Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
Private-by-Design AI Agents & Omnichain UX
Recorded: Aug. 26, 2025 Duration: 1:27:52Space Recording
Full Transcription
Thank you. Thank you. you Thank you.
I feel you.
I feel you.
Not a bad idea.
You coming back up at V-Smoke?
I'll give you this key.
Yo, yo, yo, guys.
Can you hear me?
Hello.
Awesome, awesome.
I can hear you. welcome to the team tammy
we have such amazing speakers today just uh
once again um one two one two we have fallen network from uh we have dylan kowalik from fallen network and then uh we have
zeta chain matthew karsten guys please say uh a word if you can if you hear and um just want to
be sure mgm everybody i can hear you i just i don't know on my screen i can't see that your
speakers for some reason but that's just probably my end maybe that's all good yeah we can hear you. I just, I don't know, on my screen, I can't see that you're speakers for some reason, but that's just probably my end.
Maybe.
That's all good. Yeah, we can hear you.
This is, we're on
Twitter, Dylan, that's why.
Yeah, yeah, it's a new platform.
Also, throw up your green
hearts. What are we doing with these
red hearts here?
I'm gonna do purple,
you guys do green, and you do... That guys do green and you do acid green
But I'm gonna do the green one as well
I actually don't have the green one for some reason in my
How to get the green heart?
I have only purple
You don't have green?
What?
Press and hold and more colors will come up.
Okay, okay.
Makes sense.
Makes sense.
Yeah.
I'm going to send you some green hearts, guys.
For sure.
Just all the best love and, you know, best green hearts.
Dylan's always teaching me new things.
I don't know.
And follow network.
Awesome, guys.
Yeah, so I'm going to start really quick.
A couple of minutes we took to warm up.
I invite Zeta Chain to speak. awesome guys yeah so i'm gonna start really quick a couple of minutes we uh we took to warm up uh
i invite in zeta chain to speak i'm also gonna invite uh follow network to speak just in case if
uh you guys would like to speak from official accounts you're welcome um um just again my name
is ilia paskov i'm a founder of sind um i'm super happy be at this AMA with such a cool guest.
Matthew, Dylan.
Matthew from Zeta Chain and Dylan from Fallon Network.
It took us a little bit, guys, when we've seen each other last time.
I think it was just a few weeks.
It feels like tomorrow.
We just met in SF.
We did this amazing event in temple sf um uh during uh berkeley uh
blockchain conference and uh we can like started this topic um you know about tees and uh privacy
and confidentiality uh in ai and then uh cross-chain uh universalchain universal I'm gonna mention because Matt you always correct me there is no
cross-chain there's only universal universal possibilities for the agents and we just started
this topic so pleasure to continue you know and pleasure to discuss two next steps because I think
it was very successful we We had an amazing stream.
We touched a super important topic in crypto and Web 2 and Web 3 in both of them,
and then specifically in AI.
I really like that direction because I think that's the next step of evolution,
pretty much, for the AI and Web 3.
We're currently in New Yorkork uh in mail in soho uh so uh i just came from uh
alliance dao accelerator i've been i've been there we did pretty cool hackathon on this weekend and
currently in soho new york doing meetings we met with zeta chain and met you here at malin and
zeta chain and matthew here at malin and um looks like we got uh approved by uh zeta chain hackathon
uh with synth the zeta chain hackathon and google uh that uh happening right now so we have kind of
like good updates i think it was pretty cool yeah exactly but um guys how are you doing like what's
new what's what's happening on your side?
Just a quick, quick, let's start with quick intro, quick thoughts.
Where you at?
What are you doing?
Quick introduction for people who don't know who we are.
And yeah, we're going to start our deep dive.
I'll go first.
Hi, everyone.
My name is Matthew Karsten.
I am the developer relations engineer at Zeta Chain.
I am also here in Soho, New York City, working out of our main office and have been participating
heavily in developer activations, events, and hackathons since ETH Global.
and hackathons since ETH Global.
And that's included touching Solana HQ
and Station 3 and Brookland
and all of these developer hubs in New York City
to really find when people are cooking up in the trenches
and bringing that over to Zeta Chain
to build universal applications.
We have a super large grant program.
We have a robust bounty program.
Zeta Chain is fairly new,
fairly recently funded with a Series A.
So it was a really good time to be building a file
in Zeta Chain.
I'm really happy to be here with Ilya Paschkov.
He's such a talented developer.
We've worked on several initiatives in the past.
And so I've had my eye on Sint for a while.
And I know they're doing incredible things,
especially in the trenches and hackathons
and really finding what is next and what's new
and what people care about and how
to make those things
as accessible as possible so i think that's really neat and i'm happy to be here with one of my
dearest colleagues in web3 dylan kowalik dylan i'm past the microlouge why don't you introduce
yourself to the crowd and tell them what you're all about and where you are. Cool, guys. Thanks, Matthew.
And, yeah, thank you, Ilya.
The grant program, I'm really excited for that.
That's really cool.
And something new that's happening in the whole fallacy is that we've recently released an audit for DSTAC,
which is the – I would say it's the best TEE framework
in the market right now.
And as it stands, you know, it's 100% zero trust. We've decentralized the key management system,
which is phenomenal, meaning that all of your deployments are now 100%, you know, yours,
they're confidential, no pseudo, no pseudo administrator can peer view on your data.
confidential, no pseudo, no pseudo administrator can peer view on your data.
Dun, dun, dun.
And that's been a thing.
And so anyway, we just released that audit.
That was super cool.
And yeah, so Phallus churning and burning with confidential AI.
And so we support all developers with both CPU support, GPU support with NVIDIA, and
it all runs on Intel TDX and beyond.
So I'm leading developer relations at FALA Network, and we're happy to basically start
onboarding the next, let's say, million users for TEE.
There's a huge underground wave happening right now for TEE.
It's been happening actually quite slowly for like the last year.
Just no one really knows what is a TEE.
I feel like we did this thing a few years ago with zero knowledge proofs and we called it the ZK scene and everyone was all about privacy.
But now when it comes to confidentiality, we have to push this new thing now that it's available to developers.
And it's so much easier now to onboard and TEE.
It's insane.
So everyone can basically make a private app.
And so private AI or confidential AI is on the rise.
So if you don't like your AI knowing everything about you, if you had a digital twin and you
didn't necessarily believe that your digital twin was your best friend, it's because it's not confidential.
So I'm happy to talk about that, the ins and the outs of all of the inner workings of that.
And yeah, so with Synth, it's really cool because if you have a digital twin that's like a virtual reality-like, you know, meta-like experience of an AI, but then it exposes all of your secrets, it kind of sucks.
like experience of an AI, but then it exposes all of your secrets. It kind of sucks.
So I think this is a perfect case example of how any AI, when it comes to relating with you,
that if it knows about you, your data, the messages that you send, the privacy that you're
requiring for your AI, we're entering this sort of different era of AI when it comes to AGI,
we're entering this sort of different era of AI when it comes to AGI or artificial general
intelligence but it's not so general and it's not so artificially intelligent if it's exposing all
of your data so I really want to talk about this the stats the statistics and and so so much more
so I'm happy to answer any questions and thanks everyone for you know joining in and being here
it's super early it's a it's an early morning so I know all of you are waking up early for this.
So thank you so much.
I'm so excited for this.
Awesome. Thank you, Dylan.
Thank you, Matthew.
I'm so excited.
I'm so happy, to be honest.
I'm so happy to have friends and colleagues like you guys
with which we can build cool stuff.
