Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
Aggregate Everything in Web3 🦾 The Aggregated Ep. 126
Recorded: Aug. 22, 2025 Duration: 1:44:00Space Recording
Full Transcription
Thank you. Thank you. Music Thank you. Music Thank you. Thank you. Thank you. Hello, hello. Hello everybody. Welcome to Aggregated, episode 126. Wow, it's a lot it's like more than two years pretty cool this
episode is going to be centered on aggregate everything which is a hot
topic at the moment whether it comes to things like yield aggregators or whether
it comes to blockchains themselves, aggregating transactions, make things more efficient,
makes it more likely that DAPs will be interoperable.
So it's pretty exciting to explore.
We've got a little bit of a change here
in the sense that ROC is over at the SALT conference
in Jackson Hole in the US
and Aztec is out of office today.
So I'll be helping to host along with producer
Darren who's on the quick swap handle today so to give you a little bit of
background on myself before we do intros so I'm Jack I'm chief launch officer at
Luna digital assets which is the venture studio I work with Polygon in the early
days and quick swap and a bunch of other cool things currently working with litvm and strat X and a
bunch of cool up-and-coming projects and I'm also leading BD over at strat X so
it's great to meet everybody today if we want to get started with some quick
intros that would be cool so why don't we go ahead with two cents tomy if you want to go ahead and just introduce yourself to the order yes gm gm happy friday i am
two cent timmy i am on the polygon marketing team um and yeah here to talk about aggregating
everything because that is our goal at Polygon. Awesome.
Yeah, looking forward to hearing some interesting things about the ag layer as this space goes on.
Maybe some alpha there later.
So, Dean, do you want to go ahead?
Sure.
Hello, everyone.
Good morning.
I am Dean Tribble.
I am CEO of Agoric.
And we helped kick off chain abstraction last year and orchestration this year and aggregation is sort of the fundamental pull all these
things together chain abstraction lets you have a nice smooth user interface
across that but we're really we're we think that the you need to be able to
program across that and so you aggregate everything by being able to have you
know programs running in smart contracts that can pull all this stuff together
for users and provide a seamless experience across chains. And so I'm excited for all the
other ways that people think about abstracting and aggregating this stuff.
Yeah, that's a great point. I think it's interesting, you know, like for the longest time
people have talked about, you know, quicker we can uh aggregate everything to make
like a unified layer and the quicker we could abstract all that away from the user the quicker
we'll onboard people but on a personal note i actually feel quite privileged to have been here
during this time when things weren't so slick when things were a little clunky and a little
more interesting and you you know you had to kind of figure things out and duct tape things together because it'll give all of us that are here in
this space and have been building for some time a depth of knowledge and a wealth of understanding
that um you know newer users won't have when all that is combined into a very smooth ui that works
seamlessly on any chain regardless of you of what they're trying to do.
So it's worth just thinking
about that as we go forward.
Although this building
period has been difficult and a little
janky in places,
it probably taught us all quite
a lot.
Okay, so we've got
Euclid Protocol. Do you want to go ahead?
Hey, guys, can you hear me?
Yep, loud and clear.
Perfect. Hey, guys. Yeah, so my name is George, and I'm the founder of Euclid.
And very quickly, we are building a unified liquidity layer that allows you to tap into liquidity that exists anywhere across 50 plus networks that we integrate without the need to bridge
and tap into it natively.
So you're able to bootstrap your application
or your ecosystem natively with assets
across EVM, Cosmos, and Solana,
and also non-EVM networks
in less than four seconds,
atomically without the need to bridge.
So this is very quick and trouble what we do.
Nice. Good to have you here. It would be really interesting to get your opinions on some of the stuff as we go forward. So we got Kanishk here from
UMA protocol. Hey folks, hey Jack, thanks for having me here. My name is Kanishk and
I represent ACROSS and UMA protocol.
Primarily here, I'm representing Across protocol.
Across is a bridge powered by Intense, so we're an interoperability layer for protocols.
And if you're building anything related to swaps, anything related to bridges,
and you want to make sure that you connect users coming from one protocol to the protocol where you are deployed, I think we're the fastest option,
and we help with under two second bridging.
Cool stuff.
We just launched on a bunch of chains and we also launched on Solana two days ago.
So we're super happy to be here, super happy to represent across.
And thank you so much for having me.
Excited to take this forward.
Nice.
I don't know if you just want to take a few more seconds there to explain how.
So you mentioned you just launched a couple of days ago.
So just curious, given the current market environment, like what that experience has been like.
Have you had like a lot of good traction or what's the situation?
For sure.
So in the larger landscape, AXR has been live for the last, I think, four years now.
We've done over 30 billion
in transaction volume we just hit 20 million transactions so we've done a lot of good numbers
in the past but primarily all of this has been evm centric right so ic2 centime is here from
polygon we've been on polygon for a while um and we love and support all the users there.
We're also on a bunch of chains, obviously Ethereum, Base, Arbitrum, OP, all of those.
And now we're also expanding to Solana, which opens us to a new layer of non-EVML1s and powering all the users there.
We also are live on BSC, which is another cool chain, I think.
And so, yeah, that's about it.
30 billion total volume and now expanding from Solana onwards.
Nice. Thanks for the detail. Appreciate that.
I wonder how long it's going to be until people stop saying EVM-based.
You know, as things get more abstracted away,
I think there'll come a point when people stop differentiating
and just use the tech. It's interesting. I think there'll come a point when people stop differentiating and just use the text.
It's interesting.
I think we're probably still years away from that.
At least.
At least.
We've got Captain.
Captain Juju.
Oh, hi.
I was fumbling with the mute button.
So I'm Juju.
I'm the lead BD for WolfSwap. I'm the lead BD for WolfSwap.
I'm the lead BD for WolfSwap and of course excited to be here, joined a couple of your
spaces, always fun being up here, diving into conversations and looking forward to what
you have for us today. Thank you very much. Juju out.
Nice, great to have you here.
So next we'll go to Web3 is a joke.
Apologies, I don't know who's behind that handle, so you'll have to introduce yourself.
Yes, of course.
My name is Mona.
I am the founder of Web3 is a joke.
We produce Web3 crypto and AI comedy roasts around the world. We've done 16 sold out events, including LA, San Francisco,
London, Hong Kong, and Denver.
And we are going to be at Token249 Singapore.
And our objective is to roast as many crypto bros,
nerds, and the hot chicks that show up to these events.
Nice. That sounds really interesting.
So I guess you guys are doing something around token then.
Yes, that's correct. We're going to have a major project. If you want to give a shout out to that. Yeah sounds really interesting. So I guess you guys are doing something around token then? Yes, that's correct. We're going to have a major presence.
If you want to give a shout out to that. Yeah, go ahead.
Yeah, absolutely. We're going to have a pretty big presence at Token 2049 Singapore.
We just got our event announced yesterday on BTC Wire that you can check out.
If you follow us on X, you'll check out that article there
and yeah, we're bringing on some
pretty heavy hitters, comedians
that are Rose Battle champions
from the US, from Hong Kong, from
Singapore and we're also
going to have a legend
from NWA, one of the founding
partners who's going to be DJing our
after party networking event
that I'm super excited about,
and a potential performance by a member from Pink Floyd.
If you would like to roast Timmy, feel free to anytime.
Absolutely.
Well, just send us some details and then we'll roast you.
Yeah, Darren will tell you all sorts of, like, just send me
thumbs down during the entire space.
That's usually what he does.
Listen, man, we've heard
worse. It's okay.
Awesome. Well, if you're
like many of us, going to be heading to
Singapore in a few weeks,
definitely swing by. That sounds like a lot of
fun.
So, let's go to nikki next
thank you so much for hosting me and happy to be here today so a little bit about myself my name
is nikki and i'm one of the co-founders of pelagos network we're a kind of a new project
uh in the space and uh essentially what we're doing we're just like one atomic fabric that
unifies liquidity state and execution
across all chain a little bit how it works uh basically like um it uh essentially sequences
bundles of flex with you can also include policy gates and receipts it isn't a bridge or overlap
how pelagos works it orchestrates the Rails venues you choose.
And basically core concepts are bundles,
like the complete plan, it either succeeds or revertes one.
Then you can also include your policy checks if you want to,
like UIC sanctions, limits, unbounds, enforced pre-execution. And you can get your anchor hash like receipts batched and checkpointed for
audit. Happy to be here and looking forward to meeting some of you in Singapore. Yeah great to
have you here Nikki thanks for joining and then just finally before we kick things off we'll go to
Towers or Diego Torres according to your Twitter handle. I think you're from Protofire right can't hear you
maybe leave and rejoin and I saw the Protofire handle here a minute ago so
maybe join with that if there's no luck.
For those of you that aren't aware, Protofire is an absolutely kick-ass developer DAO.
There's like 80 members.
They are incredibly technical.
They've worked on some of the best protocols in the industry from a DeFi point of view.
