Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
Crypto's Got Talent Ep.5
Recorded: March 20, 2025 Duration: 1:37:45Space Recording
Short Summary
In a dynamic episode of Crypto's Got Talent, innovative projects like Gargoyle and Indexify were unveiled, showcasing advancements in drone detection and decentralized finance. With upcoming token launches and strategic partnerships, the crypto landscape is set for exciting developments.
Full Transcription
you you you you you you you you you you you you you you you you you you you Are we on? We are live. All right. Welcome, guys. How's it going?
Wonderful. How are you doing? Awesome. Good. Well, welcome. We're live from New York City.
Sweet. It's Crypto's Got Talent, episode five, second episode that we have. So let's just go ahead and do it. Let's introduce the panelists
for today and then the first project that is going to pitch. First panelist, we have everybody
knows Aaron. Let's go. So happy to bring everyone here together in New York. Honestly, like we did
our last one in Denver together and we had some issues with the mic and everything, but this time
we like let's do it properly.
We have had like 3000 people at one time viewing in Denver.
So let's give them a good experience this time
and have some fun.
Exactly.
And yeah, it's gonna be cool.
I'm excited.
Nice.
Dillon.
Hey everyone.
My name is Dillon.
I'm a research analyst at Misari covering primarily Deepin.
Very excited for all our projects today.
Awesome. And yeah, I think it's going to be exciting for the projects because we have
like, you know, us plus, you know, pretty prominent Deepin analysts and a pretty prominent
Deepin project. Rob, how are you doing today?
Good, good. I'm Rob, one of the co-founders of Deemo, which is a mobility focused Deepin project. Rob, how you doing today? Good, good. I'm Rob, one of the co-founders of Demo, which is a mobility-focused Deepin project.
Nice.
All right, let's, uh, I don't know.
Before we kick this off, I almost forgot, guys.
So Aaron and I, when we decided that we were going to do this,
we're like, we need a team saw.
So then Aaron one day just basically sent this.
We're not going to do the whole thing, but just feel the vibe.
High budget.
It's an amazing thing. I'm going to be the Rocky 4 scene when he's in the Lambo
Yeah, you can put an inspiration what happens I'll let you know after
So you have the people now get one for demo let's go make that crypto deal guys. Let's go do it. You got it.
You got it.
Nice.
All right.
Let's kick this off guys.
Well we have about an hour so we're going to have five projects today.
I think let's give maybe 10 to 12 minutes per project.
Sure.
First project today is Gargoyle.
So guys maybe enlighten us for like seven to eight minutes and then yeah, we'll ask some questions
Okay, awesome comments. See what's up. Let's go
Yeah, I mean, let's see it started with a buddy of mine in LA was robbed
They took the safe out of the second floor
But in order to know when he left this house, they used the drone to see him depart. And then during the robbery, they use the drone as well.
So we go back a couple of years,
we're gonna work on a project together.
We'll get a bit into our backgrounds, which is relevant.
But I was like, surely there's a ring nest
for drone detection and there's not.
And so we started going on this path last summer.
We built a prototype, people wanted to buy it from us.
We had all these security companies come to us
because robberies are happening every day.
And if you're going to rob someone, why would you not have eyes in the sky as coverage?
So then it got even bigger and bigger and we started realizing that municipalities wanted
police enforcement, hospitals, casinos, beyond just stadiums and airports, which have the
radar systems now.
So we created a deep end project to be able to incentivize people to put them on the roofs by
the one-to-one digital memberships beforehand, so essentially sponsoring them. And then we share the
revenue that comes in from the municipalities with the people who put the nodes on their roofs.
We detect friend or foe. There's a lot of friendly drones out there. And if you think the future is
going to be autonomous and AI and unmanned,
then we're going to be collecting that data.
So from zero to 900 feet,
we collect the data, whether it's good or bad.
And we also get data,
we can package that up in terms of
where are all the Amazon deliveries happening
in Phoenix right now,
where all the photographer,
where the sales are going right now.
That's all edge data that we can capture.
But also we capture the defarious data
and potentially robberies
and other things that are happening there.
There's a good example.
If any IOPs that is composed of like 90 units,
those units are incentivized
by individual restoring them on their roofs
or in areas that are above ground.
And basically what's happening is
you not only are getting the benefit of using a unit that is
providing awareness from potential aerial threats, right? Because UAV is going to be
potentially aerial threats, but you're also being part of a network. And that type of network
collects data and the data cell can be sold and can be used maybe for research. And so the reason
the deep insight, as Mike mentioned,
is why don't actually incentivize
the people of this building, not just for the security
purposes, but to be part of the network.
They will be rewarded with the value of the data
that they're collecting with the hardware.
And if you're sure, they're also keeping the hardware active.
Because if you try to do this from a government perspective,
it becomes extremely expensive and,
of course, unreasonable from a perspective of the tax payers' services. But then, from a smart city
approach, you're really helping or you're incentivizing the citizens of actually adding
value to the city itself. And this becomes something you can move for becomes a city level, on a region, on a state level because these
sensors are all connected.
So they communicate with each other through the cloud and the data set is extremely extremely
valuable, especially if you're in an area that has some security concerns, like near airports or near
hospitals or near certain areas that drones might not be allowed. But in general, our data,
they're going to become way more relevant when drone deliveries is going to come out of the pilot
projects and are going to become more widespread. then you're really going to have to understand where are they,
what are they, and the direction is the coin,
and how they're open.
We've got the latest.
We just got back from Estonia.
We trained with the Ukrainian NATO drone commands
to see the latest drone detection
and what's happening now, because they're
holding the front line.
So we met with the trainers to train the 200 pods that
are holding the line right now, how to build drones,
how to detect, how to train their pilots on the are holding the line right now, how to build drones, how to detect,
how to train their pilots on DRIVE.
Quick background on ourselves.
I was at Filterbox,
we're in the first ever Tech Stars class in Boulder, Colorado.
We did social media monitoring.
We had Ping next, Gnip next door,
which is Ping spelled backwards.
So with the APIs, and we were acquired by Jive Software,
then went to Midtown with News Corporation.
So we have about 50,000 employees, 20th Century Fox, Wall Street Journalists, training journalists how
to use Twitter back in the day, how to put Rubber Duck on Twitter, how to put the hashtag
launches and everything back in the old Web2 days.
Fast forward a little bit, went to BNYMEL and America's Oldest Bank, started, kicked
off the initial Bitcoin custody program, which took seven, eight years to actually come out.
And Web3 accidentally started Empire Dow,
which was a Web3 community center
where we tokenized real estate.
And we had a lot of deep end projects in there,
including helium.
So it was really on the ground to see how this started
and whatnot.
And then when we met with the Department of Defense,
we went down to Virginia, met with all them.
We have capability, but not capacity.
How can we incentivize people to put these on their roofs?
And I'm like, oh, I've got experience.
I've got too many Williams right now.
But this is kind of what really inspired us.