And collaboration with Zeta chain and uh follow network was uh nothing but
insightful brilliant and pleasant for me because you know like that's you know pretty much through
the tech that we're building together i'm executing my goal and you know my ideas that i was
thinking many many years already and right now we go into the direction of cross-chain confidential confidential AI agents so thank you guys again I'm just gonna go
straight into the you know our MA so the whole the whole topic today is Dylan
already mentioned you know agents are getting personal they have memory and
now they can send transactional transactions they can spend uh
uh crypto they can spend funds and today they run on you know
infra infrastructure data that can be locked leaked or uh even used against the user and
that's kind of like the whole biggest problem uh what we can see on the market. Omnichain, cross-chain, universal UX, it's also still, it's a big solution right now that solves all these problems.
Because right now the experience must be pretty fragmented.
So if you want to execute some simple cross-chain swap, you have to use multiple DApps.
You have to connect multiple wallets you need to wait for a long time and yeah just try it and you will see it so what Omnichain and
Universal swap from Zetasol that immediate swap that we can do with the agent uh that has his own memory and that runs on private
secure environment and as Dylan mentioned like having the um your personal assistant or your
hologram of yourself your avatar that store your private data so what unique component that will
We differentiate in different agents between each other.
be differentiating different agents between each other it's that unique component that
It's that unique component that your knowledge base, your skill set, and probably you don't
want to link that information to others because this is your identity.
So identity storing is the key component and it should be private, it should be secure,
it should be unhackable, and ability to go into multiple chains just with a voice command or a click that's where we're going
with the tag that we put it together and that's why i'm more than excited about it so um let's go
to the questions guys i mean i'm i'm kind of like uh excited about everything and i want to start with the topic um uh who want to go first matthew or uh dylan
i'll go first um i think something that dylan said it's got me over here searching chat
for the answer and it's the difference between confidentiality and privacy they are not the same thing so confidentiality is a duty so like it's
like a business-wide organization-wide duty to perform you know from vendors and people who have
access to this data and a duty to keep the secrets a secret. Privacy is the difference. It's like an individual's right.
So if you have privacy or privacy features, you have the right to control your data and to manage
what it does. It doesn't necessarily mean you make great choices and that you are keeping your
information confidential, even though you have the right to privacy right so one is a right
you have the right to privacy which is just an option and then confidentiality is a duty
and the other thing that i kind of picked up here is
dylan is really at the at the intersection of what comes next because not only does it have
this like real bust history with zk proofs and zero knowledge but now moving on so like hardware vendors and confidentiality and to
unless it's kind of over one one another's weaknesses like you know a te the vulnerability would be that you trust the hardware vendor, like Intel or AMD or whoever is providing
the hardware service, right?
And using a TEE and then kind of using maybe ZK
inside of a TEE in the future would be a way
to cover each other, one of their vulnerabilities, right?
And so I really think with a combination of like
since and zeta chain phala you know that's that's confidentiality security uh you know
potentially zk in the future i'm thinking but also um these these things are, you know, the next layer of what comes with Web3 and with modern blockchain technology and a stack.
And I think that, you know, these kind of AMAs are really what allows us to showcase, like, to the community what we're really working on, you know, what we're coming up with and the types of applications we're trying to develop.
on, what we're coming up with and the types of applications we're trying to develop.
And they really are based on forward-thinking applications that we would want to use knowing
the current vulnerabilities of the systems that we're building.
What do you think, Helen?
I'm going to pass it over to you because I know you have way more information about ZK
and TEE, but I know I'm usually under something
and you help me fill in a lot of those gaps.
Are you out there?
Can you hear me?
Hello?
Yeah, Dylan, you hear?
Can you hear us?
Yeah, maybe.
He must be in motion at the time.
Well, that's my thought.
And my other thought on SYNP is that basically being a voice AI, I think, is really super interesting in the sense that it covers accessibility.
So I think that, you know, the major topics that we're addressing by putting these things together is
Accessibility, interoperability
Confidentiality, you know, three big words that just means you can use it. It's yours and
100%
Hey, I'm so sorry. I don't know what the hell just happened. I just had to like my my my internet just dropped
I was just speaking. Could you guys hear me now?
Now we can hear you.
Now we can hear you. Yeah, it's happening.
It's MA. It's always like that.
Twitter, not usually, you know, it's perfect.
So, you know, sometimes connection or something breaks.
No worries. You rejoined. I added you.
So everything's great.
I'm going to answer your question and then coming back to Dylan then.
So I think I'm on on many ma's on many
uh speaking exam always mention an example of jarvis and that's what sin for me is so like
jarvis experience that's where you can communicate to ai with your voice and then you have dynamic
context-based elements that you can see on a screen no matter what device you use you can
use projection mapping you can use hologram you can use uh virtual reality screen you can use hologram you can use uh uh virtual reality screen you can actually see that you know
interface and then you can see that agent as a character and sometimes you don't need the
character at all but most of the time you can hear the voice and you can communicate with your voice
to complete task and that's the you know whole direction where all the industry moves uh currently
we actually building and jamming with that voice possibilities for the past eight months, utilizing all the
latest, you know, API's from WAPI, 11 labs, deep gram, and
there's huge ton of them. And we, you know, mixing them
together, trying to define which one will be used better. We
build in the our solution as well, that can solve many issues.
And there is actually
a huge ton of issues. It's not fully done yet, if you guys understand. If anybody tried
it, yeah, you can make voice to text and text to voice. But actually, if you're talking
about real nice conversation where you want to talk to AI like a real human being with pauses, and then maybe a few people talking in the same time in one room, right?
Or some, you know, different use cases like that.
There is still a big gap with voice that should be solved.
You're totally right.
And some systems, you know, integrate better with their API,
like for the MCP, like the ones that we use,
but then other tools that often, you know,
run in the background and are more resource intensive
and require, you know, multiple models
to be running at the same time
to give you sentiment analysis and, you know,
motion detection and all of those, you know,
deep grand type features
to where you're getting more data from the voice, more context.
Those things have a lot of room for improvement.
But it's also going to come with having more resources available or these things becoming less resource-intensive.
Which will happen over time.
Yeah.
It's all in all the cool integrations that we actually, we just completed this Zeta cross-chain.
I think you've seen it, Madhu.
I just demoed it to you.
So right now with voice and synth, you just say, hey, can you swap 0.01 ETH to BNB? And it's happened immediately.
Transaction is done.
Just because you have delegated wallet, you already delegated all the rights.
You just need to confirm it on the wallet for now.
The implications of that are so serious.
You know what I mean?
It's touching on a lot of different advanced technologies and combining them into a very simple user experience, which I think is powerful.
You know, and it makes you, you know, beg the question.
It's like, okay, we spent months and months and months becoming prompt engineers and learning
what questions best to ask to get the information that we want to receive.