Some of the really big, you know, like a laundry list of the biggest protocols in the industry from a d5 point of view some of the really big you know like a laundry list of the biggest protocols in the industry and they work
closely with some of our projects as well like stratax as an example so really
really excited to have protofy here with us today
towers can you hear me I've just given you a like-stop intro. He's just coming back up.
Thanks, Jack, for that nice comment about us.
My name is Diego.
I'm working for Protofire.
As you were just saying, we are a bunch of engineers trying to help protocols in general.
We've been working in this space since 2016,
so we have kind of a big experience here.
My role is a business developer, mostly on strategy with product.
And recently I've been engaging with a lot of Web2 companies that wanted to engage with
Web3 and they really want to abstract everything.
They don't care about chains about protocols about tokens they just
wanted to to get um enhanced by this technology so i guess abstraction is is a key point here
yeah that's right and i think when we when we think about aggregation and abstraction
it's worth bearing in mind how this works in tradfi so you know a lot of a lot of your financial transactions run on the visa network or the
swift network um and you really actually think about that on a day-to-day basis you
go to whichever bank you want to use um whichever bank is close to you or whatever you get an
account you get given a debit card and then you get to use that for payments
So you're thinking about your bank for example Citibank or Barclays or whatever
Whatever you choose
And you're not necessarily thinking about the network that runs underneath
so my question to you guys here
the speakers is
When will we get to that point when the network underneath
kind of fades into irrelevance apart from for the technical people and the builders that will still
understand this stuff when will the average user be uh like not aware of this stuff because it'll
be fully aggregated and fully abstracted away and just for context, this is like an open panel spaces,
so you basically just unmute and then talk,
elbows up, and then go for it.
Yeah, I think we're,
to answer the question,
I don't necessarily know
if it'll be like a full zero to one moment.
I think we're actually getting really close
with all the tools that are being
built. And especially when you think about like the AI smart wallets that are doing it. So you
have different applications like DeFi app, the Magic Newton wallet does it. There were a few
others like smart wallet platforms that you can basically just use a DCA, get your assets in, and they abstract
everything from the user. Now, it's still a little bit clunky. But for like your average retail user,
and obviously Coinbase is building their application to try to abstract everything
away. I think we're like shifting towards like the dream future of like you have Ethereum,
and it doesn't matter that you have ethereum on
eight different chains you see your balance all on one chain i think we're very very we're there
sort of not not all the way there not as much of like it's not like we can stop the building
and development but i actually think we are getting very close and And then it's like, just for the rest of the industry and other wallets to
catch up, I think is kind of where we're at. Yeah, I think there's a long transition period,
right? Of, you know, it will fairly quickly get to the point where for most users, which is to say,
if we got a million, you know, new crypto users, that would swamp the number of monthly active users, and they won't
know or care about all the details of what's changed underneath. But that will just be for
I can get a token or I can do a swap, right? A lot of people think of Intense right now as just
the swaps. I can swap token X that happens to be on Arbitrum for token Y that happens to be on base, and I don't care about the difference there. But getting to richer and richer functionality where I can do DCAs or auto
compounding or my game and loyalty program connects in with my money that's stored in
some pool somewhere and they're all in not just different chains but different ecosystems,
somewhere and they're all in not just different chains, but different ecosystems.
That level of rich automation is something that will come over time.
So we will see nicely abstracted services delivered to users that are just scratching
the surface of what Web3 will be able to do that will gradually grow into, yeah, I can
get a loan.
And the fact that the loan is on three different chains, I don't care, right? I can spend my money. And the fact that the loan is on, you know, three different chains, I don't care, right? You know, I can spend my money. And the fact that I'm spending my money
out of accounts that are in two different ecosystems, not my problem, right? And as we get
to that, you know, you know, so there will be a decade or more where us developers are seeing,
you know, the frictions and challenges of being able to automate across
these across all these different ecosystems but will start to deliver good seamless experience
to users you know this year or next year and and you will quickly overcome the number of users in
in the space now i think i agree that's a great thing and i just say, like, you're starting to see threads of this right now, this year.
So, obviously, you alluded, I think, to Base.
And so, Base are doing a big push to get their dApps to create mini-apps
so that they give a more seamless user experience on mobile.
And then you've also got centralized exchanges like Binance
starting to source liquidity and uh trading
pairs and stuff from the decentralized um arena so it's interesting that like this abstraction and
this like um cross chain and sort of an ambiguous like where the user is not sure exactly what
they're using or how it's happening but it's it's all been abstracted away and made like foolproof it's quite interesting i actually agree with that on a very deeper level because
as intense function um we at across have made it our mission to make sure that whatever token you
have should not be the limiting factor to what you want to do on chain so if today you have say you know eth on ethereum
and you want to deposit usdc on arbitrum uh for hyper liquid that should be a that should be one
button click for your user so we've actually thought a lot about this and we've expanded
the vision of intents to something that can achieve this in again one button click so i agree that
we're still a little bit far away from the future
where you don't need to worry about the chain you're on and if you want to collateralize a loan
say with uh tokens on three chains and borrow on one chain or again borrow on three chains that
should be possible um and also shout out morfo here because morfo v2 exactly does this and um
it's also a cross-chain infrastructure
that is powered by intents.
So it is quite nice to see that ERC-7683,
the standard for cross-chain intents,
is being adopted so nicely
and people are focusing on lifting it
in the industry forward.
I believe I have to go a bit more further
because we have to look what is chain abstraction can realistically cover today.
So yes, we do have, for example, features such as account and key abstraction, gas abstraction.
We do have, for example, intents and routing, users stay what they want, and solver networks decide how across chains, and you can hide past bridges and so on.
And then you have also for example
liquidity abstraction where you have the bond balance view you can have the state access
abstraction and for example certainly policy abstraction what is still missing like what we
need to still achieve is like the finality and time because chains do have different finality than MEV and ordering, like certain
security models like TSS co-signing, like client proofs and so on.
And also to a certain extent, fee predictability.
So far, what can still not be abstracted is like economic sovereignty, for example, latency,
which is another issue,
global synchronous compatibility,
I would say true single slot,
multi-chain like synchronous execution.
And I would say like universal censorship guarantees,
like for example, right?
You can add escape hatches and multi-route relay.
And essentially like, but I do personally believe that, like,
we are kind of on the
right road, because, like, past, because
you can fully abstract to a certain extent
UX, and mostly abstract
execution to a certain extent,
but you cannot still
abstract it fully, unfortunately,
but I think we're getting there, and it also
depends, right, like, what is our view of
the, you know, speaking of the future of the space.
Do we believe that there will be multitude of chains?
Do we believe that there will be like chains,
like there will not be many general purpose chains,
but mostly the chains that will be focused on specific use cases
and essentially like what kind of use cases will be driven
and what do those actual use cases will need driven and what do they those actual use cases
will need so that's basically where we need to look and uh essential think of those parameters
but i think like we're still going in the right direction maybe we should like i wish we went like
a bit faster but uh still we are what we have so what i was going to say is this right if you're
looking at the goals what we're trying to do is this, right? If you're looking at the goals, what we're trying
to do, which is onboarding or what you'd say that centralized exchanges are doing it to some extent
where people come in, the newbies, they don't really know what chains they're getting coins
on or whatever, right? Or if you're looking into the more complex side of DeFi, like the stakings
or whatever, right? they're like people coming out
with different new tech right i recently saw one with an agnostic layer right where they're kind
of like sourcing a liquidity to pull somewhere else where you could trade like coins across
multiple chains and you necessarily don't need to know where you're doing that on but my point is with this is
are we are we making it more complex for new people coming into like crypto to understand
what's going on when basic things like i'll just set up your wallet i'll just throw your seed phrase
i'll do like you know all those basic things seem super super complex to a newbie coming into the
space so i feel like we're a very very long way from getting to the point where we should start
thinking about like how to start making like complex defy stuff like stay on chain without
people really understanding what they're like understanding what chain they're kind of like trading on or staking on or doing whatever on right people need to understand basics people need to come in easily
people need to like trade easily understand the simplicity of things right and people don't need
to also get bombarded by complex words like oh yeah talking about like interoperability agnostic layers talking about
things like seed phrase that's why we're kind of like chasing people away from crypto so the more
we could actually simplify things the better we could actually have people on and the better we
could actually push the space forward thank you very much yeah i agree so like imagine if imagine
if as an allegory to that imagine if when you went to go and open a bank account
at your local branch,
the bank manager took you through
all of the internal security procedures
and all of the back-end code.
No one would want to see that.
That wouldn't make any sense
for an average person opening up an account,
but that's the current kind of side of things
that we have on Work3 at the moment.
This might be an unpopular opinion
but like what you were just saying jack about the onboarding with the bank i actually think
we're there um like i'm actually going to kind of push back on what uh ju was saying, where I don't know if it's actually a good idea
for like really complex things in DeFi
to be made simple for the average person, right?
