A Discussion on the Future of Onchain Finance with Hayden Adams and Phil Daian

[Music] All right, we are ready for our next panel. It's going to be a very interesting one. Uh we're going to bring on two people, Hayden and Phil. Hayden from Uniswap, Phil from Flash Boss, and they're going to talk about how we think the future of onchain trading is going to evolve. So, please give them a big round of applause.

Hello. I think our plan was mostly to talk about oil rigs, but we're gonna have to pivot.

Yeah. Yeah. We're gonna have to talk about trading on chain and DeFi and L1's. So, hopefully you all are excited about that.

Some some context. I think we both thought we were going to be interviewed and now we're interviewing each other. Um, but good news is that we happen to have launched something today together, our two companies. Um well actually I can introduce you know Hayden from unis swap, Phil from flashbots. Um and you know together our two teams uh rolled out uh flash blocks for uni chain today which is basically giving us 200 millisecond swapping for the first time ever and so it's been a huge win.

Um yeah I we you know I I would love to kind of maybe one place to start actually would be curious to hear from you Phil like what is a TE how do you think about it? Yeah. So I think um you know probably many people in this room already know but tees are kind of this really exciting new tech that lets us do all this new privacy and integrity for computation. Um and I think that's really interesting for ETH for L1's and L2s. So Uni Chain is kind of an L2 trading product um looking to have a really good trading UX.

Maybe that's one question for you Hayden. Next is like what is Uni Chain and like what are what are the design goals? Um, but I think it's been really nice to to collaborate with uh Uniswap Labs and kind of build out this cutting edge tech. Uh, it really is on the frontier in many ways of technology. Like there aren't very many widely deployed low latency TE apps out there.

You've mostly seen this tech used for kind of ML training and other kind of data privacy things like that. And really having this rigorous data privacy foundation on which we can build trading systems, at least for me, it's like super super bullish. So I'm glad we managed to work together on it. Yeah. So I think that like you know the the journey of how we came to building uni chain um you know unis swap started out as this decentralized trading protocol on top of Ethereum and as this very simple interface to trading through that protocol.

And I think that like both of those components were actually really key. Um I think if we just launched the protocol we never built an interface to make it easy. I I don't know if it would have ever gotten adoption. Um and every step of the way for the past eight years since then uh you know the journey has been how can we make this better for users? How can we make it you know more efficient as a marketplace?

We had unsoft v3. How do we make it uh easier to onboard and access? Um and you and that's really where for example our work on our wallet came from. And you know when it came to building uni chain I think it was just like the next iteration of how can we make unis swap as as you know not just like a a a protocol not just a front end but like the holistic product of interacting with it as a decentralized exchange. How can we make it you know better and easier to use for users and have better outcomes for them.

Um and when you look at uh one one intuition for the chain actually comes from our decision to build the wallet. One thing that was starting to happen um is that people that sat you know or different you know things that sat between uh our users and using the protocol were were kind of getting in between and making the experience worse. Uh whether it was you know the the user experience of some of the wallets that most people were using or whether it was even the like the way that like the business deals those wallets made with private market makers to kind of uh you know extract uh ME from them. Um and so you know having our own wallet helped us kind of uh be able to provide an experience for our users that that you know improved it. And then I think from here like even like the the way that blocks are built at a chain level really has like a pretty profound impact on our users.

Um you know one one like intuition is that like something like what like one to2 billion dollars a year in value uh is made by you know me searchers and and validators through through block auctions based on the the value that is generated by by you know unis swap traders. And so, Unichain was a project and uh that that essentially allowed us to build a chain that really optimized um for for the the use case of unis swap and and trading and defy um uh while still kind of retaining the core value proposition of decentralization on Ethereum.

Yeah. So, I guess uh one interesting point about uni chain I think one thing we collaborate on a lot is like TE's and how to fit this into like the sequencer design into the OP stack things like that. Um, and at least I consider kind of Uni Chain like kind of a a leader in this in the L2 space, kind of pushing a lot of new research around things like transaction fee policies, me redistribution, um, you know, faster blocks, uh, with te guarantees on top. So maybe one question I have for you is like what made this kind of tech attractive to you and like how do you see it as maybe differentiating uni chain and like kind of uh also maybe Ethereum as a as a whole.