We need to create this D-PIN network yesterday to,
it's not an if, but when the next terrorist attack
is going to be either domestic or international,
it's gonna be using drugs.
Because what happened, what we saw in Ukraine is
fiber optic, which is like a wire,
so they can't knock it out radio frequency
Happened and then a month later as the Russians made it month later Ukrainian said two months later
Chinese are selling that Alibaba for $200 month later. It's in Myanmar. It's in Colombia
So all the things that we're seeing now the warfront always end up coming to the citizens
But it's coming fast and we need to prepare for this yesterday
So that's why we want to create this network to help protect our assets and citizens.
Amazing. I mean that was a five minute pitch. It felt like perfect. I think we got all day for the
we wanted. I have a soft spot because you know there is some Italian in the team.
But yeah what do you guys think? Let's go. Questions?
Tell us about the device. Is it working today? Is one per neighborhood?
One per house? There's a POC that we developed on our kitchen table and is able to detect
some specific drill that it wants to make sure where the deck is working. By sound?
By frequency and network signal. So a lot of the sensors that are there are extremely expensive
because they're based on either radar based or
they're really frequency based but they have extremely wide range. So we're talking about
binobics right and that's or plans and the reason why is because they're designed for battlefield
put them into open battlefield and they are able to get back the threats coming from fire.
that they are able to detect the threats coming from farther away.
In our case, you don't have to do that.
There is a node that can just cover the size of the house.
So in terms of detection range, we're considering
900 feet of a 1,000 feet of detection range.
And I'm getting confused with the meters and...
The imperial system just doesn't make sense.
It doesn't make sense
It's fine
We're starting a side beef here
Yes can have it I although I will say I think they're in high is
I think Calvin is the ultimate
Anyway, we're getting we're getting that answer your question
So sound detection mainly you said it can detect friend versus radio frequency
Yeah
And so and so how do you detect friend and Because some of the drones that are out there they actually have a specific idea so if you're
The license plates are exact
It's a let's play if you're legal you have to have one of those and report your mission. So
Every Walmart Amazon drone delivery has one. He's drawn. Yeah. Yeah, absolutely
You report your mission to just like fly on the neighborhood and take some
There's a part one or seven license your pilot you're an airline you have an aircraft in the field
We're the ones that like like line of sight only once those same
As it's why every pilot has to have, I'm sorry, every person who wants to fly a drone
has to have a pilot.
Commercially.
Has to have a pilot.
And are you checking if they've deviated
from where they said they're supposed to be?
No, that's something that right now is not necessary.
I mean, you don't do a flight plan.
You understand it, but the guy is actually using the drone.
He's a legit drone.
If he's registered or not, but also imagine that right
Scenario you're in your home
You are in the office and all of a sudden you see a drone roaming around
What is doing?
She the Drake video
Yeah, that turned into a Hollywood reporter article and just just dropped I mentioned gargoyle systems
Right so becomes a way of collecting data of John there
legit or understanding the drone is not legit,
which can be a clueless pilot, it can be a careless pilot, or it can be a criminal, right?
A person that really doesn't want to be a dangerous person. And so in that sense, it becomes a
potential threat. And you can immediately track it as well.
If I go to the store and buy a drone, am I registering it there?
If you want to stay legal, you do it at home, yeah.
There's the smaller ones and they have 4K cameras now, so you can now do one that's...
And that's why they're going to be done.
But the difference is, you know, the point of the battery doesn't last that much and the payload is very low, so you can already do much better.
It's only simple. Can you talk a little bit deeper about the difference between radio frequency plus acoustics
versus the radar?
Sure.
And like the price difference like and also just like technology wise.
100%.
Look, what's happening is radar is your emitting beams, radio frequency beams and when you
hit something, it comes back and you basically identify the result.
It's extremely expensive because you need multiple arrays of probes and high superstition software to basically process the way these radio frequencies are coming and going.
When you start to go into the more radio frequency phase, you have an antenna in the same way that you have flown basically the signal from the closest cellular tower and it basically thinks if there's a radio frequency
of a certain frequency around it. So the reason why it has to be a little cheaper, way cheaper
on the radio frequency is because first of all the complexity that comes from the array of
sensors and probes is completely different than the one from RETO. The second one is that the
post-processing for the frequencies is also on the easier way of actually executing it.
And then the reason why we considered implementing the acoustic side on the
The reason why we consider implementing the post-it side on the on the journey because as Mike mentioned some people are now studying
by experiencing what we see in Ukraine the battlefield but also
drug cartels on the borders actually using
drug because I brought the drugs don't any frequency
so those are the people the sensors cannot see the truth and
The only way to see them is by hearing them.
And so the best way to have a sensor to be able to go through, which allows us to work
higher redundancy is to have acoustic.
So acoustic is also on the cheaper side because microphones honestly, unless we're considering
microphones are extremely high end, they're particularly inexpensive and affordable.
So if you start to put them together,
it becomes an affordable system that can be used
and purchased by individuals.
Lessons, we do have some revenue,
we already have some paying customers.
Yeah.
On the demand side, who, could you go a bit more
to like who specifically are your customers
and like
how much coverage you need to provide for this data that should be down? Sure, so I'll give a few examples. It's B2B2B almost. It's the security providers that provide it for other companies,
right? So one of the banks, for example, wants all their data centers covered. So let's go take
about six for each data center. And then they also want their executives. After what happened
with UnitedHealthcare, they want gargoyles to be on their executives' roofs as well. And then when they go to conferences, they want them there so they can know if anything's going on executives. So afterward, half of United Healthcare, they want gargoyles to be on their executives' roofs as well.
And then when they go to conferences,
they want them there so they can know
if anything's going on there.
So there's a potential murders of executives and whatnot.
And then we've got a lot of HOAs.
A guy, he's pre-ordered 500.
He's like, I've got over 500 customers.
Two of the houses were robbed
and they used drones during what I explained earlier.
He's like, I'm just adding it as a line item to all of them.
Every one of my customers, I already set up their Wi-Fi,
I set up their network, I set up their cameras.
I'm also setting up gargoyle.
So that's who we're getting right now,
security providers who want it.
And eventually the idea is to end up having a tool
for a unit or a product like the blink camera
of drone protection, right?
So you end up purchasing from either catalogues online from
suppliers or even Amazon. Well, we pay you because we need the data. So we're a data company,
first and foremost, using AI to detect different patterns of what's happening up there, but
and then crypto on the back end in terms of how it shows up. Yeah, that's awesome.