But now, like, translate that into natural language, like how it's going to be a matter
of, like, directness, you you know now you've been able to
imagine but now kind of refine that into what do you really need and because that's what you're
going to get specifically with you when you ask your ai it will interpret what you say so that's
that's what i think um 100 yes we'll be we'll be the blocker in the industry there'll be a lot of garbage just like there is now
that's generated until
the language that we use
becomes more direct
to elicit the response
we desire
yes yes
also another cool thing like we just minted an NFT with voice
that was another cool component
no way you minted what?
yeah we completed the hackathon
what hackathon was that
i think it was uh eth global and it was openc openc hackathon openc hackathon yeah right now
like since avatar just literally with voice can get fetching information from openc
mint and nft with voice and then give you all the information about it i didn't realize you could mint i thought it was really just you could view you can fetch the information yeah um i have to double check
with the meeting part i know that we can we fetch the data but i think we minted it already uh so
just don't want to create like extra confusion but nfts are making a comeback and i say that
every summer people are sleeping on nfts they're like oh nfts are dead it's like you guys have been
saying that since 2018 nfts are not dead yeah actually and actually nfts are coming back but
not coming they're coming back from like utility perspective because right now uh we're integrating
this dynamic nft and synth where all the stats uh from your profile from your agent actually score
scoring in your card in your dynamic nft card and then you can share this card with with the
community to showcase hey this is amount of funds my agents are uh generated for me that's roi they
made that amount of agent i launched and this is my wallet wallet score this is my score of social
media and this is all the stats for uh for the
dynamic card that you're gonna have as an nft it's just not as you know hyped anymore but you know
it's very cool i want to come back to do what you do what you do so well ilia is really the it is
the design of the thing you know the models will change they'll update they'll get better. But the quick integration of new features like
MCP hackathon, Google, all these different models, and the voice, those two simple interactions that
rapidly, you know, increase your, your speed to a prototype to um a demo is is something to be you know not only witnessed
but replicated and i think there's really something to that and i think there's a lot
to be learned from the way that you build and um you know we're finding that into into a process
because i mean literally you flew from l or from san francisco to new york and did two
hackathons in a weekend like that's yeah yeah yeah it's not just me it's also my team so like
be kudos to my team because you know i'm i'm leading that development but you know they all
also put in you know hard work and i'm pretty open to bringing new devs to my team who actually
sharing the passion for innovation so anyone who's looking to collaborate on new edge stuff, especially
on XR AI NPCs, where you can educate those 3D characters and NFTs, flash NPCs,
make things for you. That's kind of like where I'm going to, you know, I want, I
want to make those minions that also link to the hardware, you know, like, you
know, and this is like the, the, the humanoids that have the digital representation.
I need the software to educate them with.
Right.
I'm going to get back to follow because I think we took longer and Dylan
had to kind of be quiet.
Dylan explain in a simple words to non-devs who's on this AMA, what is TEs and CVMs.
So just like, as you will explain, explain to your grandma and what's the
difference again and why people need it awesome yeah so um i mean not not to like take away from what you guys are talking
about it's actually quite it's really cool because like sin you know like in my understanding what
sin is is that you have a ton of attachments that are inside of the dashboard so you get to do things
like call on to your wallet, like through Zeta chain,
and then make the command through your voice.
And then you actually initiate the actual workflow,
which is built inside of the actual AI itself.
And so this is known as AI inference.
And so if you have a model that is Sint, for example,
and then you're trying to, you know,
agentically workflow with like your, I don't know, your Quinn or your DeepSeek or your GPT, for example, and then you're trying to, you know, agentically workflow with like your,
I don't know, your Quinn or your DeepSeq or your GPT, Gemini, whatever the attachment is,
you're going to have a workflow, but then the programmability behind that workflow is going to
then not have end-to-end security. And what we mean by end-to-end security is that the actual
operating system, which is effectively the minimal
hardened operating system that's actually running on the TDX itself. And so when we say TDX,
you have to imagine that it's like a secure virtual machine. So we all know about Ethereum
virtual machines. We know that it's actually exposed. It's not necessarily private or
confidential. However, you can run virtual machines inside of what you can imagine is like a black box.
But we're also trying to get away from this idea that it's a black box because usually
a lot of different cloud providers make their security solutions or their cybersecurity
solutions in a black box where it's proprietary workflows and you cannot necessarily verify end-to-end like what exactly is going on within that sort of confidential virtual machine.
And so this is known as zero trust, which is we have end-to-end security, but also the analytics
that provide you the verifiability of what's actually occurring on the operating system. And so when we say what a
TEE is, it's actually broken down into a modular sort of architecture where there's an operating
system, which is effectively the way that we reduce the attack surface. And then, you know,
we break that into a distributed key management service, which is known as the KMS. Okay, so the
key management system is very popular. Like every
single cloud provider, you got GCP, Cloudflare, Amazon, you name it. Right. And so basically these
central vaults, they act always as your sort of root of trust for any program. But on Fala,
we've decentralized the key root of trust or that vault specifically so that we can issue cryptographic keys to the virtual machine after you verify that the virtual machine is actually inside of a trusted state, which is the confidential virtual machine.
And what's nice is that we can actually, in fact, the hardware and this KMS together combined are the actual root of trust of the entire system.
And what they do is that they provide what's known as an attestation.
And so with zero-knowledge proofs or with any kind of cryptographic proof, even on transactions for Zeta Chain and Ethereum and so on,
you always get transaction hashes that prove, hey, you signed this with your public and private key,
right? Like you go into a block explorer, you get to actually see all the public transactions, but then what if you can't actually see a transaction that's happening inside of a TEE?
What do you get back? Well, what you get back is something known as a remote attestation,
okay? So the cool thing about FALA, but also what we offer, which is DSTAC,
The cool thing about FALA, but also what we offer, which is DSTACK, is that we open up a very specific channel called a gateway, which is the HTTPS connection.
This is the secure network gateway so that it allows users to really run their application, not only through the CVM, but what it is really cool for us is that it's effectively, it's a verifiable domain that's managed and it uses the remote attestations through what's known
as TLS. And so TLS is like the basis of a protocol, like for the internet, for example,
you can think of HTTPS, right? Anytime you go on the internet, you see the HTTP link,
you add the S, that means that it's been secured, guarded by actual certificate authority like
CertBot. And this gives you the actual proof
that you're working and talking
and knowledgeably working with a workflow
that is attesting to the secure virtual machine itself,
which is built with a zero trust networking model,
which is what we're using,
which is usually like WireGuard, right?
It's a VPN like service under the hood for connectivity.
And this is actually how we are able to achieve what's known as RATLS or remote attested TLS for that actual handshake between you and then the systems that you're working with.
And lastly, I'm breaking this down in terms of what is the architecture because if you want to understand what a TEE is, it's important to know what it does, right? And so the most important thing is sort of the orchestrator, right?
The octopus, right?
It's the one that's got all of its hands and all the cookie jars.
It's making sure that everything moves across all these modular systems.
This is us.
This is Sint.
Yeah.
So this is like you can imagine Sint is working as an orchestrator for all the different MCP sort of servers.
And, you know, it's working with all the different workflows.
And it's able to give you all these different attachments.
But then under the hood, each of those attachments are working inside of a TEE.
And then they're working through a virtual machine manager as well.
So the service is actually on the host side.
So this is an orchestrator that manages the entire lifecycle of the confidential virtual machine.
So it creates your virtual machines, it loads in all the measurements of the operating system,
and then it's basically responsibly, sorry, responsible for all of your encrypted inputs.
So it's a lot of minor operations, but the most important thing is that it's,
the big picture is that it's don't trust, but verify
because you want to have end to end verifiability from the hardware up down to the application
level, where then you basically can verify, Hey, like I want to make sure I verify every operation.
So when I send you a message to my MCP server to tell sin to send a transaction on Zeta chain,
I want to make sure that no one in
the world heard that. And no one but like Google, I don't want Google to peer view on my data,
and I don't want to give this to an MCP agent that's running on like GPT. I want to make sure
that when I send the data, A, that it's encrypted, B, when it processes the message on the MCP server
through a private inference model, I want to make sure that it can verify that, hey, this was actually encrypted,
and that the text that you actually sent
was also not exposed.
But we're going to give you a remote attestation
that proves, hey, you can trust this.
Well, not necessarily trust, you can verify.
You never want to trust,
because that's sort of the back days with GCP,
which is basically trust me, bro, security.
