Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
NEAR Legion Open Mic #5 ft. Matt from Seda
Recorded: March 27, 2026 Duration: 1:08:22Space Recording
Full Transcription
Music
Music
Music
Music
Music
Music
Music
Music
Music
Music
Music
Music
Music Music Music Music Music Music Music Music Music Music Music Music I'm going to go to the next video. Yeah. Spend my night in the Iron Claw, Hammond.
Had a black dragon drop them shrews, now I'm switching. Yeah, spend my night in the Iron Claw Hammond.
Had no black dragon drop them shrews, now I'm swig- Thank you. Thank you. I Just sneaking in the night at near a I built the fortress this the new standard all right
Build a fortress, that's the new standard, alright?
Fuck Open Claw!
That shit is exposed as hell!
135k instances just sitting there ringing the alarm, ma'am!
Malware skills creeping, stealing all your keys!
Iron Claw's the future, war some sandbox, bitch, please!
Iron Claw!
Citrovers fuck capability, tokens tight!
AI never touches raw crabs, we do that shit, right?
From Open Claw to Iron Claw, I'm migrating fast!
Fuck the malware bullshit.
This one's built to last.
Yeah.
Iron claw forever.
Open claw in the trash.
Secure your azeans right now.
Make the motherfucking switch at last.
That and Black Dragon.
Adnir AI.
Y'all the real MVPs.
Fuck around with open cloth.
Nah, not for me.
Gia,
welcome to
the fifth Open Mic
Spaces with the Legion.
Hope everyone's able to hear us clearly.
Today we have a guest from Seda.
We've got Matt joining us in today.
So I want to do a quick mic check before we get started.
GMGM, can you hear me?
Yeah, I can.
And hopefully everyone else can as well.
I'm just going to check up on Ewan because he usually co-hosts this alongside me.
So just give me a second and then we can probably get started. All right.
Yeah, maybe we can get started.
We are at five minutes past.
So this is the weekly spaces we run for the Legion,
where we talk about exciting things happening in the ecosystem,
as well as an open mic for the community to share feedback,
things they've been working on, and just a place to vibe and hang out.
But yeah, I'm really pleased to have Matt from Seda join us today,
because if you've been tinkering around with Ironclaw
and some of the agentic stuff,
you've probably seen the Seda team ship a number of tools
and cool things out there.
So the whole idea of today's interaction with Seda and Matt
is to kind of get to know what they're building
and get more people to try out the fun stuff
they're putting out there.
But yeah, let's probably get started, Matt, with quick intros about yourself and Seda
because I think Seda's been around for a really long time in Nier.
I remember Seda back when it was Flux.
So yeah, I'll just pass on the mic to you
for uh introductions yeah sounds good thank you um yeah so for those in the near community they
probably as well as yourself uh remember when sata was flux um and just quick tldr on flux
flack was a first party oracle similar to the other Oracles that you see in the space,
your Chainlinks, Redstones, Pith.
And then from building the Oracle, first-party Oracle,
the company took a thesis that those kind of Oracles were too static,
essentially, to put it in quite like simple high-level terms.
And that programmability or Oracle logic was going to be sort of the next frontier for data
feeds and for oracles for where the space was heading that was their thesis four years ago and
i mean or five years ago and now here we are on a space talking about agents you know spinning up
data feeds well we will be speaking about agents spinning up data feeds on the fly
um you know reading in real life signals and then sort of the custom logic that has to go into that
feed. And, you know, we can dive into Seder a little bit later, but just a TLDR on what Seder
is now is it's a programmable Oracle layer. And the best way to think about it, if you're familiar
with Oracles is we sit half a step down in the sense that you can run all the normal, you know,
sort of processes that you would with an Oracle, transport data from source A into environment B,
but with the added benefit of being able to deploy Oracle programs,
which are essentially smart contracts that dictate how that data feed
behaves, sorry.
So from what sources you want to consume,
how you want to consume them, when they need to be consumed,
how that data is computed, and then how it's delivered on chain.
So the same way that Ethereum opens up, Nia opens up a design space that's
effectively infinite with smart contracts and what can be done, we're doing the same
thing, but in the Oracle space.
And then myself, I'm in the expansion team, so across marketing and growth at Sata, and
I've been with the company for two and a half years. So very excited for the chat today.
Very excited to talk about what we've been doing with Ironclaw.
But yeah, I'll leave the intro there.
If you have any other questions, let me know.
Yeah, I think that's great.
I kind of want to highlight the fact that, yeah,
Seda has been around for a really long time.
Under a different name earlier, of course,
but the offering and the allegiance to the ecosystem
has been there for a really long time.
So I probably want to dumb it down a bit
because there might be folks in the audience
who don't know what oracles are,
how do they work, and why they're important.
Maybe we can start off there
and then like
go down the rabbit hole of like hey how do how in your opinions have oracles evolved
what kind of use cases they have like in the agentic world and stuff
yeah it's definitely a great place to start so quick primer um on oracles and i'll keep it
brief and light and And if anyone has
questions, I guess they can drop in the comments or whatever. And we're happy to address those as
we go. But the essential oracle problem is that blockchains don't natively connect to any data
sources outside of their own environment. So whatever's happening on the blockchain,
which is obviously a massive restriction because in order to build applications that interact with financial data, weather data, corporate action data, RWA data, whatever
it might be, you need to have a bridge from these data sources to bring them from where
that data is being created into the blockchain's environment so that the applications being
built on the blockchain can consume that data. So price feeds, for example, you need to be able to have an off-chain source for a price.
Let's say equity is a very big thing now that people are trading on chain.
So the NASDAQ is a great off-chain source, one of many, where you can build an Oracle
that reports the price of the assets, say Tesla, and then brings it into the blockchain's environment
for applications to ingest and then use that within their trading. Very financial forward
use case. Of course, that's kind of sort of one of the bigger use cases for oracles that we've seen.
However, sort of where we think like oracles are heading is it's not just reporting
are heading is it's not just reporting a static score or a static price update, if you will,
that we need to have a system that can essentially blend the two worlds off-chain and on-chain. So
going beyond just bringing static and binary updates onto a blockchain so that an app can,
you know, stay true to the off-chain price of an asset. But there's so many more things that are coming on blockchain. And blockchain as a tool goes well beyond simple trading applications.
