March 21, 2025 • 44 mins
What if launching your own Layer 1 blockchain was as fast and seamless as spinning up a cloud server? In this episode of The Index, Jin Kwon, Co-founder of Saga.xyz, shares how Saga is redefining how developers scale in Web3. Rather than relying on rollups or months of coordination, Saga lets you launch your own Layer 1 chain—called a chainlet—in just minutes.
Jin explains how his computer architecture and finance background helped shape Saga’s vision and why automation is key to building scalable blockchain infrastructure. He breaks down how Saga’s validator system works, how pricing is dynamically optimized through competition, and how chainlets come pre-integrated with bridges and liquidity—all with minimal developer effort.
The conversation also explores the future of application-level sharding, using Uniswap as a real-world example of how splitting liquidity across multiple chainlets could unlock massive throughput without compromising user experience.
If you’re a developer looking for faster, cheaper, and more modular ways to build in Web3, this episode is packed with insights into where blockchain infrastructure is headed.
Connect with Jin Kwon and the Saga team:
Saga: https://www.saga.xyz/
Follow on X: https://x.com/sagaxyz__
Follow us:
The Index (https://www.podpage.com/the-index-podcast/)
X Channel (https://x.com/theindexshow)
YouTube (https://www.youtube.com/channel/UCP-VGLfPNwianYhY-oqq5zg)
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Hey everyone, it's Alex Cahaya from the Index
Podcast.
I want to tell you about Mantis, a groundbreaking platform
that's simplifying the way weinteract across blockchains.
If you're a developer or justinto DeFi, you'll want to pay
attention.
Mantis enables trust-minimizedtransactions across different
chains, letting you trade orexecute actions seamlessly while
(00:21):
getting the best possibleoutcome, all without the usual
complexities.
Imagine being able to moveassets and settle transactions
across blockchains easily, withmaximum value extraction, all
while staying secure anddecentralized.
That is what Mantis is bringingto the table.
Mantis is an official sponsor ofthe Index podcast, and their
(00:42):
founder, omar, and I regularlyhost a new live stream series on
X called Everything SVM.
We have these live streamsweekly, and if you want to keep
up with what's happening in theSolana ecosystem, especially as
it relates to the new innovativedeployments of the Solana
virtual machine, you should tuneinto this live stream.
Check them out at mantisapp andfollow them on X at mantis
(01:08):
M-A-N-T-I-S At the Index.
We believe that people areworth knowing and we thank
Mantis for enabling us to tellthe stories of the people who
are building the future of theinternet.
We'll see you on the other sideside.
(01:55):
Welcome to the Index Podcasthosted by Alex Cahaya, and today
I'm excited to have Jim Kwan,who is the co-founder of Saga,
which gives visionaries thehigh-end tools and support they
need to turn their web3 dreamsinto reality.
Saga is and I'm just going topull this up off of your website
here the complete web3 platformmade for developers, and I want
(02:18):
to learn more about what thatmeans.
But first tell me a little bitabout yourself and kind of how
you got into the space andreally why you're building what
you're working on.
Speaker 3 (02:29):
So I was originally an engineer a long time ago,
studied computer science andthen did my master's at Berkeley
, essentially doing researcharound like computer
architecture and systems design.
Essentially my work was aroundlike silicon, photonics.
So it's like when you replaceelectrical channels with light
channels on chain, how do youre-architect how computer
(02:51):
architecture looks like?
Because a lot of designdecisions in computers have been
optimized for electricity overyou know decades and decades of
design essentially.
And so this was like kind of aninteresting kind of pivot point
where people were like, hey,let's see if we can bring light
channels all the way down to onchain and see if you can replace
that, and then what kinds ofbroader architectural decisions
(03:12):
can you make to change that?
So worked on that during mygrad school years and then went
to Intel and I was doing, youknow computer architecture stuff
for a little while.
I worked in engineering forjust a little bit and then I
quickly realized that I wantedto learn a little bit more about
how the world functions in ahigher level.
So what I learned is, as you godeeper into engineering
(03:33):
generally, you go deeper anddeeper into like a singular
vertical and you get you becomemore of a master of like a
singular kind of topic, and Iwanted to go broader essentially
, and so I shifted into likekind of more of a like a
business role within Intel.
Afterwards I went to businessschool and then I popped out on
the other end as an investmentbanker.
So I did technology investmentbanking for a few years, just
kind of learning about how youknow giant mergers and
(03:55):
acquisitions function in thetech industry and how all this
works essentially as human andas financial kind of numbers in
the background.
And I actually discoveredblockchains and crypto kind of
accidentally.
So my brother is the founder ofCosmos, so he's Jae Kwon.
He always had like reallyinteresting crypto startup
(04:17):
projects that he did, butusually they were like one or
two people or pretty small.
And then he came to me and waslike hey, do you want to come
help build Tendermintessentially, which is the
company that built Cosmos and alot of the Tendermint protocol
underneath?
He was like, oh, there's likeabout 10 to 20 people
essentially working on this.
I was like, oh, this is actuallykind of something serious
essentially, and so I lookedinto that and I had no idea what
(04:39):
any of this was.
I was like, okay, well, itseems like it's something
serious.
He's built something kind ofreal here and so let me fold my
career essentially as aninvestment banker and then I'm
going to come help tenement andjust as like I want to come and
operate a startup.
Essentially, that was like themotivation behind it and then
through that, that was like in2018.
I just like learned about thiswhole world of crypto and
(05:01):
blockchains and all of it, andit turns out my background kind
of in systems design andarchitecture, as well as all the
financial work that I did, isvery relevant for this topic
essentially, and so it slotsright in.
Speaker 1 (05:15):
Yeah, that's exactly the combo you need, especially
if you're going to found acompany in the space as a
technical lead.
You know, cto co-founder typeperson.
Speaker 3 (05:23):
Yep, yep, yeah, so I'm technically not the
technical lead, I'm more of theRebecca.
The CEO likes to talk and callme the master of coin.
Essentially, I do everythingaround economics, everything
around incentive design,everything around kind of
broader strategy of how weshould think about things and so
.
But people like to makeblockchain sound like it's more
(05:43):
complicated than it is.
It's really just a database,right, and it's a distributed
database that's not controlledby one person, and there has
been architectural decisionsthat have been made over tens of
years to optimize and makethese things better, and so
everything that people talkabout today whether it's DA,
whether it's like all theselittle like buzzwords that
people talk about it's beendiscovered and already kind of
talked about in the web 2 world.
(06:05):
It's just now done in a muchmore decentralized fashion, and
so from my perspective, it'slike bringing all those kind of
knowledges of computerarchitecture and also bringing
kind of the financial side ofthings from my banking days
essentially, and bringing it allin into one cohesive package is
like kind of a dream come truein my in my eyes, and so that
that's what's really beenexciting for me is like how do
(06:25):
we like bring all theseknowledges from these industries
that has gone through this kindof motion before, and how do we
make it so that crypto isbetter, such that we're not just
reinventing everything fromscratch?
Speaker 1 (06:37):
essentially, well, it's really interesting because
I feel like it's rare that teamshave someone like you on their
team, especially, you know, inyour position, with the
intersection of experience andskills that you have.
Because, yeah, blockchains arejust databases, but the game
theory around how peopleinteract with that and the
(06:58):
programs that are running onthem is not simple, and being
able to identify the attackvectors and even think about
that when you're talking aboutsomething like token design
takes a lot of skill andknowledge.
Right Like I'm.
I'm CEO of my company of ABKlabs, but I'm not the guy who
can tell you what the attackvectors are for something like
(07:18):
that.
I mean, I can talk about themintelligently after my partner
in CTO has told me, you know,these are the things that we got
to like prepare for and designaround and this is how we do an
incentive correctly and this ishow we create a stick, you know,
to ensure performance orsomething like that.
And I do find that it's likeoften people are so focused on
(07:39):
the protocol and like the actualtech or they're really focused
on like the business and like goto market and they don't
consider the token design andhow and how it fits into the
technology and the go to marketCause.
It makes a ton of differencethe way you do this.
We have a sub show calledeverything bagel and it's all
about the intersection of opensource software, ai and crypto.
(08:01):
It's specific to like AI andopen source and crypto, and we
had Greg Osuri, the founder ofAkash, on.
I don't know if you know him,but you must yeah, you must know
each other, because we'retaking
a connection Awesome guy.
So so that was one of my firstangel investments in crypto, by
the way, was into Akash, and Isaid this on the show yesterday.
It was one of the best angelinvestments I've ever I've ever
made.
I was betting on Greg, and onething I love about Cosmos is
(08:24):
just like the hardcore opensource ethos.
I live and breathe that.
My partner, brian I was tellingyou before the show helped
start the open source movementback in the day, and it's like a
core thing.
But anyway, another reallyinteresting thing and you've
been here for a long time, likein the space, really long time
is like the design space hasshifted and there's a lot that's
known now that wasn't known.
Like when a cache launched, thethings they were doing around
(08:47):
incentive design were brand new.
That was like we're going towe're going to fuck around and
find out if this works, and itworked, but it could have not
worked right, it could have notworked disastrously, and we've
seen other ideas like completelynot work.
So can we talk about thatthough?
Like what is the thing you guysare building?
How does Saga help Web3developers?
(09:08):
It sounds like it like you candeploy multiple L1s is kind of
what I was reading.
How does that work?
And then I'm really interestedto dig into the let's like use
your example on talk.
If you can talk about tokendesign and stuff, I'd love to
just like dig into your brain alittle bit about how you guys
got from the idea to like youknow what's getting executed and
tested out and talk aroundthings you're confident about
(09:30):
versus like I don't know, fafelfor this idea.
You know we're going to seewhat happens.
Yeah, yeah, yeah.
Speaker 3 (09:36):
Well, what you bring up is like really the crux of
how I deal with everything aswell, because, as much as I
would say, hey, I have thisintersection of these like
knowledges or whatever I don'tknow everything and the
landscape of how things shift.
There's no way I can knowenough to predict everything
that's going to happen.
So what's important in thisspace, in my opinion, is the
(09:59):
experimentation and the pace atwhich you can kind of try things
out and pivot your direction,and so that's always kind of the
way that at least the mantra ofwhat I follow, which is let's
try something, if it doesn'twork, we can scrap it.
If it does work, great, likewe'll keep making it better
essentially, and at the end itwill feel like we designed this
exactly the way it should befrom the beginning.
(10:20):
But it's not.
It's like it's an organicprocess that every product
designer really goes throughright, which is like you just
treat the token economics as aproduct and you iterate, you
constantly iterate, you make itbetter, you improve it, you
transition it, and that's theway that at least I operate
within Saga ultimately, but Iguess we'll talk a little bit
more specifically.