Like, if we're going to go into that, it's basically every layer of the stack that Google has on you.
They cannot confirm or deny whether or not they're actually peer viewing on the data,
so that they have their own assurances for their insurance policies in terms of conditions,
because, you know, there's a lot of things that they can't actually host on GCP,
and Cloudflare, and so on on like AI inference models that might be
destructive and so a lot of people are scared of AI so they're like oh well it
all has to be public we have to see exactly what's going on with the AI well
bro now we know exactly what's going on with the AI if it's a workflow if it's
just sending a freaking transaction on Zeta chain sorry for my language but if
it's like you know if it's like if you know what it does then why do I need to trust it just verify i don't need to trust it i just need to know that it works
and that it's not going to expose me for what i'm actually doing so i hopefully get that gives you a
better picture of the difference between what a public versus a private ai will do and the
difference is is that you don't have to trust you verify you get to be more compliant potentially
if you're going to be more compliant. Potentially,
if you're going to be working with your own medical data or GDPR or HIPAA compliant data,
like that's important for compliance reasons. And if you're going to be issuing bonds, then you're going to want to make sure that you're issuing bonds where then it can have a private
transfer between one remittency to another. But then if you're going to be issuing the transaction,
then how do you do this where you're actually no longer doing it with a co-processing unit that is actually making the data confidential? You have to be doing this with a TE. That's the only way that these major large-scale systems for finances, for like true DeFi, like RWAs, these sorts of activities that are actually new and groundbreaking beyond NFTs, right? There's NFTs and then there's like
RWAs and then there's like, you know, identity protocols that require like, you know, ERC 3643
and these new types of primitives on Ethereum. Like these are the big new topics now, but they
all require co-processing if they want to be like actually competitive with the real world market
of today. So AI, if it's going to be like agentic on web
three, then how is it going to do the big fancy financial transactions if it can't do like
literally anything that's actually useful, useful work. Okay. So that's, um, I'm bullish, right?
Um, I, I think like if you integrated a workflow that was doing a private transfer, I would never
trust your model, right? Ilya, if it was never doing it where I got a remote attestation. And that's the nice
thing about DSTAC with FALA is that you can host the workflow, and then it can literally send you
in the UI UX the actual remote attestation that proves, hey, like, this is the code that's running
your, you know, like workflow on Zeta. Hey, this is also verifiably not sending data to any other
provider. Hey, the actual message was encrypted. So you have a private chat with this thing.
And the D stack image is also verified by the TEE because we have this get commit hash because
you've open sourced it. Or maybe you have the code posted somewhere on Docker, right? Like you
provide the actual proof that says,
hey, like this might not be a smart contract on the blockchain, but this is a program that you
can literally go and read. When it hashes itself, it comes to this git commit hash, and you can
verify it with the sig store that's inside of the TLS message in this application right now running.
And so it goes beyond HTTPS. Now the billion dollar question for AI and just
really most applications in general is what is the UX UI of confidentiality? This is actually
a huge question that no one really can answer right now. We've been asking Microsoft, Google,
like Anthropic, I've been asking all of them. They don't know. Why? Because they're not actually
prioritizing privacy. And they're also not
prioritizing confidentiality. If anything, 81% of confidential, like, what are their names?
Chiefs information security officers, 81% of them don't actually even think that AI data pipelines
today are secure at all. And that's true. Like yet 55% of all Americans, I think the statistic is higher now,
uh, are actually using AI on a daily basis yet little do they know how basically, um, how bad
it is. I mean, it's just as bad as when we were using Facebook back in the day and everyone
finally realized that Zuckerberg was taking their data and looking at their kids and, uh,
and their Facebook photos and like figuring out what their like favorite ice cream was,
which is basically the issue.
Now it's actually even worse.
Now you're literally talking to an AI,
and it learns everything about you in a breakneck speed,
and then it infers things about you,
and then it tells your sister, your brother,
and then your mortal enemy all this information.
So yeah.
You just pretty much pay your own funding.
It's the same concept as why your own AI is important
and why it's important to learn how to train a model
is because eventually you want to be the trainer
and the owner, the sole owner of your own knowledge base
that you're keeping and that you're using on a daily basis
so that
your truths remain your truths as a
by google or facebook or whomever um
losing you matthew i think your connection is breaking but yeah pretty much um uh if you guys
can hear me so everything is good can you hear me doing yeah i can all good yeah i think we lost magic because he's uh but i mean that's why i like
our conversation and like our amazing events that we put in together because we like i think we're
going into really really cool direction uh solving real um use cases because i want to talk to my AI privately, but I want to still have this experience.
I want to have a full Web3 integration and I want to pull any data sets I want. I don't want to
worry about what chain it is or all the additional details. I just want to execute the command. I
want to get all the benefits from chain from the user perspective and i would say like when i'm looking on rama is like s zp is like since
that's a paula i think it's it's such a good combination where uh now agents can actually
run super private ai and then execute all this cross-chain universal agents.
Because, like, you know, I like dynamic context-based interfaces.
I want to execute AI with the voice command and, you know,
put it in the best user-friendly UI possible.
I want to be sure that it's secure, private, and, you know,
working around my company, my business, or around me around me like as a user privately but then i want to
give it a bridge you know to execute uh any command on chain uh just with with one command
with one voice uh command or just a text or one button and that's how that bridge how the next
steps i don't wanna i don't wanna uh share all the private information we have in our group chats but uh i think that
uh the partnership itself what we put in together between these uh three amazing companies and
projects and you guys you guys can hear the level of expertise and you know the uh deep thoughts
that we put in together into building these things um i'm gonna reach out to the uh audience and uh if you guys have any questions please uh ask them
in a chat send a comment in twitter i know we have many many smart people here who actually
listeners listening us from uh around the world we got this timeline so i think we we hit even
azure a little bit so i know azure is probably the latest one, but Europe is not too late still.
I'm on East Coast, dealing on West Coast.
So ask your questions, guys.
We'll be happy to answer.
You can ask anyone or three of us together, like what's the use case.
I would like to come back to utility that could be done with AI agent and why actually it's needed so many people
still don't understand agents they still don't understand blockchain maybe and they still don't
understand the privacy component right from this perspective but I want to like combine together all these three components and uh try to
model uh the best use case so it's going to be clear for non-very tech savvy users what actually
could be done uh with the tech that we are putting together um i mean from my from my example uh
and i already mentioned that like that's the demo that we actually just executed with Zeta.
You literally just say a voice command.
With voice, you can say send 25 USDC to my L2 wallet and tag as August Expenses and email me the receipt.
That could be a command that you can say
just with voice so on the on the scene side you will see the uh dynamic context base uh ui ux
um where you know you're just gonna prompt that model that scopes you and show you that
transaction is complete uh we want to get a verifier from Fala.
Like attestation is verified that, you know,
whatever you say and whatever you send is actually...
Well, not even that.
Not even that, though.
But also, Ilya, is did the model that you choose,
was it the model that ran?
Not just, oh, was the message verified?
But we can actually verify the exact model that you're trying to infer.
That level of detail.
So yeah, exactly.
Exactly.
That level of detail that, you know, we, the information wasn't sent to OpenAI LLM directly and you didn't see that information.
No, but like, so by the way, you keep it privately, right?