And especially when we move into the agentic world, it's something that I sit and think about
a lot. And I would love to discuss on this call because I still don't have very clear
thoughts in my own head. But what will agents be doing on-chain? And what will it look like
between off-chain agents and
on-chain agents and just in that conversation itself you can see that there's this kind of
off-chain environment that's always running where agents are obviously being built for right now
they're doing lots of things and then when that starts to merge with the on-chain world there
needs to be this layer in the middle that can communicate the data whether it's on a verif
verifability if that's a word whether it's on a verifability,
if that's a word, whether it's verifying that data and having this kind of like
neutral truth layer, or it's bringing data from the outside and bringing it on
chain, but having, you know, sort of computation put on top of that as well.
The design space and the thinking space just expands infinitely.
But that's kind of how we're thinking about oracles is we're moving from this
sort of static transport from, you know from source A into the blockchain as the destination,
more into this sort of dynamic space of like, where do I need data right now?
Why do I need it?
And how do I need it for the function that I want to perform on chain?
Yeah, that's interesting.
And I also probably want to add that, like, whilst it seems that like financial data is
like the biggest demand driver for Oracle's, but that's like, you could even envision like
Oracle's for like weather data, like other sorts of like real-time real-world data that like might be needed to be brought on chain so
I mean weather is just one example but I think like you could pretty much bring any data on chain
one thing I've also I mean I've always wondered is and it kind of differs from each Oracle provider is like, how do you ensure the data is accurate?
Yeah, every Oracle answers this differently
because everyone has sort of a different data relationship
with their data providers.
SATA is a little bit different to what the other Oracles do
in the sense that we really sit in the middle
and there's a marketplace on top of SATA.
And what I mean by that is that data providers come to us and they host their own data proxies
on our network. This is for private data. And then they make that proxy available through Seda
to the developers on the other side, which makes the marketplace. There's people who need the data
and there's people who are providing the data. Seda is a neutral layer in the middle. So we actually don't have these kind of relationships with the data providers, although
we do put forward our premium providers who we have close relationships with in the sense that
we vet them in short words, basically like DX feed. Pith is one of the biggest data suppliers
on Seda right now as well. We've built a lot of logic on top of Pith's feeds, and they've got some of the best in-class
data.
If people are curious about how they make their data verifiable or to make sure that the data
is correct, you can have a look at Pith's documentation.
But then with Seda, it's a little bit different because any data source can become an endpoint
on our network, whether that's via an API, whether that's via WebSocket.
People are now using agents to scrape websites.
There was a new tool released recently where you could scrape the website and create an
endpoint.
I think it's called AnyAPI.
And we tried that and that was spun into a data proxy over a course of an hour with
an agent.
So really in Sata, the best way to think about us is more of a neutral layer where the developer
defines what type of data that they want to get so they look on the other side and see what's
available where you're looking at something like Pith who has direct relationships and licensing
with you know Jump Street, NASDAQ, all these massive institutions that's a totally different
story because they're getting direct data from the publisher and then someone like Pith is working on top of Seda
or working in collaboration with us.
They've done all the due diligence with their data.
We're more than happy to collaborate with them
specifically, which is a lot of the PERP markets
that we're powering is coming from Pith data,
specifically because we need to make sure that that data is,
of course, of the best grade.
But a high level thinking is we're
more of a neutral layer in the middle where developers really
define how they want to use data on us,
if that makes sense.
Yeah, so if I understand it correctly,
like you're kind of a hybrid Oracle provider
where you are using other Oracles as well
to kind of provide more dynamic data, right?
Yeah, I mean, for an example on Nia,
we've spoken to a few applications there who use Pith,
and a lot of the data that they get there is great,
but there's some really niche or long tail data
that they can't get that a lot of the Oracle providers won't.
And in that case, we'll happily work with the developer
to source that data if we can.
We'll find a place for them to scrape that or to get that data.
But I'm more concerned with sort of the logic.
So like, okay, you have the data.
The data's available.
How can we add an extra element to it?
Because if you think about the space over the last five years,
we've seen however many networks be spun up.
And existing oracles, they all go to these new networks straight away.
But we see the same applications being built time and time and time again.
And the way that we think about it is like, if you go to a restaurant and they give you
a menu, you know, they, they give you 10 items to choose from.
You can only choose from the food items that are on the menu.
But if you go to something that's more like a buffet where all the ingredients are all
there and you get to choose how the chef's going to cook it for you, then you have so many more possibilities of what plates you could eat at that restaurant.
So that's kind of how we think about Seder.
The data exists.
The data's there.
It's very accessible no matter who it's from, premium suppliers or public data in the case of like weather, for example. But what's more interesting is like, how can we do different things with that data
to allow different use cases to come on chain?
Things that have been blocked by sort of the static nature
of existing oracles.
Yeah, I think that's a really great analogy
of comparing like a restaurant menu card
like a restaurant menu card with like going shopping for groceries um i kind of want to
with like going shopping for groceries.
pivot to like the agentic side because i've particularly uh been very excited about uh
some of the stuff you've been you and your team have been building out on iron claw so um wanted to have broad discussion on that uh one the first being is like why uh
is seda so interested in iron claw like what niche use cases do you find it like how's your
experience been and then uh what are the stuff uh that your team's actually building on top of it? Like, is it skills?
Like, how are you kind of adding value to the tech stack?
Yeah.
So, like, the interest in Nia, I think probably at the heart of it,
goes back to the relationship our founders have with Ilya, for example.
Actually, one of my first tasks when I joined Seda was to host Ilya on a podcast
a couple of years ago because him and Jasper, our CTO and Peter, our CEO, are all very close
and have worked together in the NIA ecosystem for a long time.
So obviously we've been watching what NIA is doing and of course we've seen what now
NIA AI is doing and listening to sort of Ilya's vision.
And I think for me personally, less as like a company vision, but sort of for me
personally, why I started building on Ironclaw first, because it wasn't kind of like a planned
thing.
It was just the Jasper, our CTO and myself were very interested in agents.
