Hello, hello. Hey Jay, I can hear you
Awesome. Are you all doing today? I'm doing well. How are you?
doing well. Nice and early today here in San Francisco. Where are you guys calling from? Yeah, I'm calling from UK. Okay, so it's like early afternoon free or like early evening for you. I don't like
up again. So 4pm something. Yeah. Oh, nice nice. Yeah. And let's just wait for one more minute. Yeah. For more audience to come in. Awesome. Sounds good. Yeah. Partition.
Great, and I also saw your official account joins. And my colleague will set up, set, say, official accounts as a co-host. Great, I think we can start now. Thank you everyone for joining today's 5-bit Twitter space.
I just meet you Jay. Thank you so much for your time for coming to buy it today. Yeah, so you'll be great if you can talk about how I was about yourself and then our audience, then more about you.
Of course, well, great to meet you everyone. My name is Jit and I'm a co-founder of SayLabs. So SayLabs is the US-based development entity that contributes open source software for the state blockchain. In terms of my background and my story, I guess I'll start early, so I grew up
in the San Francisco Bay Area, Don'ts Locont Valley. I was surrounded by technology my entire life, so I got into programming pretty early. And so when I was in high school, I used to do a ton of speech in a bit, and that's actually where I ended up meeting my co-founder. So both of us were
in competing high schools. And we met each other through a practice debate when we were sophomores in high school. So this was over 10 years ago now. And then afterwards both the plus just kept in touch. He went off to UC Berkeley and then I went off to UCLA to study computer science. And I personally ended up getting into crypto back in 2017.
My roommate at the time, he was starting a crypto company who was going through Binance Labs and both of us ended up tinkering on a couple of projects together. And afterwards ended up joining Robinhood. So I spent almost four years in Robinhood. I saw the company 10X and I was an engineering lead.
when the entire game stop saga happened two years ago. So I don't know if you were following along with it. I don't know how much the listeners remember about that. But what happened during the game stop saga was there was a set of like 10 to 15 stocks that were these meme stocks, right? Socks like GameStop,
I think Blackberry was another one. And these stocks were not necessarily sexy stocks, but they had been just pumping to the moon. And that was because retail in the United States was really rallying behind these stocks. And one interesting thing that was happening is it hedge funds were shorting these stocks, but because retail was able to help
push the price higher and higher hedge funds were being short squeeze. So it was a true Robinhood moment where you were essentially taking from the rich and distributing it to the poor and middle class. So this was kind of the backdrop for I guess January 2021. The primary place where all of these folks were trading was on Robinhood right because
Robinhood is the first place that most new, I guess, users in the United States will open their first brokerage account. So everyone was using Robinhood and then one day just completely out of the blue. Robinhood, without telling anyone, just turned off eyes for these meme socks.
So the entire world was like, yo, what the hell just happened over here? Because Robin, who was the primary place where this buy pressure was coming from, and then Robin could just turn off buys. So first of all, customers were pissed because they could no longer buy these stocks. But secondly, because there was less buy pressure, the price of the stock also ended up falling. So people were actively losing money from that.
So yeah, I mean as you can imagine that was handled exceptionally poorly by Robinhood because they just there was a complete lack of transparency right complete lack of communication with everyone that was Using Robinhood at the time and even as an insider it wasn't any better because the leadership did not tell us
anything either. So the data that that decision was made, as you can imagine, it was just incredibly chaotic because I had a bunch of people just reach out to me, so like some old friends, some old colleagues, we're just really mad like, yo, I'm losing money from this, what the hell is happening. I had nothing to be telling them, right? My team was also pretty confused and
and in many ways upset, because I was an engineering leader at the time, right? And I didn't have anything to be telling them either. I was just like, yeah, let me get back to you. So after going through that experience myself, I became much more of a decentralization Maxi, because anything that happens on chain is inherently
And rather than just one or two dudes in a room pulling the shots, making decision, anything that happens in a trustless way, in a decentralized way, does have much more transparency. So one of the friends that had hit me up at the time when this entire meme stock kind of stock was happening was my co-founder, Jeff.