(10:40):
So one of the things that we'retrying to do.
And it's really hard when Idescribe what Saga is to people
because they have all these likepre-existing kind of biases of
how scaling kind of functions inthis world.
So today I would say there'stwo kinds of scaling,
predominantly One scaling iskind of what Solana is doing,
(11:02):
which is, hey, I'm going to makemy one fat pipe as fat as
possible.
Essentially that's like theirapproach.
And then there's, like in theworld of Ethereum scaling it's
like this like quasi horizontalscaling thing that they have
with roll ups and like DA andstuff like that.
But no one really talks aboutlike when I say we're more of a
(11:22):
horizontal scaling architecture,everyone just assumes that
we're like, oh, you're justdoing what Ethereum is doing,
which is rollups and L2s and DAand stuff like that, when in
reality we are horizontal.
But we like the fundamentalapproach of how we think scaling
should happen in blockchains isactually very antithetical to
the way that Ethereum is doingit, and so I'll just like float
(11:44):
that out there and then let medescribe kind of what Saga is.
So Saga is a layer one thatlaunches layer ones, and
fundamentally it's like kind of,I would say, what Cosmos was
supposed to be.
But the issue with Cosmos ingeneral is that to launch a
layer one using Cosmostechnology takes, I would say,
somewhere between six to 12 to18 months of coordination time
(12:07):
and effort.
To bring a chain up,essentially Develop it.
Bring the chain up, get enoughpeople excited about it so that
you have a staking token thatpeople have value for you
attract validators.
It takes about six to 18 months.
That is not enough time tohorizontally scale applications.
Your turnaround time for how tohorizontally scale an
application should be minutes,should be seconds.
(12:29):
This is how AWS works, right.
Like when you go to AWS and say, hey, I want two cores.
You run some website and you'relike, oh, I need four cores.
You just go to AWS and be like,hey, I want four cores and it
like instantly comes essentiallyright.
It's not like as if it says,hey, I want four cores and
suddenly Amazon says okay, we'regoing to need to go and build a
data center that has two morecores over there.
It's going to take you like 18months.
We're going to charge you like$800,000.
(12:50):
Like it doesn't work like that,right, and so fundamentally.
Speaker 1 (12:53):
Those are the security audits, yeah.
Speaker 3 (12:58):
Yeah, and a security audit.
There's like so many thingsthey don't have enough people to
even do it.
Speaker 1 (13:01):
They're like, yeah, pay us a million and we'll look
at you.
Speaker 3 (13:04):
Yep, yep, yeah, yeah, exactly.
Because they're so busy Exactly,and this is true all throughout
every stack.
So this is in block production,this is in, like bridges, for
example.
It's the exact same thing.
You go to like layer zero orsomebody and you say, or Axelar,
(13:26):
and you say, hey, I wantbridges.
They're like, oh, go into thisBD pipeline.
We'll talk to you in like eightmonths, essentially.
And so our perspective is no,this is not efficient.
We need to automate all ofthese pieces.
And so automation and makingessentially like a product out
of automatically launching thesekinds of chains is
fundamentally what the beginningof Saga was.
Let's take the Cosmos stack andsimplify it such that somebody
(13:48):
can go on a chain and requestusing a transaction saying I
want this kind of chain, andthen in the background,
everything comes online thebridges, the validators, the
nodes, everything.
And then you just get kind ofwhat you configured in the front
end essentially.
So that's like, in a technicallevel, what Saga is.
Yeah, no, I really love this.
And then you just get kind ofwhat you configured in the front
end essentially.
Speaker 1 (14:06):
So that's like in a technical level, what's up.
Yeah, no, I really love thisand it's something you probably
didn't know because I haven'treally talked about this that
much.
It's a core problem that Iactually started working on in
June in the Solana ecosystem.
So I was at the SolanaFoundation before and I left
because they're to start workingon enabling people to build
permission environments usingSolana as a stack, and what I
(14:27):
very quickly realized is that inorder to launch a Solana
network, you needed really deepexpertise.
It's even worse than Cosmos insome ways because it's not
designed for that.
It's a giant monorepo whereOnza knows their way around the
main team building andmaintaining that.
But operationally to run it asa network or even a validator
(14:48):
for, like, a new person comingin less so much a validator,
cause if you have a certainlevel of skills you can figure
that out, but a whole networkfrom Genesis block really
difficult.
You know it takes a really goodengineer weeks to just get one
up and running and that's liketime that most enterprises or
like other people, like indiedevs or whatever, like trying to
make an app chain or something.
(15:09):
They just don't have that kindof time or resource, and so we
ended up building for ourcustomers this thing we call SVM
kit, which is not as deep aswhat you guys have done.
I can already tell you,basically TLDR, it's like
operational tooling.
We use Pulumi we don't supportthis yet, but you could like use
Terraform or any other kind ofcode is infrastructure, is code
(15:30):
tooling if you wanted to.
But it automates everythingfrom like going from Genesis
block to a full cluster to justlike a single node, that voting
or non-voting RPC node, whateveryou want and like you can
launch a whole network in 10minutes.
And the only reason it takes 10minutes is because that's like
literally how long it takes forlike the infrastructure's code
to get through its processes.
Like the commands are like fivecommands, and so it is a
(15:53):
fundamentally very similarproblem.
And I found the same bottlenecksthat you just described, like
bridges.
There's all these, all SVMchains and like how are they
going to bridge?
And oh wait, we have to waitlike a couple months for layer
zero or wormhole to like supportus, like that's not a viable
option.
There are some people and it'sactually all come from Cosmos,
right, the Mantis team.
(16:14):
They built an IBC connectionfor Solana, for like an all SVM
or network extension or L2,whatever you want to call it.
To like Solana to Solana or IBC, to like Cosmos to Solana.
They built a bridge using IBC.
What I'm really curious about,though, is how does the
validator stuff work for youguys, because that's a piece
where I was like man.
Wouldn't it be amazing if Icould just launch a Solana
(16:37):
network and all the validatorson mainnet were just like
validating my network?
Yep, that's exactly how itworks here.
How does that work from a tokendesign Like?
Walk me through that.
Speaker 3 (16:46):
Yeah yeah.
So what you described isexactly how it works.
We're taking a differentapproach than most other layer
ones.
So a lot of later ones,especially Solana and Ethereum,
even Cosmos they're like let'smake the validator set ginormous
and make it as big as possibleis, when you employ something
like what you just mentioned andyou want to replicate that for
(17:08):
every kind of like sub chainthat you launch, you're
multiplying every infrastructurecost and effort by 100, 200,
50,000, whatever number ofvalidators you have.
So, from the very get go, rightnow in our main net we only
have 21 validators.
We're going to increase thateventually, but we need to get
(17:29):
all of the infrastructure piecesready to automate all this with
21 first, and then we willstart opening it up to hundreds,
hundreds of thousands orwhatever, if you wanted to.
And so generally, the way itworks is the 21 validators that
are validating our mainnet areessentially watching the mainnet
for these transactions, saying,hey, launch a chainlet.
The one big problem essentiallythat ICS and the interchange
security which is like kind ofsimilar to what we're doing and
(17:50):
Cosmos has is there isn't adefined economic structure that
is set to launch an interchangeblockchain.
Essentially, using Cosmos, youhave to individually negotiate
exactly what the rate is thatyou're going to pay, and a lot
of the time it's like uncertain,it's like it's like in other
tokens that you don't know.
Our model is we universalizethis whole thing.
(18:12):
It works exactly like AWS,where the validators come
together and they like kind ofset a price.
Essentially, let's say it's youknow, 10 saga tokens per month.
I don't know what the it'sprobably a lot more than that,
but they all come and agreethrough an auction amongst the
validators saying, hey, I'mgoing to set the price and
universally, this is what theprice of the chain.
That should be per month,essentially.
(18:33):
And the only thing the user hasto do or the developer has to
do is acquire that many Sagatokens and put it in an escrow
account and say launch.
And then essentially itlaunches that chain and the
escrow balance reduces over timebased on how much, what the
rate is essentially on a dailybasis.
Oh, that's super interesting,yep.
And then when the escrow runsout, the validator shut down the
(18:55):
chain essentially, and then youcan add a future time if you
wanted to bring that back onlineby paying for the escrow again.
Speaker 1 (19:01):
Essentially, Does this mean that if I'm a
developer and I'm building myown chain I've got some custom
features to it right, likedifferent economics things chain
, I've got some custom featuresto it right, like different
economics things and I've got myown token my chain could
generate fees.
I could use some of those feesto buy saga tokens, automated
right, and pay.
The validator set yep, yep,that's exactly right, and then,
like this is exactly how itshould work.
Speaker 3 (19:23):
You figured it out exactly like and also like this
is kind of okay.
This is like if you squint alittle bit, this is kind of okay
.
This is like if you squint alittle bit.
This is kind of how Polkadotworks, but it's flipped Minus
the pair of Jane auctions yeahit's flipped in our model where
the validators are bidding, andso the way that our protocol
functions essentially is iseveryone bids some price, saying
hey, I can validate a chainletfor X dollars per month,
(19:46):
essentially, and what theprotocol does is it goes across
all the validators and thenselects the cheapest 21,
essentially, and then whoever isthe most more expensive than
that point just kicks everyoneelse out essentially, and then
the way it functions is theprice is set to the most
expensive person within that 21set, so everyone who bid gets
greater than or equal to theamount that they bid essentially
(20:08):
, and so this auction model isnot in yet Right now.
This is like the model thatwe're building towards.
Speaker 1 (20:13):
It's interesting because there is a Web2 example
of this and it's a classiccustomer validation, market
validation tactic of A-B testingpricing.
The rule of thumb when you'redoing that is you're kind of
flipping on this idea and saidbut the rule of thumb is like if
I'm selling you a t-shirt and Isell a thousand and I AB test
$10 for $15 or $15, if themarket picks $10, then I refund
(20:36):
to the $15 people.
If the market picks $15, though, all the $10 people get $15.
Yep, you know what I mean.
It's kind of the same thing.
Speaker 3 (20:45):
Yeah, that's exactly right, and so, from our
perspective, in the long run, weactually don't think this is
going to be the value driver forSaga.
So what we want to do is createthis economic model where
validators are constantlypushing down this price as low
as possible because they'recompeting right.
Essentially, the ideal positionwe have is there's like 25
(21:05):
people constantly fighting for21 spots, if you kind of think
about it this way, and everyoneis just constantly trying to
drive like, oh, I think we canessentially offer this a little
bit cheaper.
And let's say, you lost a bid.
Essentially, you're like, hmm,how come I can't run my
infrastructure cheaper, butthese guys, these 21 guys, can.