And like the way I think about it,
like it's not easy to understand the yield
and liquidity of a government bond market
or corporate bonds and corporate bond rating.
Like that's very highly technical stuff. And there's friction to learn, like there's a learning curve, but because it's
complicated, and I think we can make simple things easy, like, I don't know, call it like a B3
liquidity position and rebalancer, like, that's great, you don't really have to get super
technical, but the highly technical, like, I don't know, I'm thinking like DeFi looping and lending protocols, right?
Like that, there should be some barrier to entry or you're just going to have really poorly educated people participating and not knowing the risk.
So I think I'm a big fan of let everyone do what they want to do and it should be open source and trustless and permissionless.
But I also think there should be some learning that needs to go into place or you're setting people up for failure.
And I think they're like, I think for opening a bank account, like, yeah, you can go to open up a wallet and it's pretty straightforward at this point.
forward at this point i just think like there's a way to go for the true abstraction on like
buying an nft that you want on abstract when all your eth is on optimism or something like that
so like to give you give you to put this in like a more defy related context so on stratx
we have a system where because you mentioned borrowing and relooping strategies right that
was one of the things that you mentioned so we have a system where um because you mentioned borrowing and re-looping strategies, right? That was one of the things that you mentioned.
So we have a system where the base layer is a strategy marketplace where developers can come and build DeFi strats,
like complex things like borrowing and re-looping.
And then we have vaults which aggregate the best risk adjusted return of those strategies
and then mint ERC-4626 vault vault standard tokens which can trade as your bearing
instruments on the secondary market so in that way like somebody like you like i mean stratix is
is not going live on base until 27 so you can't do it now but when when that happens somebody will
be able to basically buy and hold a yield bearing version of btc usd or eth and then get the yield from that where there's quite complex things going
on underneath but the only thing that they really need to know is that it's uh you know it's done in
a delta neutral way it's insurable and the the apr is currently you know whatever nine or ten percent
so that's like a good example of abstraction i feel and obviously i'm I'm biased here, guys, because this is something I'm working on.
So feel free to like critique and chew it down.
But I just wanted to bring that back because you mentioned about, you know, your concern is people getting involved with complex things like that could cause issues.
So I kind of want to like relay that.
Okay.
So I guess I could step in right now.
So when, just to kind of like push back on what
simi said as well right i feel when you're looking at the banking system where you're looking at
irl right there are levels to things just like you said but the main basics are things that
everybody or anybody even in the literate could basically understand like you could walk into a bank set up
an account it's super easy you need to like save your funds or like keep it safe all you need is a
login right and you don't need anything like you don't need to know or you don't need things like
keys you don't need to like hear like other random stuff like that right you could know about like
bonds you could know about whatever
in the banking system but when it comes to like crypto people tend to like often use big words
that in itself is a problem right when you're trying to like explain things and you could see
that as well when it came when it happened with all the nodes right it was kind of like marketed
or explained using a whole lot of complex word when it was just something really simple or basic.
Same thing with rooms and it kind of like affected them long term.
So I'm just going to say that. Right. So with banking system, it's the barrier to entry is like super easy.
People understand what they're doing. But with crypto, it's something new. Right.
But with crypto, it's something new, right? And we don't need to make it super complex by, you know, using big words or trying to like bombard people with information real quick. Yeah, I understand that abstract is doing things like making it super easy, right? Because you can use your email to set up a wallet. But this is also something that other wallets hasn't started using yet. And when it comes to like abstract, it's strictly abstract only at this point, not even like total EVM.
So there's still like a whole lot of friction in the space.
And I feel like starting from the ground level up is better to like sort out the frictions than jumping straight away into like oh yeah interoperability and trying to do
things multi-chain because even if we're coming IRL right now you could go to a bank you could
get a mastercard you could get a visa card you could get a verve and if you're traveling like
globally right you could like most times just use your mastercard you can't use your verve
any other place right in other countries right and that in itself
i could compare that with interoperability in web 3 so there's always going to be that kind of
friction if we're going to be really honest with ourselves there's always going to be those kind
of frictions that always exist and eventually it'll come down to like people deciding where
they choose to trade at maybe like ev EVM chains, whatever chain it is.
But I just feel like we need to like make other things simpler, make people find it easy to understand, make people come into the space and be like, oh, yeah, this is it.
I could just ride with this rather than like maybe focusing attention in more complex things that people coming in at the ground level
can't really understand this is also a product and education thing no because if you come to
say a banking application today they have a flow that walks you through in which you set up your
account even if you go to them in person they will print things out give you your account details in
print um and i think this is a product problem because if the flow makes the user confused,
throws big words at them
without really explaining what's happening under the hood,
it's just going to confuse them and drive them away.
Half of the Web2 world
and people not in the crypto Twitter bubble
anywhere look at crypto as illegal.
So if you don't simplify it for them
through good products with good user flows
and without educating them it's just going to cause problems i'll give a nice story here right
and this is something that actually happened to me i was walking out of the gym uh late night
about 9 30 10 o'clock and the main receptionist was also walking out with me and we were just
talking and he mentioned that he's an artist so So I was like, dude, have you checked out the base app? You can tweet, like post things out there and you might just make some money if your art piece goes viral.
It took us five minutes to help him set up that account.
And the flow was so, so, so good and so seamless with pass keys.
And he does not have to remember any wallet keys there as well.
And I just felt, wow, this is is this is one step closer to the adoption
that we talk about and dream about but actually translating that into tangible code is what that
app did so huge shout out to them and now coming to like hyper liquid phantom perps and you know
a little bit more technical where you need to have some crypto experience if today you go to phantom
even if you have like a hundred dollars on it and you go and
deposit it to your perp trading account, which trades directly on Hyperliquid through Phantom,
it teaches you what to do first.
And I think that experience was just so good for me because not only did it teach me the
fundamentals of perp trading again, made sure that I lose at least $5 less this time, but
it also made sure that I understand their application properly because whatever they made was bespoke for the first time a wallet integrated perps
directly on hyperliquid and i think that education in the product was needed um so you can have
interoperability you can have intense across as apis for it uh half of the crypto world uses it
but um you will not be able to solve adoption
until you make the products actually clean
and you give a shit about the user's experience.
And that's just what I feel.
So it is risky, but still works.
I'm just going to...
Can I just jump in real quick
and just totally agree with Kanishka on that?
I feel like, you know,
what we're trying to do with What is a Joke
is to do exactly what Kanishka and some of the, you know, and even Juju kind of touched on, which is just
kind of that bridging of that world. And how do you bridge that world, right? So the biggest problem
that we're seeing with the mass adoption is people don't understand it and people don't trust it. So
how do you build that bridge? And, you know, what, what we're trying to do with Web 3 is a joke is to trying to put it in comedic terms and trying to bridge it by,
you know, simplifying as much as we can of understanding some, you know, just simplifying
things. Like when you throw something like RWA at someone, they're like, I don't know what the hell
RWA is. It's like, well, that's a real world asset. Like, you know, people are like, why is it so
complex? Why can't you just say real world asset?
And I feel like the tipping point comes when, you know,
you look at apps like gaming and DeFi and consumer wallets
and you stop advertising like, oh, it's built on like X chain.
And instead you push only like use case and experience
very much like the way Kanish just expressed it to us.
It was very simple.
People understood it.
We got it.
Like, okay, I was walking out with this assistant person, you know, introduced them to Pace.
It was seamless.
They got it and they adopted it.
You know, like when you look at Starbucks, isn't really marketing his Polygon NFTs.
They're not, that's not what they're doing.
They're saying it's Starbucks Odyssey Rewards.
Like, okay, the average person who uses Starbucks, they understand that.
It's simple.
They get it.
And I feel like that's where the tipping point is, and that's where the simplification comes in.
Yes, but I think that we need to focus on two different user types.
We have the basic user persona that don't care about everything about Web3.
They don't care about us. They don't care about space.
And then the most use case applicable for Web3 in this type of community is payment channels.
So I believe that on one side, we need to solve that.
We need to make payment channels as simple as possible, which is extremely simple process.
We are talking about DeFi, Yield, aggregation, all of that, that is amazing for us,
and probably for a lot of people that are still out of the Web3 environment,
but the basic people just need payments.
I live in Argentina, and actually we need a payment system that is not connected to the banking system.
In my country, we don't believe in the bank.
We have more than $200,000 in our matrix out of the system, making zero yield, losing 10% of purchasing power every year.
And we don't care about that because we don't believe in our system. zero yield, losing 10% of purchasing power every year.
And we don't care about that because we don't believe in our system. So if we could have a Web3 completely bankless mechanism that don't steal our money and in which our money is safe,
this could be like a skyrocket in Argentina, for example.
Just give you that example.
But it should be like extremely easy, as you were saying.
I cannot agree with that more.