He was just checking in on me and then a little bit after that both of us basically had a conversation talking about what happened with Robinhood and we realized that if Robinhood had been built in a decentralized way, things would have been a lot cleaner. So that was the original inspiration for us and we set out starting to build a decentralized exchange.
in 2021. So we're going to build a Dex and this led to us looking into all the infrastructure that we could use to build a Dex. We realized the infrastructure was lacking and that pushed us to start building safe. So say is a general purpose lera went for trading and yeah that's necessary for how we got started.
That's cool. And it would be great if you can tell the audience what you guys have been
building. I can what is this I for?
Of course, so CAA is a general purpose layer one for training. And internally, it's a, we have just one core thesis, which is that the fundamental use case of blockchain is the ability to exchange digital assets.
Right. Every successful app in crypto right now is either directly or indirectly a trading application. For example, if you look at Uniswalt, if you look at OpenC, those are pretty clearly trading applications because you're trading either a fungible or
non-punjable tokens on them, right? If you look at Steppen, Steppen is also a trading application. So Steppen is a game, but one of the core components of the game is their in-game marketplace where you can go and trade in game assets, right? And
That makes it the trading application by definition, and that's where a lot of the activity for stuff in comes room. Even metamask is indirectly a trading application because the primary reason that people go and install metamask is to then be able to
go on chain and speculate on different cryptocurrencies on chain. So even putting aside this swapping component for Metamask, Metamask is also a trading application.
So exchanges are the most critical application in crypto. And this is not just true for DeFi. That's a common misconception that people have that trading is DeFi focused. That's absolutely not true. In reality, trading is critical for DeFi, NFTs, games, and every other type of application that is
currently being built in crypto. There's also several tailwinds that are pushing more and more activity to start happening on chain. First of all, there's a much more, I guess there's much more focus on regulation.
with everything that happened with FTX and everything that is also happening this week from Binance and Coinbase as well. The end result of any government action over here is going to be that the training activity, the desire for people to trade isn't going to go away.
and end is not going to suddenly go straight down to zero. But the places where they can trade and see decentralized ways, those are going to become much more restricted. So there's going to be a portion of this trading activity that's happening in centralized exchanges that is going to then start happening on decentralized exchanges. And we're already starting to see more of this happening right now.
Beyond that, just year over year, there's more organic on-chain activity happening. There's a greater focus on tokenization. There's more digital assets that people want to be trading, which is leading to a greater need to exchange these digital assets.
Clearly exchanges are important. There's more and more demand, more tailwinds to start using these exchanges on chain. The question then becomes how do you help these exchanges grow? How do you help them scale? And that's the core problem that we, let's say,
helps all. How do we solve the exchange scaling problem? And the core promise that we make to teams is that you can focus on user acquisition and say we'll focus on building the most performant infrastructure for you because say is a general purpose layer one and we're
is purely on building the infrastructure. And this core promise has resonated with a lot of teams at this point. There's over 150 projects that are building on say and yeah say is gearing up for a more public main at launch in the future.
Thank you so much for giving the information about the mental information I'd say. And you describe say as optimized like layer one or
trading and exchanges of digital access. So is it just for design for trading and exchanges and whether you have like more use case?
So, say is a general purpose, uh, layer one. It supports every single type of application that you would see in any other ecosystem, such as Ethereum or Solana. Um, with that being said, there are optimizations that have been made at the chain level.
been enabled better performance through the consensus mechanism, through the protocol improvements such as parallelization, and also through the native matching engine that give exchanges building on say substantially better experience than building anywhere else.
Thank you so much. Where do you see trading ahead in towards over the next few years?