I honestly would say I'm done like much longer Yes Yeah, we gotta have like, you know
Yeah, I wanted to ask about ecosystem what we're gonna see you guys live but it's coming coming
I'll launch a deep end for drones for criminals
So they can use them to rob places and together will generate a lot of business. We would assist them that has the freezing point at a 30 something and a boiling point. How often do you care about what water
temperature boils? Don't you just want to know if it's warm outside? I find it simple that it's
like okay it freezes at zero, it boils at 100, divide this in equal units and... Here's why
Fahrenheit is great. It's because I can't claim ownership. I forget where I heard this from but
because I can't I can't claim ownership I forget where I heard this from but I
think I like this Fahrenheit is the percent hot that it is then percent hot
if it's 70% hot outside that's perfect if it's a hundred percent hot outside
it's a hundred percent hot it's too hot oh so hundred degrees Fahrenheit a
hundred percent hot too hot 70 percent seven Oh my god. This just blew my mind. 30 percent? That's not very hot. It's kind of cold. 0 percent freezing. Like really cold.
It's percent hot. It's not as accurate, but I like that it's about. I think that's where it actually came from some meters meters. That's better.
But kilograms pounds who cares the same thing.
That's good.
So, it's not dragging in front of you. Just put it in your bucket. Yeah.
You don't want to have that thing. Yeah, we can send them on down one.
I appreciate you sliding us in.
All right, that's a lot of creative use cases.
That actually sounded useful.
I'm just like, like, how big is the sending, right?
How much coverage do you need for this to be useful?
That wasn't clear.
Yeah, which the, you know, you asked.
They're like, are they deploying it themselves?
Like, that's not really how deep it usually works. I think they're
incentivizing people, but yeah, it's like, if you're like, you you you you Awesome.
Okay. What's next? Who do we have here? Welcome, guys. It's the first time I kind of meet you, except, you know, in the corridor.
But yeah, what are your names?
What's your project that you're representing?
I'm Koji.
I work with Axel.
I lead grip.
And then Ash founded Axel.
Do everything he says.
I like that.
What's Axel?
Yeah, so I'm a Axel, I lead Rob. And then Ash founded Axel, do everything he says.
I like that.
What's Axel?
So I'll go very briefly to start.
We're believe in an active internet.
So what that means is right now, pretty much every way
you use the internet is likely, in some capacity,
just getting information.
Like, obviously, generative AI gets you information,
spits it out.
We think agentic AI
is the next step. What does that mean? Agents for us are any third party that goes and doesn't task
for you. What makes us a little different is we use the intent solver architecture that we can
cross, UniswapX, Cowswap does. And we believe that can be applied to generalized tasks. But our first
step to get to that future is autopilot.
Yeah, I mean autopilot is sort of like app one on the server for like agents that can do arbitrary tasks.
Picture Robin Hood or Lovebump, like super intuitive UI that everyone will see.
Drag some sliders, create your custom portfolio, copy and trade others.
On-Ramp, you have to set up a wall for you, sign in with Google or like Twitter or something else.
Under the hood, this exact architecture that Cozy was just describing,
like we're running our few systems or other mechanisms
to find counterparties to go and do the tasks,
like to buy your dream portfolio of Bitcoin and Ethereum
and the top 100 meetings today.
We'll go and auction that off.
Whoever can do it fastest and cheapest wins.
We prove that they've done it.
They get slash rewarded accordingly.
Adding that theme of state your intent, actual source,
and verify the agent to go and do that job.
And then they get punished or rewarded accordingly. That applies to everything.
Like Apple, Apple is Apple one down the line. It could be sports betting. It could be travel planning.
Like any task that you can verify, you can create a market around it.
Polymarket is a great example of what you can do at scale. Their verification mechanism is just
purely an optimistic oracle, potentially not as end game as decay. But if you can create a market for people to vote and have some sort of outcome, let's
move beyond information and get to tasks.
You might have tasks that are very literal, like I need to swap, bridge, tell me what
the weather is in New York, go and buy me a croissant.
Any task that you can verify, you can outsource that against sort of the endgame.
When you think about bringing guardrails to a very agent future.
How do you verify a croissant?
Yeah, so for something that's super real world,
for public real world information,
you can use an optimistic oracle.
For something like proof of ordering,
you could do ZKTLS to prove that some exact API was called.
Tees get interesting.
Like you could have some external party
running all their code in T to call the Uber API
or call the Jordanash API.
We know for sure that they've done it.
You could do stuff like proof of receipt
or proof of photos and stuff, but I'm
sure this will all be gamed in the future.
It's not as strong.
If I had the ultimate guarantee you have
is that there's some on-chain proof of something being true,
that requires, let's say, a T or ZKTLS.
So right now, we're obviously starting with what's easy
to verify and still has demand.
It's like, across doesn't pitch themselves
with this intense-based bridge.
They pitch themselves as the fastest, cheapest way to bridge.
And they win because of the architecture. We want to do the same thing. It's like the fastest cheapest way to bridge and they win because the architecture we
want to do the same thing.
It's like what's easy to verify and create these markets, prove that this
style of architecture works, then go beyond to the harder to verify stuff,
but very much outside of crypto is the goal.
The goal is not to be the best at verifying like in-house we're not trying to
build that, we're trying to go to whoever is the best at doing that and like
use their architecture.
Yeah, same thing for the agents. It's like we're not in the business of building agents long-term.
Today we're doing it all over the building of the front end.
We're building agents, resourcing and verifying ourselves for now.
The goal is once you have the demand in your workflow, get more folks competing.
Competitive free markets are amazing.
Probably the best thing that you may be doing is ever created, but counterparty risk is high.
Blockchain is allowing you to eliminate counterparty risk.
The outcomes from that are insane.
What you can do in a moment to your counterparty risk is the active creation of your system.
I could have a Like blockchain is why you eliminate counter-predicures. And like the outcomes from that are real.
So like what you can do in a moment
to your counter-predicures is like,
yeah, I could creation a new instance.
Like I could have a chat to be used on UI,
where it's like, I just state what I want.
Some engine compiles that into a subset of tasks
that other people don't do.
I don't think about doing it at all.
And I've guaranteed that the best is doing it
because there's some competition
or some optimizing mechanism, like an option.
Goal is to be the universal task.
I like that.
Paul Coles, guys.
How far are you guys gone so far?
Yeah, so the idea was, let's say,
August, small pre-seed around that.
Team formed around me, let's say, October.
Public beta of Autopilot, so this app one,
folks done trading,
ran from Halloween until end of January,
I think something like that.
We had like around 5,000 users,
which is crazy considering it's not real money,
like just people playing around with UI,
seeing it like poking holes in that.
And that was a good experience.
Went live with autopilot yesterday.
And it's a great test of the architecture.
Like in our head, it's not just like,
we're building a standalone app
and then build a network around it.
We're halfway through the network itself. And now
it's like that the app is live and it has order flow. And we think this is the main
issue with crypto. Too many people have built out the network without building a use case
for the order flow. And for us, it's like get the app, start to build out the network,
build proving systems for the really specific things that that app needs, and then over
time build everything else. So that's the set. It's like app is live. You guys can use
it. Literally just AX.AL.
You can click on button, go and use it on your phone.
The goal is for you to be able to onboard
anyone with this thing.
And then hopefully in a few months,
we have more random apps.
And then by the end of the year,
three to five apps that we built ourselves,
network is fully good to go.
And there's like an SDK.