Yeah. So I think that like you know when we look at where value like I think something that was I think the tech kind of followed the the user needs right and so like you know we don't look at something say like what is like you know like like what is cool tech we we think like what do our users need and what our users needed one of the things people needed was like lower block times lower block times is good for for traders uh because it you know reduces the time from when their trade is made to when it's executed which reduces value that leaks from them. It also creates a far better user experience. Um, uh, lower block times is also extremely good for liquidity providers. Uh, because liquidity providers, you know, they they kind of lose money in some ways off of like long like price divergence, uh, when they're when they're in a position, but then they make money off of all the trades along the way.

And so the the lower the block time, the more like intermediate arbitrage and trading volume that they're seeing, the more kind of money they can make off off of that. And so, you know, low block times was was was good from a like a price perspective. it was good for liquidity providers. Um and it just makes it like you know when people interact with blockchains I I think that one thing that we've you know I've been feeling a lot is that uh you know that like to actually grow massively from where like we've DeFi has grown so much from where we were in the early days but to kind of grow to the next stage it can't just be like you know well like like decentralization can't be the only benefit like like you need to kind of meet people where they are in terms of user experience and so like having something look and feel instant which um I confirmed before I came out stage like if you use our wallet right now on uni chain the swaps are instant like you it you can't there's no percept uh we we the number 200 milliseconds didn't come from nowhere it came from like that is the the rate at which the human eye notices these things more or less and um and so you know I think that like the reason that like and then when it when it comes to like tees and transaction ordering um again like the whole supply chain of of when someone makes a trade to when it's executed is like extremely important and to us like you know one of the core value propositions of Ethereum um was you like sort of like the decentralization and the and the transparency that that can come around it. Um, and when it came to like trying to run this at a at a greater, you know, have in the L2 space, there's often like historically been like a trade-off people make, which is like by moving more execution offchain.

Um, you're able to scale up greater, but then what you lose is often like some of the transparency or the guarantees. Um, and so things like TEES are are ways to create better, you know, uh, an example would be like the the way that like a sequencer orders transactions. um the uh you know by doing it inside of TE we can create like a a very strong guarantee around um and and you know there's a there's a there's a future you know thing called flattestations or something like that.

Yeah. Flash testation.

Flash testations.

Flash flash testations.

Flash test station.

Yeah. Yeah.

We got to work on our naming. Um but yeah but that that will give us you know like a cryptographic proof of the logic that was used to order the block which which can help you know uh great better outcomes for users essentially.

Yeah. Yeah. So making this whole supply chain kind of both verifiable but also faster. Um it was kind of the design goal from from Uni Chain is is what I'm hearing. Um, how do you think about that in relationship to like I guess the L1 wars because like one of the hot Twitter topics these days is like you know different chains.

What should we do for ETH L1? Should we even have DeFi on on L1? Um, how does this TE fit into that narrative? Because at least to me it seems like in some ways a lot of the ways that other domains and other chains have been able to advertise from a trading perspective is to kind of make claims around latency or UX or something like that. And this to me sounds at least what you've described is a pretty good UX.

You go in your wallet, you press, you get like an instant confirmation, your trade is done. Um, is there anything missing or or or how do you think about I guess the positioning there?

Yeah, I think it's a good question and I think that like you know the whole like should D5 be on L1 conversation is an interesting one and I think that like you know I think that like probably one way to put it is that I don't think that crypto scales by having like a single monolithic chain. Like I think that like one really like funny way thing like like pattern that I repeatedly see in the world is people being like we need something to decentralize. How do we create a single thing that is the most decentralized version of it? And it's like actually I think the way you decentralize is you have many versions of the thing um that that kind of have some and and then they all come together right and that like when you look at like how the internet scaled the internet didn't scale by having like one globally owned computer. It came by having like many computers that could talk to each other.