Yeah, so in the case of trading, I would say that there's definitely going to be significant tailwinds that help push trading activity to be happening on chain and I mean kind of taking a step back, right? Like I was at Robinhood. I saw how things are incredibly
just bad when there's decentralized entities that are responsible for essentially every step of the financial journey, which is what the current financial system is like. So what I think is going to be happening is that they're going to start being more and more activity that's happening in a decentralized way, just because trustless
is one of the most fundamental value propositions that anyone can rally behind. And I think on chain activities point to start permeating many more parts of the financial stack in the future. And I mean, that's what kind of stood out to us at the very start when we got started.
And I guess a little bit more about our journey when we initially got started. So we were trying to build an exchange rate. This led to us looking into every layer one, every layer two, and all of the other infrastructure that we could use to build an exchange.
And that's when we started seeing what we called internally as the exchange trial model. So between decentralization, capital efficiency and scalability, every exchange out there right now is only able to get two of the three. Right. So any centralized exchange,
is not have decentralization by definition, but building things to a centralized way, you do get scalability and you do get capital efficiency. Now, on the other hand, if you build something like you must all be two, you get decentralization and you get scalability, but you don't get capital efficiency.
This lack of capital efficiency is what led to uniswap creating uniswap v3. And uniswap v3 gets capital efficiency and decentralization, but it's unfortunately no longer scalable because it's much more complicated. So we don't think the solution here is to keep iterating on exchange
mechanism design, like going from minus 12p2 to v3 to v4 isn't the solution. That's what people have been trying to do for the past several years. And I mean, it has led to improvements, but it's not really the solution. Instead, we think the solution is to do a rewrite of the underlying infrastructure, because that's how innovation typically
works. There's this idea of application infrastructure cycles, which are basically that there's some new infrastructure that gets created. This leads to new applications. And then some of these applications need much more specialized infrastructure to help them grow. And
We think that we saw something like that happening in the database industry. For example, you started off with Oracle, which then led to all of the web one and web two applications that we use. And now we're starting to have things like Databricks Warehouse, which is Hilaramade for AI and MLG.
And we think a similar thing is going to be happening with lockchains starting off with Ethereum. This led to all the decentralized applications that we've used in the past several years. And now for the ones that have found product market fit like
exchanges, there's going to be much more specialized infrastructure getting created. So that's the approach that we took with today. That's why we decided to build say where we optimize every single part of this stack to help give exchanges the best possible experience.
Great, thank you. And also, we would like to know, like, why did you decide if you would say as a layer one rather than a layer two?
Yeah, yeah, that's definitely a great question. So yeah, I mean, the way that layer 2's work, especially when we initially about started with this is there's some computation that happens off chain. And then afterwards you take the transaction data that happened off chain on
that roll up, you compress it, and then you write it to the base layer, whatever the data availability layer is. In the case of Ethereum right now, the way that roll ups work is you have to write this compressed call data, but this compressed data is called data to Ethereum BL1.
This means that the bottleneck for performance is throughput. Like, is how much data you can actually write, which limits the amount of throughput you're able to get. And the reason for this is that every byte of data that is written to Ethereum costs 16 gas.
And there's a target of 15 million gas per block, right? So if it costs 16 gas, if you do the map around it, if the rollup has only Transverse of ether happening between different accounts, you'd be able to get around 6000 TPS if all of the block space of
the layer 1 is used only for relapse. In reality there's going to be other types of activity happening on the layer 1 as well. There will be open c, it will be uniswap, there will be a ton of other things that will be competing for boxlaces. So the actual throughput that you've been able to see
of Dinksharding and then Protodinksharding. Protodinksharding isn't really going to help very much, at least from the throughput side because it allows you to add in these blobs of data every single block. But the throughput ends up being closer to 6.7 KTPS. So roughly a 10% improvement, which is
good, but it doesn't really solve this scaling problem. And Dinksharding, which is a much more long-term vision, is extremely complex, but if it is implemented, it might lead to somewhat higher performance from Ethereum sides. But overall, we don't think that
building a roll-up is scalable at all right now. So the better solution is to just build a layer one and you're able to get much better performance and you can focus on really solving those core problems around the exchange tralema and help exchange a scale by building a specialized layer one rather than building a layer two.