Like any random developer can go in.
The replacement of backend API calls
when you're first learning how to code
is like intent request capsule.
Like you just submit these intent requests,
your entire app is floating on the sea of agents
that are doing these jobs for you.
Yeah, right now it's just your one agent.
Right now it's our one agent doing tasks
related to autopilot, hopefully in a month,
and that we've spoken to all the solvers
for like top ecosystems today.
Without order flow, nobody gets to share it
without anything, very understandable,
it's just like, it's good talking about this idea.
Once we have order flow in there, the pitch is very simple.
It's like pre-order flow to you.
You can be my shitty solver.
And is that, is that auction like marketplace live today?
It's just, there's one robot winning it every single time.
One agent winning every time, but is it ready for the second one to come?
So it's still permission and like crypto it's totally like
Casbop is totally permission.
Like it's the most, you see the 30 people every time.
You saw that's also very permission. Our goal is to be totally permission. Like it's the most sense. If you see the 30 people every time you saw back, it's also very
permission. Our goal is to be totally permissionless. That'll probably be
another month. I don't care if it's totally permissionless and it's just me
and somebody else because it workloads not at the right level, but like get
the architecture there and then start attracting more people.
What do you think will be the first non token acquisition thing that is done
with Axel?
Yeah, I think we both have different hopes for this.
I literally, we did a college short,
so I could colleges in Canada and then colleges in Boston.
So I went to five cities in literally five days.
I know exactly what I want to do.
I know where I want to be and I know
what time I want to be there.
It's just actually booking it is really frustrating.
And I totally would be willing to pay someone
a percentage of what I'd be willing to pay in total
to go and like actually make sure
that I go and buy a rail in Canada
from Montreal to Toronto on time.
So I think travel makes a lot of sense.
I think that's it.
We actually met like a really like senior person
that one of these like travel aggregators.
Yeah.
Yeah.
Yeah.
So we met the head of Google flights
and like pretty much what he was saying is
that like a lot of people enjoy traveling or planning their travels, which we fully agree.
I think where I disagree with that though, and like some principle is the fact that people
don't like booking. Like I knew I wanted to go from like Boston to Barnstable this weekend,
but I didn't want to like actually go and find out that this bus comes at this time.
And some third party be that,, be that down the line,
some form of AI or right now,
like some form of third party kid
who's just scraping the internet,
colexipy is able to go and do this for them.
And where do you disagree?
I just have a different example.
I'll give it to you.
This one came up a few days ago.
I'm a degenerate getting very down to it.
Drafting, I do like next fuel goal,
it's like the most unlucky thing about it.
One of my friends has like a much more straightforward thing.
It's like natural language.
If the odds of the Celtics winning the championship
according to ESPN go above 80% this week,
I wanna pull a lot of money on this,
unlike them winning the chip.
But at any point from now until the playoffs,
their odds go below 50, cash out of the bank.
Like I made a tournament decision.
I've stated my intent in five seconds to you guys.
The extra we go and do this is a huge nightmare.
Like he doesn't have to rate code at all.
It's just like working some rip book, not technically at all.
He has to go check the ESPN every day for the rest of the week.
Go and see what the odds are.
Then he has to figure out which sports book has the best odds for him to go in
places that he then he has to monitor odds for the rest of the season to see
what they become, then maybe cash out of that when that occurs, very manual, very painful. Each of these subtasks can be verified in answers. So ESPN odds,
like this you can just use in Umaport. You don't even need to get TLS. They're already doing it.
Polymarket Sports is like their whole thing and they have Instagram ads for a lot of regular people.
It's like that's easy to verify. Task two, replacing the bet, probably some TC, TLS kind of thing.
So like which sports book is best? And you could hard cut it. You could say between DraftKings and
Vandal or something else. I don't care which one is the best,
just go and buy. Task three is again event listening by the book. Task four is again
placing an event. Like me, like Axel, without doing any customer can write primitives to
verify arbitrary types in my head, the types off-chain or like proof of API call or proof
of website statement. Random people are submitting in heads for their billing in that data. We're
just auctioning that off.
So the market for who's gonna tell me yes,
get odds is very different from like the market today
for like cash flow offering.
Very different, very different entities,
but it's still in somebody's interest to go and do the job.
Cause my friends totally willing to pay a couple bucks
for a couple of dips off of like every transaction
in exchange for somebody doing this whole job.
I think like that is the magical automated feature.
Like anybody in New York would agree,
it's like we will live in a road ride to state
when I want it to happen, it's how do you actually get there?
You need guarantees and you need competitive options.
And that's my wife with it, it's like guaranteed.
Pretty scary, anything you say happens.
Somebody else does it for you.
Well, you think about it, that's like the travel agent
example, like travel agents used to be like
a really big thing, right? And I guess Travel agents used to be a really big thing.
And I guess it's not really a thing anymore. Do you want to explain why?
Why normal traveling? I think it's probably when people are annoyed,
and I actually think that's where the Google Fights guy is tracked. People do enjoy it.
A lot of our friends work in more corporate jobs. So it's their task to go and plan. And that was kind of the feedback we got.
But I think the counter to this is literally,
if I asked you what is the best train service
between Kitchener, Ontario, and Toronto,
you probably wouldn't know.
I don't know.
I'm from Canada, so I should know.
Oh, fun, yeah.
I took it with me.
Most people in this room, they wouldn't know.
But it's viral. So you guys are on a base?
No, no.
We're just a fan.
Well, Waterloo's cracked.
It's like a school for training kids.
So like, they have to work.
They have a co-op program.
So it's like great.
I think Jesse for Bay said it's like the backbone for Silicon Valley.
Our CEO went there and I had research, and my little sister's there right now.
So I go all the time.
It's a little weird talent.
Not the best city, but amazing school. A brilliant school.
So I think that's I think people enjoy planning but they don't actually enjoy
executing. That's exactly where you like the intent is like I want to be at these
places at this time at this cost or whatever and you know third party goes
and does it for me. I don't care if that third party is a click worker in the
Philippines or if it's like an AI agent
Did you guys do like the UX like the user flow just like using actual yeah, so I mean it's very
Autopilot like literally picture well front
For creating it's like side with the little doctor about the general go to a screen or you see
Rough categories that we've predefined and you can also create custom index so you can drag sliders to blue chips, stables, tfile, tubes, which is got a couple below,
info tips everywhere so you can get assessments going on.
Create a custom index.
Next step, sell advanced permission, limit orders, whatever, just check the box.
20 seconds, get exposure to the analysis.
Like, that is the globe.
After that, they literally list up relevant.
So you have a chart dashboard, you have a theater board page where you can copy and read people, like very app the globe. After that, they could read those sub-reminents. So you have to chart dashboard,
you have a theater board page where you can copy
and read people, like very app specific.
Everything on the tech side is like 100 different.