So you go back in there and yousay, hey, how do I make this
(21:25):
more efficient?
Let me fix this a little bit.
Let me fix the costing.
Oh, I think we can get it downto, you know, a lower than that
amount.
Then you go and bid back inessentially, and this process
continues essentially, and itequilibrates to a position until
it gets to a point where youchoose the 21 cheapest
validators in the world,essentially the most efficient
(21:45):
operators ultimately.
And this is the model that weare like pushing towards, which
is, how do we make block spaceas cheap as possible?
Because most chains out theretheir economics is such that
it's like gas based, which meanshow do I maximize the amount of
money that I can get from users?
It's like very flipped.
We believe that infrastructureis a commodity item.
(22:07):
There's no such thing asinfrastructure that needs
bidding to win and to driveprices up.
It should be the inverse,ultimately.
Because hardware is cheap, awsis relatively very cheap in this
world, and so why is it thatwe're paying a hundred times
multiple of what the cost of theactual infrastructure is, when
(22:28):
you can set that price lower andyou can extract value in other
directions?
Speaker 1 (22:32):
essentially, ultimately, Well, it depends on
the large zone.
The hard respect too, though,right Like Solana, the hard
respect is high, it's expensiveand, you know, as the network
has grown it's gotten harder,especially for RPC, for
Validator less so, but still hasgone up.
But if you're running RPC nodeson Solana it's tough and that's
actually a core problem thatABK is working on solving.
(22:53):
We're building a protocol and anetwork to drastically reduce
the cost of running RPC onSolana.
One of the other problems thatwe identified in Solana, that's
Solana specific, and I'm justsharing it with you because I
think you'll think it'sinteresting.
And Cosmos largely solved this.
And this is where I think likecross-pollination in an open
(23:14):
source manner between, likeCosmos validator tech and I
forget what you guys call thatif it's the Cosmos SDK,
technically speaking or what,but anyway Cosmos validator tech
and Solana.
It's a monorepo right With amillion lines of code, and as
soon as you start launchingother chains and making
modifications, you diverge fromthe mainnet Agave repo and it
very quickly becomes nearimpossible to maintain it and
keep up.
So now you become Anza and youlose all the security benefits
(23:39):
of, like an open source projectvery quickly and your upgrade
path, like if there's a zero daybug, anza patches it.
You're fucked because you'renot going to be able to patch it
.
Speaker 3 (23:48):
Yeah, yeah, yeah yeah .
There's so much to fixessentially that you've modified
.
Speaker 1 (23:52):
Exactly, and this is largely the path that Linux went
.
Okay, we believe that you need amodular client and what we
proved you can do is you canstrip out.
There's these things calledpolicies that define how the
validator behaves.
You can strip them out and makethem plugins that live on the
chain and they load from eitheryour site, your chain, like your
fork, or they can load frommainnet, like however you want,
(24:15):
and then, once you get that inplace, then you can potentially
start doing some of the stuffthat you're talking about with
Saga, where, like a node on anetwork because it's just
running the kernel of Solana canvalidate multiple chains,
potentially as a single node ormaybe like a cluster of nodes,
but it can like I don't knowexactly how the architecture
would work out, but that waslike that's our vision, like
(24:37):
long-term, abk Labs wants tobuild that as an open source
project, not even.
I mean, I think we can make alot of money doing it.
I think there's some tokeneconomic structure we could do,
but really just because we thinkit needs to exist.
Speaker 3 (24:48):
Yeah, yeah, yeah, I agree, but it's one of those
things where Cosmos started asthis modular framework,
essentially there's problemswith Cosmos too.
There are modules out therethat literally every Cosmos
chain uses, because once you tryto modify that one thing, you
have to modify like 15 modulesin the background, and so
(25:10):
there's other issues as well.
Speaker 1 (25:11):
Yeah, so it starts to feel not so modular at that
point.
Speaker 3 (25:13):
It's not so modular because, like everything has
like these interdependencies inthe modules essentially.
But, as I said, it's a goodgoal to have at least.
So, like the nice thing, it isreally nice building in this
Cosmos world because from ourperspective it's like whenever
we generally think, hey, likethere's this like thought of how
to improve like certain partsof our stack Somebody has
(25:34):
already done it before we havelike lots of reference
implementations that we can gothrough and see, hey, is this
what we want, is this not whatwe want?
And compare and actually pickand choose essentially which
implementation is closer to whatwe want, and then we can also
modify ultimately.
And so that's kind of the nicething about cosmos is it's so
decentralized.
And this like entire modularframework has existed.
(25:55):
I think, like with solana, asmentioned, like the, the goal
from the very beginning was likeone fat pipe, essentially right
.
So it's like, yeah, the fatpipe's awesome, but if you want
to make parallel pipes, it'slike it gets a little bit more
complicated, right I think oneof the interesting things I want
to point out about Solanathough.
Speaker 1 (26:11):
Yes, that was the goal right Blockchain at the
speed of light.
You know NASDAQ, you know onchain right.
But one of the cool thingsabout Tully is he is an open
source maxi right.
Like he believes in open source, and so his, I think his vision
has always been, and he hasencourage people to experiment
with Solana and do differentthings, and I think it's exactly
(26:32):
that ethos that will lead to animproved, more modular Solana.
That will improve mainnet.
It will enable, like, if you gothe way of Linux and you have a
modular stack, it actuallymakes it so that, like, you can
get more and more granular, andthat's what you have to be in
order to optimize, to make thosepipes bigger.
You know, ultimately, I thinkmaybe I'm totally wrong, but
that's my intuition slash,experience from other people who
(26:54):
you know were there when Linuxwas built- you also want like,
more than just Anza, like doingthis right.
Speaker 3 (26:59):
You want, like multiple people looking at this
at the same time.
Speaker 1 (27:02):
Well, and they are now and that's the cool thing
Like it is really cool thatthat's happening.
There are many clients nowbeing built, many really
interesting talented teamslooking at it.
One thing I'm really excitedabout this conversation with you
is just how we have we havelooked at the same problem, just
from different ecosystems andprobably with some different
needs, which it's like cool totalk to somebody who really like
dug into that, like I do thinkit's going to work.
(27:24):
Like I don't know how far alongyou guys are in this experiment
, but everything you're sayingto me it's like you've looked at
the problem correctly.
You thought about it from firstprinciples.
You designed a system thatcould work.
How far along are you now?
Like where are we today?
On whatever?
Today is sometime in February,mid February.
Speaker 3 (27:42):
So right now the basis is all there, so you can
go right now on saga, appsagaxyz, essentially put a transaction
in to launch a chain.
Uh, permission, if right nowit's permission, it uses this
like test token called psaga,that we control the supply of,
because we don't want right nowthe economic activity of
(28:04):
actually like paying for thechain is not implemented yet.
But as long as you have youknow these psaga, you just ask
us for it, we'll justimplemented yet.
But as long as you have youknow these P Saga, you just ask
us for it, we'll just give it toyou.
As long as you have that, rightnow it takes about two minutes
to launch kind of a chain.
It's fully decentralized, itcomes with an RPC endpoint, it
comes with an explorer and thechain kind of comes online.
We don't really do anything, itjust happens magically amongst
(28:26):
all the validators that we have,like for mainnet essentially,
and so the thing that's a littlebit more complicated with this
is so in the beginning what wewere thinking about was okay,
people will launch their ownchainlets, the projects that are
building on Saga, and they willattract liquidity themselves.
So you know, our big focus forstep one was just the validator,
(28:48):
the RPC endpoint, some of theselike very infra heavy things.
And then what we quicklyrealized is, like this liquidity
issue and the fact to bringlike bridges and all that is the
exact same problem that we hadwith the validators, essentially
.
And so we're in this stagewhere we're kind of launching
mainnet, essentially 2.0, wherewhen you launch this new
chainlet, it actually comes withwhat we call liquidity
(29:11):
integration layer, which isshort, for you can shorten it to
LIL.
We call it LIL and essentiallyit's like a bunch of not only
when you put that transaction into say, hey, I want to launch
my chain, it launches the chainitself with the validators, but
it also stands up all thebridges that connects it to
every other chain that and toconnections to like all the
(29:34):
Ethereum ecosystems, but also itautomatically stands up all of
the infrastructure needed toautomatically route between them
.
So you don't have to go to likesome bridging website and
bridge between the chain.
That it's all automatic.
Yeah, that's really interestingwebsite and a bridge between
the chain.
That it's all automatic.
Speaker 1 (29:47):
Yeah, that's really interesting.
So it's like Jupiter swapswhere it like routes everything
automatically, but you're doingacross multiple chains.
Yep, do you guys use Relayer?
Do you know what I'm talkingabout?
Is that who you use for someparts of this?
The bridging piece?
Speaker 3 (30:01):
Our bridges because we're using Cosmos is all just
IBC and essentially in thebackground we have all the
infrastructure that spins up therelayers needed to make the
bridges all function properly.
So it's like a little bitcomplicated.
Speaker 1 (30:14):
So the way it works is Okay.
So it is a Cosmos fork.
Speaker 3 (30:18):
It's your building, it is Cosmos.
Speaker 1 (30:19):
It's all built on Cosmos.
We never actually explicitlysaid that.
I kind of assumed it.
Speaker 3 (30:31):
Yeah, yeah, explicitly said that.
I kind of assumed it.
Yeah, yeah, yeah, I mean likeI'm I'm wearing, like here I
have a cause.
Yeah, I can't see that.
Yeah, yeah, we're on cosmospeople.
Yeah, yeah, yeah, but but it'slike we're we're like a weird
subset because we're like evmcosmos, so we're like aligned
with a lot of the ethereumcommunities, but our tech in the
background is kind of invisiblyall cosmos, which is like kind
of cool in my opinion, and soall the bridges automatically
get set up.
So let's say, we go and launchan ABK chain you request an ABK
(30:53):
chain In the background, asmentioned all the validators
launches a node, connects to oneanother and then makes that
chain for you as well.
All of the bridging pipelinesto bring up all the relayers
between this new ABK chain, therouter that routes all the
traffic across Saga isautomatically stood up at the
(31:14):
same time.
So it literally takes about twominutes for you to spin up a
chain, have all the bridgesnecessary to go connect to
things that you need essentially, and all of the routing is
stood up automatically as well.
And so what we want to do is beable to for someone to go and
launch this, you know ABK chainand then immediately using our
API, and say, hey, bringEthereum into my chainlet
(31:37):
instantly.
And then you just put atransaction in use our routing
tools.
It will just suck up thatEthereum all the way from
through Axelar down into ourrouting chainlet.
It will relay that routingchain Ethereum all the way from
through XLR down into ourrouting chain that.
It will relay that routingchain that back to your chain
that and that Ethereum will justkind of magically show up on
your chain and this is like alldone.