Um, and so I think when you think about um uh the the like one even like the value proposition Ethereum trying to bring to the world, it's like this like ex maximally decentralized um version of of blockspace and I think that's extraordinarily valuable and I think that like um you know things like uni chain for example can leverage that right right you know uni chain posts its data to Ethereum L1 uh but you know putting every single part of every single computation on a single like if if if the goal of Ethereum is to have you know be able to run on a single node then you're inherently not going you be able to run every piece of computation in all society on it. And so I think that like figuring out how you can kind of leverage different pieces, you know, different tools um to kind of and and have them, you know, like the you have them all kind of come together using standard uh you know, interfacing like like standard um you know, like there's there's a kind of a reason that like like computers can talk to each other and and information can kind of seamlessly flow through them. It's because we have like standard information sharing protocols and and standard formats for representing data. And I think kind of what we're doing in DeFi is just that for value. But like you know the internet has many servers and they're optimized for different use cases.

And I kind of see that playing out here. But um generally speaking like you know part of the value proposition of DeFi is the decentralization of it. And so we just have to figure out how to make you know many of these different pieces uh decentralized in their own right. And so I think things like TEES give us the ability to like give us some of the benefits of Ethereum while not using you know every single piece of computation in one place.

Yeah I think that's a fair point. Uh so one of the reasons like I and I think flashbots is super excited about TE's from a trading context is like once you start looking at how do you secure a trading protocol I mean let's even say like we both want to participate in a darkpool um and you want to do that from like maybe a technical rather than a legal perspective that's kind of the ambition of of DeFi right it's like we can do this technical stuff it'll be way more efficient it'll be way better um well you have all these privacy tools at your disposal a lot of them even let you program them in various ways um only give you privacy in various conditions or use the privacy to kind of leverage your information in certain ways. But then there's many different ways to like actually do these things. So there's like MPC, FHE, many kind of technical objects that will let you achieve this computational privacy that's like super useful for trading. Um I think TE's are a really interesting trade-off point because they give you kind of a lot of integrity and uh privacy properties um almost for free.

they're almost as fast as just running the thing on a server um rather than some of these other protocols which maybe require back and forth communication between parties or like parties in the servers. Um so in some ways I think I think what you said about the internet topology is like super interesting. We can kind of using TE's iterate to mirror that because the computational kind of workload profile um is very similar. Um yeah, I also want to say one interesting thing which your which your answer made me think of. It's like I think DeFi is actually in a very special place right now.

Uh we've actually built up a lot of the infrastructure that's been promised like really since 2017 I would say when things really started popping off and the hype really started blowing up. Um, and we have these exchanges, these trading systems that run at internet scale across the world. Uh, with many of them with like fairly good UXs and different trade-off guarantees, many of them with rigorous technical notions of privacy that don't exist, uh, on Wall Street and Trafi. Um, so I think we should maybe once in a while also take a moment to recognize that we're in a pretty special place in like kind of the history of finance. Um, and uh, yeah, that's my rant.

Uh I'll I'll I'll leave it to you to rant or ask me questions. But

yeah, I I I do agree. I think that like DeFi has come so far so quickly. Um I remember when you know when unis swap initially launched in 2018, I I think the word was only just begun to be being used for the first time. It like maker dow had recently launched, compound launched in the same month as unis swap and like the word defi kind of became a thing just like six like seven year seven eight years ago. Um, and now it's kind of like currently being hotly debated in in DC for like a law that's like how do we carve out the language to make sure that the the you know the rules that apply to centralized systems are not the same as for decentralized ones.

And that's like a it's kind of like a crazy trajectory and and you know there's like unis swap in the past fiveish years has processed like three and a half trillion dollars in trading and um has kind of operated at like a pretty insane scale. And I think that part of that is like it's like not really possible without the sort of technical like like the there's sort of like a scale that you could reach with decentralization um that that is is pretty insane. Uh like we kind of put this thing out into the world and then like the fact that anyone could integrate with it, anyone could build on top of it. Um and and sort of uh there is sort of like a one one kind of uh like angle that I sometimes think about it is almost like you you compare like unis swap the protocol as a as a product relative to like things like Twitter or things like Reddit. um where like Twitter and Reddit uh are two very massive internet companies that I get um I guess X now but uh massive internet companies that like had these like really powerful APIs and uh and then both of them uh basically shut down their APIs at different times to like um because either they didn't like some of the things people were building on top of the API or they they wanted to compete with the companies that were building on their own APIs um or you know they felt like it was like the business optimal thing to do um and the kind of idea that you can like build on top of something like the unisoft protocol uh and you know combined with uni chain and ethereum to like create an application that like in some ways can't be rugged like I don't have the ability to kick someone off of off of off of uh unis you know off of Ethereum I don't have the ability to kick someone off of unis swap um so there's sort of like this notion of like um and because of that people can actually build on top of like like there's it like dramatically lowers the risk and makes it like a better counterparty I think That's kind of like generally the thing um when you think about like how uh crypto will kind of interface with the broader world.