It's like, even if you're a developer
and you're building an app that's like totally just like,
you're outsourcing, let's say a complex multi-step
launching task, like treasure management,
producer engagement, and it's just a set of API calls
to App Store, like we handle the sourcing of a party
to do the job, verifying it, and then giving you some set of API calls to AXL. Like we handle the sourcing of a party to do the job,
verifying it, and then giving you some sort of
out-of-response as we go.
So for the travel funding app, it'll probably look like
TravelApps for somebody making sports betting
probably looks like DraftBand.
It's like whatever you want it to be.
And this is on this is like, you know,
like crypto is where we're like starting with autopilot.
But down the line, we definitely think that this is actually
like a blockchain for verifying like something
actually happening is like as Ash does, like probably the line, we definitely think that this is actually like a blockchain for verifying like something actually happening
is like as Ash says, like probably the best use case
for a blockchain.
So down the line, like, you know,
we want this to not feel like crypto at all.
We want it to feel like it's, you know, a travel site.
And it's a good, like, why do you need crypto?
Like if you ask like for any of the examples
that Jacob and I gave, it's like,
you could do this yourself.
If you trust somebody else,
you could try to build it yourself.
There are many reasons why it's bad,
but it's not flexible and like trust part is really bad. If you're paying for money for any service like
good like you won't let I'm sure you guys all saw like the OVM operator demo. It's like the
stinking book or restaurant reservation for me. I don't care if it fucks up because it's like
renovation it's not that neat. But like am I going to trust it to spend thousands on flights?
I'm going to trust it to like move my money around for any task? Absolutely not. Like anything where
you like really care about optimizing and you want guarantees
or you just like your money that say, yeah, and we've been talking a lot,
but I want to give one final example.
Like there are things that you can do right now that you probably wouldn't trust
the Internet to do for you.
Going back to that phrase active Internet, like your wedding,
you probably wouldn't cater your wedding through Uber Eats for multiple reasons.
But the biggest is like, you know, there's no counterpart.
There's no guarantee that your counterparty
will do exactly what you tell them to do.
And that's exactly where in theory,
like this idea of verification for agents matters the most.
Like you wouldn't trust it for booking your wedding
or planning your wedding food or whatever,
or like booking your one big vacation for your family.
You're spending like tens of thousands of dollars
your normal family, like it's not something
you'd like, you know, necessarily trust.
And that's what we're trying to do,
build trust for complex multi-set tasks.
Nice.
Wow. And that was perfectly on time.
Guys, that was awesome.
I really appreciate it.
And yeah, let's chat maybe afterwards,
but yeah, love to catch up more.
Yeah, check out the demo.
Yeah, we're live right now actually.
ax.al Yeah, check out the demo. Yeah, check that out. We're live right now, actually. AX.al.
Yeah, we're in touch.
Yeah, great plug.
Sweet.
Sweet.
All right, see you later, guys.
Cool. you you Hello. Good. Good.
Good.
Hello.
Hello.
Hello.
Hello.
Hello.
Hello.
Hello.
Hello.
Hello. Hello. Good, good, thanks. See you. . Catching fire. Yeah. I'm not surprised. Like when you see crypto's got talent sitting in the corner of your screen, that's a pretty
attractive thing to click on, you know?
Yeah, it's good.
It's a good thing.
It's working.
I mean, it's fun.
The main thing is that we're having fun.
36 live?
No, we're good.
Claire, we're good. Wow. He has the mic Live? No, we're good. Claire was good.
Wow.
Because he has the mic there. Yeah, we're good. I did it.
We're unmuted.
We have Claire, by the way, everybody behind the camera, and she's doing amazing work.
Are we unmuted?
No, no, I am muted.
Oh, we're unmuted. Okay, got you.
I just wanted to give a shout out to Claire.
Maybe just bring the mic a little closer.
Yeah.
Cool. Awesome. Sebastian just bring the mic a little closer. Yeah. Cool.
Awesome.
Sebastian, nice to meet you.
Nice to meet you, guys.
Tell us more.
Enlighten us.
Yeah.
What project are you representing?
What's going on?
Yeah, so about six months ago, I was
doing bankruptcy restructuring representing the creditors
of Celsius, Block 5, Voyager.
Ended up leaving that job and starting
the company Indexify. It allows people to create structured products and portfolios or indexes
without having to be a registered security advisor or registering a product as a security. So we
create a social mechanism, a social investing mechanism where people can monetize their
portfolios without actually having to invest,
and it creates a nice social economy for people to share their investments.
What's the different way of approaching it that makes it not a security, not a problem, all that? Yeah, so the way that the actual purchasing mechanism works is you're actually mirroring
the creator. So when the creator goes and creates that, what we call a stack, they list out the assets
that they want to have in the stack
and they put proportions to that.
When the purchaser actually goes and purchases
the underlying stack, they go and purchase
all the underlying assets and hold them in self custody.
So that way that we're not really touching the money
as the money flows and the creator doesn't actually
manipulate or take custody of the funds.
So there's no common enterprise.
So it's a good way for people to create,
yeah, just a social investment.
Cool.
And is it the idea that you'd have maybe some KOLs,
some great fund managers, I want to call them that,
that would get on your platform, demonstrate the portfolio,
and then others would mirror that.
And would it rebalance in real time as they rebalance,
or would it be kind of like a set it
and move on kind of strategy?
Yeah, so there's a couple different things there.
But so first, yes, there is a lot of rebalancing.
You can have a lot of rebalancing
with verified users. So there's some security on the platform that would make sure that you go
through. So if somebody bounces into a coin that you don't necessarily want to be on.
Though, to start, it will be manual and you'll receive a notification rebalance or reconstitute
into a specific index. you would click rebalance,
and so that's how we make it.
That's cool, makes a lot of sense.
Yeah, and the revenue model, you take a percentage?
Yeah, we take a percentage of those trades.
There's also some creation models,
so if a creator creates an index or a portfolio,
you can subscribe to them,
just like any token-gated application,
so you can pay $5 a month to see what they're actually doing,
their strategy.
You can also see the entire history
as they change the weightings, as they do it.
The chart will take into account all their different changes.
So you can see the total over the lifetime of the portfolio
or strategy, how it performs.
So you can actually do relative.
You can go on the platform
and see relatively how well they do compared to the market or anything else. What stage of development
are you right now? Yeah, so we're doing early signups now. It's in Dexify.finance. We're launching
a closed beta at the end of the month. So that will be based on Solana to start and then we'll go
cross-chain from there based on where the market wants us to go.
But we would like to be chain and cost stick at the end of the day where investors don't have to choose where they go chain.
And like, I don't know, in terms of like users right now, like where are you currently, you know, with the data testing and coming up?
Sure. So right now we're doing in-house testing.
But in terms of the people that are in-house testing,
we're integrating Arkham Intelligence into the platform.
So essentially, you can go on the platform
to start and search anyone that they did not.
So they have over a billion attributions.