You can launch this in likethree minutes and you can
parallelly launch 300 of thesein the background and suck
(32:00):
Ethereum into each one of themindividually and it will all
come essentially instantly.
I would say we're like reallyclose to this.
We're releasing liquidityintegration layer that's coming
live like this week, next week.
Essentially, our first instanceof this with Uniswap and Oku is
going up immediately after that, and so really close to all
this.
And then there's some likeimprovements we need to make in
(32:22):
the user experience front tomake this happen.
But, like this kind ofkickstart, a lot of the defy
activities that are happening inin the chains that people have
been asking for is like hey, howdo I get liquidity onto these
chain lists that I'm launching?
Speaker 1 (32:33):
essentially, so what's the scalability like from
a transaction you know tpsthing standpoint for you guys?
Speaker 3 (32:40):
okay, I would say tps is almost irrelevant, but it's
like let's assume, for I don'tknow what the actual metric is,
but let's say we can shove likea small number, let's say a
hundred TPS, onto a chain.
The way that we think aboutthis is an application should
actually start sharding itsapplication across multiple
(33:00):
parallel chains, because if itonly takes two minutes to launch
a chain, essentially you candynamically figure out when your
application is gettingcongested, right, and start
spinning up backup chains in thebackground.
And so this 100 TPS eventuallywill start incorporating a lot
of innovations that arehappening in Cosmospace.
That brings that up to you know, a few thousand TPS easily,
(33:23):
right?
But in the meantime you canjust scale this as much as you
want, right?
And so one example that I'vebeen talking about I just talked
about Uniswap, right?
I'm working essentially withthe idea that we can start
sharding Uniswap instances inthe background in Saga.
So Uniswap is going to go onlive on a singular chainlet
pretty soon in Saga.
But you can imagine like aneasy ways to do this would be
(33:46):
like you take one pool, let'ssay USDC, usdt, and you stick
that into a separate chainlet inthe background, right, and you
can start parallelizing everysingle one of these pools that
you have in Uniswap can now goon individual chains in the
background, right, and then youcan also.
That's like step one.
Step one is just like one poolper chainlet like sharding,
(34:09):
essentially.
Step two is you can startactually sharding the same pools
onto multiple chains.
It's super easy to do.
The reality of it is noteveryone needs the entire huge,
ginormous pool liquidity that aUniswap pool has, right, you can
.
Some people are doing like oneETH transactions and other
people are doing like 100 ETHtransactions.
The one ETH transactions can gointo a Uniswap instance that
(34:32):
has a smaller pool and it won'taffect the slippage as much.
So you can actually startsplitting out one pool into
parallel versions of this.
Like think of it as one smallpool and one big pool.
So now you have doubled theamount of capacity, but you only
increase the liquidity by like5% or something tiny,
essentially Right, and you cando this in multiple times.
(34:54):
So now you can have 10 or 15small pools that people get
routed to for transactions andthen only the big transactions
go to the big pool, essentially,and you can do these like
dynamic routing.
Essentially.
That allows you to shardUniswap almost like infinitely,
and this is like kind of thevision of what we have.
It's like look, this is howscaling should happen in
(35:15):
blockchain.
This is the design philosophythat we have, and it's like how
do we enable this to happen?
And so we're going to startwith Uniswap.
Other stuff is going to happen,and I'm hoping that we can
create an SDK that allows peopleto just parallel shard
applications automatically.
Speaker 1 (35:29):
Essentially, Is the Uniswap like a formal
partnership with them, or didyou just fork it?
Speaker 3 (35:35):
This is the funny part.
No, we're just.
It's not really forking, sowe're going to shard them
regardless if they want to ornot.
Uniswap, the DAO, essentiallyhas ways to deploy Uniswap
contracts like canonical Uniswapcontracts and they own the
actual contract itself.
And so, from their perspective,when there's two Uniswap
(35:56):
contracts that are canonical onSaga, they'll just treat it as
two different chains in theirheads, but like it doesn't
really matter if it's the sameecosystem or not.
Now they have double the amountof instances that they need to
think about.
But they have a process forthis.
So we'll start with a singularUniswap chainlet and then later
(36:17):
down the line we're going toidentify one pool to split out
and say we're going to deploy asecond Uniswap chain.
Then we'll apply again.
They'll say, yes, this is nowthe canonical Uniswap chain,
except now there's two, two SAGAessentially.
And then we'll just keep doingthat and actually, in fact, I
actually prefer not having thispartnership.
Speaker 1 (36:35):
But you still have to ask them for permission to use
it, though I mean Uniswap is anopen source software right.
There's just controls aroundlike that yeah, that bsd license
thing right with like part ofit right, wasn't it?
Speaker 3 (36:48):
and so then it becomes open source later, or no
, no, I'm pretty sure uniswap isfully open source.
Yeah, so it's like anyone canuse it.
Speaker 1 (36:55):
I feel like there was a thing where they changed it
after a while or something.
Speaker 3 (36:58):
I can't like I might be totally off on that yeah,
yeah, but, as I said, like frommy perspective it it's like
anybody who believes has an opensource software, that's a
Solidity smart contract.
We'll just be able to work withthem to deploy one version and
then shard it to like hundredsand different chainlets in the
background, and there's somecomplications, right.
So like the front end needs toknow that when they receive a
(37:21):
like a USDT to USDC liketransaction, to route it to the
correct chain, and so we justneed to put some piping in the
front that allows them tomultiplex that, and and then
it's like pretty simple.
After that, you just, you justkind of shard it to like however
many that you want and then, uh, direct traffic in whichever
direction you want to.
Ultimately, it's.
Speaker 1 (37:40):
It's super cool.
We are at the top of the show.
The last thing I ask everybody,everybody, is you know what?
Have I not asked you that Ishould have asked?
Speaker 3 (37:47):
I really like the fact that you're thinking about
how to make Solana more modularand more parallel, and I don't
think people quite understandthe power of horizontal
scalability People always talkabout, talked about, like
horizontal scalability.
It was existing, like Cosmoshas existed since 2018.
(38:07):
It's technically a horizontalarchitecture, but having it
available as something thatcomes up in minutes is not
something that has existedbefore, and I think we're going
to see a pretty big transitionessentially of how developers
think about applications oncethis paradigm exists in this
blockchain world.
(38:27):
And so I don't know,question-wise, what would I say?
Maybe if you asked more abouthorizontal design.
There's some interesting stuffwe can talk about, but I feel
like we don't have that muchtime.
Speaker 1 (38:39):
Yeah, it's a rabbit hole, I mean.
What I'll say is I think thisis why Cosmos people and Solana
people like each other so muchis because we share this true
open source ethos.
Speaker 3 (38:49):
Yeah, yeah, I love Anatole.
I met Toli before Solanalaunched, actually, and I sat
down in the office in SanFrancisco with him and I think
it was him.
He was explaining how theSolana block production and the
catch-up mechanism for verifyingthe blocks, how there's two
different points.
That moves along in time.
(39:09):
I was like, oh, this is reallyinteresting.
I think ethos-wise we're veryaligned.
It's just slight variations onhow we think we should
prioritize the scalingultimately.
But yeah, I think you're right,it's pretty much the same thing
.
Speaker 1 (39:21):
My point about that is that it just makes these
conversations so much more fun.
Like, the why for ABK Labs isjust build together.
That's what I want to do.
I left.
I was working at the SolanaFoundation.
I had a great job, but I wantedto go build something important
.
Like I was like kind of workedmyself out of the job there and
I loved it there, but I wantedto do is build open source
(39:42):
software with great engineersand people that I like.
Speaker 3 (39:52):
You know how our system is L1 to launch L1s
essentially.
So the way that our tech stackis kind of implemented actually
is it's all containerized,essentially, like the chainlets
are all containerized, and sothe consensus algorithm and the
technology stack that powerseach chainlet actually doesn't
have to be the same.
So as long as each of thecontainers speak the same
language in our case it's IDCand there's like certain other
things that it needs to adhereto but if we can design,
(40:14):
essentially, a Solana stack thatadheres to those languages and
models, we can actually launch afull Solana stack chainlet
essentially and parallelize it.
And so, as I said, it's likeit's exactly as you mentioned.
There needs to be work to makeSolana a little bit more modular
, where it's easier to do this.
But once that step happens, wecan deploy, you know infinite
(40:38):
numbers.
Speaker 1 (40:38):
That would be really cool to explore, cause I and I
think we could do that with you,because we, you know, we've
already got SVM kit, which islike that operational tooling
for the deployment part, but thepiece you have is this like
interop and the chainlet ideaand the economics piece.
That's like the piece that weare really missing.
Nothing we've done has reallybeen about making money or like
(41:00):
building a token or anythinglike that.
It's literally just been like Iwant to build open source
software, we're going to figureout how to fund that.
That's my job.
I'm CEO.
That's like literally how Iview.
My job is fund engineers makingopen source software.
Yeah, yeah.
Speaker 3 (41:11):
That's it.
Yeah, we should explore that.
Yeah, of course, I mean, as Isaid, it's a.
Speaker 1 (41:20):
Did you say it's already built?
There's one built, so there isa way to do a alt SVM chain or
network extension bridge toanother SVM chain, main net or
another SVM chain using theirstuff.
Like it hasn't been, like Ithink there's, just I don't know
if it needs to be.
Is it if it's as automated aswhat you have, where we could
(41:40):
just like easily configure it todo what you're talking about?
Because the thing that you cando really easily right now you
mentioned about like thedifferent consensus stuff and
like the modifications andthings like that, that's the
piece that's still kind of hardon solana, that needs to be like
innovated on, I think, toreally make a solana saga
(42:00):
product yep work yeah, yeah yeah, you gotta you gotta cross the
chasm there and meet in themiddle somewhere to like, really
be like.
Okay, you know we've already gotit down so you can launch
chains really fast you know, butnot where they're like
automatically connected with thebridges and the routing and all
that stuff.
Yeah, yeah.
Speaker 3 (42:18):
Yeah, yeah, I mean, even in Cosmos, which is
designed to do this.
It's like there's it's a lot ofwork to do this.
We've spent significantengineering cycle because, like
there's like really dumbproblems that come when you do
this, like, for example, whentwo chains come online,
sometimes the tool will makelike three bridges that are like
parallel and doing the samething, and then all the tokens
(42:40):
that are going between the threebridges it's like not fungible
with one another, and then it'slike everything breaks right and
so, and so we have to make sureto for it to check that like
there's only one bridge thatit's using essentially, like
there's all these like likethings that that that that come
up that like are problematic toautomate essentially, and so
this is like this sounds to melike something that we can work
together on, which is likereally exciting, in my opinion.