And then from a blockchain perspective, we we're gonna get the message that uh transaction is complete
and for example if i send 25 eth uh or i swap the 25 eth um you know to uh stablecoin uh that could
be easy done on the zeta cross-chain possibilities and then we get in the
proof receipt that automatically we'll be sending to user to the email to the log that's going to
be generated um i have a few questions like i mean i've been asked it a million times and i like that
that was the whole idea i just want to explain to the listeners why we're doing and what and you
know maybe you guys can share you know dylan and
matthew like any agentic ideas you think what what the use cases we actually bring into uh
to the uh to the communities and what they can do with our tech when we're going to launch uh
since that of all uh agent well i'll go because i don't know if he's here right now, but there is, I think there's a lot of different use cases, like higher level precisely.
I think in general now that TEEs have been lowered, the barrier to entry has actually been lowered by quite a lot.
It's the utilization of attestations now.
So that's, I think that is the true use case.
And this is actually kind of a new field.
A lot of people are trying to basically figure out right now.
It's pretty seamless.
Like as so long as you're a developer who knows how to Dockerize like containers,
like using Docker and you can compose it into a YAML file, which is actually quite easy.
And if you don't know how to do it, just vibe code it.
Like just take a Docker image, you know, figure out like, Hey, I have this Python program or whatever it is that you're trying to run as a
workflow for the agent.
And then you can Dockerize it, you can launch it to FALA.
And then, you know, basically you get all the attestations out of the box, but
like what attestations do you use and what's the use case?
Right.
I think that's the kind of the, the key question is just so, um, so, so for
example, you have a RARPC.
So, like, you have the remote attestation of the RPC, which is, I think, it's super interesting for use cases because internally, it's the same thing as a GRPC-based API.
So, this is actually provided out of the box by the DSTAC guest agent, which is the end-to-end HTTPS connection between your agent and the actual interweb.
So you're able to prove that the agent is running inside of the secure virtual machine and yada yada.
But then the actual application or the container itself asks the TEE to perform an attestation related app, you know, to the actual program.
So for example, you have an app and you like, you want to call the RARPC to get like a fresh
quote, you know, at that very moment.
And then you want to derive a cryptographic key that's tied to the TEE's identity.
So you want to prove that the connection that you have with the agent is also tied to the TEE itself. And so this is how you can then you can emit basically like a custom event that extends what's known as the registry. So this is known as RTMR and its purpose. So all of this RARPC calls themselves are basically protected by the RATLS already.
So the communication between the application and the guest agent is actually attested as
well. So if you have an endpoint and an application that's hosted on a domain,
but you want to verify, say like the front end,
that's a really useful thing because a lot of people, they can hack the front end.
And you have an agent that might be a decentralized agent because we're moving into this strange era of like democratization of agents.
So you have near shad agents, you have Gaia, you know, decentralized agents, you have Nanda democratized agents, you have obviously like the workflows of Synth.
You have X, Y, and Z things.
So all of these servers that are running basically untrusted code,
you want to be able to verify that not only is the front end,
but also the application that's running and hosted on that domain
is attested by the same attestation package
that you want to be able to receive.
So I can almost imagine the use case looking like a color pattern. I've been demonstrating this for quite some time now,
but it's basically the idea that, hey, like if I get the entire attestation back and it turns green,
well, then that means that the KMS, right, has actually attested to the hardware root of trust,
which also can verify, hey, it turns blue and green. So it's not just
green, but it's also blue and green, which attests to the fact that, hey, the attested endpoint and
the KMS are actually tied and linked to the same package. So I can verify, for example, the SIG store,
which is very popular when it comes to signing certificates online for domains and extensions
and so on. So it's like whatever app I'm running, literally, I don't trust the front end. So I want to also verify that the actual
domain is hosting the correct front end. So you can verify the actual console itself. So no one
can actually front run and hack your actual front end. And then you can also get like a purple
letter that says, yeah, but like, this is 100% Olama, not DeepSeq, because you can prove the operating system level
as well. So this is what I mean by remote attestation. So you literally can prove this is
100% Olama that comes back with a verifiable front end. You can see the git commit hash in
this attestation quote, and then you can verify that it was a decentralized root of trust. So
the hardware host can actually see the data.
So who's hosting that?
Fala.
But like Fala right now is the lead of Dstack.
Anyone can run Dstack.
Like if you're an infrastructure provider, this is like a huge call out, but like everyone in the world can run Dstack.
It's open source.
So if you want to like basically go through the same thing that Fala did, which is very
difficult, but it's sort of like Kubernetes moment that Google had for Kubernetes.
And then the entire world attested to like the strength of kubernetes and like pods and all the
stuff for launching like large scalable applications um the same is going to be true for confidential
virtual machine loads that's super cool that's super cool let's talk about a little bit where
this where this comes comes from because that's what i'm thinking about is like the why the why the why
it's like you know the new word is is vibe coding and we're calling it that but it really is you
know if you're a traditional software developer these are sprints very quick sprints they are
co-creation opportunities we are iterating through hackathons through build-a-thons, through meetups, through workshops, through all that.
These ideas come from personal experiences and pitfalls of things that we personally have
encountered in the development process while we're trying to maybe build something a little
bit new outside of the normal system, something that doesn't exist yet, right?
There's always these barriers of things that you have to remove
when you're building things that don't exist yet,
whether it's interoperability, accessibility, confidentiality.
All of these different functions are things from personal experiences of Ilya
and personal experience of Dylan, of myself,
of things where we tried to build and we hit a wall at some point.
And these are our little tricks of the trade that we've picked up along the way
that are really defining the processes as we develop these things.
And it's to solve a problem that we've personally encountered.
Why do we want these things?
Well, if you can iterate as fast as you can imagine,
and you can deploy and scale faster than your competitor,
well, then you have the real advantage.
And so that's why these developers are the ones that are on the stage today the ones
that we've got our eyes on as far as is why these builders why these products deeper than the product
is is the development underneath and that's awesome guys i met you you you a little bit a
little bit sometimes losing connection with you so maybe if
you can find a little better connection will be great but uh right now um you can like started
losing you again at the end but uh i have a question to you by the way that's coming from
the community what um and why how you would explain what is the universal um um universal ux
is the universal UX.
I think that would be the easiest thing to explain today.
Compare to the bridge.
Or like the regular bridge that you see
or cross-chain possibilities that people got used to.
Would be great if you can explain.
Sure.
So Zeta Chain uses native assets.
They don't use wrap tokens or bridges
to avoid vulnerabilities that are happening in other cross-chain solutions.
So this is definitely not the only cross-chain solution.
The other cross-chain solutions are called multi-chain.
And Zeta chains are called universal.
And it's universal because we're connecting many more chains than before.
This is a next generation of multi-chain.
We call it universal applications.
And so the difference is that you can do anything to anything.
We're building out active integrations with Solana and TUN and Sui, which are all new.
New experiences to be interoperable as a complete program, as a complete platform.
And so Zeta Chain, how is it different?
It's different because we're using the Layer 1 blockchain Zeta Chain.
They have a native gas token.
It's called the ZRC20 token.
So if you're a Solidity developer, you're coming from the Ethereum ecosystem that is
a directly comparable asset to an ERC20 token on Ethereum. And so we can use ZRC20 as our guest
token. We use our blockchain Zeta chain, and we have different interoperability solutions. So
it's not the only interoperable solution. It is comparable to something that you may have seen in the past with competitors like Wormhole or Layer Zero,
or maybe even Chainlink at this point. Cross-chain transactions are what's happening in Web3 and
crypto right now. And this is a different solution. It's a different team. The startup
is launched in a different way.
They've made it all the way to a Series A.
They're launched out of San Francisco, run very much as an agile startup with a lot of experience.
And so as far as what I've seen in Web3 and experience-wide, they're growing at a much more accelerated rate
than blockchains previously.