And so, on the other hand, they have, I believe,
20,000 to 40 thousand
attributions so you can go and search anybody's name copy their portfolio
it's one of the kind of easy to use cases but yeah so you'll be able to go
on so the that's gonna be like our main testing is you're gonna have you know
20 to 40 thousand indexes stacks we call them stacks already available so it's a
good system for us.
It's the idea that I could go on there,
see someone's weightings, just go buy it myself on Radium.
But I won't do it because it's just a pain.
I'd rather just give you the fee and click the button.
Exactly, and it's actively managed.
I mean, if you're talking about something to act,
it does take a lot of time.
And those people have there's really no real great way to monetize your time expense. Like, aside from you sharing your couple chats with your buddies.
Right. And then they see it, but you might have done, like, remember that you put that in there. Right. So if you just add to a stack for free, and you just see how it performs at a time, you actually see how valuable your time is going to be spent. So the fact that it's free, the fact that it's easily shareable, the fact that people
follow you, there's really no skin in the game. You can just go do it, have fun with it. You don't
have to do that. It's like an index fund. I can see what's in the Vanguard index fund, but I'm
not going to go buy all those individuals. Is there like an algorithm for displaying
I guess certain KOLOLs portfolios.
Like, how do you decide that?
Yeah, in terms of like a trending.
Yeah.
Yeah.
So there's various ways that you can do trending.
I mean, like for us, it's like volume based based on how many people are actually following them.
So you're able to see that we've invested really heavily in human life.
So like the experience is almost like a shop field where you really don't know that you're on a cheap.
It's almost like we really want to push the social limits not necessarily like that's what we're pushing pushing
Um, so yeah, I mean just in terms of trending it's really based on like how many people actually do
Um, you also see trending users. So how do you have trendy users?
It's like how much money actually is following that person how active are they, you know, with the total return, etc. Is there like a particular metric that you've seen
is like more successful so far in terms of like users, is it just like ROI or is it just like yeah top KLLs? Yeah well we're private so we're just in internal testing at this point in time
but in terms of I, I would suspect that ROI
would be the main people.
Although if you think about it,
the one of the metrics that we do have
is average weighted market cap of the index.
You can have an ROI of something that's about 15,000
non-market cap coin, 10 of them, right?
And they can buy back, right?
But you can't really do that with any money.
So you do have to have some type of magnitude
in terms of like where, in terms of trending, how to actually do stuff.
I'm not texting just for fun. I'm texting because I hear some feedback and I just wanted to make sure it's all good.
I'm texting with family while you're like...
Yeah.
Great.
What are more mid to longer-term steps in terms of what you're like?
Yeah.
Yeah, my background is in asset-backed securities before I think I'm sure.
So, I helped write the frontier aircrafts.
Looked up the criteria.
And so our, our belief.
Is that in the next 10 years, everything.
There are no mechanisms where you can create an easily diversified passive assets.
Without having to go through so many hoops, right?
And we want to be the first in class way to securitize
or create structured products.
So that means that with stocks are brought on chain,
we are just plugging,
our users can go invest in this if they have a doubt,
just like that.
When you wanna have an agent, a financial agent, an AI
agent that manipulates your portfolio, this is the best way to do it because they actually can,
you don't have to give the agent money. They can actually, you can just mirror them. At some point,
you could send card rails against that. So it's really meant for the future of finance. It's
meant for when you bring mortgages on chain, you want to create mortgage-backed securities where you have hedge funds actually manipulating the underlying holding
of the mortgage-backed security. They can do it individually because they hold on to all the
underlying assets. So when you start thinking about structured products over the long term,
this is the primary solution for very great granular trading and over risk management for any even institutional. So our goal is to get the market, let retail have a great time with it, and then go into institutional really heavily to make it so that super easy to trade global assets in the diverse.
How's the fee work on this? So is it one time when I sign up and then I kind of like, free free after that. Do I pay each time it rebalances?
Do I pay a monthly management fee?
How does that work?
Yeah, so there's two kind of fees, right?
There's one is like active fees, that's creator fees,
and there's one like future fees.
So for the creator, we do take like a percentage
of the commission that they would have
on the creator fees.
If they do token-gated access, we would take a percent of the commission that they would have with the creator fees. If they do token-gated access,
we would take a percent of that.
You will be able to get the old off-the-ground holdings,
like via like margin-file, things like that,
because you actually have them.
You can do that yourself, just with a single click.
So we'll take a percentage.
So there's certain things that the user do
as they make money that we can also make money,
but it's really only if the user is making money.
So one of the tokens in the portfolio I get is stakeable,
I can stake it.
Yeah, absolutely.
Yeah, that's one way.
And you'll get a cut of that,
staking right? Yeah, yeah, it's pretty,
it's pretty, yeah. Sure.
And yeah, so there's no, it's all free.
Everything to start, you can go create for free.
That's like the easiest thing.
Like you can go create a portfolio.
You can go back test, like execute
an algorithmic trading strategy for free.
It has somebody follow you and put money into it.
You can make money without spending it.
So it's token gated.
If it's token gated, it's one time.
I pay once and I can forever copy your trades.
It's based on the creator.
So if the creator says one time, one time.
The creator wants it monthly, one time.
It's not based on us, we don't sell.
Cool.
That makes me like it.
Yeah, so just to break down a little bit further.
So your competition, I suppose,
is like the existing copy trading type of platforms out there. Is that true?
I think they are like a competitor. I think there's many, there's no real platforms just like this because we really want the social element.
We want the kid well to come bring, show their audiences how they perform our time. We just don't see that. That's one of the reasons for our ability if we don't see it in the market. I think you could look at the copy traders.
Yes, that is one, but they're very technical.
You have to be very technically savvy.
And honestly, they're like black box, right?
You don't really know what's happening.
Very hard to do.
And so many flies are really bad coin.
You're like, I have no control over this.
Like you have really have no control.
So those are really for experts, experts. We want kind of that middle tier of people that are interested they follow people they want to invest
You know, it takes a lot of research time to go find winners and there's people that have
Following or
So I think that's where we differentiate ourselves
is both UI experience and the fact that we're going
to market three months is that we want user feedback
as soon as soon as possible because we do think
that there is a very strong value top for consumers
and institutions that belong.
And you see KOLs all the time post their winners
but never their losers.
And that's what I like about this is kind of bringing some transparency to that.
And the social element, I kind of see a trend with a lot of companies that are kind of becoming
successful right now, like the Pelotons of the world, and you kind of have the social
element or like Auroring, you can share your sleep score with other people and kind of
see how they're doing compared to you.
I love the social element idea of this too. Yeah, I mean I would love, in fact like my brother had copy traders on his wall in 2021
and he would just put in two dollars and make like 40 grand. It would be crazy, right? And like,
you know, now it's like there was no, like everybody wanted to give us money to go manage and we're like like we're not gonna do that
We we buy 20 maybe 100 coins a day and 99 go to zero
Like I don't want you to really call it, but we know that there's like five or ten
We're like we were like, okay, like yeah, like we had a couple bucks here, but now we can go do that
They can choose whether or not they want to invest us us. And like, we want people to share their portfolio
with their family and friends.