Speaker 1 (43:00):
Let's do it.
That's why I love doing thisshow.
Speaker 3 (43:02):
Thanks for coming on.
I appreciate you.
Yeah, thanks, man, appreciateit and let me know if there's
any other questions and I'mhappy to answer.
Speaker 1 (43:08):
I will do, I guess the last thing is how can people
reach out to you?
Speaker 3 (43:11):
If you just go to our website, sagaxyz, there's all
the contact forms.
Feel free to reach out to ourteam and somebody will contact
you.
Speaker 1 (43:17):
So awesome Sagaxyz.
Thanks, man, bye-bye, availableevery Friday.
Thanks for tuning in.
Hey everyone, it's Alex Cahaya from the Index
Podcast.
I want to tell you about Mantis, a groundbreaking platform
that's simplifying the way weinteract across blockchains.
If you're a developer or justinto DeFi, you'll want to pay
attention.
Mantis enables trust-minimizedtransactions across different
chains, letting you trade orexecute actions seamlessly while
(00:21):
getting the best possibleoutcome, all without the usual
complexities.
Imagine being able to moveassets and settle transactions
across blockchains easily, withmaximum value extraction, all
while staying secure anddecentralized.
That is what Mantis is bringingto the table.
Mantis is an official sponsor ofthe Index podcast, and their
(00:42):
founder, omar, and I regularlyhost a new live stream series on
X called Everything SVM.
We have these live streamsweekly, and if you want to keep
up with what's happening in theSolana ecosystem, especially as
it relates to the new innovativedeployments of the Solana
virtual machine, you should tuneinto this live stream.
Check them out at mantisapp andfollow them on X at mantis
(01:08):
M-A-N-T-I-S At the Index.
We believe that people areworth knowing and we thank
Mantis for enabling us to tellthe stories of the people who
are building the future of theinternet.
We'll see you on the other sideside.
(01:55):
Welcome to the Index Podcasthosted by Alex Cahaya, and today
I'm excited to have Jim Kwan,who is the co-founder of Saga,
which gives visionaries thehigh-end tools and support they
need to turn their web3 dreamsinto reality.
Saga is and I'm just going topull this up off of your website
here the complete web3 platformmade for developers, and I want
(02:18):
to learn more about what thatmeans.
But first tell me a little bitabout yourself and kind of how
you got into the space andreally why you're building what
you're working on.
Speaker 3 (02:29):
So I was originally an engineer a long time ago,
studied computer science andthen did my master's at Berkeley
, essentially doing researcharound like computer
architecture and systems design.
Essentially my work was aroundlike silicon, photonics.
So it's like when you replaceelectrical channels with light
channels on chain, how do youre-architect how computer
(02:51):
architecture looks like?
Because a lot of designdecisions in computers have been
optimized for electricity overyou know decades and decades of
design essentially.
And so this was like kind of aninteresting kind of pivot point
where people were like, hey,let's see if we can bring light
channels all the way down to onchain and see if you can replace
that, and then what kinds ofbroader architectural decisions
(03:12):
can you make to change that?
So worked on that during mygrad school years and then went
to Intel and I was doing, youknow computer architecture stuff
for a little while.
I worked in engineering forjust a little bit and then I
quickly realized that I wantedto learn a little bit more about
how the world functions in ahigher level.
So what I learned is, as you godeeper into engineering
(03:33):
generally, you go deeper anddeeper into like a singular
vertical and you get you becomemore of a master of like a
singular kind of topic, and Iwanted to go broader essentially
, and so I shifted into likekind of more of a like a
business role within Intel.
Afterwards I went to businessschool and then I popped out on
the other end as an investmentbanker.
So I did technology investmentbanking for a few years, just
kind of learning about how youknow giant mergers and
(03:55):
acquisitions function in thetech industry and how all this
works essentially as human andas financial kind of numbers in
the background.
And I actually discoveredblockchains and crypto kind of
accidentally.
So my brother is the founder ofCosmos, so he's Jae Kwon.
He always had like reallyinteresting crypto startup
(04:17):
projects that he did, butusually they were like one or
two people or pretty small.
And then he came to me and waslike hey, do you want to come
help build Tendermintessentially, which is the
company that built Cosmos and alot of the Tendermint protocol
underneath?
He was like, oh, there's likeabout 10 to 20 people
essentially working on this.
I was like, oh, this is actuallykind of something serious
essentially, and so I lookedinto that and I had no idea what
(04:39):
any of this was.
I was like, okay, well, itseems like it's something
serious.
He's built something kind ofreal here and so let me fold my
career essentially as aninvestment banker and then I'm
going to come help tenement andjust as like I want to come and
operate a startup.
Essentially, that was like themotivation behind it and then
through that, that was like in2018.
I just like learned about thiswhole world of crypto and
(05:01):
blockchains and all of it, andit turns out my background kind
of in systems design andarchitecture, as well as all the
financial work that I did, isvery relevant for this topic
essentially, and so it slotsright in.
Speaker 1 (05:15):
Yeah, that's exactly the combo you need, especially
if you're going to found acompany in the space as a
technical lead.
You know, cto co-founder typeperson.
Speaker 3 (05:23):
Yep, yep, yeah, so I'm technically not the
technical lead, I'm more of theRebecca.
The CEO likes to talk and callme the master of coin.
Essentially, I do everythingaround economics, everything
around incentive design,everything around kind of
broader strategy of how weshould think about things and so
.
But people like to makeblockchain sound like it's more
(05:43):
complicated than it is.
It's really just a database,right, and it's a distributed
database that's not controlledby one person, and there has
been architectural decisionsthat have been made over tens of
years to optimize and makethese things better, and so
everything that people talkabout today whether it's DA,
whether it's like all theselittle like buzzwords that
people talk about it's beendiscovered and already kind of
talked about in the web 2 world.
(06:05):
It's just now done in a muchmore decentralized fashion, and
so from my perspective, it'slike bringing all those kind of
knowledges of computerarchitecture and also bringing
kind of the financial side ofthings from my banking days
essentially, and bringing it allin into one cohesive package is
like kind of a dream come truein my in my eyes, and so that
that's what's really beenexciting for me is like how do
(06:25):
we like bring all theseknowledges from these industries
that has gone through this kindof motion before, and how do we
make it so that crypto isbetter, such that we're not just
reinventing everything fromscratch?
Speaker 1 (06:37):
essentially, well, it's really interesting because
I feel like it's rare that teamshave someone like you on their
team, especially, you know, inyour position, with the
intersection of experience andskills that you have.
Because, yeah, blockchains arejust databases, but the game
theory around how peopleinteract with that and the
(06:58):
programs that are running onthem is not simple, and being
able to identify the attackvectors and even think about
that when you're talking aboutsomething like token design
takes a lot of skill andknowledge.
Right Like I'm.
I'm CEO of my company of ABKlabs, but I'm not the guy who
can tell you what the attackvectors are for something like
(07:18):
that.
I mean, I can talk about themintelligently after my partner
in CTO has told me, you know,these are the things that we got
to like prepare for and designaround and this is how we do an
incentive correctly and this ishow we create a stick, you know,
to ensure performance orsomething like that.
And I do find that it's likeoften people are so focused on
(07:39):
the protocol and like the actualtech or they're really focused
on like the business and like goto market and they don't
consider the token design andhow and how it fits into the
technology and the go to marketCause.
It makes a ton of differencethe way you do this.
We have a sub show calledeverything bagel and it's all
about the intersection of opensource software, ai and crypto.
(08:01):
It's specific to like AI andopen source and crypto, and we
had Greg Osuri, the founder ofAkash, on.
I don't know if you know him,but you must yeah, you must know
each other, because we'retaking
a connection Awesome guy.
So so that was one of my firstangel investments in crypto, by
the way, was into Akash, and Isaid this on the show yesterday.
It was one of the best angelinvestments I've ever I've ever
made.
I was betting on Greg, and onething I love about Cosmos is
(08:24):
just like the hardcore opensource ethos.
I live and breathe that.
My partner, brian I was tellingyou before the show helped
start the open source movementback in the day, and it's like a
core thing.
But anyway, another reallyinteresting thing and you've
been here for a long time, likein the space, really long time
is like the design space hasshifted and there's a lot that's
known now that wasn't known.
Like when a cache launched, thethings they were doing around
(08:47):
incentive design were brand new.
That was like we're going towe're going to fuck around and
find out if this works, and itworked, but it could have not
worked right, it could have notworked disastrously, and we've
seen other ideas like completelynot work.
So can we talk about thatthough?
Like what is the thing you guysare building?
How does Saga help Web3developers?
(09:08):
It sounds like it like you candeploy multiple L1s is kind of
what I was reading.
How does that work?
And then I'm really interestedto dig into the let's like use
your example on talk.
If you can talk about tokendesign and stuff, I'd love to
just like dig into your brain alittle bit about how you guys
got from the idea to like youknow what's getting executed and
tested out and talk aroundthings you're confident about
(09:30):
versus like I don't know, fafelfor this idea.
You know we're going to seewhat happens.
Yeah, yeah, yeah.
Speaker 3 (09:36):
Well, what you bring up is like really the crux of
how I deal with everything aswell, because, as much as I
would say, hey, I have thisintersection of these like
knowledges or whatever I don'tknow everything and the
landscape of how things shift.
There's no way I can knowenough to predict everything
that's going to happen.
So what's important in thisspace, in my opinion, is the
(09:59):
experimentation and the pace atwhich you can kind of try things
out and pivot your direction,and so that's always kind of the
way that at least the mantra ofwhat I follow, which is let's
try something, if it doesn'twork, we can scrap it.
If it does work, great, likewe'll keep making it better
essentially, and at the end itwill feel like we designed this
exactly the way it should befrom the beginning.
(10:20):
But it's not.
It's like it's an organicprocess that every product
designer really goes throughright, which is like you just
treat the token economics as aproduct and you iterate, you
constantly iterate, you make itbetter, you improve it, you
transition it, and that's theway that at least I operate
within Saga ultimately, but Iguess we'll talk a little bit
more specifically.
(10:40):
So one of the things that we'retrying to do.
And it's really hard when Idescribe what Saga is to people
because they have all these likepre-existing kind of biases of
how scaling kind of functions inthis world.
So today I would say there'stwo kinds of scaling,
predominantly One scaling iskind of what Solana is doing,
(11:02):
which is, hey, I'm going to makemy one fat pipe as fat as
possible.
Essentially that's like theirapproach.
And then there's, like in theworld of Ethereum scaling it's
like this like quasi horizontalscaling thing that they have
with roll ups and like DA andstuff like that.