And that is a really strong signal
and a really strong market.
So it's definitely a great place to be building.
I think there's a lot of opportunity at Zeta Chain
to start building out those custom solutions,
those new solutions on this chain.
You know, chains do get a lot, a lot, everybody at some point gets a narrative and it's hard to break away
from those things.
And I think that is something that's being defined at Zeta Chain right now, which is
a really strong signal.
I mean, you don't want to build in an overcrowded ecosystem.
Like if you tried to do NFTs on Solana right now, you'd be a drop in the ocean, you know?
And I think that that's really where you want to be in Web3 is building on what's new, on the latest, greatest, most advanced solutions.
I would say this is, you know, the new next generation modern blockchain.
Yes, yes, yes. I'm just just gonna leak a little bit of information like
give some uh insights of what we are building we sent we have new grants programs we have new
bounty programs for building our active integrations with solano we have a super team bounty yeah i
think do you think do you think it will be okay if i'm gonna share a little bit with the community
what exactly we're building right now um i think it's going to be pretty exciting for the listeners, right? So the latest stuff,
I'm just like a simple agent that we put in together and kind of like utilizing
universal possibilities from Zeta chain. We're building an agent that, you know, everyone is in crypto, right? Paying the employees in crypto.
And some of the employees have ETH wallet.
Some of them have Solana wallet.
Some of them BNB wallets.
So it's like a multiple payments that you want to do in one payroll, let's say, right?
Or multiple vendors.
And not all the vendors can accept, let's say, or don't want to accept usdt or c20 they won't accept uh usdt on
back 20 like bnb for example so we launch in an agent where you can just create an invoice
and that uh invoice will be paid to multiple vendors on a multiple chain utilizing universal possibilities
from Zeta and then post it in one receipt.
It's one agentic flow, one UX, and there is no breach juggling anything.
So we just get it in LinkedIn.
It takes about 10 seconds.
So we got the developer experience is much improved than blockchains previously. And coming from Solana and Polygon before that,
I've experienced all of the hardships,
a lot of the hardships that developers face
when you're building without documentation or resources
or trying to discover what's possible with your technology.
And I think ZetaChain has really got it down to a new CLI
that is really quick to install and even faster to use.
So you can deploy a universal application
or a smart contract on Zeta Chain in about five minutes.
And we're actually showing people how to do that tonight
at Brookland, in Brooklyn,
we're Vibe Coding After Dark at Brookland.
And so devices are provided,
the partnerships with Bolt.new,
they're providing Vibe Coding resources
so that we have pro versions of their tools, among others.
It's with Venture Miner,
who is the local New York City.
Well, I think they're global, actually.
It's a global boot camp, coding academy.
They do an awesome job,
and so they're helping facilitate
these workshops tonight.
We've got BitBasil in the house.
It's going to be a packed house at Brooklyn.
So if you're in New York City,
go down to Brooklyn.
Go to Brooklyn.
We got Swags and Grants and Bounties. And I'll be there as well. It's going to be a packed house at Brooklyn. So if you're in New York city, go down to Brooklyn, go to Brooklyn.
We got swags and grants and I'll be there as well. So since, since coming to wipe code with you guys, so like, definitely let's,
let's do it. So guys, I'm going to go to the, to the questions from the audience,
a quick Q and a, let's try to cut it short. Uh, we have to wrap up the M a soon,
even though I want to make it recurring I want to make it
we'll drop the link in the comments for them yeah maybe a weekly one or bi-weekly one because we
have we can we can go into so many cool questions together I love that group and next MAs we want
to bring another devs from different projects you you know, touching the points of, you know, about three problems and then security issues and actually solving the real world problems, real human problems that, you know, they have to deal with using I would just recommend to answer pretty short and fast because we have plenty of questions so what's the difference from what you were saying on
the same follow collaboration confidentially I compared to grok or
chat GPT private mode so well the difference is is that like with synth
and these workflows they're your own usually So you get to set up your own workflows. Like programmatically, you get to host your own AI rather than having to trust the, you know, super general intelligence of Grok or GPT.
You get the privacy out of the box with like a democratized stack that allows you to have an AI inference model hosted on your own TEE CVM.
And then, you know, you can basically locally host a DeepSeek or a Quen or a GPT for like six cents an hour on Fala versus, which is basically like the equivalent of $15 a month versus, you know, $40 or $20 plus for the same thing with GPT.
And also like with GPT, the difference is that if you're on OpenAI's like, you know, platform,
you're going to basically have no, you don't have any verifiability of what they're actually doing with your data.
And if you really think that they are keeping the chat private, well, they're training on your data. They've even admitted this in open, where they said that we
don't actually recommend you sending any legal compliant data or anything when it comes to legal
court cases or your own personal court cases, because we cannot confirm or deny that we won't
actually use the data. The information can be used against you in the court.
Yeah.
So Sam Altman actually already said this.
So that's even with private mode on.
Okay.
So like, and so, yeah, no, like, how do you think they're able to train the smartest AI's in the world?
It's because they use your data.
So it's not, it's not private.
The difference again is between privacy and confidentiality.
So yeah, like privacy is not actually private when it comes to cryptographic sort of, um,
you know, rhyme reason.
Um, privacy is like, yeah, it's, it's, it's committed and revealed, but then in the witness
log, you can still see the plain text data.
Okay.
So that's the thing, um, with Fala and like confidential compute, um, the plain text data
is basically hidden behind firewalls, and basically it happens and gets processed within a confidential black box container running on the processing unit ran by an Intel TDX machine, and you get the verifiable attestation back that says this is 100% the GPT that you're looking for.
These are the droids you're looking for. You know, these are the droids you're looking for.
You don't have to trust us.
You can verify it.
And so that's the major difference.
Yeah, I'm going to go to the next question
because we have a couple of interesting ones I want to ask.
So let's go to the next one is
how do I add my own skill MCP to Synth?
Good point.
We're going to open Docker.
We're going to open, we have opened GitHub.
First of all, you can always participate, send us requests if you want to participate as a dev.
In the future, the ability to activate skill is super easy.
You just open Synth, you go to to the marketplace you activate skill and uh now your
agent can do some additional functionality uh if you're talking uh as a dev who build an mcp the
whole idea we build in the open it's called sin bridge it's open bridge where if you have an mcp
you can connect to synth and we can connect to you through the SYNT as well.
Another question, do I need to be in crypto to use SYNT?
No, it's Web2, Web3, so we can have just Google account or X account just to log in.
Next question, where are the audits stored and who can see them?
Audit logs stored and who can see them? Audit logs storage and who can see them?
I guess it's a question to Dylan.
The audit logs are basically the logs of the,
like what the quote is.
So the stored logs are in your profile.
So you get to see them, you know,
because you have a decryption key on your account
using FALA.
But in D-Stack in general, they come back to you as logs.
So they're stored in what's known as DSTAC.
So DSTAC is the OpenTE framework that fetches the logs,
and it returns it to you.
So these logs only prove that the bootup had occurred,
such as the KMS, the VMM, the CVM, and so on and so forth.
So these logs come back to you as a report.
The report then gets detached to you as a report, the report then gets
detached to the attestation. And so you get both the attestations and quotes along with the audit
logs. I got my favorite one. And you know, I do and I told you I've been building decentralized
storage already in the past, right. So we had like a whole web three IPFS based decentralized
storage solution that I was building before. So I'm quite familiar with security and Web3.
But that's my favorite question I asked a million times.
And the user is asking you right now,
what if a court of government asks for data?
Well, there's nothing I can do for you.
Sorry.