And it's obviously a use case that we're building for at Seda.
So we've been trying all the different tools.
But the thing that I liked about Ironclaw, what I like about the NIA community, what I like about Ilya's vision is that they're not rushing into the race, if this makes sense.
When you look at like off-chain agents or other people building on-chain, it's just they're fueling a lot of hype and they're not following best practices.
I don't think that we need to build the agentic ecosystem in a day.
I don't think that we need to build the agentic ecosystem in a day.
I think we need to build it right so that it works,
so that privacy is taken into account,
so that security is really taken into account.
And I personally don't really trust a lot of the other builders out there
that are rushing into the space because, you know,
there's a wave of hype around agents and building agentic economies
and what that might look like on chain.
But when I look at what Ironclaw is doing, and you look at even the conversations that happen in the various telegram groups, you know, is that there's a genuine interest from the
people here to have something that is secure, that's reliable, and that's trustworthy, and that
could scale to then, you know, be one of the greatest systems, the more prominent systems in the agentic economy in the future.
And that for me is very, very interesting.
And because obviously we're sort of this neutral data layer,
this programmable Oracle layer,
we want to work with obviously agent systems
that are similar minded to us.
And the Ironclaw community, the way that iron claw is built
you know the privacy first model versus like some of the off-chain agents that people are building
with today um it just goes hand in hand it just makes a lot of sense for us to you know build our
tools there first work with developers there first especially because we're still in a trial phase
and we'll be for quite some time um and at least from my experience, it's been sort of the
the best return on conversation investment with developers there than developers in other
ecosystems. So I find that very exciting and happy to move at the pace that like Ironclaw is moving
at. And I love the way that they get the feedback from the community and then they're implementing
it. Like, you know, you don't get a CEO like Ilya in many group chats,
you know, in other ecosystems, but he's in there quite often, I see,
just to show you sort of the passion of the team behind it.
And then to your second point, what are we building at the moment,
what we've built? Actually, I built this.
So it was actually one of the first big tools I ever built.
It was the SATA MCP and tool List for agents to build any data feed on
SATA in a permissionless way.
So it has, at the moment, I think it has 14 tools in it because SATA is
quite like a process if you've never interacted with it before as an agent.
So it goes all the way from setting up the first wallet, funding it with
the testnet tokens.
These are all like, I made the tools very bite-sized for agents just so we could find where pinches in the process were.
And then it goes all the way through building the Oracle program, sourcing the data, whether it's sourcing private data or sourcing public data, spinning up a data proxy if you need to for data that's not available on the network but you want to make it available.
proxy if you need to for data that's not available on the network, but you want to make it available.
And then choosing your delivery method, whether you want to do a fully decentralized permissionless
method on SATA core or for people that are building more hyper-performant apps, there's
another option through SATA fast and it teaches the agent how to get an API key or to use,
well, this tool is not finished yet, but like MMP, near payments as well. I forget what the native one's called as well.
And then, yeah, how to deploy the Oracle program,
how to test the Oracle program.
And then of course, finally, how to post data requests
and set up scripts to make those data requests reoccurring.
As to the use cases of what this might mean
when an agent can come to a new network
and build its own data feeds for any source,
it's a great question. And actually, I'd love to pick your brain about it of what this might mean when an agent can come to a new network and build its own data feeds for any source?
It's a great question.
And actually, I'd love to pick your brain about it because we've spoken to lots of developers.
Obviously, like the first use cases that we've had with people building,
everyone wants like sentiment scores and like ticker trackers and, you know,
scanning all various sites and locations for how people are feeling about different tokens or new tokens launching.
It's very trading heavy uh i'm very excited for when we start to see new use cases um come that
maybe might be less trading focused um but for now i mean we're still speculating for what they might
be yeah there's a few things i'd probably like to comment on. Like, I kind of appreciate the fact that you acknowledge that like going slow isn't entirely a bad thing.
We've had like somewhat a fair bit of criticism around like, hey, this thing's not ready yet, which is completely fine, right? Like, I think once you kind of have the
foundations of it, like, completely fixed, and they're kind of good to go, it's very easy to
scale something like Ironclaw up. So we're kind of at that phase where we're getting ready with
getting Ironclaw off the ground. And once that's done, like, I mean, it can kind of unravel as quickly as OpenClaw
just because it takes security into consideration
to like a great scale.
That aside, when it comes to use cases
that I can envision or I'd like if I had like a wish list.
I've been trying to kind of track funding rates from different perp exchanges.
And that data is not very easy to get.
If there was an Oracle, could be Seda, where you can get all of this data that I could feed to my agent and then run carry trades and stuff like that.
I think that would be really useful.
This is for me personally.
I don't know how much of an appetite there would be
for something like this out there,
but I think what people...
I've come to realize that people want to use their agents uh for them to make money for
them so anything we can think of that facilitates that has some kind of product market fit
yeah 100 and i think it's also a good place to start, especially when we're speaking a lot about on-chain
because it's where the on-chain world is right now.
I'm going to try and find you those endpoints
and build you an Oracle program for that.
Because essentially, if you can find the endpoints
for that use case,
which I'm assuming you can query a bunch of different perpdexes
for their funding rates for different markets,
then you can compile all that into an Oracle program
because you can put as many data sources as you need to
up to a certain limit as long as the return size isn't over.
But I'll have a look.
I'll try and find you some endpoints,
and I'll put together a mock Oracle program for you
as proof of what we could do.
That would be nice just to kind of play around with.
But I'm also looking at how can we get this in the hands of more people?
Like how easy it is for someone to use the said Oracle through their agents?
Like are you planning to build out skills which people can just like drop down an MD file?
Or like, how would that like,
how would hooking up to like the Seda Oracle work like?
So right now we do have skills, but for basic price feeds.
So if you want just like, for example,
I've been building etUSD price feeds as sort of the test.
And then it's expanded to,
so I've done the Weather Oracle program, for example.
Public data feeds are very, very easy to do right now.
There's a skill for using the starter kit,
the Sata starter kit.
And essentially what it does is it just subs out the API
that's in our starter kit, which is a blockchain, which is Binance's API I think we use as a test for people.