Thank you for sharing. I believe that there's quite a lot of trade-offs and considerations when you are building say. So I would like to know how do you think about decide decisions and trade-off as your team building say?
Yeah, so when we got started building a little one, we obviously looked at what is there on the market right now and we saw what other chains like slona, avalanche, ethereum were doing also saw what app toss and sweet were
planning to do at the time. And it became clear that there were some core issues across all of these layer ones that were particularly negative for exchanges. So some of these issues included the time to find out the throughput
and lack of specialization. So time to finality is how long it takes for a transaction that is added to a block to actually become part of the canonical chain, right? So for it to be finalized. And it's basically, I mean, from a, from
an exchange point if you are similar to the latency or it trades to be finalized from the web to world. But basically the lower the latency there is over here, the better the user experience ends up being. And in other chains like Ethereum for example, on Solana as well, there's this concept of multi-slot finality.
So rather than a block being added to the network and being finalized, you need to have a chain of blocks and the longest chain is the one that ends up winning. And if one chain is there and then another chain becomes longer, then the first chain will get re-orbed and it'll no longer be part of the canonical.
chain. And we also are particularly bad for exchanges. And the reason for that is that, I mean, let's say you're a market maker and you want to open a position on Ethereum and then you want to hedge it off chain, let's say on a centralized exchange. There's two ways that you could do this.
The first way you can do this is by opening the position on Ethereum and then immediately as soon as that lock is added to the network, so before it's finalized, hedging that position option. If you do that, the downside is that your Ethereum position
the block that your position was in could get re-orbed. And then afterwards, that trade might not get filled on the Ethereum side or it might get filled at a different price. And then you as a market maker are taking on losses from that. So the way that you account for these losses and this additional risk
is by offering wider spreads users. And that just leads to a worse trading experience. The other way that you could do things to the market maker is by waiting until the block is finalized. If that takes 60 seconds, then you're just going to wait those 60 seconds and get exposed to that price volatility from those 60 seconds.
If during that one minute window of time there was a 10-bit change in the price, that's losses that you are taking on. And there's essentially the way that you would account for that is by offering Writers, Writers users as well. So rather than using a multi-slot
a functionality approach, which leads to complications like that. Say use is to enter both consensus. So a couple of really big benefits from doing that. The first benefit is that it has single slot finality. So for a block to be added to the blockchain, it needs to go through full consensus from the nodes in the network.
And what this means is that you're able to get, you don't really have the concept of review on getting more. So you're able to get a block added and then it's finalized immediately. So from the perspective of a trading application and from the perspective of market makers, this is huge, right? Because suddenly you're able to offer much tighter support.
because there's less risk you need to take on from the consensus mechanism design. The second thing that's really substantial about twin turbo consensus is the time to finally. So we're able to get around 450 millisecond time to finally in the current public
test net. This is really, really fast. Because for context, most Clare1 blockchains right now that actually go through decentralized consensus, they end up having multiple second time to finality. Even Sloane for example, that might have 600 millisecond block times.
those blocks aren't finalized for maybe three plus seconds. So it ends up being a pretty meaningful difference between twin-turbo consensus and the mechanism, the consensus mechanism, employed by other letter ones. The second thing that's is parallelization.