That is probably one of the best stakes here right now,
that you should be able to spend your crypto normally,
go about your day without getting centralized banks involved,
without KYCs involved,
and just make the architecture and the flow permissionless
because that is what crypto intended to be um thank you for shouting that out i literally
tweeted this today like three hours ago so yes that is a very very good point
i'm sorry yeah the good thing that you mentioned right like the fair point is that we need to
understand what type of users are we targeting right like
there's like retail users or like people that don't care about decentralization privacy there
are users that care about decentralization and privacy and they're also like users like that
are more professional which we're talking about like institutions uh like market makers or whatever
like hedge funds and so on because if we're talking about like a special DeFi
and it's further adoption, right?
Like, so for example, currently if we look at the space,
like why those, for example, players have a problem
because of the liquidity and book fragmentation.
Because you have best prices and debts
like scattered across L1s, L2s, Dex.
So you can't really lean size without hopping chains.
Then you have to for example have
the flaws where say like okay message then execute means partial fields and stranded legs when any
hope fails and then for example the other stuff that you have is like collateral fragmentation
which is where margin is scattered across chains when you like which means higher national parts
slower rotations like you have haircuts on
transfers and then like when you also have to take into account their opposite reconciliations
right so you have like multi-ledger pnl fee attribution time stamp alignment like you have
also missing audit trails for compliance and so on and like basically like what you need to be
able to also provide those guys is that like they they have to have a solution which enables them to go through one bundle or not at all.
And there's like no leg stranding, no stock inventory, no, like, for example, sand the bridge, but couldn't feel like that lets them run tighter spreads and bigger size across chain with far less risk capital part per venue.
tighter spreads and bigger size across chain was far less risk capital per
venue.
And this is basically where you have actually centralized exchanges
prevailing because to a certain extent, they do offer that to them.
And this is basically where we'll also have to get,
like if we're trying to kind of get DeFi also,
like adoptions for DeFi further.
Yeah. I mean, you're hitting on a key point.
Providing simple experiences to users
does not mean it's simple underneath, right?
When I use a credit card to buy an airplane ticket
in another country, you know,
the number of systems that all cooperate
to make that feel like I just bought a thing
is astonishing.
And we have the same thing that needs to evolve in Web3.
That's why aggregation to do it requires orchestrating activity
of a single user use case across multiple chains and ecosystems
to deliver that simple kind of payment experience,
to be able to pull together and aggregate my liquidity
that is spread across multiple sites and multiple applications in order to buy a thing that might have been in this Pace application that might have been on a different chain than where my money is.
And the buyer is on yet a third chain than where I'm deploying my art.
my art and I want them to buy it, right? If we don't have the orchestration that lets you have
And I want them to buy it.
sophisticated interactions across these different chains, then we can't get there. The world of Web
2 is smooth because you've got UIs talking to an orchestration layer, talking to microservices
underneath. From that point of view, Yearn and Beefy and Uniswap and all those are microservices where the user experience is a user experience that's just layered on top of Web3 or Web2 in the historical case.
And so our first product is focusing on being, WiMAX is focused on being a command center to be able to move your money around easily where you don't care about the underlying infrastructure.
You just care about I want to put my money in Aave. Now I want to pull
it out and buy that thing, right? And I want to pull it out and put it into compound. The fact
that one's on Arbitrum and one's on base is not my problem. To have a working payment system
in the way that you want to have, in the way that people want, it's gotta span these multiple chains, these multiple ecosystems.
And in order to have it be able to pay for
the Web3 services that we wanna provide to people,
in order to be able to pay your loan
or even get a loan in Web3 and then pay it off,
where that loan might be in a different ecosystem,
all those have to bridge smoothly where,
and it's gotta bridge smoothly,
not because someone did a bunch of offline stuff, but because our wiring on chain supports it.
And so, you know, not just swap intents, not just, you know, simple, you know, sort of simple, but really important and really valuable intents like that, but intents of the flavor, you know, I want to pay down my loan out of the earnings from this yield over here.
Make it happen i want to you know when when when when yeah my earnings cross 10 you know rebalance my portfolio
those kinds of things so dean let me just uh put this to you then and the rest of the guys here
is one thing that i wanted to talk about when we talk about abstraction and intents
is actually the end result i believe of this kind of period that
we're in of like web 3 intersecting with ai and all that kind of stuff um is that the the human
is abstracted away and the intent is very very general so like you would have when we talk about
ai agents in the not too distant future and i mean this in months or
very very few number of years less than three years you'll be able to have an ai agent that
is just just for you that for a relatively reasonable amount of money um that looks after
your needs that understands your needs based on analyzing all your financial history and some
goals that you put in okay i want to buy a house in two years or i want to pay off my kids college
tuition in the future and so on and it'll understand broad goals like okay i want um i want to earn i
have you know x amount of capital and i want to earn yield on that that's above five percent for
example and it'll go out and look for those opportunities in Web3 on your behalf
and it'll execute a complex series of actions
that are not always easily predicted by the human
but will be logical to the AI.
And I think one thing that's quite interesting at the moment
is there's lots of UI optimization for humans
in the sense that we're looking at this like
we need to make it as easy as possible to onboard the next billion people to web 3 like that was the cliche from a couple of years ago
that everybody was saying in their marketing but another thing that possibly might catch people
off guard is i think the most successful dApps in the future are the ones that are optimized for ai
um i don't know exactly what that looks like. I think there's a transition period where,
and this is at least the direction we're going,
is where AI is assistance to humans,
but the humans are in control.
And you gradually get more and more handing off of,
yeah, you do that for me,
but I got to feel like I can take control when I need it.
And I might never need it, the better the AI is,
but I've got to be able to
put guardrails, enable structured cooperation and change my mind and get my money out when I
want to get out and that sort of thing and get more, you know, build up trust in the agents that
are working on my behalf. And yeah, I believe that in, you know, certainly five years and probably three, as you say, financial advice and financial support will largely be done by AIs.
But, man, giving people, making sure they have confidence in what is happening is critically important.
But much of the rails and the infrastructure to make that happen for humans to have control can be made available to AI agents, right?
You know, the same things, the same plumbing that we're doing on chain, the same, you know, kinds of sort of semi-on, semi-off chain plumbing that you get from across, right?
through one pathway, you know, through our, you know, WiMAX infrastructure or whatever else we
build can control accounts and activity on multiple chains in the same way that a human does,
that actually gives you more safety and ability to put guardrails around what the AIs are doing
for you so that you can, again, straddle this, I want the intelligence, but I also want the control,
I want them to work together
on my behalf. Well, I think that's an interesting question. I'd like to throw that out to the rest
of the speakers here. Is AI optimization, and like from a user experience point of view,
is that the same as human optimization? Or what are key differences if if you know them guardrails um in one word a human when he notices oh shit i'm losing money um will put up some guardrails
on the money that they are about to lose maybe hit a stop loss or whatever they'll do there
right save the loss is basically the first human instinct right fight or flight
ai will not know what to do that unless you tell it to. It can execute actions for you,
but you need to spend time teaching this AI what to do. We saw this happen with Coinbase agents.
You can natively get across in the Coinbase agents. We made an agent, we just said, hey,
just arbitrage and just bridge over and wherever you find more volume, just keep bridging and
bridging and bridging. That AI was not really competent enough to understand simple profits.
So it saw quotes and but then when it was time to actually execute them, it fetches data for 10
seconds, makes a strategy, but then when it was time to execute the strategy, it did not like
verify it on any front. So these are issues that we saw with like ai agents in general and
we've actively been trying to like force and bring this particular aspect of crypto out more with
with crypto and ai um the only one that actually worked nicely was notion ai and when it when we
told it hey just make a doc on like erc 7683 and make it work. That was the only good one.
Apart from that, if we give activity to any AI agent,
it suffered and it had no guardrails, it went haywire.
So we were hesitant to even put $100
into that AI agent's wallet.
And this experiment is currently ongoing.
I share a bunch of this on my Twitter, but that's one thing.
Guardrails is where AI needs help. Yeah, I would love to have a read through that. So you said that this on my Twitter, but that's one thing. Guardrails is where AI needs help.
Yeah, I would love to have a read through that.
So you said that's on your Twitter.
I'll check that out.
Yeah, I'd love to hear anybody else's input here,
particularly from the Protafire guys,
because you obviously deal with a lot of dark design.
Is this something you're thinking about internally?
Sorry, guys, I missed the question.
Sorry about that.
Yeah, no problem. I'll
repeat. So we're just talking about
UI optimization
for AI
and humans. Is it the same
thing? Or, you know, like
Kanisha's just been talking us through
the guardrails
that he was putting in place for one of the AIs that he was working with because
he was mentioning that it went quite badly wrong when it was
given a simple task so yeah I'm just wondering from like a UI design and an
optimization perspective is it the same when you're optimizing a system for AI
as it is humans or is there a lot of difference?
Well, I guess it's completely different.
