try to compute basically the the common clearance advisory or you unseal it by using a using a group or some multi-party computation um and that gives you like a like a more decentralized or Fair potentially fairer um auction you can match it against on

chain uh liquidity using it using a particular algorithm um you can uh put it in trusted Hardware uh which I got in a moment but like you can you know it's much easier potentially for uh trusted Hardware to operate on orders because they didn't conclude basically all the information you need about them right in the order without having to involve interaction with a bunch of ethereum State um like for all these reasons they're just more tractable and so I I think like in the long run I have no idea actually what these things will look like um but I think I know I'm pretty sure it's going to look like some kind of

intense some kind of um when users Express their their desire to trade it's going to look more like an intent I think than a particular on-chain transaction and that's because they're just so much more flexible so there's just a lot more room to explore yeah it's the flexibility of the expressivity of an intent as in it's very very open ended I think maybe another way to help Define an intent is it just provides like kind of a the bowling alley guard rails as to like here's what I here's everything I don't want but anything inside it in this specific specific landscape of what I do want you're

allowed to give me and so basically saying you just write the rules of like hey you can't give me less than this and then the mechanism will naturally give you as much as possible and but not codifying it I think allows for a lot of expressivity and I think that's one half of the magic um that I think you're you're talking about the other half is that like since it's off chain the cost of computation are zero if there's no there's no gas costs to matching orders of user a to user B to user C we you can compile

every single order together all at once and all of that computation that it takes to clear those orders doesn't have to actually run on a blockchain that can run on some local market workers server and all of a sudden we get really good order execution from not having to be encumbered by blockchain computation and that's my intuition is this right is this right that's right I think just off-chain computation is incredibly powerful um and we actually saw just with with