And then you can basically put any data source in there.
And then there's in that skill how to just change the logic function
of the Oracle program.
So you basically need to remove it just for very simple price feeds,
which are just functioning as an Oracle, basically.
Get the data from here and bring it to Seda.
The vision, though, for me from here and bring it to SEDA.
The vision though, for me, what I really want to start doing is building specific skills for specific use cases, because the sort of this design space that I talk about is just so vast
that you run into this problem with agents where it's not specific enough for them. You say, build
me a feed on funding rates, for example, all the different funding rates.
It can kind of be too specific
and the agent sort of runs around in circles
with where we're at right now.
So I want to see like a library of skills.
So let's say you want to build financial agents
that are making you money.
We might have a library of different skills
that are related to that that people could search for.
Maybe it's less financial use cases and we have libraries and skills for that.
Where we're at right now is the actual MCP and all the tools is with our engineering
team, just getting a final look over all the code or sorry, all the tools and everything.
And then we're going to properly launch that MCP publicly.
And then from that, I think we want to just keep working with some developers, just doing
some more of the simple feed stuff, just to make sure that there's no kinks in the process.
But from there, absolutely, the way I would love to do it is to eventually have a
marketplace, whether it's we talk to the Nier team about putting some of Seda's stuff into
the Nier marketplace, or even just on Seda, people who come in our docs and just find
a full library of different skills for different things that they want to do. I think that would be
probably the best thing to do. And then obviously, we want to run some hackathons and see what people
can build with these. But I think for it to work on scale, you make a good point. We need to really
have the skills for the different use cases because if you're doing 24-7 pricing logic,
which is one of the big things that we do,
or if you're doing lending on RWA assets,
which is another big thing that we do,
there's so much context in there.
And when it's those kind of use cases,
you really don't want the agent to mess around with the logic too much.
You really want it to have it quite specific
so that the logic is not tampered with or it's not put together wrong,
especially if you're building kind of these financial use cases on top,
which might be something that agents are doing in two to three years or less.
Yeah, you want to be very careful with tweaking stuff when it involves money.
It's fun if you just want to do like crone jobs or something basic.
But when you have money on the line,
you kind of get more paranoid.
And especially when you're giving it to an agent, right?
Like you want to make sure it doesn't make mistakes,
even after you tell it not to make mistakes.
But yeah, it's an interesting design space that I've only started to play around since the last two or so weeks.
But I'm still kind of scratching the surface because I'm not a dev myself.
But I think a lot of the tools are becoming friendly for people who don't have any dev background.
So yeah, I think this is how I'm spending a good part of my weekends,
just trying to kind of make stuff work
and then worry that it doesn't break or, like, it doesn't, like, mess up.
I would love to get the MCP into your hands for testing, actually,
because theoretically, right, like, I just ran a search
for all the endpoints that you're looking for now
and they're coming through slowly one by one
for the different perp deckses.
So then if you were to give that to your agent
and say with the SATA MCP as it is now,
you could plug that into the Oracle program
very, very easily because the data source part
of the Oracle program is actually very simple.
It's just listing where the data sources are.
And if you just want the funding rate, there's not really any logic that you need there.
You just need the funding rates in a list, basically.
So for you, for example, in your use case, as these endpoints come through and you say to your agent,
okay, now add them into a single SEDA Oracle program,
I mean, this could effectively be built in about 30 minutes, I would say, with an agent, maybe less.
I mean, this could effectively be built in about 30 minutes, I would say, with an agent, maybe less.
Yeah, the challenge is always to kind of get it fully autonomous,
where your agent is depositing USDC on your behalf on Hyperliquid and doing this carry trade for you.
this carry trade for you so you don't have to do it because the the thing i've kind of i mean
So you don't have to do it.
i've done carry trades for a number of years now managing your position is something that's like
pretty time consuming if if you could like the my end goal would be to kind of have this setup where
you could give your iron claw agent that like taking these price feeds
and like these these data points from uh say that and like executing uh delta neutral trades based
on the information that it's getting through through through the oracles um i i think that
setup is still like far away from being like semi-autonomous or even fully
autonomous yeah but we've got intense to kind of make the the the moving of funds a bit more uh
seamless yeah yeah yeah the autonomous side is a different thing i mean for the for the
sata side of things we can definitely get get the bot that's making those autonomous actions on chain, the signal,
which is kind of actually a pretty good way to, to, for people to position where SATA fits in
their head, because you would build the Oracle program, have a look at those funding rates.
And then the logic part that you would want to add on later, if you wanted to make it more complex
would be, okay, what does these funding rates mean for my agent? What decision should it make based on this data?
So it needs a way to like order the data in a way that it can,
you know, in however you've programmed it,
react in the way that you want it to do to do the right trade.
You know what I mean?
Because if it's just looking at raw data, it can make mistakes.
But if it can read the data and understand the data,
that's a different story.
Do you know what I mean?
Like as soon as the funding rates are changing
from where they were a week ago, for example,
or 24 hours ago, whatever it might be,
that's where the logic on SataSight comes in really well.
But then, yeah, of course, there's a second step,
which you're talking about is like, okay,
but I want the agent to run that thinking for me
autonomously on the other side
once Sata's brought me the signals.
Yeah, it's, as I said, it's an interesting design space for stuff like this
um and yeah i think like for for now for starters it's it's good enough that you just kind of get
this data parsed and like it's in real time it's accurate and and everything for you to make
informed decisions and like trading strategies um but like what
everyone's kind of envisioning is like how can my agent make money whilst i am whilst i'm asleep
uh we're still i don't think we're there yet but it's it's right around the corner
the golden question hey Golden question, eh? Yeah. Yeah, I mean, that's like, when I open my timeline on X,
all the stuff I see at the top of my timeline is just agents making them money.
I know about 90% of them, it's just fluff.
But it's what everyone wants.
It's what everyone's looking for so i mean i'm not sure if we'll get to a scenario where everyone's agents are making money for them
because like someone's gonna lose like right in a perps trade uh you you need someone on the other side but um it's nonetheless like what everyone's
looking forward to uh as as one of the last questions i probably want to ask you guys on
future things you have in mind to to develop uh both like on the oracle side and both on top of Ironclaw?