But literally, I think, because if the data itself was processed within a TEE,
then you can prove to the court that the data was actually ran where the data was once processed on using defense in depth technology.
And so they're going to know what this means, because if you if you had to have been HIPAA compliant,
then the government can't go, do you have the data that we're looking for?
It's like, well, we're also HIPAA compliant.
So A, no, we don't store medical data.
So that's proven by the TEE, which attests to the fact that the data doesn't actually get stored on the TEE.
The TEE is not meant for storage.
It's meant for processing.
So the data is just co-processed by a TEE unit and then it might be issuing or, you know,
fetching requests, but it's doing this in a confidential way.
So you can't give data that you don't have and you can't give data that you can't see.
I guess, you know, everyone is going to the question.
Yeah, it is.
It is everything you're saying.
It makes sense.
I guess everyone is going to the hot topic in that direction that what if
the you know i i i know how i answer in this question because i've been asked as many times
because we push and synthesize the confidential ai but the question is what if the terrorist or
some bad actor or a hacker want to use llm to do bad stuff are you guys not trying to
use LLM to do a bad stuff.
Are you guys not trying to kind of like make them private to do
their own shady things?
Well, no, like, I mean, like there are insurance, there are, there are
assurances in terms of conditions like that falla as a cloud provider has for
itself, because you know, it's an American based company and we can't like host VPNs.
So there are rules in terms of what you launch just because like you can't like, there's a difference between the law and confidentiality.
Like you can't do something against the law.
Right.
And try to be confidential at the same time.
Then go to like Dubai.
Okay.
So that's, that's, that's something that is ill-advised, right?
right? No one should be able to launch crimes against humanity, for example. And so I think
No one should be able to launch, um, crimes against humanity, for example.
anything regarding like FALA is synonymous with crimes against humanity is like totally,
well, it just doesn't make any sense. Because A, we have terms and conditions that also prove that
if we find that there are services that are listed on our CVMs, we have the right to actually push
the image, meaning we no longer have to actually host it.
You're using our services.
You can, of course, use your own DSTAC.
You're more than welcome to.
That's what I've been saying.
So it's an open-sourced TEE framework that anyone in the world can host.
So if you wanted to launch your own crazy platform using DSTAC and you wanted to have remote attestations and running confidential compute,
if you can get your own hands on your own hardware, like a TDX, go right ahead. Like
I believe in freedom. You know, I'm an American. I think everyone has the right to basically
do as they see fit. But, you know, this is known as unconditional privacy or unconditional sort of
like wavering means of permissionless architecture. And so I think a lot of people in Ethereum and
crypto really vibe with this, but then a lot of people are also scared of what are the ramifications
of AI and also, you know, people who do bad things in the world. Well, there is like, you can't,
you can't innovate if you, if you're consistently stopped, you know, and to really push the needle
forward. And so I think that's really kind of what happened recently with AI. That might not be everyone's cup of tea, but that's kind of how
the world works. Like the whole world basically is racing towards AI. What we're saying is that
you need to be able to race and then also ensure that there's secure parameters around that AI.
And then we have terms of conditions where we can actually list private applications and
confidential processing units for your AI. But then there are some rules and conditions that follow the law.
Of course, you can't break the law. So yeah, just that that needs to be really clear, because just
because something is confidential does not mean it's illegal. It was the same thing that people
thought when it came to zero knowledge proofs. Oh if you're using tornado cash then you must be money laundering it's like no that's not true yeah it's like i can use tornado
cash to send money privately but does it mean that i'm trying to send you money privately
for a bad reason no maybe it's because i'm trying to pay you and i need to make sure that it's
confidential but then it's on chain using ethereum so or public goods where i need to be able to basically
uh not admit to how much money i have because i'm trying to keep my um my corporate compliance
standards up so that my competitors don't see how much is in my bank account things like this like
there are billions of use cases in the world defense yeah i usually i actually you know i
build my own solution because we build in a layer which identifies what activities you do at some point and then it's actually have to stop you of doing something that's going to be
illegal uh similar to what you can see in gpt but we have to have our own layer identifying those
because if we want to run it in private confidential tees then we need to uh take a warning you know
or you know some sort of uh layer of uh at the station for uh the system showcasing that
you know the user trying to you know build the bomb right now with his synth and we don't want
to do that so that's kind of like you know what i what i'm putting together in the syntagic right
now a question to zeta um i met you if you're here if you can hear us uh or um can you speak
yeah i'm here i'm listening okay awesome yeah so how does that the
chain plan to attract non-defi builders like ai gaming or social apps to attract or to track to
attract defy and gaming builders yeah yeah non-non-non-defi builders like gaming social
and ai man they're just lining these questions up for
us we're doing a series of builder building events and workshops starting with vibe coding and
tutorials for for zeta chain basically showing people how easy it is to get started building
on zeta chain especially so if you're a solidity developer, you've built on previous EVM-based chains in the past,
it will be very familiar.
And I think you'll enjoy the experience as far as the ease of use.
We have some fantastic protocol engineers
that are really working on the user experience.
We call it the developer experience.
We actually have a whole team just for that.
So that's what I'm out here focusing on.
So how are we going to attract non-DeFi users?
Well, I think that we're doing so right now through AI.
I mean, I think there's a big intersection
of really smart people who like the difference
between SF and New York.
In SF, everybody's talking about AI
and worry less so about the Web3.
And in New York City, people are very much like Web3 and stable coins.
And AI plays a lesser role here in New York City.
Right.
The way I understand the question, yeah, AI is definitely because, I mean, we're already talking about DeFi.
We're talking about AI.
The question was from a gaming community, the people who play games.
Is that a change or something that could be interesting for them?
Stay tuned for that because I'm not going to blow any whistle too soon.
But we've got some really awesome stuff in the pipeline that I think you guys would be really happy to see these communities that you're very familiar with.
From DeFi and gaming and web3 and nfts and i mean we're having
a lot of crossover a lot of people are interested in our for ability with their project and so i
think you'll see some familiar faces in our ecosystem very soon um i'm gonna i'm gonna ask
the last one the last one for you as well uh when i can expect the first killer universal d app on zeta well who's talking step right
build it we're waiting we're waiting for you so so we we've got some cool apps coming out but um
you know we're really waiting to be to be blown away but and i think there's some there's um great
contenders in the pipeline right now like people who are on this call, you know, that's who's at bat right now.
So I would say stay tuned for what's coming from SYNC
and from FALA.
I mean, I know these guys are cooking every day.
And so just, you know, what's gonna come from Zeta Chain
is we're really trying to create the experience
that the builders of the next killer apps want to be part of so if you've got these ideas come on
over and cook yeah awesome yes that's great a couple of people from community i just added
to the speakers you can have ability to ask an rma in real time if you like just raise your hand
if you're shy it's fine yeah may if you want to ask the question you're welcome you can speak
all right all right I'm not shy thanks so um I got in this kitchen from since cause I thought
the post he did before so um what is it to say is not like a question but i'm just trying to
um because okay let me just say what i posted so i and a friend please we made a
uh an onboarding okay because i'm it was also so we started an onboarding okay because it was also so we started our onboarding space or into the
touch last two weeks the last we made or two were some few questions that was being asked and we
were not in the right position to answer them so I one of the questions was bringing complaints
that the Zeta Chain rocked them
due to some test nets and the likes.
So that was one of the questions they asked.
So, and we didn't want to just bring false hope.
I don't know if you get what I'm trying to say,
cause we were just trying to like tell people
what Zeta Chain is building.
Cause I just found out about Zeta Chain three months back.
And I've been making my research.
So when these questions came up, I was like, okay, you should calm down.