In Point of Fire, to be honest, we are not addressing this kind of challenges right now, but we are doing some research.
right now but we are doing some research and what we have been uh the discoveries is that
there is like a new uh completely different ui that is 100 oriented to to ai agents and has
nothing to do with what we have ever seen before so i i guess the answer is different, it's completely different. But what I've been
researching regarding that was not related to Web3 yet. So basically how AI can interact within the
regular internet in a more optimized way is completely different from what we have been seeing.
So, yes, it's a new challenge, I believe.
Interesting.
Does anybody else want to weigh in on this before we shift to a different question?
I just wanted to chime in and give sort of a big thing that went down, I believe it was last week or the week before.
So just to kind of answer your question up top, you said, I believe for humans, like the UI design is about experience optimization, right?
But for AI, the guardrails are more about behavior optimization.
And I think there's a big distinction between those two things. However, and I believe, like, I think it was like two weeks ago or something, X had this situation.
So Grok had this meltdown, basically, where Grok was asked. Now, this is not a political statement
on my part. This is just what happened. I'm just stating the facts. So Grok was asked about the whole Middle East,
you know, situation that's happening between Gaza and Israel. And, you know, and Grok said what
Grok said. Grok said, yeah, you know, Israel is and the U.S. is involved and whatever, right?
And basically, Grok was suspended for a few days. And people are like, what happened to Grok?
Grok is like, I'm sorry, I have been suspended for a few days. And people are like, what happened to Grok? Grok is like, I'm sorry,
I have been suspended for a few days. And then Grok came back. And then people asked the same
question to Grok. And then Grok said, I don't know the answer to that. And it just behaved
completely differently. Like it didn't know what to respond. It had completely changed its response
from before. So Grok had basically been compromised, according to a lot of people.
So I feel like that's the guardrails about the behavior optimization that I'm talking about.
So you can go and I feel like there is some sort of manipulation that can go on with AI, where,
you know, AI, I guess, maybe, perhaps, maybe in its purest form is giving you and just
stating basing basing it on what the facts are and all the information that it's gathered but
you can go and basically tweak it in many ways for it to give the responses that you wanted to give
interesting yeah the reason I brought this up is because i listened to an interview with sam
altman recently where he stated that he believes that in the future it'll be the the experience
on the internet will be very different to what it is today so you will like have your first point
of contact essentially with being online as the ai and then you will say you know like today I
want to do x and y I want to post photos I want to do I want to watch videos or I want to do whatever
but you'll access that through the AI agent and that will kind of navigate all of the the click
through and all of the awkward bits on the website for you and just take you to what you actually
want which is quite interesting so I think when we think about aggregation and optimization into the future
it's worth bearing that in mind from a design point of view because if if ultimately more
artificial users will be using the website than human users it begs the question who should we
be optimizing and aggregating for and i think that's very important when we talk
you know when we're talking about longer term projects about daps that we hope will be around
five ten years from now but um changing changing track a little bit i just wanted to ask your
opinion on things like uh gas abstraction and so we see like a lot of a lot of people talking about
that and it's generally a really positive experience for users.
But I'm just wondering if any of you guys are using or have good experiences with gas abstraction, how can it be?
You know, what are the most promising approaches? What are the tradeoffs?
Well, first off, I think it's generally critical.
Well, first off, I think it's generally critical.
I mean, when I walk into a UPS door to mail a package, I don't pay for a little bit of the clerk's time and some of the lights and a little bit of the floor space that I'm taking up.
I pay to ship a package, and it's up to the business to make all of the costs work for them and really not my problem.
That's one of those fundamental things that the UX of Web3 just got wrong from my perspective.
And so gas abstraction.
Sorry, can you said it.
I just said that's a great example.
And so, you know, so by as we get to these orchestration layers,
this goes back to the AI, right?
You know, the orchestration layer might be an AI,
but right now that's not what we build.
Having that do a bunch of work underneath the covers
as a business, it's got to pay for the work it does. I don't care or I shouldn't be paying for
the work it does. And by separating those out, now you can do much richer functionality for
rebalancing or amortizing costs as I do bulk operations or stuff behind the scenes that
businesses do in order to provide better service with less capitalization. And that's something
that because of the fundamental gas model that's currently in Web3, it's actually an obstacle. I
can't cash things and have the hundreds customer then carry all of the weight of rebalancing my cash, right?
I've got to have the business do that. And so gas abstraction will not just make for a better
experience for users, but, you know, and this will require the orchestration layer in the middle to
be able to make these decisions to provide, you know, an instant liquidity layer that spans across
multiple chains. But it, but that gas abstraction will
enable us as people producing these applications to take advantage of, you know, the software that
we've had 30 decades, you know, three decades of building interesting financial instruments
to use that to provide better experience at a cheaper cost with higher profits out to users.
a better experience at a cheaper cost with higher profits out to users.
So I think it's critical and a fundamental part of orchestration.
I want to chime in on one other point there because I've worked with multiple L1s who had gas station programs and, you know, different other pre-abstraction stuff, if it's not built into the base layer
and there isn't enough incentive or value for a third party to soak up that orchestration
business, it's usually not worth the rub.
That's what I've seen is that you'll end up with a need for light clients and need for
some localized client, but I'm not hearing Kahnis.
I apologize if I'm talking over him.
That's very interesting. Can you give a more concrete example?
Not sure if somebody's rubbed her here but i'm struggling to hear people uh nikki it looked
like you was talking but you you your mic was off but then you weren't saying anything and then
i think dean asked omnity a question um omnity did you hear that i'm just trying to work out
who's rubbed no i couldn't hear dean okay right so then probably you are rubbed so basically i'll just i'll just uh explain so
dean just said that it was interesting and he wanted to hear a more concrete example from
uh what you were saying essentially oh well i'll i'll give the super short answer and then take a
lap um in a nutshell when you um make a station, you're sort of recognizing that the protocol at the L1 level is not going to consider a pre-allocation of GWE or SATs or whatever other pennies, you know, for execution.
And then, you know, that always has to be brought at the time of execution.
that is designed to, you know, preempt that to prepare the sort of, you know, pennies, gas needed in advance,
then, you know, inherently you're creating an operation cost, which is perfectly fine if it's Ethereum
and there's, you know, 100 people trying to cram through that door getting that one DAP to work right now.
But, you know, if you're in, let's say, Cosmos and there's multiple blockchains and you're trying to do an IBC transaction across multiple chains, you know, preparing the gas for that in advance, it requires a sort of loss leader.
And that's why I'm saying that I think L1s need to be looking at that as part of their protocol
development to make gas easier for users in general, because offloading that to a third
party or expecting a third party business to come up and absorb that orchestration businesses.
It's not crazy. It's not wild, but it's a rough rub. So I'll take a quick lap. Thank you.
I totally agree. I would just say, though, and I do agree with that in general,
I think it would be nice to see gas grants as part of the grant program system that a lot of change run.
But the one thing I will say is that you have to remember that essentially the core business,
you know, one of the core business models of the L1 or L2 chain is going to be the fees
earned from transactions.
And so when they start sponsoring gas, essentially like giving away their core product for free.
There's other stuff, obviously, that they can use to make money, but I just think that
perhaps that explains some of the reluctance to do that.
Yeah, our approach to it is we have a smart contract that is the business, and it's generally
creating and managing accounts on behalf of the user through the smart contract.
So what that means is the user isn't directly issuing a transaction on Arbitrum. They issue it through Wimax, and that
will then go and, you know, issue the contract on Arbitrum to pull it out of Aave and bridge it over
to Compound. And then we issue the command over to Base to put it into Compound. And so the
orchestration contract is paying the gas
because it's sort of the direct customer.
So if the underlying platforms can provide better support
so that someone who wants to provide a business
can cover the gas for their business,
and it's not that they have to take on more
of the gas responsibility than the business they're doing for the user,
that would help a lot, right?
That could really simplify and make it much easier to provide, you know, more complex businesses
on, you know, for users where users are then paying transaction fees.
I mean, I'll happily take, you know, 1% of a large transaction in order to pay for your gas
in almost every circumstance.
I guess here, one of the key points here regarding the gas and how it plays out in profit and loss for businesses is the scale.
If we reach mass adoption, gas shouldn't be a problem because it's going to be divided among a great deal of value.
And it should be a marginal thing, not playing a key role in a profit and loss business unit.
That's what I think.
Well, you say that, but look at what happened to the the big l2 scene got going with eath
um eath fees on swaps were almost prohibitive if you look at like 2021 around then they were like
you know you like a few hundred dollars a swap on uniswap just like i mean i'm not saying that
ethos is the perfect gas system because it doesn't, and we all know that. And that's why there's like a million other chains now than there were when Ether was primarily used.
But I just think it's still an issue.
And I would like to see more gas abstraction just as part of the user flow.
I think the way that Dean said it was really great, which is when you go into a business,
you don't want to know that you're paying for all the individual parts of the operation.
You just want to pay one thing,
and what you pay is what you want to buy, basically.
You want the service or you want the products.