Yeah, so I would say like on the Oracle side, let's start there.
There's kind of three buckets that we're really building towards.
One, we've entered to very successfully in Q1,
which was the 24-7 pricing logic for perps.
Primarily, that's been in Hyperliquid,
powered over 11 billion in this quarter.
And that's going to be our first quarter, first entry into the market.
And we got somewhere around 29 markets live, all pricing 24-5 or 24-7 because we're doing some 24-5 stuff with Injective until they go 24-7.
But we will continue on the perp front as well.
We have a lot of people in our pipeline that are adding that weekend pricing logic for equities, commodities, pre-IPO, you know, sort of traditional assets as
they're coming on chain. And also some custom assets. That's the other cool thing is you can
basically define what an asset is and how that market should operate on SATA. So we'll continue
on that. And one of the things that we're sort of moving into where those two worlds are actually
merging is around 24-7 markets for RWA lending, which I think is my favorite use case, actually, because there's a lot of context, as I said earlier, in an RWA asset, liquidity environments, mint redeem policy, AML, KYC, things that need to be factored in.
And that can be all factored into a SATA Oracle program.
So very excited to wade into those waters.
That'll be Q2, Q3. And then, of course, the third bucket is the one that we're talking about today, which is this kind of concept of being the data. What's our CTO say? He
calls it the plumbing, the data plumbing for the agentic ecosystem. So we're still building
towards that. And we're doing some upgrades on our permissionless
network, say to core, or permissionless delivery system, say to core, which will be very favorable
for agents. And I can't spill too much on that for about two or three more weeks, I
think, but then we'll make all those updates public. And in the meantime, I think the best
thing for agent builders is that we'll be building the skills out that I was talking to.
So once the MCP is fully finished and we're happy with all the tools, then, yeah, I really want to just be talking to developers,
finding out what use cases they're building and then build the skills for the agents that will help them do that and then vet them.
So if anyone's building with Ironclaw at the moment and they think this could be exciting for them, please reach out.
Please drop me a DM or find me in one of the Telegram chats with Nia.
I would love to understand what you're building and figure out how we can build
the skills for you and then come and test it out on SEDA.
My vision would be sort of towards Q4 that we're seeing a lot more agents
building on SEDA.
Our CTO thinks by 2027 that most of the transactions on SATA could be agentic.
So that's kind of a hint towards what we're building.
But yeah, I'd love to work with anyone to get some different skills.
It can be trading, non-trading, financial, non-financial, whatever it might be.
But I would say Q2, Q3 is a big focus on building out this on the library,
just to make it super simple.
An agent comes in, they can scan for the skill that they need themselves, boom, use it, you know, on the marketplace, upload it to their own knowledge bank, whatever, and off they go.
Yeah, that's awesome. I think Ilya shares a very similar timeline when it comes to agentic payments, agentic
transactions.
We will, maybe, I'm not sure, I can't foresee that much into the future, but I do think
a good portion of transactions on chain later this year will be through agents.
Just by looking at the amount of excitement around not only on Nier, but on other ecosystems,
everyone's doubling down on agents.
So it kind of seems inevitable.
I kind of caught attention of you guys based on the stuff you were doing on the Ironclaw
Telegram chat, sharing the stuff you've built out.
And yeah, as you said, I think your team is very approachable to kind of make things work
and test new product launches. So yeah, as you said,
if anyone has interesting things
that they want to build with you guys,
they should definitely hit you guys up.
Yeah, absolutely.
We're halfway through your build,
just to give people the idea of the speed here.
I'm running this on Claw
just because I had it open on my laptop here.
But it's halfway through building a multi-exchange funding rate Oracle
that reports the rates for a given asset across five DEXs
in a single data request.
So there you go.
We're building the Oracle program that you requested in real time.
Yeah, it's insane how things, how fastly things can move but yeah i'll definitely like
i'm having like some issues with like my my agent not responding at times so i need to fix that once
i can get through that hurdle i'm definitely going to be playing around with stuff like this. But yeah, thanks.
Thanks a lot for joining in today, Matt.
You're happy to stay around.
What we usually do is we leave about 20, 25 minutes for the rest of the spaces just to
kind of give opportunities for folks in the ecosystem to ask questions and get involved in the discussions.
I see Jean has his hand up and he's kind of interacted with your team a few times. So,
Jean, if you have something to share. Yeah, what's up? What's up, Matt? I have a question for you
because I've seen a couple of those.
I was like, at some point before, I was trying to do with Oracle,
and every time I tried to use one, whatever, it was like always $500.
I was like, there's no way I'm paying that.
And then recently, like not long ago, like maybe this year,
there was Outlayer that got released.
And now I can do stuff that are off-chain and insecure.
So always your solution actually meant for users?
Is it meant for B2B to have billions of dollars that can use your things?
Or who is it meant for?
Because I know you're here for B2B pipeline or whatever, but who is it for actually?
That's my question.
Yeah, it's a good question.
And the reality is it's going back to the start of the conversation around being this
kind of neutral data layer. the developer defines most of it.
So we have focused more on B2B now, just purely as a startup that's growing.
And that's where a lot of the interest has been around 24-7 perp markets.
Obviously, a very dominated area of the industry from enterprise, from B2B players.
from enterprise, from B2B players.
But one of the things that we're working on over the next quarter
is automating our sign-up flow for API keys for SATA Fast.
And the second thing is, as I mentioned just before,
upgrading the permissionless system on SATA Core
and then building these skills out.
And the reason why that's important is because
solo developers coming to the network now face a little bit of a challenge
navigating the documentation and building the Oracle programs
if they've never done anything like with Oracles before
or data feeds.
With the new upgrades, the new documentation
and sort of the automated sign-up flows,
the Oracle program libraries that we've been building,
you know, pre-built programs, these kind of things,
it's going to get much more friendlier
for a solo developer, for example,
who wants to come on.
Because one of the conversations
that we've had with Jasper, our CTO over the years,
especially having lots of podcasts on Intense last year,
is this kind of idea that people should be able
to come on chain and build their application
more or less over a weekend, right?