I I think that like people like the reason one of I think one of its like structural like one of the inherent advantages that I think will help DeFi platforms win or like gain much greater adoption is just that nature which is like people will want to use them because they know if you know they can never lose access to to their like the assets within them or they they will or if they're a developer they'll want to build on them because they know that you know they'll always be able to to kind of self-s serve their their own you know uh interfaces with it. So I just find that like a really interesting dynamic.

Yeah. It's almost like an unstoppable application I guess in some ways. Um, yeah, I'm curious like one interesting thing I was meditating on a few days ago. So, this is kind of like a I guess founder journey at least panel or conference or whatever, right? And and you just kind of mentioned and I agree that in many ways we have DeFi going mainstream like we're discussing it in Washington.

Uh, we have like really big conferences around the world. We have many kind of institutions quote unquote getting in. Um, that's kind of not how things were when we started working on these things, I would say. Um, I don't know. Do you see any any kind of unique challenges of that?

Because at least for me, like thinking about some of these issues like privacy. Okay, we're going to have programmable privacy. It's going to let us build super efficient trading systems. uh that seems great, but also like in some ways we as a community are maybe going to be in the center of conversations about like how does privacy impact the world and like how far should we really go with these technologies and who should and shouldn't run them. Uh so I don't know.

I guess my broad thing is like do things also feel different for you now that they're more mainstream because at least for me some of these things feel higher stakes.

Yeah, I think that's right. Um they definitely feel higher stakes. Um, at the same time I think that like it's it sometimes it still feels like we're like early like we got it like DeFi got so big so fast it's kind of crazy. Um, and definitely like the risk is always like dilution in the like the values or or like um but I I I think the thing that I like keep kind of keeps me pretty optimistic about it is just that like there's like this like fundamental I kind of like compare it against the internet pretty often. Um, right like like what did the internet do for the world?

The internet basically democratized access to information. Um, uh, you know, most most information now is on the internet. Uh, it's mostly digitized. It's because by representing information in standard formats and and in this place, it kind of is instantly global and and accessible to more people. And I think that like there's this like inherent nature to some of the technology where it's like we can build like I think what like DeFi is trying to do is to democratize access to to the creation of markets to to access to markets um to the kind of transmission of value uh in you know to make it like you know most value is available kind of instantly globally and and certainly that puts it in like the center of certain things and and agree there's like challenges to it but like I think that like I don't know like I don't think any like no one will argue that the internet was purely a good thing.

No one will say the internet has only done good things in the world. Um but I think most people think that the internet is like overall like democratizing or or um you know like uh expanding of access and I think that like that's like there's like a similar nature to the the technology we build um at least from my perspective.

I think it's an interesting question. I was actually discussing this with someone at the ETH foundation yesterday at dinner. I won't dox them uh chatam house but uh um uh they they had asked me something like for the block building market and how blocks are produced and kind of trades are reconciled in these systems how decentralized do I see it being in like 10 years and my answer was kind of like I don't know 10 years to me feels like a short time in some ways because I also love comparing things to the internet it's like a little cringe but I think it actually does work and it's true uh and so I stand by it uh as well um and so I was like look compar it to the internet right? That was invented in like the 70s really. Um uh and and really didn't go to everyone's house until like kind of the early 90s.

And at that point, you had this big explosion, a lot of capital, these hype bubbles. Every company that was built was was being built, etc. That was already 30 years ago. We've seen like several waves of business models come and go, a lot of nation state politics come and go. But at least for me, like it feels like on the internet, the jury is still out.

like whether we end up in like a totalitarian kind of controlled internet environment uh where things increasingly break into FFTs where nation states kind of concentrate this power or whether maybe the internet pushes these social forces more towards something decentralized and like cooperative that's at least my kind of better world. Um I think it's still an open question. So, I don't know. Uh, I'm I'm just kind of meditating on the fact that I think things things often happen slower than we want them to. And I'm hoping maybe crypto can help the internet uh stay decentralized and and make sure there's enough need for cooperation among all these parties that we don't lean into our own kind of uh FFTs and nationalism and things like that.