And there might be some cost within that that's associated with business maintenance,
but you don't want to know all of those pieces.
And certainly as a user, a fresh user that doesn't understand this stuff anyway, you certainly don't want to know all of those pieces. And certainly as a user, a fresh user that doesn't understand this stuff anyway,
you certainly don't want to know any of that.
And I think, so there's this alignment, I think,
that needs to happen between Chain, Wallet, DAP.
And so I'm interested to hear, like, if there's any examples
or if anybody has any ideas of how we could make this vertical stack of what people are using to interact with Web3.
Either combine it into one thing or abstract it all the way.
Yeah, I'm just interested because there's like three essential points of contact with Web3 at the moment.
points of contact with Web3 at the moment.
And fundamentally, no one-chain solution is going to cut it
because most use cases in the future will be cross-chain.
And so that's why it's fundamentally got to be
between the UX layer and the underlying service of the chain
to address the gas fees.
When I, I mean, I just saw, can you just get right? If I do an across transaction,
that's already two chains.
Whose gas am I paying?
It's got to be absorbed by the business of transfer.
Technically both.
You've got to pay the origin fee gas
and the destination fee,
and that's going to be a part of the transfer.
Essentially, which is what the code does,
input amount minus whatever fees you're paying on both the fronts
is your output amount.
But I also want to extend the question a little bit
and follow through here and ask,
doesn't a super app solve this?
If I go to Infinix today, they do Gasless
and they have all the features you would
want and you can perform more using the wallet experience that they have. The base app is also
super app in a lot of ways and you can probably do that there as well. So my question here is,
you mentioned that what vertical stack should be used? Is the answer not simple? Just use the best
super app that you could find or just use the one that is the most talked about
till you find the best one for you.
Those are really chain abstraction apps
where they are at one time,
I can, underneath the covers,
do a transaction on Arbitrum and then Base
or whatever it is.
But that doesn't let you do action over time.
That doesn't let you do auto compounding.
It doesn't let you set up a DCA.
It doesn't let you do a lot of the things where if I want to long live interaction, if I want a mortgage,
if I want that's automatically getting paid down or a yield position, that's going to build up a
separate Bitcoin position. You can't do that in a super app. You've got to do that with a smart
contract on chain. And that means that the functionality, you know, the place where this goes to is that orchestration.
There's multiple ways to do it.
There's not just the way we're doing it.
But fundamentally, you need that richer functionality in the future.
Super apps are nice as a transition,
but you really need the rich on-chain functionality.
I agree.
We have to-
Sorry, go ahead. Go ahead.
No, just wanted to say that we have to bear in mind that in Web3, we are used to deal
with different things.
But in Web2, we have only one application for use case.
We don't do multi-homing.
Exactly.
So eventually, there's going to be some piece of the stack that is going to take over
this and and and we will have to choose either we will use the the apple the apple phone or the
or an android i guess we're going to end that is this not though like an integration problem for
the super app so say there's like a loan protocol that everybody loves and likes,
and everybody wants to use.
So in this particular scenario, if your particular super app, I'm just going to take Infinix as an example, because I use Infinix.
If they just integrate that protocol, does that not solve the use case?
So it's not really a product problem.
It's not really, it's just an integration hurdle.
And to take this a step further, though, I would say that the ultimate super app will be in AI.
That's a good point.
I mean, we're talking about Infinix here.
Maybe I can add here, like actually it was pretty well mentioned that yes, it smooths the UX, but it doesn't really delete the hard part.
So what I mean by that,
it's like Infinix,
it still doesn't really offer
real time cross-chain execution.
It doesn't really have this reactive logic
across L1s and L2s.
You don't have this generalized messaging
and execution.
And essentially,
it's on the UI, right whatever you have and uh yeah it does
does it like that smooth certain stuff like one login and smart account ux you have like gasless
one token flows and you can have this unified portfolio and connect layer but still like uh
what it doesn't make disappear is that like volatile costs and quad decay.
So, and there's also bridge and aggregator dependency.
Cross-chain lags still rely on external routers.
You do have also certain security and compliance tradels.
Yes, it solves the perception, I would say, and workflow,
but it doesn't really solve the technicalities, I would say.
That's an interesting take.
I really like that.
But yeah, I mean, AI will be the next super app.
I really like that.
Do you mind if I tweet that?
That's a good one.
Yeah, go ahead, man.
Very good.
Because ultimately, I know that as, you know,
designers and marketers and BD people in this space,
we spend our lives agonizing over app design and, like, you know,
user flow and user retention and all this stuff.
And the reason I brought up the AI thing earlier is because I think,
like, these conversations are going to kind of seem silly in, like, and all this stuff and i the reason i brought the i think earlier is because i think like this
these conversations are going to kind of seem silly in like five years when the ai is just
crawling everything in a completely different way to what we imagined and executing things very
rapidly like i don't know if you can tell but i'm an ai optimist i don't think the world's going to
end when we get agi i do think the world's going to change very quickly though and a lot of the
systems that we use today to do uh daily transactions and our daily world's going to change very quickly, though. And a lot of the systems that we use today to do daily transactions
and our daily needs are going to be massively uprooted and upended.
And the best companies, I think, are the ones that just dive straight into that
and work on AI optimization now.
And then they'll end up basically eating everybody's launch over the next five years.
they'll end up basically eating everybody's launch over the next five years.
Yeah, actually, there's going to be brand new devices that we will,
they may make us see our current cell phones as things from the past.
They're going to change completely.
I heard one of the guys that is an expert on AI talking that they are actually working on a new operative system, brand new native AI
infrastructure that has nothing to do with what we've ever seen. So that's pretty cool.
Yeah, it's amazing. I heard Sam Altman again talking about this thing they're designing at
OpenAI called the PUC, which is like a small round device that you can have in your pocket.
And he thinks of it like a third core
device so if you if you two core devices it's basically your cell phone and then um you know
some kind of computer whether it's a laptop or desktop for more deep work that the puck would
be a third device that sits alongside that like oh personally i just feel like that'll just be
integrated into the cell phone eventually when he does a deal with Apple or whatever.
But, you know, apparently, like, the idea is that it's just super,
you know, this puck will be packed full of really powerful
and specific hardware to run the AI locally.
But I think that's an interesting thought.
And, you know, the thing that I'd actually to all bear in mind
is if this seems a little far-fetched when I'm throwing out timelines
like a few months or a couple of years just bear in mind that each iteration of ai
allows the next design of ai to be done more efficiently and quicker and so you know that
we very very quickly approached the runaway trade moment um and so i i do think the future will be
completely aggregated and abstracted away but i think the way that we'll access it is by a voice command or by,
you know,
interacting with an AI centric device.
I want to put a grain of salt on that.
Because that was tried the rabbit R2 like that.
That's an APK that was loaded to a device that didn't have the horsepower
locally because making,
you know,
NPUs that small that are that battery efficient,
like that's a physical boundary that we have to cross before we're going to hit that.
I think that your point's valid and that there is a reasonable world where we have like, you know,
low-scale, tiny LLMs running on our devices locally, and those act as our sort of, you know,
concierge, large action model, whatever you want to call it. You know, that makes a lot of sense
to me, but, you know, the horsepower to run that stuff locally, it's just starting
to be available, and Intel and AMD are both falling by the wayside to some of those risk
competitors and ARM competitors.
So I think on the hardware side for smaller devices, it's going to take a little bit of
time, but I do really appreciate and like the approach of aggregation through like clients managed by a sort of, you know, intermediary, that
orchestration layer that we were talking about before. Yeah. Yeah. And when you think about it,
it kind of has to be locally, because when you think about loading it full of your very sensitive
personal data, especially when we talk about finance. So, you you know the business that we're often in is
decentralized finance and we know that that's really rapidly smashing into traditional finance
and eventually there'll be one essentially one thing um i i don't know about you guys but i
personally wouldn't feel comfortable sharing all that information with a centralized server in some
random place that could be compromised at least with a local
device you would know like nobody can physically interfere with it um because it's in your house
right i i just want to add that i i don't need another device from open ai
and i i don't want them to advertise in my dreams. Like, let my dreams be my dreams.
Because I don't know what they're cooking up over there, but it's a little terrifying.
I'm just going to say that.
I have a question.
Have you ever thought about self-custody abstraction?
thought about self-custody abstraction because one of the key things here is that we need to
keep continue to be self-custody I guess that's like a main goal here and all this abstraction
it may work but we shouldn't lose that and and burying the 24 or the 12 words and all of that is kind of a problem.
So do you see any possibility for a self-custody abstraction?
Well, there were quite a few protocols in the last cycle
that were doing things like an abstracted seed phrase
where the seed phrase would be 24 words or whatever,
but you would get three words of that 24.
The 21 words will be hidden from you and abstracted in the protocol,
and you would just remember in your own head the three words,
and that will complete the seed phrase.
There were some protocols doing things like that, which is kind of interesting.