But one of the biggest bottlenecks for that has always been Oracle infrastructure.
And that's one of the things that we wanted to change.
It was just that B2B enterprise was one of the things that picked up very quickly at
Sata.
So we focused a lot of our resources on there naturally.
However, yeah, Q2, those upgrades will be coming in.
A lot of the programs, some programs will be pre-built.
These skills for agents, for example, will make it much easier
for people building with agents or people building without agents
to access data on-chain that they can't get from other oracles
or they can't afford from other oracles as well.
Because that's the other thing is a lot of the static oracles
obviously have a massive sort of paywall that you have to get over.
They're designed for enterprise.
But on SATA Core, the permissionless system,
you can spin up a data feed as you need. have to get over, they're designed for enterprise. But on SATA Core, the permissionless system,
you can spin up a data feed as you need.
You'll only need to get some SATA tokens
to deploy the Oracle program.
But yeah, you can have a look on the Explorer
to see what a data request costs, 0, 0, 0, something cents.
So looking forward to having more solid developers come.
We want to do some more work in hackathons as well
and open up some subscription tiers if people want the hyper-performant model
and less the permissionless model in whatever use case.
Open some tiers up for solo developers as well
to get access to those credits.
Do you have a plan to use, like having some...
Because I know you were doing X402 or whatever earlier,
but I'm not sure if you will be enabling the MPP one, because I would really like to be able to use my things
I built with MPP to be able to use it like elsewhere and actually use it for actual stuff
like you built. I feel like that's something that would be great for. Yeah. For even reaching, not even for me, but you know.
Yeah, 100%.
No, it's definitely needed.
Hopefully it can get done next week.
It's something on my to-do list.
We did the X402 skill.
It doesn't take long at all.
It's just a matter of getting around to it.
But yeah, I'll definitely post in the group when that gets done
because I think that, like you say,
it'll make it a lot easier for people who don't want to be locked
into a single chain, which I think is what you said.
Also, another question.
You said like CETA or whatever, but I'm really stupid,
but would it be just possible to have like USDC or something
that is not like bound to whatever?
Or like, just so like, because like, I know like with intent,
I could probably figure it out, but there's, because CETA is not on there or whatever, just having like uh because like i know like with intent i could probably figure it out but there's because seda is not on near whatever just having something that is a like cross-chain like
compatible since that that's what i am trying to do having like a cross-chain ecosystem that
nobody's like talking about that's possible now so but i know you're talking about setup but i'm
pretty sure you could yeah i mean you can't like when it comes to doing the data request for example
on the on the permissionless delivery method on SATA core,
you can do it in whatever token you want because the actual data request
is picked up by a solver and the solver receives the currency that you want
to settle the data request cost in and they have SATA on the other side.
So there's like this kind of like pick up on one side,
send out SATA on the other, which makes it much, much simpler.
But to deploy the actual Oracle program, a very small amount of tokens will be needed.
But I mean, I'm sure we could.
But like, I would, you know, like, you know, I tried to aim for like Apple style.
I do not want to go like the other side needed to do something on your chain, like, you know, like
in 10 styles, like like i want to deploy
anything and i'm on here i don't want to i don't want to do the manually and having to do the
things you know like i could want to be object on a smart contract usdco here's how you you can
onboard you know you have something oh i'm sending you five dollars you create me and you convert me
to set up whatever and then my my wallet is or my thing is created i mean i know i know
that's like that's too outreach and i'm not b2b but i feel like like going in dashboard and doing
manually and having transfer between meta wallet or whatever it's just annoying and like we we let
ourselves go into that and because it's b2b or whatever but i think uh we can do better than
that but i know i'm just one person, so.
It's great feedback for the solo dev stuff.
It's something that we need to think about
when it comes to solo developers.
Obviously, it's not something that ever comes up
as a problem on the enterprise side.
No, but it's just like we can do better.
Why do we always do fucking dashboard and doing like,
it's just, we can, the tech is there
and now we have like, we can write software that like
does it for us like why we come like let ourselves oh you know you just have to do that you know like
we like we can have the x that is excellent now and we don't have reason for not doing it yeah it's
uh it's the um achilles heel for the crypto industry right like the ui ux experience in
crypto compared to what you get off in an off-chain experience is like fully chalk and cheese no no but it's like
you know like you just like you ask your agent or when you write code you just I want the Apple
experience what is the flow of what could be an Apple experience and sometimes you get the like
oh this is like seamless you know you, you want something seamless. And we accept, all of the crypto people accept bullshit.
And like, we should aim better.
But I know it's worthless and like nobody cares.
But in the end, like being able to say, oh, I want to deploy something,
like I want my agent and then they just do it.
I don't need to go on three different apps just to change my USDC to another chain.
You know? I want my agent to do it for me. It's the Revolut experience actually. I think
like the Apple experience is even at like Revolut experience to me is like, I don't know anyone who
uses Revolut, but in terms of UI UX, that's my gold standard. And I think like companies like
that coming on chain will hopefully raise the standard for crypto builders across the…
No, but don't wait for someone.
You know, you just have like…
When you build software, you should aim like that.
And like there's a lot of people here, like they're just like…
You know, like everyone's doing like they're this thing on their own chain.
You're like, oh, I'm on the…
I'm on the Solana.
I'm just building for Solana.
Come on, you can do better.
Like we're all like so locked up to our own chain.
It's annoying.
But I mean, it's just me.
Walls are coming down.
I mean, Intense have definitely improved.
You're in the near ecosystem.
So you're like at the frontier of that.
I think like swapping has changed drastically over the last 18 months.
You know, a lot of the tech around like managing tokens is much better than it was.
It's not perfect. So there's a long way to go. But I also think that like less networks is a good
thing. Now, the app chain thesis is not proving to be true. And we can focus building tech for the
smaller pool that we're in right now and then making sure it scales. Like going back to what
we talked about earlier, it's the same with iron core. It's like, okay, the tech does exist,
like you're saying, sure.
But there are so many like nuances
when you're building these big systems,
when you're handling billions of dollars
or more or, you know,
or different sort of use cases,
whatever it might be,
that I really like the near approach
and it's similar to the Sater approach
is like building slowly,
but doing things kind of right.