So we saw that the processing from other chains, especially chains like EVM chains that are single threaded was not performing enough. Because when you're doing things in a single thread way, when you're subconsciously executing, you end up getting much worse performance. And I guess for listeners,
The quential execution is, as the name implies, executing things one after the other. Parallel execution is where you, let's say you have 10 transactions, you can execute all 10 of them at the same time and you're able to just finish executing it maybe 10 times faster because you're able to
the multiple cores on your machine to handle workloads in parallel. So say I did in parallelization as well and this helped say get substantially improved throughput. Say is able to process around 20,000 orders per second, which is a magnitude more than what you see on most other ecosystem who get closer to one to the
3000 orders to can be processed every second. And the last thing they say has is a native matching engine. And the really cool thing about this is because we've added in specializations in the chain level, we're able to batch together trades, which then help us prevent front running. So front running
is illegal in traditional financial markets and front running is a practice where someone with privilege information sees trades that are coming in and then they sell assets to whoever is buying at a worst price or they buy assets at a lower price. So if you're doing sequential execution and every
everything is transparent than preventing front running entirely is difficult. But in case case, because everything is batched together, every trade in a block, so every market order gets filled at the same price within each market. And that makes it so that you can't, automatically, see someone's trade coming in and then buy that asset.
it up before them, sell it to them at a higher price, and profit off of that. So this native matching engine helps event negative MEV like front running. And yeah, just helps improve user experience for market makers as well. Through this concept of order bundling, where a market
make or if they're trying to submit 50 separate trades to object 50 separate markets within one block, they don't need to submit 50 separate transactions. Instead, they could just submit one transaction, hit the gas fees for one transaction, and then have the chain automatically parse all these different trades and route them to the correct markets.
So yeah, those are a lot of the optimizations that we ended up making around today.
Thank you so much. There's very insightful and you talk about lots about trading infrastructures and besides define idea and the other types of apps on say ecosystem.
Yes. So a common misconception is that say is focused exclusively on D5. Like I mean honestly even in the past couple of months I've heard this dozens of times and that is absolutely wrong. Say is a general purpose layer one because its core focus is on
trading. And trading is the most fundamental and the most general purpose use case that exists in crypto right now. So say is not focused exclusively on TFI. Say is a general purpose chain. And with that being said, we're seeing a lot of different types of applications building on set. At this point, there are over
and some of the more interesting ones to highlight would be from more of a gaming and social side. So from the gaming side, one of the projects that's really interesting is Fable. So Fable is an on-chain eSports lead. So it lets people create different teams and then compete.
against each other and win rewards, which I think that's really, really interesting and has really, really strong network effects that could come into play if it's able to take off. Another really cool project is Tatami, which is building an on-chain game publisher. And then for more
on the social side, one of the projects that's really interesting is Senate out, which is building political tokens. So there's a lot of pretty wacky experiments that are happening on say right now. There's all the traditional applications that you would see, like different types of exchanges, different types of core, like de-fiber
But there's also a lot of very different types of applications and say is lower time to finality means a really snappy user experience compared to other chains, right? Like a 450 millisecond time to finality means you're essentially having a web to like experience and that leads to
a very different type of founder that becomes interested in building on set because they understand the benefits of this user experience and how that results in greater engagement and greater retention from users. Like there's been a ton of studies that have happened in Web 2 around how once you pass like the one second mark people just tend to zone out right so if you go to Twitter
You try to like open your, I guess you just wait for the homepage to load. If it takes too long, then you're just going to go to a different website. So people's attention spans are super short. And that's why like having a lower attack to finality just greatly improves the user experience across every single type of project that you'd want to be building on chain.
We're very exciting. It seems that quite a lot of different types of project will come to say to build. And also personally I heard like
like Suzy Swap, they already kind of were coming to say and is there any other apps that you're excited for coming to say?
I would say sushi is the biggest announced one. So I mean for listeners, sushi is one of the biggest exchanges in the Ethereum ecosystem. And this is going to be their first foray out of the Ethereum ecosystem. They're going to be launching a perpetual exchange on say, which is
Because there's not that many examples of Ethereum projects that have explored outside of the Ethereum ecosystem. So I think it's going to be huge. And they're interested in not just building on purpose, but rather building an entire suite of products.