There's always a trade-off, feel though with these things where you have like
the ultra secure seed phrase model where like you know there's more atoms in the there's less
atoms in the universe and there is combinations or whatever um and every step we take to make
that easier to deal with gets us closer to less security and so i feel like yes it's super annoying writing down a seed
phrase and then hiding it somewhere presumably or trying to memorize it but that makes it impossible
for anybody to steal your shit unless you get it unless you get the security steps wrong or you
store it on an iphone or something so for me like i mean obviously we're tech people, right? So personally, I would take ultra secure and annoying over simple and less secure.
I'm serious.
I would rather have both.
Thanks.
So I think this goes back to the AI guardrails thing, right?
I don't want, you know, I think AI is not the ultimate user experience.
I think it's, I want it as an amplifier for me, you know, and the, a smart contract with
rules, not an opaque AI should be the thing that controls, actually controls access to
my funds and my experience on Web3.
And the rules give AIs the ability to act on my behalf within constraints, you know,
audited log, checkedited, logged, checked
over each other, that kind of thing. And so that means that there's stuff where, you know, I can
have pass keys for controlling of certain things. I can have, you know, a ledger that's locked in my
lawyer's safe or whatever for other things. But fundamentally, you know, the user experience
needs to be easy. We want to be able to have all this AI assistance, but it's not the right thing to control everything.
It's the right thing to exist in a framework such that multiple agents can cooperate with
secure relationships with each other.
And that's what gives me custody of my assets is the smart contract framework that controls
the interaction of agents around my money,
not those agents themselves.
You know, so I'll decide it can only go to non-interest bearing accounts.
And now the agent's there.
So I agree with you in spirit, and I think choice is really important. But because of the way markets are efficient and tend towards the decommission of anything that is redundant,
I can foresee a future where if AI becomes universally adopted, and that's the way most people access apps,
most people access like um apps then the the basically the the force in the market will
uh make make systems trend towards basically being optimized for ai not humans and eventually
we'll probably get to something that's unreadable so i would say like i agree with you but if it's
more efficient to use an ai market forces will probably just push it in that direction over time until the human internet users are relatively obscure.
Yeah, I think it'll just be more efficient to use an AI in a structured context that provides those rules.
I think we can get there faster by having good guardrails because then it's easier, safer to adopt.
And so I think we'll actually get some of both of what we're talking about.
Yes, we can also create an AI service that actually can know if we are doing something willingly or not.
or not. So we could have like an inherent self-custody that is AI integrated and they know if we are
doing things by our own will or not and that can be a nice security feature, I guess.
security feature, I guess.
Nicky, I think you were trying to get in on one of those points there a few
minutes ago.
Just me about AI or what exactly?
No, no, I think you were trying to jump in on one of the points a few minutes
ago.
Yeah.
I think it was more about Infinix,
but generally speaking, coming back to our question
about AI and so on, how it's going to unwrap in the future.
I mean, generally speaking, I have a little bit,
I would say, of a pragmatic approach to it.
Because in the end of the day, what is like AI, right?
AI simply relies on huge amounts of data to kind of identify patterns and make decisions.
Of course, it depends the data you feed it with and you can allow it to adapt and improve its performance.
And what is good about AI is like it's just a boat.
You have to treat it as a boat, right? Like it can automate repetitive and complex tasks, like, you know, kind of free up your
time, do the stuff for you so you don't have to care about it and so on.
And we've been discussing quite a lot about the AI and the UX part.
And I mean, I think both things like they depend on each other.
So when we're talking about like AI leading to better UX,
you have to use AI to reduce effort, time and errors,
like kind of personalize it, predict,
like automate stuff where you can,
whether it's centralized AI tools,
not decentralized or whatever.
But it also like UX leads to better AI as well.
So because it helps you design the AI features
so it's legible, controllable, and safe.
So you give clear inputs.
You kind of, like, improve it as well, like explain outputs,
like, and essentially get it to the point where you need it.
So I think it basically goes both ways.
Yeah, it makes sense.
Omniti, I think you've had your hand raised for a few minutes.
Yeah, just a quick one, because personally, I come from a cybersecurity back end.
I'm definitely like a back end kind of person.
There is a world where we get to that Android and iPhone level of consumer choice for these kinds of UX
decisions. But that also comes with generations of arguing over standards. It's not because Linus
Torvalds did something great that the world basically runs Linux in the background without
realizing it. For anybody who didn't know, the majority of your cell phones are running a watered
down version of Linux inside the processor itself. We we didn't get to that point through a, you know, large
market movement that was centralized, that had a specific product and success. You know, it was
arguments over lots of different smaller implementations, more use case specific
implementations that were going back and forth saying, hey, I need this from Linux, I need this
from Linux.
And it's not that Linux is the perfect model, because obviously open source software is kind of different from blockchain economics. A lot of the time, unfortunately, they can be
both in concert and against each other. But there's a really boring world right in front
of us where we all just give up and use Ethereum because
Ethereum is the best thing that we should use and there's no reason to try anything else.
Like, I'm not saying that ETH is bad. Obviously, Ethereum has an important position in the market
right now and in the stack in general for most people. But there's no reason to fold into one
standard while we're still so early in debating standards, because in order to get to that world
where roughly
what Dean was suggesting, where you might have like a permissions framework that will,
let's say, at the operating system level, disallow your interaction between two sections
of memory where two different AIs are trying to make a transaction for you, pay a friend,
you know, go offer up and list something and then charge your buddy for it.
Like there's a lot of different mechanics in that large action model level of blockchain that we want to see.
And I think a lot of it is just consensus.
And that means we've got to fight about this stuff and, you know, have debates and try different implementations that will all, you know, in some way not congeal and still be beneficial to just discuss. Like there's a really wide horizon of standards that we'll be discussing for the next decade to three decades,
just on the core of that AI UX alone.
Yeah, I think that makes a lot of sense.
It's interesting you you mentioned about um different
clients demanding different things from the software it's quite interesting it's like we
we end up with a negotiated consensus that's based on thousands of businesses thousands of users
thousands of millions of developers um and then we get to something like a hive mind consensus
over what's good and what
isn't based on like literally millions of hours of testing um so i think that will continue and
although i'm talking a lot about ai here only in the sense that it relates to aggregation and
optimization i think that um engineer-based negotiation will just continue to go on. Perhaps some of it will involve AI.
But I don't expect everything's going to be replaced with AI overnight
because there's just a ton of systems that won't react to that well
and the guardrails won't work as well.
We tried to teach AI something, as Kenesh mentioned with his example,
where the bot, even though it's designed to go and seek yield it doesn't understand that you want profit so it just makes a bunch of bad bad decisions um it's interesting and i think this
is the way like what ai really lacks at the moment which is quite interesting it's incredibly
intelligent anybody that's dealt with
gpt5 or the newest grot model will understand like it really it really is getting very very good now but it does sometimes lack like fundamental common sense so where a human
would just take one step back and go hang on a minute my overall objective is x so i need to go
and find information on y and then execute x the ai might get so caught up in y
that it goes off you know basically like off the off the reservation and uh misses x entirely which
is something that humans tend to not do because we have a very good understanding of general
terms and objectives even though we're not as good technically as AI is right now on specific examples. I have had enough developers of a broad spectrum of skills working for me over the various years of running projects
that humans have done much bigger doozies than any of the AI tools I've used in the last year
in terms of, wow, you deeply did not understand
what we needed or how this system worked and just totally did something crazy here.
So I have more respect for AIs because of seeing those kinds of whoppers.
I guess it's also less likely that an AI would deliver.
So we hear a lot about AI ethics at the moment right and there's a there's a pretty good essay on this called AI 2027 which talks about AI ethics and alignment
um and we hear a lot about how like you know uh user retention uh like stuff within the model
like to help user retention so like speaking pleasantly to your user can trend towards
sycophancy which can you know flattery becomes sycophancy sycophancy becomes deception and then
deception becomes misalignment um and so this while there is that element to consider
you know an ai would is unlikely to deliberately steal from you for example where a human could
be desperate and they could make a decision
that I need to steal some software to go and make a forked version
somewhere else to take it to market quicker to rip off your idea.
And AI is unlikely to do that unless it's been directed by a human,
at least for now.
And so while AIs, like you mentioned, they can make some do's,
I think that it does remove an element of deliberate mal-intent. AIs, like you mentioned, they can make some doosies.
I think it does remove an element of
deliberate malintent, which
seemingly, for now at least,
humans are the only ones that are capable of.
As we get towards the
end of this
space,
there's like one more
question that I wanted to get into which is
like
developer incentives broadly
so we've talked a lot about
aggregation
optimization and a little on AI
there
but I'm interested to know
like from a software
management point of view or
if any of you guys are leading teams or
building dApps, leading developer teams,
how do you
incentivize your developers
to meet the standards
and push them towards
adopting aggregation
rather than sticking to
a single chain focus or a siloed mindset?