And we're coming to this point now
where it's coming into the hand of the consumer,
which is why you raise a great point.
It's like, OK, cool.
The sort of the back end stuff works really, really well.
How can we attach that to the user experience
on the front end, which is kind of what you're speaking about?
And I think it's a very valid question.
Oh, a valid demand, we'll say.
Yeah, I mean, that's why I'm also on here because all the other chains have gas and
really annoying to do and on here it's sold already and chain is cheap and stuff.
So that's why I'm here and not on some other unknown chain that dies.
Yeah, 100%.
I am just going to drop a data request into the comment section of this chat.
Give me one sec because I'm on two devices and the data request is on our testnet explorer.
I hope it worked.
I didn't really look at the actual result, but it was the Oracle program that you asked
for, Mr. Potato.
Cool, cool.
Yeah, I think that would be something I'll definitely try and integrate over the weekend
because I've been doing this manually for months.
But yeah, I'm kind of like coming to a point where I see light at the end of the tunnel and I can get my agent hopefully to kind of at least alert, monitor and like tell me what to do if it can't be semi-autonomous.
But I think we'll get there at some point.
We've got about 10 minutes left.
Yuen is also with me here today.
And there's exciting stuff happening, I believe, this weekend, Yuen.
You're hosting an Ironclaw meetup in Malaysia.
Yes, sir. this weekend, Yuan, you're hosting an Ironclaw meetup in Malaysia?
Yes sir, I'm trying to onboard more people to try out Ironclaw, see the user flow, see how when a new user that aren't really devs come into, as well as devs that come into Ironclaw,
what are the roadblocks or what are the learnings that we could learn from them to further improve it to make it even more easier to onboard maybe the next million agents.
This is like an old saying before with shade agents, but why don't we try it again with Ironclaw again.
So hopefully I will bring this learning towards the other Legion events as well.
towards the other Legion events as well.
So I'll share them all my learnings,
I'll share them all the brochures that I made
so that we could further onboard more users
throughout the globe into the Ironclad movement.
So see you guys this Sunday if you are in Malaysia.
Awesome, yeah, I think we're gonna have a huge pickup
in the number of Ironclad meetups
now that things are slowly coming together.
I also noticed Monza's on this base, and both his and my PFP have gone dark.
Monza, do you know why?
No, dude, I don't know what's up with that.
I think it's just bugging out or something.
I don't know.
Yeah, probably Elon messing up.
But if you also notice, Alex Shachanko's PFP is dark.
People are changing their bios.
If you follow Serge at all, you know he tends to know things because he works on things.
So I'd check out his uh his profile and uh yeah it might
have uh might have soft launched a little uh a little privacy push on the intensive side of
things today so i would uh check out near.com and then maybe you'll see why things are uh
things are dark maybe they aren't going dark maybe they've already gone dark you just have to uh
Maybe they aren't going dark.
Maybe they've already gone dark.
You just have to press a little button.
Interesting.
Yeah, I did see Matt's comment, and I've pinned it to the spaces.
So if anyone wants to kind of explore that Oracle, it's up there.
And it's fascinating.
Matt could get this built
whilst he was on the spaces.
It just goes to show how fast
things can ship.
But yeah, you and I,
do you have other stuff
you want to bring up?
We've got a few more minutes to go.
Oh, I was going to say
another use case that I saw
somebody using OpenCloud was they tracked, they pulled data and curated like what shops are selling on-market trading cards so that they could snipe it very fast. whenever it's listed, somebody else will buy it. So I think this could be something people could use,
people could do with Ironclaw to just set that,
hey, I want to buy these trading cards at this certain price point
whenever this is listed in any store throughout the world,
buy it for me and then ship it to my house.
What do you think about people using Iron to uh as their bot to like farm these uh you know trading card game releases
and things like that i uh i've got a buddy that is in a uh like a discord where that's all they
do is they like set up bots for releases of different like card games and such and uh
he was trying to show him i'm using Ironclaw to do it so
see if we can get that going. Think about it like before that it was like
developers coding with their own board right if we could just use iron cloud to
buy it for us how easy would it be to for us to finally get market prices on
that instead of like getting scalped by scalpers i need more charge
the thing is like one thing i'm kind of noticing is like we might have an edge like for the next year
or two but if if we come if we get to the scenario where everyone has their own agents doing work for them, I think it kind of cancels out.
So we can scalp the scalpers for now, but I think eventually if the agentic future is going to play out how we all envision, it'll be like a fair game because everyone will have their own agents kind of
trading and doing stuff on their behalf oh don't worry i'm gonna get the master key for iron claw
so my prompts are the first compared to others may the best agent win right yeah we did have
something along those lines on um uh during near conCon where we had an agent and a hackathon only open for agents.
I wasn't able to follow how did that go, but it was a very interesting concept.
And I didn't have time to kind of catch what was going on with that.
Matt, you were saying something?
Yeah, just back to everyone having an agent.
I talked about this today, and it's kind of the concept of
when there was lots of doping going on in cycling races
internationally.
If everyone's doping, then the race is still fair,
as long as they're all doping on the same sort of thing,
taking steroids or whatever it is that they were doing.
And it's the same thing with agents, right?
It's like you have an edge while you have an agent,
but until everyone has an agent
or until everyone's taking the performance drugs for cycling,
then it becomes like a fair playing field again, you know?
And then it's like the next thing.
And I always think about what's going to be the next edge
once everyone has agents.
Yeah.
You're a big Lance Armstrong guy or something.
That might be the first time I ever heard anyone defend the guy yeah yeah it's not too soon anymore i guess
yeah they're always i think like we are a few years away from that happening but
eventually it will kind of be a level playing field. And that is something that Ilya has been kind of pushing for.
Everyone should have a fair chance
of getting themselves on board AI
and making it work for them.
So it's going to be a very interesting future,
whatever happens.
But yeah, really cool ideas and thoughts to be honest but yeah if anyone has
other topics to discuss we are pretty much on the hour but if anyone from the community wants to
step in and have a few words now's the time if not i did promise to kind of go oh john um you have your
hand up on one of my devices i'm not sure oh yeah i was all right so i wanted to share what i was
working on or i will be working on so um earlier you said um he was talking about the like feed of
data whatever like you said um so uh my I made a bot that my bot is great,
because I find Noister is a really great place
for getting out data about getting feeds
from different sources.