But no one really talks aboutlike when I say we're more of a
(11:22):
horizontal scaling architecture,everyone just assumes that
we're like, oh, you're justdoing what Ethereum is doing,
which is rollups and L2s and DAand stuff like that, when in
reality we are horizontal.
But we like the fundamentalapproach of how we think scaling
should happen in blockchains isactually very antithetical to
the way that Ethereum is doingit, and so I'll just like float
(11:44):
that out there and then let medescribe kind of what Saga is.
So Saga is a layer one thatlaunches layer ones, and
fundamentally it's like kind of,I would say, what Cosmos was
supposed to be.
But the issue with Cosmos ingeneral is that to launch a
layer one using Cosmostechnology takes, I would say,
somewhere between six to 12 to18 months of coordination time
(12:07):
and effort.
To bring a chain up,essentially Develop it.
Bring the chain up, get enoughpeople excited about it so that
you have a staking token thatpeople have value for you
attract validators.
It takes about six to 18 months.
That is not enough time tohorizontally scale applications.
Your turnaround time for how tohorizontally scale an
application should be minutes,should be seconds.
(12:29):
This is how AWS works, right.
Like when you go to AWS and say, hey, I want two cores.
You run some website and you'relike, oh, I need four cores.
You just go to AWS and be like,hey, I want four cores and it
like instantly comes essentiallyright.
It's not like as if it says,hey, I want four cores and
suddenly Amazon says okay, we'regoing to need to go and build a
data center that has two morecores over there.
It's going to take you like 18months.
We're going to charge you like$800,000.
(12:50):
Like it doesn't work like that,right, and so fundamentally.
Speaker 1 (12:53):
Those are the security audits, yeah.
Speaker 3 (12:58):
Yeah, and a security audit.
There's like so many thingsthey don't have enough people to
even do it.
Speaker 1 (13:01):
They're like, yeah, pay us a million and we'll look
at you.
Speaker 3 (13:04):
Yep, yep, yeah, yeah, exactly.
Because they're so busy Exactly,and this is true all throughout
every stack.
So this is in block production,this is in, like bridges, for
example.
It's the exact same thing.
You go to like layer zero orsomebody and you say, or Axelar,
(13:26):
and you say, hey, I wantbridges.
They're like, oh, go into thisBD pipeline.
We'll talk to you in like eightmonths, essentially.
And so our perspective is no,this is not efficient.
We need to automate all ofthese pieces.
And so automation and makingessentially like a product out
of automatically launching thesekinds of chains is
fundamentally what the beginningof Saga was.
Let's take the Cosmos stack andsimplify it such that somebody
(13:48):
can go on a chain and requestusing a transaction saying I
want this kind of chain, andthen in the background,
everything comes online thebridges, the validators, the
nodes, everything.
And then you just get kind ofwhat you configured in the front
end essentially.
So that's like, in a technicallevel, what Saga is.
Yeah, no, I really love this.
And then you just get kind ofwhat you configured in the front
end essentially.
Speaker 1 (14:06):
So that's like in a technical level, what's up.
Yeah, no, I really love thisand it's something you probably
didn't know because I haven'treally talked about this that
much.
It's a core problem that Iactually started working on in
June in the Solana ecosystem.
So I was at the SolanaFoundation before and I left
because they're to start workingon enabling people to build
permission environments usingSolana as a stack, and what I
(14:27):
very quickly realized is that inorder to launch a Solana
network, you needed really deepexpertise.
It's even worse than Cosmos insome ways because it's not
designed for that.
It's a giant monorepo whereOnza knows their way around the
main team building andmaintaining that.
But operationally to run it asa network or even a validator
(14:48):
for, like, a new person comingin less so much a validator,
cause if you have a certainlevel of skills you can figure
that out, but a whole networkfrom Genesis block really
difficult.
You know it takes a really goodengineer weeks to just get one
up and running and that's liketime that most enterprises or
like other people, like indiedevs or whatever, like trying to
make an app chain or something.
(15:09):
They just don't have that kindof time or resource, and so we
ended up building for ourcustomers this thing we call SVM
kit, which is not as deep aswhat you guys have done.
I can already tell you,basically TLDR, it's like
operational tooling.
We use Pulumi we don't supportthis yet, but you could like use
Terraform or any other kind ofcode is infrastructure, is code
(15:30):
tooling if you wanted to.
But it automates everythingfrom like going from Genesis
block to a full cluster to justlike a single node, that voting
or non-voting RPC node, whateveryou want and like you can
launch a whole network in 10minutes.
And the only reason it takes 10minutes is because that's like
literally how long it takes forlike the infrastructure's code
to get through its processes.
Like the commands are like fivecommands, and so it is a
(15:53):
fundamentally very similarproblem.
And I found the same bottlenecksthat you just described, like
bridges.
There's all these, all SVMchains and like how are they
going to bridge?
And oh wait, we have to waitlike a couple months for layer
zero or wormhole to like supportus, like that's not a viable
option.
There are some people and it'sactually all come from Cosmos,
right, the Mantis team.
(16:14):
They built an IBC connectionfor Solana, for like an all SVM
or network extension or L2,whatever you want to call it.
To like Solana to Solana or IBC, to like Cosmos to Solana.
They built a bridge using IBC.
What I'm really curious about,though, is how does the
validator stuff work for youguys, because that's a piece
where I was like man.
Wouldn't it be amazing if Icould just launch a Solana
(16:37):
network and all the validatorson mainnet were just like
validating my network?
Yep, that's exactly how itworks here.
How does that work from a tokendesign Like?
Walk me through that.
Speaker 3 (16:46):
Yeah yeah.
So what you described isexactly how it works.
We're taking a differentapproach than most other layer
ones.
So a lot of later ones,especially Solana and Ethereum,
even Cosmos they're like let'smake the validator set ginormous
and make it as big as possibleis, when you employ something
like what you just mentioned andyou want to replicate that for
(17:08):
every kind of like sub chainthat you launch, you're
multiplying every infrastructurecost and effort by 100, 200,
50,000, whatever number ofvalidators you have.
So, from the very get go, rightnow in our main net we only
have 21 validators.
We're going to increase thateventually, but we need to get
(17:29):
all of the infrastructure piecesready to automate all this with
21 first, and then we willstart opening it up to hundreds,
hundreds of thousands orwhatever, if you wanted to.
And so generally, the way itworks is the 21 validators that
are validating our mainnet areessentially watching the mainnet
for these transactions, saying,hey, launch a chainlet.
The one big problem essentiallythat ICS and the interchange
security which is like kind ofsimilar to what we're doing and
(17:50):
Cosmos has is there isn't adefined economic structure that
is set to launch an interchangeblockchain.
Essentially, using Cosmos, youhave to individually negotiate
exactly what the rate is thatyou're going to pay, and a lot
of the time it's like uncertain,it's like it's like in other
tokens that you don't know.
Our model is we universalizethis whole thing.
(18:12):
It works exactly like AWS,where the validators come
together and they like kind ofset a price.
Essentially, let's say it's youknow, 10 saga tokens per month.
I don't know what the it'sprobably a lot more than that,
but they all come and agreethrough an auction amongst the
validators saying, hey, I'mgoing to set the price and
universally, this is what theprice of the chain.
That should be per month,essentially.
(18:33):
And the only thing the user hasto do or the developer has to
do is acquire that many Sagatokens and put it in an escrow
account and say launch.
And then essentially itlaunches that chain and the
escrow balance reduces over timebased on how much, what the
rate is essentially on a dailybasis.
Oh, that's super interesting,yep.
And then when the escrow runsout, the validator shut down the
(18:55):
chain essentially, and then youcan add a future time if you
wanted to bring that back onlineby paying for the escrow again.
Speaker 1 (19:01):
Essentially, Does this mean that if I'm a
developer and I'm building myown chain I've got some custom
features to it right, likedifferent economics things chain
, I've got some custom featuresto it right, like different
economics things and I've got myown token my chain could
generate fees.
I could use some of those feesto buy saga tokens, automated
right, and pay.
The validator set yep, yep,that's exactly right, and then,
like this is exactly how itshould work.
Speaker 3 (19:23):
You figured it out exactly like and also like this
is kind of okay.
This is like if you squint alittle bit, this is kind of okay
.
This is like if you squint alittle bit.
This is kind of how Polkadotworks, but it's flipped Minus
the pair of Jane auctions yeahit's flipped in our model where
the validators are bidding, andso the way that our protocol
functions essentially is iseveryone bids some price, saying
hey, I can validate a chainletfor X dollars per month,
(19:46):
essentially, and what theprotocol does is it goes across
all the validators and thenselects the cheapest 21,
essentially, and then whoever isthe most more expensive than
that point just kicks everyoneelse out essentially, and then
the way it functions is theprice is set to the most
expensive person within that 21set, so everyone who bid gets
greater than or equal to theamount that they bid essentially
(20:08):
, and so this auction model isnot in yet Right now.
This is like the model thatwe're building towards.
Speaker 1 (20:13):
It's interesting because there is a Web2 example
of this and it's a classiccustomer validation, market
validation tactic of A-B testingpricing.
The rule of thumb when you'redoing that is you're kind of
flipping on this idea and saidbut the rule of thumb is like if
I'm selling you a t-shirt and Isell a thousand and I AB test
$10 for $15 or $15, if themarket picks $10, then I refund
(20:36):
to the $15 people.
If the market picks $15, though, all the $10 people get $15.
Yep, you know what I mean.
It's kind of the same thing.
Speaker 3 (20:45):
Yeah, that's exactly right, and so, from our
perspective, in the long run, weactually don't think this is
going to be the value driver forSaga.
So what we want to do is createthis economic model where
validators are constantlypushing down this price as low
as possible because they'recompeting right.
Essentially, the ideal positionwe have is there's like 25
(21:05):
people constantly fighting for21 spots, if you kind of think
about it this way, and everyoneis just constantly trying to
drive like, oh, I think we canessentially offer this a little
bit cheaper.
And let's say, you lost a bid.
Essentially, you're like, hmm,how come I can't run my
infrastructure cheaper, butthese guys, these 21 guys, can.
So you go back in there and yousay, hey, how do I make this
(21:25):
more efficient?
Let me fix this a little bit.
Let me fix the costing.
Oh, I think we can get it downto, you know, a lower than that
amount.
Then you go and bid back inessentially, and this process
continues essentially, and itequilibrates to a position until
it gets to a point where youchoose the 21 cheapest
validators in the world,essentially the most efficient
(21:45):
operators ultimately.
And this is the model that weare like pushing towards, which
is, how do we make block spaceas cheap as possible?
Because most chains out theretheir economics is such that
it's like gas based, which meanshow do I maximize the amount of
money that I can get from users?
It's like very flipped.
We believe that infrastructureis a commodity item.