There's a lot of interesting ways for the SAE ecosystem, and so she took a lab rate moving forward. Another one of the larger projects that has launched in TestNet recently is Astroport. So Astroport was one of the biggest exchanges in the Cosmos ecosystem.
And it's now launching on say as well. In terms of other kind of publicly announced things, I guess there's nothing really that I can share right now. But just see on the lookout, follow this A Twitter handle and there will be many more announcements coming in the future.
Let's go look forward to it. And we would like to know what's the reason for say, what's your next step this year or in the future?
Yeah, so the value proposition for say simple. A trading application that is built on say will offer a better user experience than on any other type of infrastructure.
That is the core value proposition that stood out to all the teams building on set. And I would say that it has been resonating with a lot of teams so far.
So the reason that this is super significant is because the experience of building on chain right now is extremely clunky, both from a builder perspective and from an end user perspective, right? Like building a decentralized application is
So it's much more difficult than building a Web2 application. And these are experienced right now from the onboarding perspective and from the actual usage perspective around things like congestion and slowness around the chain.
and around meetings like custody your own fund, for example, those are all very difficult experiences. So what we imagine is that if say successful, the tradeoff between
decentralized applications and Web2 applications will no longer exist. Like if save does its job at executing well, then the difference between a decentralized app on save will be either identical to that of the experience of building that as a Web2 application or as close to that
So longer term, that's what say is really getting towards to help decentralized exchanges to help any other type of application just have much better experience than building on any other ecosystem. Now what this will end up meaning is that say we'll end up becoming the primary place where digital assets and trading
to be will happen. Because if say it does lead to better user experiences, then it's kind of a no brainer to be building your trading applications and your other type of applications on say. So long term, we think that say it will be the primary place where a majority of training activity can crypto into happening if say does a good job of executing.
Thank you so much for sharing. And I believe that for most audience here, definitely, I think quite a lot of us will be very interested in how to get involved with say and kill up. So it would be great if you can share some of the wise.
And how we can. Yeah. Yeah. So from say standpoint right now, the biggest focus point is developers. The way that say become successful is by having really good developers learn about say.
Start playing around with say and then eventually building applications that will become the killer applications for say right So around that one of the things that say recently started offering is an entrepreneur and restrooms program So this is a program we'll say labs
We'll cover your salary for some period of time and be an operating partner for founders as they get their web three apps kind of ID, build out and then help them go to market with it as well and
I mean, four founders, this is kind of a no brainer because they currently has a $120 million ecosystem and liquidity fund. So this is money that has been set aside by investors and market makers to help both invest in and provide liquidity to projects that are building all
So the entrepreneur in restaurants program would be a really good way to get kick started in building on set. They also is going to be offering its first hackathon in a few weeks. And this is going to be an AI focus hackathon.
So it'll be a fun way to explore the design space between AI and crypto and see what kind of either hybrid or fully on chain applications can be built that helps solve some really that helps solve some real user problems by using AI and besides that
I mean in terms of getting involved with say you can follow the core Twitter handle. I believe say yeah say is co-host for this event so you can just click on the say Twitter handle and then follow it and there'll be a ton of announcements happening in the next several months and I guess last
The last thing that I would say is say it does have a public testnet going on right now. So if you're interested in seeing what the actual product is, how things work, you can go ahead and play around on the incentivized testnet. And it would be a good way to just get your hands dirty, what's in.
Thank you. I believe many audience may join and join our test nests and explore the products and the upcoming products on say and also we are really looking forward
to the coming updates of C. Yeah, so thank you so much for sharing all this insights to our audience today and really appreciate our time, Jay.
Thank you for coming. Yeah, thank you for having me on and is there going to be an AMA component over here? Are we going to top off right now? I think that yeah, because I've kind of go through the key questions, I think yeah, there's soupy and
of the Twitter space. So yeah, so thank you everyone for joining by the Twitter space today. So hope you guys have a great day and yeah, take care. Awesome. Thank you for having me on everyone. Bye.