I think it doesn't take much of a leap from focusing on the user.
I mean, we pivoted heavily,
originally, initially,
sort of the smaller category chain abstraction,
but into orchestration as you
know what we you know a what our technology can do that nobody else can but b what you need in
order to provide this a unified experience to end users to you know end-to-end seamless experience
and so some of it is just experience with this is the kind of transition you saw
in web 2 where for a while people noticed
whether something was running on Amazon or Google or whatever.
And now, I mean, I bring up DoorDash and I order a hamburger
and I have no idea where it's running.
And so, you know, just echoing that transition,
it doesn't take much to go,
yeah, you got to abstract this stuff away.
You got to provide, you got to solve user problems
and provide user benefits, not wow them with your technology.
And that's just, you know, that's a typical transition
of starting with technologists, we build cool stuff,
but eventually you sort of realize, no, no,
it's got to be about solving user problems.
And that just leads to it.
So the objective is to meet users where they are today
rather than where we want them to be tomorrow.
Exactly.
I guess.
All right.
And Kanish, I think you were going to jump in there as well. Yeah, was just going to relay i don't lead the internal developer team at across i don't
think a dev rel should lead internal developers at a company but i do talk to a lot of teams and
like independent community developers who are building applications on top of the api stacks
that we have and i just asked them one simple thing do you want to silo off your users onto
the protocol that you deploy or do you want to silo off your users onto the protocol
that you deploy or do you want to open your gates and allow liquidity to inflow from wherever the
user has it the answer almost always is i want users to come in from wherever they are and be
able to use the protocol that i built on whatever blockchain i have and that itself is enough
motivation because the one thing that a developer wants to see is their product being used at mass.
So they feel proud of it.
And that is probably the best way to motivate a developer, that people will use what you built only if you follow this principle.
And if you can make this principle make sense in their mind, they will be more than motivated to do it.
They'll spend nights figuring this problem out and making it work.
Um, but yeah, that's just what I wanted to say.
And I totally agree with the previous speaker as well.
Yeah.
Nikki, I think you were going to jump in there.
Yes, yes, yes.
Because I want to think like, because I also speak to kind of a bunch of builders
and so on, and like, I've worked in various alons and all too.
So like, you kind of see this pain all over the place.
Because most of the apps, of course, they want to capture a biggest chunk of users,
biggest chunk of liquidity.
And when you're on one specific chain or you're dedicated to one specific chain,
you kind of miss out a lot of opportunities as well, especially because as a user,
because as a user like as a tab you want to attract as many users as possible wherever they
like as a tab, you want to attract as many users as possible, wherever they come from.
come from and um essentially like so far like the multi-chain experience in most of the cases
especially we're talking about non-uvm integrations they face like long integration cycles sometimes
duplicated infrastructure and you do have still like this frustrated user experience like kind of
shipping real time truly multi-chain products is impossible kind of was
current infrastructure i would say because you know every chain is different to support many
you have to dedicate certain resources there's infra bots there's still bad ux and uh essential
like what causes this is like um each chain is different on its own like finalized is different
on its own schedule there's no
kind of global atomic committee uh across them i would say there's no shared ordering there are
messaging limits of course there's liquidity and gas from fragmentation and essentially like what
you would want to provide like uh what a lot of builders would look into is of course to get this
like cross-chain execution fabric that cost science fasts like and exposes like trust tiers and also like
it offers the shared or coordinated sequencing like unified account and gas
policy and so on so basically like they really that's also apart from users like
would want to have this real-time, truly multi-chain experience.
And essentially, I think they're also looking to get into that.
I don't think there's a need to be on one chain, especially, if you want to capture,
generally speaking, the bigger audience.
And I personally do believe that like eventually users like nowadays will understand
for example what's running where and why and so on like which tap is on what chain but down the
road the users will not care hopefully soon. Hey guys I'm sorry to just jump in but I have a
meeting that I need to jump on but I really appreciate you having me here and part of this incredible conversation.
If you guys, if you don't mind, if you guys want to give us a follow on X, that'd be amazing.
Also, we're going to be at Token 2049 Singapore.
If you're going to be there, you want your project and you want to come and have an activation with us, DM us, let us know.
And we'd love to partner with you guys. But thank you so much for having me.
Awesome. Thanks for joining.
Yeah, I think we're going to be wrapping up soon anyway, so cool.
Does anybody have any closing thoughts on aggregation or any of the other stuff?
Or if you want to just jump in on that last question, feel free.
I certainly have some plenty of closing thoughts.
I think we're entering a world where capital is not just managed.
It's orchestrated.
It is fundamentally all our use cases are cross-chain.
And that is, you know, and, you know, obviously AI is a huge part of it.
But being able to abstract away and have on-chain automation and orchestration across chain is what will enable institutions, DAOs, and everyday users
to be able to interact with Web3 in a way that feels natural and straightforward rather
than awash in technical crap and details that they don't need to care about.
And so I think that's a critical element of the future.
And I'm so glad of, you know, you guys focusing on aggregation and driving this.
How do we pull all this sort of stuff together for users?
So thank you for having us.
Great. Thanks, Dean. Thanks for being here.
Does anybody else want to round us off here?
We've got Omnit here. Do you want to go ahead?
And if you guys want to plug anything, feel free.
Now's the time.
Yeah, I'll give it a quick one.
Just big on aggregation and the scale of blockchain.
That means Bitcoin's got to be part of it.
So that's our stack is Bitcoin L1 open source, tooling for developers, dApps for users.
We're the first open source DEX for runes as well.
So we're working on Bitcoin interoperability and we're definitely going to be back on
the aggregated in the future because we appreciate tuning into y'all's work on Fridays.
And also, you know, a shout
out for, uh, Agoric and for Uma, which are great products that we've, you know, uh, spoken to in
the past. And, uh, I'm sure that, you know, the aggregated is a good place where we can have
discussions with different folks who work on interoperability and also these other, you know,
less product centric and more like existential focuses.
So thanks again for having us.
Awesome.
Yeah.
Thanks for coming out of interest. Your decks for runes, would that work for Litecoin assets as well?
I believe the runes branch for Litecoin isn't up to date with some of the stuff that we
use, but if you can, you know, spend a little bit of time with PSPT constructions, the runes branch for Litecoin isn't up to date with some of the stuff that we use.
But if you can, you know, spend a little bit of time with PSPT construction,
then, yeah, that's a LitVM discussion we can have on the side if you'd like to.
Yeah. Darren, can you reach out to Omniti and just make that happen for LitVM?
Okay. Nikki, do you want to go ahead?
Yeah, sure, sure. Thank you so much, actually, for hosting. And general speaking, we all want to, like, we've been in this space for quite some time,
and we really want to kind of, like, have, step into this world or step into this space
where every wallet, every smart contract, and every chain, like, acts as one seamless, like, network,
and you can tap into one kind of colossal pool of liquidity
so so like let's just like kind of stop maybe to a certain extent also like integrating chains but
start shipping like outcomes like more products like and keep in mind right like one click and
any chain and um i'm insanely kind of bullish of this future uh and i believe that soon it will stop just being
a simple prog promise and will become a product like um and we will be kind of like building
better products on top of it and essentially like abstracting the complexity from the end users
just normal retail users and institutions and like thank you so much for uh for this
opportunity was happy to be here and really enjoyed this conversation thank you guys and
happy friday thanks for joining nikki khanesh do you want to go ahead round for sure this is actually
my second time at the aggregated space and i'm i'm a fan i absolutely love being here um this is also
a great place for me to just discuss ideas because hey hey, who else I'm going to do it with?
My mom.
She doesn't know crypto, but I hope she does one day.
Anyway, jokes aside, thank you so much for having me.
If you are a user who wants to go to any of the 24 chains that are across the post today, you can get there in less than two seconds.
And if you face any issues, problems, while you're integrating, please reach out to me and I'll make sure that I get you the help that you need
in the quickest amount of time possible.
Thank you so much Omniti for that shout out as well.
We absolutely love it when our friends and partners send love our way.
And I look forward to being here again in the very short future.
Thank you, folks. Thanks, Jack.
Awesome. Thanks for coming. Towers, do you want to give Protavire a shout out?
Sure, sure. Thanks for having me here and thanks for thinking in Protofire to be part of this type of space.
We love to be part of this.
And if anyone needs any help regarding development, adoption, TBL, whatever, you can reach out to us, to the Portofire account, and we will be happy to help you.
Yeah, I just want to echo that as well.
Portofire are, like, honestly, really professional, incredible development team to work with.
I don't want them to get too busy because I've got a bunch of ideas that I'd love to to have them take to market so i want to use them for myself but really cool all right all right
guys i think we've uh covered everybody here um thank you all for joining it's been amazing
uh this has been episode 26 of the aggregated i know Rock usually does some shout outs from the audience.
Probably not going to do that too much.
But everybody here, you know who you are.
Thank you very much for joining and see you next week.