(22:07):
There's no such thing asinfrastructure that needs
bidding to win and to driveprices up.
It should be the inverse,ultimately.
Because hardware is cheap, awsis relatively very cheap in this
world, and so why is it thatwe're paying a hundred times
multiple of what the cost of theactual infrastructure is, when
(22:28):
you can set that price lower andyou can extract value in other
directions?
Speaker 1 (22:32):
essentially, ultimately, Well, it depends on
the large zone.
The hard respect too, though,right Like Solana, the hard
respect is high, it's expensiveand, you know, as the network
has grown it's gotten harder,especially for RPC, for
Validator less so, but still hasgone up.
But if you're running RPC nodeson Solana it's tough and that's
actually a core problem thatABK is working on solving.
(22:53):
We're building a protocol and anetwork to drastically reduce
the cost of running RPC onSolana.
One of the other problems thatwe identified in Solana, that's
Solana specific, and I'm justsharing it with you because I
think you'll think it'sinteresting.
And Cosmos largely solved this.
And this is where I think likecross-pollination in an open
(23:14):
source manner between, likeCosmos validator tech and I
forget what you guys call thatif it's the Cosmos SDK,
technically speaking or what,but anyway Cosmos validator tech
and Solana.
It's a monorepo right With amillion lines of code, and as
soon as you start launchingother chains and making
modifications, you diverge fromthe mainnet Agave repo and it
very quickly becomes nearimpossible to maintain it and
keep up.
So now you become Anza and youlose all the security benefits
(23:39):
of, like an open source projectvery quickly and your upgrade
path, like if there's a zero daybug, anza patches it.
You're fucked because you'renot going to be able to patch it
.
Speaker 3 (23:48):
Yeah, yeah, yeah yeah .
There's so much to fixessentially that you've modified
.
Speaker 1 (23:52):
Exactly, and this is largely the path that Linux went
.
Okay, we believe that you need amodular client and what we
proved you can do is you canstrip out.
There's these things calledpolicies that define how the
validator behaves.
You can strip them out and makethem plugins that live on the
chain and they load from eitheryour site, your chain, like your
fork, or they can load frommainnet, like however you want,
(24:15):
and then, once you get that inplace, then you can potentially
start doing some of the stuffthat you're talking about with
Saga, where, like a node on anetwork because it's just
running the kernel of Solana canvalidate multiple chains,
potentially as a single node ormaybe like a cluster of nodes,
but it can like I don't knowexactly how the architecture
would work out, but that waslike that's our vision, like
(24:37):
long-term, abk Labs wants tobuild that as an open source
project, not even.
I mean, I think we can make alot of money doing it.
I think there's some tokeneconomic structure we could do,
but really just because we thinkit needs to exist.
Speaker 3 (24:48):
Yeah, yeah, yeah, I agree, but it's one of those
things where Cosmos started asthis modular framework,
essentially there's problemswith Cosmos too.
There are modules out therethat literally every Cosmos
chain uses, because once you tryto modify that one thing, you
have to modify like 15 modulesin the background, and so
(25:10):
there's other issues as well.
Speaker 1 (25:11):
Yeah, so it starts to feel not so modular at that
point.
Speaker 3 (25:13):
It's not so modular because, like everything has
like these interdependencies inthe modules essentially.
But, as I said, it's a goodgoal to have at least.
So, like the nice thing, it isreally nice building in this
Cosmos world because from ourperspective it's like whenever
we generally think, hey, likethere's this like thought of how
to improve like certain partsof our stack Somebody has
(25:34):
already done it before we havelike lots of reference
implementations that we can gothrough and see, hey, is this
what we want, is this not whatwe want?
And compare and actually pickand choose essentially which
implementation is closer to whatwe want, and then we can also
modify ultimately.
And so that's kind of the nicething about cosmos is it's so
decentralized.
And this like entire modularframework has existed.
(25:55):
I think, like with solana, asmentioned, like the, the goal
from the very beginning was likeone fat pipe, essentially right
.
So it's like, yeah, the fatpipe's awesome, but if you want
to make parallel pipes, it'slike it gets a little bit more
complicated, right I think oneof the interesting things I want
to point out about Solanathough.
Speaker 1 (26:11):
Yes, that was the goal right Blockchain at the
speed of light.
You know NASDAQ, you know onchain right.
But one of the cool thingsabout Tully is he is an open
source maxi right.
Like he believes in open source, and so his, I think his vision
has always been, and he hasencourage people to experiment
with Solana and do differentthings, and I think it's exactly
(26:32):
that ethos that will lead to animproved, more modular Solana.
That will improve mainnet.
It will enable, like, if you gothe way of Linux and you have a
modular stack, it actuallymakes it so that, like, you can
get more and more granular, andthat's what you have to be in
order to optimize, to make thosepipes bigger.
You know, ultimately, I thinkmaybe I'm totally wrong, but
that's my intuition slash,experience from other people who
(26:54):
you know were there when Linuxwas built- you also want like,
more than just Anza, like doingthis right.
Speaker 3 (26:59):
You want, like multiple people looking at this
at the same time.
Speaker 1 (27:02):
Well, and they are now and that's the cool thing
Like it is really cool thatthat's happening.
There are many clients nowbeing built, many really
interesting talented teamslooking at it.
One thing I'm really excitedabout this conversation with you
is just how we have we havelooked at the same problem, just
from different ecosystems andprobably with some different
needs, which it's like cool totalk to somebody who really like
dug into that, like I do thinkit's going to work.
(27:24):
Like I don't know how far alongyou guys are in this experiment
, but everything you're sayingto me it's like you've looked at
the problem correctly.
You thought about it from firstprinciples.
You designed a system thatcould work.
How far along are you now?
Like where are we today?
On whatever?
Today is sometime in February,mid February.
Speaker 3 (27:42):
So right now the basis is all there, so you can
go right now on saga, appsagaxyz, essentially put a transaction
in to launch a chain.
Uh, permission, if right nowit's permission, it uses this
like test token called psaga,that we control the supply of,
because we don't want right nowthe economic activity of
(28:04):
actually like paying for thechain is not implemented yet.
But as long as you have youknow these psaga, you just ask
us for it, we'll justimplemented yet.
But as long as you have youknow these P Saga, you just ask
us for it, we'll just give it toyou.
As long as you have that, rightnow it takes about two minutes
to launch kind of a chain.
It's fully decentralized, itcomes with an RPC endpoint, it
comes with an explorer and thechain kind of comes online.
We don't really do anything, itjust happens magically amongst
(28:26):
all the validators that we have,like for mainnet essentially,
and so the thing that's a littlebit more complicated with this
is so in the beginning what wewere thinking about was okay,
people will launch their ownchainlets, the projects that are
building on Saga, and they willattract liquidity themselves.
So you know, our big focus forstep one was just the validator,
(28:48):
the RPC endpoint, some of theselike very infra heavy things.
And then what we quicklyrealized is, like this liquidity
issue and the fact to bringlike bridges and all that is the
exact same problem that we hadwith the validators, essentially
.
And so we're in this stagewhere we're kind of launching
mainnet, essentially 2.0, wherewhen you launch this new
chainlet, it actually comes withwhat we call liquidity
(29:11):
integration layer, which isshort, for you can shorten it to
LIL.
We call it LIL and essentiallyit's like a bunch of not only
when you put that transaction into say, hey, I want to launch
my chain, it launches the chainitself with the validators, but
it also stands up all thebridges that connects it to
every other chain that and toconnections to like all the
(29:34):
Ethereum ecosystems, but also itautomatically stands up all of
the infrastructure needed toautomatically route between them
.
So you don't have to go to likesome bridging website and
bridge between the chain.
That it's all automatic.
Yeah, that's really interestingwebsite and a bridge between
the chain.
That it's all automatic.
Speaker 1 (29:47):
Yeah, that's really interesting.
So it's like Jupiter swapswhere it like routes everything
automatically, but you're doingacross multiple chains.
Yep, do you guys use Relayer?
Do you know what I'm talkingabout?
Is that who you use for someparts of this?
The bridging piece?
Speaker 3 (30:01):
Our bridges because we're using Cosmos is all just
IBC and essentially in thebackground we have all the
infrastructure that spins up therelayers needed to make the
bridges all function properly.
So it's like a little bitcomplicated.
Speaker 1 (30:14):
So the way it works is Okay.
So it is a Cosmos fork.
Speaker 3 (30:18):
It's your building, it is Cosmos.
Speaker 1 (30:19):
It's all built on Cosmos.
We never actually explicitlysaid that.
I kind of assumed it.
Speaker 3 (30:31):
Yeah, yeah, explicitly said that.
I kind of assumed it.
Yeah, yeah, yeah, I mean likeI'm I'm wearing, like here I
have a cause.
Yeah, I can't see that.
Yeah, yeah, we're on cosmospeople.
Yeah, yeah, yeah, but but it'slike we're we're like a weird
subset because we're like evmcosmos, so we're like aligned
with a lot of the ethereumcommunities, but our tech in the
background is kind of invisiblyall cosmos, which is like kind
of cool in my opinion, and soall the bridges automatically
get set up.
So let's say, we go and launchan ABK chain you request an ABK
(30:53):
chain In the background, asmentioned all the validators
launches a node, connects to oneanother and then makes that
chain for you as well.
All of the bridging pipelinesto bring up all the relayers
between this new ABK chain, therouter that routes all the
traffic across Saga isautomatically stood up at the
(31:14):
same time.
So it literally takes about twominutes for you to spin up a
chain, have all the bridgesnecessary to go connect to
things that you need essentially, and all of the routing is
stood up automatically as well.
And so what we want to do is beable to for someone to go and
launch this, you know ABK chainand then immediately using our
API, and say, hey, bringEthereum into my chainlet
(31:37):
instantly.
And then you just put atransaction in use our routing
tools.
It will just suck up thatEthereum all the way from
through Axelar down into ourrouting chainlet.
It will relay that routingchain Ethereum all the way from
through XLR down into ourrouting chain that.
It will relay that routingchain that back to your chain
that and that Ethereum will justkind of magically show up on
your chain and this is like alldone.
You can launch this in likethree minutes and you can
parallelly launch 300 of thesein the background and suck
(32:00):
Ethereum into each one of themindividually and it will all
come essentially instantly.
I would say we're like reallyclose to this.
We're releasing liquidityintegration layer that's coming
live like this week, next week.
Essentially, our first instanceof this with Uniswap and Oku is
going up immediately after that, and so really close to all
this.
And then there's some likeimprovements we need to make in
(32:22):
the user experience front tomake this happen.
But, like this kind ofkickstart, a lot of the defy
activities that are happening inin the chains that people have
been asking for is like hey, howdo I get liquidity onto these
chain lists that I'm launching?