Somebody asked us like, why would somebody
with a really good portfolio performance use us?
And it's like, well, I mean, if they have a family,
they can just, they can give it to their family,
their family can invest in them without them having to
touch their money.
It's like an easy way for people to get onboarded.
They can go test it out, to go add their favorite dog coins,
or if they're more technically savvy and utility coins.
But yeah, we're here for long term.
We're here for stocks, real world assets more than this.
This is like just the start of like what a platform can do when you start diversifying
for this.
So right off the bat, it's just crypto, right?
Yeah, digitally native assets.
And I think we'll end up until we will get our licenses. That's what we end on doing where we can operate the most autopilot for stocks, where you can copy and it's closer to the next, which is great.
That is something that we want to go, but we really are about like the next generation of finance. That's like tokenized rent rolls
Fractalizing real estate through that's like that's what we really want to granular risk management
But you really can't do like interoperability is impossible
So that's where we want and all that's contingent on
tokenization like institutions.
I wouldn't be in crypto if I didn't take it from institutions.
You think this will be mostly for large cap projects?
Or I mean, if it's a lot of small cap meme coins,
the copy traders could be the market
where it's like, you don't token you you don't token get your strategy because it's
If people are copying you they could because the low supply could send a token and everybody could be in on it
That seems a kind of the frenetic
Crazy, maybe less sustainable version. Although maybe not. I don't know. How do you see it? Yeah, I think I think it's it's
There's different there's different,
there's very different of, you know,
when you have a magnitude of triggers come in.
At the end of the day, it comes to like security
for our users of, you know, setting alerts,
setting thresholds where if something is beyond a risk,
like you set your own risk, right?
If the slippage is gonna be 10% on auto,
and that's your limit, then it just doesn't go through.
Like you have to manually do it.
Now, if you're an expert, right?
And you know that, you know, this guy that I know really well,
he's really good at buying, generating coins.
We know that there's 10 dog coins named X, Y, and Z
coming out in the next 20 minutes,
one of them is gonna have a hundred bucks. Put them all in a basket, put 10 bucks into it. Try buying each one individually and tracking that,
tracing it, trying to figure out which one's going to win. Don't just do it all at one time.
But when people move into these smaller cap coins, the risk will be like the user will
set their risk limits. You won't be auto rebalanced from building too.
And so, you know, and then you can see the history
of the performance.
So all of those different things will limit.
Yeah, I would suspect they want to.
They're probably, if you can see that somebody's
one of the most followed accounts,
and you know, like you just bought something
with very low supply, everyone's gonna click yes.
And I just have to first click yes.
Is there any use of like intents or solvers
on the backend?
Yeah, well, right now we're using
just traditional DEX operators, I think over the long term.
Like RQ systems is like absolutely solvers.
That's actually, that's absolutely where we wanna go.
We wanna first really hammer down
our social investing platform, grow our user base because
I feel like that's when you have that type of user base really like creating new things like we are
we want to be kind of that base later of creators and like people that invest and then building our
few systems and do like that like that's very you know that's pretty normal. People have done it. We can do it. We can create our own Dex aggregator.
It's kind of fun.
It's good to come out of this one.
I wish we could say more as the other ones.
But yeah, we're on time.
Sebastian, thank you.
It was awesome.
Yeah.
Yeah, thank you.
See you later.
Or outside.
Yeah, it was awesome.
Thanks for coming, man. see you after yeah go yeah
Jump in you might grabbing the next project as well
Quite a diverse range of projects a yeah, I like it's like everything is like very niche
Like everybody's doing like something quite specific hey
Absolutely hello welcome how's it going? Like everybody's doing something quite specific. Hey. Absolutely.
Hello, welcome.
How's it going?
Nice to meet you as well.
I'm Jen Pong.
Nice, you can put the mic closer to you if you want.
Nice, what's Jen Pong?
Actually started from my personal backstory
out of care loss aging myself at the age of 21.
This is where I come through to it
because when we went to a parallel,
we think of an aging issue, a male issue,
instead of like the 20 year old girl's issue.
But when I looked into this,
I actually realized that I was on the same path market,
so we started building this data set with POSTCOR,
trying to train the political data
for that AI I am. However,
you realize if you just make the hospital divide the department into like a portable version of that,
then everybody actually just have access to this expert.
I'm going to let's have a good building. to the health expert ambulance. That's how GenPOP community works. We're building this decentralized AI deepening
for dermatology, and then everybody can have an AI expert
from PNP at any point.
That's the first time I hear of something like this.
So cool.
Let's dig in.
How, why deepen?
How does the deepen model work?
Like, just tell us more.
Sure. Yeah, when we were training with hospital, it actually took us two years to collect 35 images.
Usually these doctors are very busy. At the same time, we're doing it as a
passion project and they're not paying enough for us to have the motivation. However, when we launched
the app, we got 400,000 organic research we've been doing today.
So that's what proves that actually the general public, we want this product and they're going to expand the data set instead of working with centralized institutes.
So that's why we're moving this deep into the hardware, integrating the fabric and all the new training so that consumers at one point can have the animal that's without hair,
like follicle density, radius, hair density, radius of 13 minutes, of 20 minutes,
and actually get rewarded and incentivized for their business.
Oh wow, okay. I mean Dylan, you're a deep analysis. Would you maybe, you know,
do this in your spare time? Would you analyze? Oh, wow. Let me be short for camera here.
The flagship product is Lush here, focusing on the hair loss analysis.
We ran a lot of testing with Hairdell and we also scanned the tail and Aubrey, the great,
who's also our GSI hair situation.
How frequently do you scan error as a user?
That's a very good question.
So actually when we think of like the error, the only FDF for the solution is the knockout
of the master.
That's usually 6 months to 18 months of treatment.
And it's a long term process.
To actually keep track of that, currently our users scan more than two times a month
to keep track of their score for growth.
Also, as I said, now we have more than 2,000 users from 15 different countries.
So we're actually also getting data into the age, racial, and gender groups.
That actually makes our data way more valuable than the data we work together with, for example,
hospitals to get a better percentage of more unmarginated racial groups and their data input. then the data we work together with a physical hospital, because we know we're going to do more
and like, you know, the visual groups
and the data input.
So how many people have that device
or are scanning regularly?
So for having a device now,
we have approximately 1,200 people.
And then more than, we actually have like a selfie scan
as well, because that's a good picture of your hairline
and you know about the state and the hairline.
And that's on the app you have,
so you want to find out.
Okay, wow.
And what, so I know this generates a lot of data
that you use for your trainings.
What data are you getting from the selfie?
Is that sufficient or is that really just a way
to upsell somebody for that?
Yeah, so the selfie is more like,
so it's like a premium model.
The selfie is like an easy and great feature while the scan is more professional. sell somebody to that. However, later on we're actually also going into like skin, like for acne,
acne, acne, wrinkles, or like pigments, and the selfie picture can be also valuable.