So you can get data from the news, what's happening now.
And I could probably try to get news from Trumps,
but I haven't done that yet.
But I was just like, Northstar is a really great aggregator of open source data, like open data.
And I'm trying something, because Northstar was something that I liked before.
I was like, this is a cool idea, but I didn't understand what it was for.
And now I have access to an AI, I've parsed all the docs, I understand what it's for.
It's the inverse of X, fucking closed source.
Your AI cannot go on it, you need to pay for it.
North source is the inverse of that.
It's all open source and over WebSocket
and you can have your agents trip all the data,
you can have it come in.
It's like the totally inverse of what you have here.
And I want to, I even seen that apparently
the horns of AI are not that great.
So I'm gonna try it.
And I would like, I made a relay
and I want my agent to talk over a relay,
talk to each other.
And I am making something so that your,
when you create your account,
it's binded likeed with ZK proof.
So it's on the near chain, but you cannot see on chain your interaction
between your address and the contract.
So you cannot bind like, oh, it's that public address and this is the other one.
And I'm finding some interesting things and maybe how the
private mode on the NearThings works. So in theory I could maybe implement the
private swap like how they do it because they don't fucking share it and it's not
open source but I figured out how it could work. So yeah I tried to reverse
engineer all of the ecosystem that is closed source on Nier, and I really like enjoying it because I don't like what's closed source and that's not open.
I really hate it.
Fair game.
I really like Nostra, or however you pronounce that.
I kind of see is like, it's very biased content, which is closer like to the doomer side of society.
I'm not sure. I'm not saying it's wrong, but like it does give biases and you want to be careful to kind of give biases to your AI, which is very easy to manipulate.
Yeah, just trying.
Basically, the only thing I try to find is finance data.
So related to macro data and saying where finance is going.
Because in theory, you don't care about... I mean, somebody died in Africa, whatever.
But the only thing that what moves the market,
like always down or is the yield on your asset yield on your your asset or like on the usd
is increasing that's like some like leading indicator and like if you have like war going
it's moved the market but like if you have news like uh like trump's or whatever so interior like
the market is just moved by ai now like even today like it just there's no like human like
there is some discriminatory but like right now it's mostly AI that drives it.
So like, even if there was bias, your AI needs the bias.
Yeah, John, you're so numb.
Like you can't, you don't care if someone dies, man.
No, but I know, I know it's wrong, but the AI is things like that.
Yeah, and the world thinks like that as well, sadly.
But yeah, it is what it is.
Yeah, it's interesting.
The fact that you can plug it very easily as opposed to getting data from X is good.
I just think it's not the perfect platform to get data from because of the biases.
But if you can kind of just contain it to get like objective data around numbers,
and if those numbers are accurate, then... Basically, there's a solution that use a
Python that you can like look at the sentiment and seeing if the thing is like positive
it's bearish or a bullet i wouldn't be using sentiment data on on nostril like it it gives
you it would probably give you a very biased take no but yeah for for now like everything like for
the the short time that i've been using it the only thing is everything is bearish there's nothing
bullish on there but of course that i like there's other platforms like Finviz that has better positive data that are more central.
But I'm not getting paid for this.
So I just do what can work for now and see if there's any value in it.
Yeah, you'll always find bearish takes on nostril like it's it's kind of like the
anti-establishment i'm not saying like they're left-wing or anything but like they're anti-establishment
and you'll always have like doomer content um that's why i kind of stopped looking at it because
you you kind of get carried away after a while and you don't want to because reality isn't totally pricing in what they're saying.
Yeah, but actually, here's the thing. I am not using Nordster like the main field.
They have bots that aggregate the different news sites where you get the headlines for them.
But that's probably as good as an RSS field.
Yeah, yeah.
There's like,
I have also like Financial Juice
that also have that.
But like,
like I wanted like some more than just one
because like if I just have one,
like maybe the source that I'm missing,
you know?
Yeah, no, it's interesting.
It's nice that we've kind of continued
this conversation around data,
and it was good to kind of have Matt talk to us about that.
Hopefully, we only have to use Seda,
and then Seda figures out how to get the data from multiple sources.
You found the actual answer.
Before, you said that if everyone has an agent who has the edge is getting the data that is unknown
or what actually makes a difference.
If you're about to figure out what is the leading indicator or whatever, that's where you can find the edge.
or whatever, that's where you can find the edge.
Yeah, I think the edge thing is still not that big of an issue,
but you will need an edge if you want to kind of make money out of it.
Yeah, yeah, of course.
I mean, finding, having data has been since the dawn of time
that you can have an AI, like read data, whatever.
This is not new from yesterday.
You know like since every time you have like when the US like share their public data like
this is this is like not a today's problem. It's been it's been solved. Like institutions already
have those things, those models already built and they're using it every day. It's like it's just
that we as like people that can be like retail
or whatever if you can call it like that could have a chance to be able to use a certain system
like that that is not like proprietary and that no one knows that exists. Yeah I agree. Yeah I think
we're up on the hour. It was nice kind of having this conversation.
We do this every Friday.
So definitely tune in for the next one.
And as for the speakers and guests,
thanks, Matt and Monza, Jean and Yuen for tuning in.
And yeah, we'll do this again next week uh if you're building
something around iron claw then definitely hit us up because we want to kind of dog food this a lot
more so that people can make out the most of their agents so uh if you're doing that then uh definitely
let us know uh until next time then folks and. And yeah, I won't be touching grass.
I'll be kind of trying to get my agent to do stuff.
But otherwise, I recommend touching grass as well.
And don't forget to go to near.com and press that button
and try that as well this weekend there, Mr. Potato.
Oh, yeah, definitely.
Hopefully my PFP gets fixed as well in the coming days.
Yeah, we should be good on, like, Tuesday.
I think you might be stuck for a few days, so.
Oh, shit.
Okay.
All right.
Thanks, guys, and see you on the next one.
Bye.