Speaker 1 (32:33):
essentially, so what's the scalability like from
a transaction you know tpsthing standpoint for you guys?
Speaker 3 (32:40):
okay, I would say tps is almost irrelevant, but it's
like let's assume, for I don'tknow what the actual metric is,
but let's say we can shove likea small number, let's say a
hundred TPS, onto a chain.
The way that we think aboutthis is an application should
actually start sharding itsapplication across multiple
(33:00):
parallel chains, because if itonly takes two minutes to launch
a chain, essentially you candynamically figure out when your
application is gettingcongested, right, and start
spinning up backup chains in thebackground.
And so this 100 TPS eventuallywill start incorporating a lot
of innovations that arehappening in Cosmospace.
That brings that up to you know, a few thousand TPS easily,
(33:23):
right?
But in the meantime you canjust scale this as much as you
want, right?
And so one example that I'vebeen talking about I just talked
about Uniswap, right?
I'm working essentially withthe idea that we can start
sharding Uniswap instances inthe background in Saga.
So Uniswap is going to go onlive on a singular chainlet
pretty soon in Saga.
But you can imagine like aneasy ways to do this would be
(33:46):
like you take one pool, let'ssay USDC, usdt, and you stick
that into a separate chainlet inthe background, right, and you
can start parallelizing everysingle one of these pools that
you have in Uniswap can now goon individual chains in the
background, right, and then youcan also.
That's like step one.
Step one is just like one poolper chainlet like sharding,
(34:09):
essentially.
Step two is you can startactually sharding the same pools
onto multiple chains.
It's super easy to do.
The reality of it is noteveryone needs the entire huge,
ginormous pool liquidity that aUniswap pool has, right, you can
.
Some people are doing like oneETH transactions and other
people are doing like 100 ETHtransactions.
The one ETH transactions can gointo a Uniswap instance that
(34:32):
has a smaller pool and it won'taffect the slippage as much.
So you can actually startsplitting out one pool into
parallel versions of this.
Like think of it as one smallpool and one big pool.
So now you have doubled theamount of capacity, but you only
increase the liquidity by like5% or something tiny,
essentially Right, and you cando this in multiple times.
(34:54):
So now you can have 10 or 15small pools that people get
routed to for transactions andthen only the big transactions
go to the big pool, essentially,and you can do these like
dynamic routing.
Essentially.
That allows you to shardUniswap almost like infinitely,
and this is like kind of thevision of what we have.
It's like look, this is howscaling should happen in
(35:15):
blockchain.
This is the design philosophythat we have, and it's like how
do we enable this to happen?
And so we're going to startwith Uniswap.
Other stuff is going to happen,and I'm hoping that we can
create an SDK that allows peopleto just parallel shard
applications automatically.
Speaker 1 (35:29):
Essentially, Is the Uniswap like a formal
partnership with them, or didyou just fork it?
Speaker 3 (35:35):
This is the funny part.
No, we're just.
It's not really forking, sowe're going to shard them
regardless if they want to ornot.
Uniswap, the DAO, essentiallyhas ways to deploy Uniswap
contracts like canonical Uniswapcontracts and they own the
actual contract itself.
And so, from their perspective,when there's two Uniswap
(35:56):
contracts that are canonical onSaga, they'll just treat it as
two different chains in theirheads, but like it doesn't
really matter if it's the sameecosystem or not.
Now they have double the amountof instances that they need to
think about.
But they have a process forthis.
So we'll start with a singularUniswap chainlet and then later
(36:17):
down the line we're going toidentify one pool to split out
and say we're going to deploy asecond Uniswap chain.
Then we'll apply again.
They'll say, yes, this is nowthe canonical Uniswap chain,
except now there's two, two SAGAessentially.
And then we'll just keep doingthat and actually, in fact, I
actually prefer not having thispartnership.
Speaker 1 (36:35):
But you still have to ask them for permission to use
it, though I mean Uniswap is anopen source software right.
There's just controls aroundlike that yeah, that bsd license
thing right with like part ofit right, wasn't it?
Speaker 3 (36:48):
and so then it becomes open source later, or no
, no, I'm pretty sure uniswap isfully open source.
Yeah, so it's like anyone canuse it.
Speaker 1 (36:55):
I feel like there was a thing where they changed it
after a while or something.
Speaker 3 (36:58):
I can't like I might be totally off on that yeah,
yeah, but, as I said, like frommy perspective it it's like
anybody who believes has an opensource software, that's a
Solidity smart contract.
We'll just be able to work withthem to deploy one version and
then shard it to like hundredsand different chainlets in the
background, and there's somecomplications, right.
So like the front end needs toknow that when they receive a
(37:21):
like a USDT to USDC liketransaction, to route it to the
correct chain, and so we justneed to put some piping in the
front that allows them tomultiplex that, and and then
it's like pretty simple.
After that, you just, you justkind of shard it to like however
many that you want and then, uh, direct traffic in whichever
direction you want to.
Ultimately, it's.
Speaker 1 (37:40):
It's super cool.
We are at the top of the show.
The last thing I ask everybody,everybody, is you know what?
Have I not asked you that Ishould have asked?
Speaker 3 (37:47):
I really like the fact that you're thinking about
how to make Solana more modularand more parallel, and I don't
think people quite understandthe power of horizontal
scalability People always talkabout, talked about, like
horizontal scalability.
It was existing, like Cosmoshas existed since 2018.
(38:07):
It's technically a horizontalarchitecture, but having it
available as something thatcomes up in minutes is not
something that has existedbefore, and I think we're going
to see a pretty big transitionessentially of how developers
think about applications oncethis paradigm exists in this
blockchain world.
(38:27):
And so I don't know,question-wise, what would I say?
Maybe if you asked more abouthorizontal design.
There's some interesting stuffwe can talk about, but I feel
like we don't have that muchtime.
Speaker 1 (38:39):
Yeah, it's a rabbit hole, I mean.
What I'll say is I think thisis why Cosmos people and Solana
people like each other so muchis because we share this true
open source ethos.
Speaker 3 (38:49):
Yeah, yeah, I love Anatole.
I met Toli before Solanalaunched, actually, and I sat
down in the office in SanFrancisco with him and I think
it was him.
He was explaining how theSolana block production and the
catch-up mechanism for verifyingthe blocks, how there's two
different points.
That moves along in time.
(39:09):
I was like, oh, this is reallyinteresting.
I think ethos-wise we're veryaligned.
It's just slight variations onhow we think we should
prioritize the scalingultimately.
But yeah, I think you're right,it's pretty much the same thing
.
Speaker 1 (39:21):
My point about that is that it just makes these
conversations so much more fun.
Like, the why for ABK Labs isjust build together.
That's what I want to do.
I left.
I was working at the SolanaFoundation.
I had a great job, but I wantedto go build something important
.
Like I was like kind of workedmyself out of the job there and
I loved it there, but I wantedto do is build open source
(39:42):
software with great engineersand people that I like.
Speaker 3 (39:52):
You know how our system is L1 to launch L1s
essentially.
So the way that our tech stackis kind of implemented actually
is it's all containerized,essentially, like the chainlets
are all containerized, and sothe consensus algorithm and the
technology stack that powerseach chainlet actually doesn't
have to be the same.
So as long as each of thecontainers speak the same
language in our case it's IDCand there's like certain other
things that it needs to adhereto but if we can design,
(40:14):
essentially, a Solana stack thatadheres to those languages and
models, we can actually launch afull Solana stack chainlet
essentially and parallelize it.
And so, as I said, it's likeit's exactly as you mentioned.
There needs to be work to makeSolana a little bit more modular
, where it's easier to do this.
But once that step happens, wecan deploy, you know infinite
(40:38):
numbers.
Speaker 1 (40:38):
That would be really cool to explore, cause I and I
think we could do that with you,because we, you know, we've
already got SVM kit, which islike that operational tooling
for the deployment part, but thepiece you have is this like
interop and the chainlet ideaand the economics piece.
That's like the piece that weare really missing.
Nothing we've done has reallybeen about making money or like
(41:00):
building a token or anythinglike that.
It's literally just been like Iwant to build open source
software, we're going to figureout how to fund that.
That's my job.
I'm CEO.
That's like literally how Iview.
My job is fund engineers makingopen source software.
Yeah, yeah.
Speaker 3 (41:11):
That's it.
Yeah, we should explore that.
Yeah, of course, I mean, as Isaid, it's a.
Speaker 1 (41:20):
Did you say it's already built?
There's one built, so there isa way to do a alt SVM chain or
network extension bridge toanother SVM chain, main net or
another SVM chain using theirstuff.
Like it hasn't been, like Ithink there's, just I don't know
if it needs to be.
Is it if it's as automated aswhat you have, where we could
(41:40):
just like easily configure it todo what you're talking about?
Because the thing that you cando really easily right now you
mentioned about like thedifferent consensus stuff and
like the modifications andthings like that, that's the
piece that's still kind of hardon solana, that needs to be like
innovated on, I think, toreally make a solana saga
(42:00):
product yep work yeah, yeah yeah, you gotta you gotta cross the
chasm there and meet in themiddle somewhere to like, really
be like.
Okay, you know we've already gotit down so you can launch
chains really fast you know, butnot where they're like
automatically connected with thebridges and the routing and all
that stuff.
Yeah, yeah.
Speaker 3 (42:18):
Yeah, yeah, I mean, even in Cosmos, which is
designed to do this.
It's like there's it's a lot ofwork to do this.
We've spent significantengineering cycle because, like
there's like really dumbproblems that come when you do
this, like, for example, whentwo chains come online,
sometimes the tool will makelike three bridges that are like
parallel and doing the samething, and then all the tokens
(42:40):
that are going between the threebridges it's like not fungible
with one another, and then it'slike everything breaks right and
so, and so we have to make sureto for it to check that like
there's only one bridge thatit's using essentially, like
there's all these like likethings that that that that come
up that like are problematic toautomate essentially, and so
this is like this sounds to melike something that we can work
together on, which is likereally exciting, in my opinion.
Speaker 1 (43:00):
Let's do it.
That's why I love doing thisshow.
Speaker 3 (43:02):
Thanks for coming on.
I appreciate you.
Yeah, thanks, man, appreciateit and let me know if there's
any other questions and I'mhappy to answer.
Speaker 1 (43:08):
I will do, I guess the last thing is how can people
reach out to you?
Speaker 3 (43:11):
If you just go to our website, sagaxyz, there's all
the contact forms.
Feel free to reach out to ourteam and somebody will contact
you.
Speaker 1 (43:17):
So awesome Sagaxyz.
Thanks, man, bye-bye, availableevery Friday.
Thanks for tuning in.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com