And I'll try my best to reach out to the ecosystem you get.
Sure.
I recently joined Zeta Chain about two and a half months ago.
And from what I can see, it's a very strong agile team.
I think they operate a little bit differently, more like a startup, more like what you would expect to see from like a Silicon Valley startup. So it's pipelines and management of vendors and resources. As far as any token trading campaigns or anything like that,
I don't know much about that.
But from what I can see, it's a strong token
that's run by a really strong team, myself included.
I definitely don't think that you have to worry about it
being a rug pull by any means.
It's definitely a well-funded company.
And there's definitely a lot of, you know, real things out here happening with ZetaChain.
So hopefully we can help you have a better experience moving forward.
Yeah, thank you for the question, guys.
We have a couple of other, I think we have Todd and we have Hermila.
Do you guys want to ask something?
Yeah.
Yeah.
I have a question.
First of all, what's up, Dylan?
Yo, good to see you.
Good to see you too, man.
Yeah, Dylan and I went to school together.
We did.
Yeah.
I always run into Dylan though at crypto conferences and now online.
So, you know, stay in crypto.
You make friends for a long time. And I also was head of operations
at Zeta Chain for a year and a half.
You were there before me. What up, Ted?
This is so funny. DM me after this. I holla at your boy.
Yeah, I was, yes, I know Ankur and he's
absolutely right. Yeah, zeta is very professional
they um they know how to get things done and do it properly um great investors too so i'm really
excited to see all the great work you guys have been doing um but i wanted to ask are you guys
doing anything with rwas or like how do you approach kind of the you know bridge between
like real world assets right now or unstable
coin regulation so before i make any announcements or like things that are happening internally i'm
trying really really hard to keep professional so just keep your eyes on you know i think there's
there's a lot of talk about what what can what the potential is and i think I think that we're in a much better position
to execute on those kinds of deliverables
than, say,
lesser blockchains or
maybe some blockchains that have already
picked their narrative.
And so I think there's a lot
of exploration happening right now.
I think the attraction
of that, you know, it starts with vibe coding
and onboarding for community to get the word out about that we're out here building.
But, you know, the attraction of higher caliber talent or maybe even institutional projects or projects that have already, you know, been flushed out on several blockchains or those.
I mean, those are those are, you know, Those are ripe partnerships for Zeta Chain as far as
business development and integrating with Zeta Chain. I think definitely they grow
much more quickly through partnerships, acquisitions, mergersgers integrations than um you know building from grassroots or the ground up right
so i would say um there's there's definitely you know some something big is coming and while we're
building community apps and portfolio apps and and i think um you'll see some pretty robust use cases
come of the new you know interoperability that's available
god if you don't mind i'll add from since i'd uh we just partnered with a project called um
nubula uh it's new york uh based uh asian new york based startup uh so pretty much i mean it's
rwa but it's more like it's not from the assess perspective. So pretty much what we do there and this type of partnership I'm going to is the region with robotics and IoT. So the whole idea that, you know, we build an agent that can consume real world data, such as like weather data, you know, like any sensor supply chain or any IoT data, like credit rankings, you name it.
And then we pretty much automate, you know, through that agent optimization, yield optimization,
or you can issue an insurance and loan contracts on that data, or you can make like a power
management, power risk models. So that's,'s you know that's kind of like the cool
direction in like real world uh data integration but generally speaking rwa rwa is like real estate
tokenization real estate invoices commodities you know this is all could be uh used by you know
inference layer that coming from the agents where agent can
handle you know onboarding kyc or any contact execution way faster than you do it uh manually
uh and you know same as i say like yield monitoring or liquidation triggers different
things that are helping in r w a to improve UX and
improve the flow that's kind of like where I want to take scent on the w a
side yeah as for us we just did a podcast actually talking about how T E's
are being used heavily for our w a s if you check my Twitter checked phallus
Twitter we just reposted our podcast latest podcast uh
called black box podcast i had a full conversation with spout finance uh my good buddy paul mirlow
who just actually broke down the entire utility and usage of rwas with erc3643 and like how
exactly tees are being used for that and um, um, so yeah, like RWA plus TEEs, perfect use case.
Um, if you don't already know, go check out that episode to understand how they used Inco.
Um, I believe there's still testnet.
And so if you have a program and you're interested in basically launching your RWA co-processing
for like AAR programs or for corporate bonds or for these
like derivatives and whatever it is. Yeah, like we're seeing some of these sort of trade fi slash
defi programs getting listed now on falla. So it's pretty, it's pretty new though. But if if you want
to hint at least to where they'll go and look, go look up ERC3643 on Ethereum to go get more information.
You'll get a lot of stuff there.
Awesome, Todd.
Are you still here or are you already left?
Because it would be great to hear from you pretty much what you do.
Quick intro because we love to to see like real professionals
and special guests on the ama so please maybe introduce yourself or what do you guys solve in
our double space oh yeah totally um can you hear me now yes yes okay cool yeah that's really
interesting thanks for sharing so i work at l Lorenzo Protocol. We're basically like an investment bank, right, for Bitcoin whales, traditionally on Babylon. But then, you know, we want like real stable yields. So we made like the first yield bearing stable coin on USD1 infrastructure. So we're partnered with World Liberty Financial and we're basically doing
a three yields for any stable coin depositor. And we mint them a USD one plus token, which
can be staked as well. So you get rewards from a quant trading desk that trades crypto
markets. And then you have treasury bills. And that's the real world asset side of thing. And so there's that stable yield that's always coming from those deaths and it's tokenized.
There's a lot of interesting companies that are like in immediate areas for
for these different T-bill funds like BlackRock has one securitized they work with, etc.
So these are bridges are really open now because of the new regulation.
But what we do also is we integrate downstream DeFi
as well for our USD one plus.
So you can use it as collateral, you can use it on Pendle.
We're working on getting integrated
with all of those downstream DeFi's.
And then also we're working with some exchanges too
so that we can offer like collateral
for those exchanges themselves.
So you can really start to like stack these yields and do it in a compliant way, like
with world assets included.
We also have a deal we're coming out with open Eden soon.
And they're a T-bill staff back stablecoin themselves.
So you could, you could, you know, have one of their stablecoins deposit into, you know, our platform, Mint USD 1+, etc.
But it's great to hear all these builders.
Really nice to connect with you all.
Hope we can stay in touch.
Thanks.
Sure, Ted.
Pleasure to meet you as well.
So let's stay in touch.
Let's follow back each other and maybe set up a separate ma just
specifically for wa topics and what you guys do because i love to find the best use cases for the
agents so agents can optimize you know not just the user experience but also uh patch data faster
and improve so many different things in the current tech and current products.
I guess that's it, guys.
I mean, I think it was like almost hour and a half.
Such amazing AMA.
I think we did such a good deep dive.
So as you can see, Zeta Chain, our network.
Thank you so much for coming.
Thank you, Ilya.
Thank you, Dylan.
Just remember, we'll all be at Brookland tonight.
B-R-O-O-K-L-A-N, Brookland in Brooklyn, New York City.
And I'm going to give you an alpha for the community as well. So app.stint.gg. We haven't posted it much yet, but this is the first alpha.
And then send me the feedback, what you think. Now you can play around with voice agents.
And we're going gonna do a launch soon
right now it's pretty much for the community for the listeners you can claim points points will be
converted to the agent energy and the tokens later and yeah let me know what do you guys think
um well done good job Yeah, thank you, guys. Awesome.
See you guys soon.
Speak soon.
Bye, guys.
Have a good day.
Bye.
Ciao. Thank you.