Yeah, the the jury still out thing is like it kind of it's true like there is like a lot of ways that like society can evolve from here. Um I think at the same time like there is like something nice which is a lot of people have like you know can can kind of look up anything and and send anything but yeah I I I I do I do see the point there. Um one thing I'm kind of actually speaking of like kind of like founder journeys and u maybe context for me and Phil have known each other for a very long time. Um we we first met uh almost a year before UNISOP launched and Phil was like a PhD student doing research on what later became known as MEV. um but was before the term MV was made up and you know and I was working on unis swap and um I think that like the early I kind of am curious to hear how you think about like AMM like you know AMM market structure uh because I remember you know in the early days you had a lot of concerns around it and how it could um you know potentially value or or or whatever and then um I'm kind of curious uh very recently saw you tweeted um that you thought AMM architecture would replace all order books forever ever.

So, I'm kind of curious about that journey and how you think about automated marketing making and and kind of expanding. Um, to me, it's it's part of the story of like uh it's it's like automated marketing like right now it's just like in in there in the traditional financial world it's so hard to create an asset and then like YouTube made it like the same way like it's like harder to create a Netflix documentary than it is to create a YouTube video and to me automated marketmaking is partially about like doing that for financial markets. But I'm curious your take on it. Yeah, my my take is actually a slightly different angle which I think is fun. Um, so yes, I did tweet that I think AMMs will dominate all trading.

For everyone's context, I like to make like fairly aggressive predictions uh about the future. Some of them are wrong, some of them are right. Um, so take it with a grain of salt. Um, but the reason I think that actually originally came about as like a technical one and it kind of relates to something you said earlier. you were talking about how decentralization kind of fueled the adoption of unis swap because people could kind of deploy things, they could deploy assets, things like that.

That kind of brought me to like some meme that we've been playing with internally, which is basically MEV as like possibly thinking of it as a scaling technology. Um, so like how your chain processes and handles its ME actually determines like what transactions and what state uh get made to kind of have the economic output that you want. Um, and in some ways I think what you said made me think, okay, maybe decentralization is also a scaling technology and we haven't kind of framed that the right way for people where they think there's this conflict between, okay, you can scale on a centralized server or you can have this like slow clunky decentralized thing. Maybe that's actually not the case. Um, and decentralization is a scaling technology.

Um, so to relate that to AMMs, I think when you think about just efficiency and you think about trading, okay, you want TPS, you want efficiency. Um those are like the first order UX requirements of trading. I think uh just as a pure technical object, if you think about like draw a picture of a of a order book as a machine with like all of the state entries it needs to update in the C program and like the way it kind of finds and reconciles the trades and and kind of the execution process and the confirmation and then draw a picture of like the simplest AMM that you can draw and like the AMM picture is just like a lot more aesthetically pleasing uh from a technical perspective. If you have like the simplest AMM which is like the V2 style like two variable AMM, you can fit the entire uh kind of context needed to execute that inside like a CPU cache which is inherently like a faster level. It's a more compressed level of memory.

Um because instead of compressing the state of the world down to like this this order book object which can be almost infinitely large uh fairly unbounded um you you're compressing into this very kind of bounded amount of data. Um and actually all these transactions and this MEV and this block building that's essentially happening on top uh is is compressing this data just into a very efficient final state uh output. So for me this just aesthetically feels like obvious as a computer scientist. It's just a better way to do it. And less like you can come up with something really compelling that's just like okay now order books have this which is better.

Um I haven't seen any argument like that. Um, I think probably the next wave of things we'll see in AMM market structure specifically, well, there's one that I'm hoping for and one that I think we'll actually see uh first. So, the one I'm hoping for is kind of better uh maybe interoperability and just cross domain notions of AMM trading. Um, so things like how do we consider you know uh uni on L1 interacting with uh uni chain L2 which is giving this very nice fast trading UX. That's a problem that I really care a lot about.

But I think in terms of what the market wants like uh it's kind of more like the hyperlquid DGEN thing, right? And like uh I think the reason like people like CLBs even have any oxygen in this conversation is people just haven't built like the right DGEN frontends for for AMMs in some way. Uh but that was kind of the justification in the tweet. It's a it's a fairly spicy and it's more of a technologist than an economic uh take. But yeah, that that that that actually I makes sense.

I I think from a um I would kind of be remiss to not at this point mention a little bit of like the interplay between the Unisoft protocol and uni chain and kind of like how they can sort of fit together in a in a mutually beneficial way. Um which is like you know and and it kind of like you know we have this system of fair ordering using tees for the transactions. Um, one, you know, one general important part to AMMs is that like, you know, liquidity providers are able to kind of sort of be profitable and and there's sort of like this like in some ways when you see like uh a lot of value leak from LPs to arbitrageers, um, it's it's like not it's usually like a sign that like the market could be improved. Um, at least from my perspective. Um but um you know one thing that like uh is very interesting is that in some ways by having this like pro uh by by doing the fair ordering in a TE and then having that like um having information from that accessible at runtime you know basically when someone is making when the trade is actually executed in a block um you know liquidity providers in some ways like sometimes it's like the reason that that uh arbitrageers can make money off liquidity providers is you have these slower blocks and then you have uh people that have real-time off-chain information that are trading into those slower blocks and and they can kind of like benefit from that that latency in information.

But by basically having the auction happen uh and then uh you know have that accessible within the the the chain you can actually liquidity providers for example can use that information to update their prices right before the trades execute. Um and and so like a simple example you know unis swap before gives us programmable pools that can do things like dynamic fees. um we haven't even mentioned unis swap before I don't think yet but recent version of the unisoft protocol that's pretty cool um and and uh more customizable but one thing it gives us is like things like you could have a custom fees and and then what you could do is you could use you know the priority price for example set paid paid in the TE auction uh as a as a input information to liquidity providers that they can use to like update their prices um and and um and in that way you know if if there's a if someone trading against liquidity provider is actually taking a lot of value from them that would be represented in the TE auction which then the LP could be like actually you know pay me a little bit more. Um and so you can sort of like kind of converge these different elements together to create you know like the best uh version of of unit. And I think that's kind of gets at the um the like different types of block space.

You have L1 DL2 is like I think that like block space inherently can specialize um and and um and and so I think that like having block space that is specialized to like really improve automated market making and improve DeFi uh is is like a is like an inherently useful thing.

Yeah. Yeah. No, I I agree. I think that's an interesting perspective kind of wrapping it in an Oracle um construction. And I kind of think this is like the next generation of AMM market structure.

It's it's like how can we take all these primitives we have this technology TEES it lets us share information among untrusting parties. It lets us build blocks in like very fast efficient ways. Uh we're doing this on kind of all these different domains. Uni is running on all these different domains. How do we now leverage all of this infra to like work really really well together?

Um, but I think there are kind of many many points that we can do that at including like you said the the kind of price data side and and uh maybe even liquidity sharing. Who knows like what the future holds. Um, so I think I'm very excited about all those innovations and like the beautiful part about the AMM is like from the chains point of view, it almost compresses all of that like very intense and very like specific to the flow like computation into a very very nice uh state update. Uh, so I'm bullish. I think we have 40 seconds left for closing thoughts.

I will say try unis swap with flash blocks in your wallet. It's very efficient and uh should be the best DeFi trading experience. uh you know uh uh on chain right now probably you know I would imagine like uh I don't know I haven't tried it yet so I have to do that also it just went live today I've been on a plane

I tried it I tried it backstage right before coming out to confirm um uh but yeah it's 200 milliseconds when you when you swap on uni chain and it which makes it feel instant um and and it feels really good um because you know yeah this has been something that's been plaguing us for a long time uh that like just latency of of using blockchains. So, we can end it there. Um, and yeah, it was great being up here with Phil, my my good friend.

Great to see you, Hayden. Um, good to see you all.

Yeah, always happy to chat about Flash Blocks, Uniswap, TE's, anything offline, you can find one of those.

All right, cool. Thanks, everyone. [Music]