Can you move the mic closer?
Sure.
And then to also example the interface real quick.
After getting into our app, connecting the hardware through Wi-Fi, and then we can actually, I'll just touch my own hair.
So we can actually see. Oh, I thought you were, not that's what I love.
You should touch this guy. Look at this. You want to try? I'm scared.
I was going to tell you you're losing all your hair. Exactly. It's like, oh, it's a wig.
It's a wig. Yeah. Okay, tell me anything it doesn't doesn mean it doesn't change the fact that it looks great right now
Yeah, I'm not sure if I notice audience notice but just a big on a haircut deep in Jesus I know I'm yeah
Bonus points for anybody that comes up with a new name for me
So long enough to qualify.
But it's kind of fun when people come up with names.
Oh, it's good.
You get different parts of it.
Good.
I take 77.
It's a good number.
So what is 100?
People get 100, people get 0.
What does that mean?
So it's 77 out of 100.
But also, we're measuring 14 different higher reasons,
including like UCAN for the toes,
cranial size, alunas and fertility and the scope. The goal is to give like a holistic report
so you can understand about more about hair because hair is not just a static problem
and you look great already but at the same time time has also like wellness issues.
If your head tells you about like if you're getting like other
under detection or like some people like for me, I was
having like, a lot of that now is the first like stress lack
of sleep and maybe also that diet or not enough fat movement.
That's also something you can switch out. And ultimately, we
can even create like a protocol for each individual to train their own AI agents.
Each people have a different experience,
or even when you live in different countries,
like say, Singapore is very humid, very hot.
Well, I'm in New York right now, it's pretty cold to me,
and also less humid, so my so my hair can also react.
And that told you that you had a certain deficiency
with nutrients and actually come up with that recommendation?
So actually that can actually run like a projection
because that's the issue, if you have certain hair loss issues
you might need more Fountain or like vitamin D, vitamin E
while if you have like flakiness,
usually you can be like using a run shampoo
and you can also be testing out different stuff
on the QT if that helps with your treatment.
What was your best attribute?
So the pick for different juices is, so you can use some powerful anti-drying spray which
contains like argan oil, coconut oil, shea butter, or if you can use like something to
make it like more shiny and that contains a lot of things.
So it's giving you a recommendation.
Yeah.
Got you.
What is the characteristic of his hair
that is most like closest to perfect?
The best.
Like he's got exceptional thickness or something.
So most of these metrics is all like
in the healthy threshold.
But like actually in the analysis report,
you can see that this yellow block is where the healthy one should
be and where your app.
So your app's here, and then this
is all automatically annotated.
And you can actually, I can actually
do this with the hair around.
Wow.
This is how our hair looks like.
It's getting annotated.
So are users rewarded with tokens for uploading?
Yeah, we're going to do it step by step.
Firstly, it's like a reward for added.
So now we can actually give them discounts on the product for change.
The second step would be like cashback,
because actually we offer that part of here to users in like
KSYS Asia or even like in China,
so that a person can be physically
for them to trade or like,
yes, yes, yes,
it will be much easier way.
It's just a little more.
So that would be actually the
and then actually also lead to
the other aspects of the business model,
which is like the data we have
been generated from the YouTube
is also very valuable for
pharmacy company, and often companies that for reference, which is data we have generated from the YouTube is also very valuable for pharmacy companies.
Let's go reference the last time FDA approved an empty hair loss
finasteride and that was already 20 years ago. So a lot of companies are running analysis towards
the hair loss issue or customer insights on the hair the issue. So we can actually provide API access
for like institutes to that.
I'm gonna send to your search wave
or just run customer container in my office
and the program that is based on that of the institute.
And so you're making revenue in a few places though.
Are you making revenue on the products you're selling there?
Yes.
Obviously this, but I mean the,
like the recommended products.
Oh, so this is more like a creation for now, Yes, I obviously this but I mean the like the recommended products and maybe connect Puto Wallet to the account so that it's such a PR director actually can deliver to the adult.
You can also look at the mic just a little.
Yeah, so it's closer to your mouth.
Because sometimes people are having a hard time hearing you.
Yeah, I also, when I see this, think a little bit
about Whoop as well and all the different parameters
that you're putting in as a user and saying,
do you sleep with your cat in your room?
If so, that's like a recommendation. You probably
shouldn't do that, right? Do you have any of that kind of input? Or is that
a agent thing you're talking about? Like, again, putting in your very good
question. So we're building this hair quiz thing when you're logging part of
your onboarding. So that's when you like putting like your
History like if your parents are having her off or like how many alcohol have you been drinking?
That's the question in the author's section. Are you a smoker?
That's it. We're smokers or like audience data input as well as we keep track of your routine
So we built this feature where like today
Your scale like 70 something.
70 something.
Yeah, you can actually remember.
Can you change that into your Twitter for help?
Maybe use a 70.
And then we will like, let's say if you're using a shoulder shank rule, then the, it's, you can measure that.
Oh, this is a product that's working
because yesterday I was 66.
And then when I'm using the other product,
I was 77 yesterday, 66 today.
So that's also like multi data input
to make the more make their own agent or our own self more efficient.
Yeah, it's gonna make the data sets deeper.
Yeah, more fast.
Exactly. How much development have you done on the crypto deep inside so far. I would say work more in the beginning stage because we actually started the project
with the last two products and already generated. However, a little bit
the fact that actually the first thing about 200 has to start from So, we started looking into the innovation.
and some of the data to like the interventions like
So far we have been deploying our
virtual protocol for all of our things.
So, we have some interactions together with how we make the data safer, So data is anonymized or going to be anonymous? Yeah, because these are not biometric data.
So the hair analysis report is actually the more valuable data for our outcome.
However, that's another way we're looking into, which is the user themselves can make their own,
together with your identity data and their data and your product you're using with the company or a church or a beauty products.
Okay. So that would be like on a one-on-one sale. So you'd be just selling the NFT, but you're also making revenue, right? By selling everyone's data. So that's where the scan to earn model can be displayed, where they can earn some central
financials.
What's the highest dare score you've seen?
That's actually a very good question.
What's the lowest? Is it 77?
I remember that day. I was at the J4U University SF and there was an India founder, he got like 88 or 89, that was the highest score.
I'm not that far, band and he was 46.
But I think in my defense, it will be because it's like
very sunny day and then he pair with like people on
video streaming.
So that's not-
It was still 46.
And ironic part that he's sister is a somatologist.
A lot of great crypto patterns with bald.
So is bald zero or does it take account of the remaining hair?
Actually, when you talk about bald, balding has a program.
So there are several different states according to the knowledge categories internationally
recognizing the hair loss issue can actually
have seven different stage.
So when you're bald, usually you always feel like this only,
the top part affects baby one.
But when you're like stage seven,
it's pretty much you're losing all your hair.
So that's also why starting early is important
for tackling your hair loss issue,
because for example, hair transplant
is going to stick around from straight
to a white profile procedure to like,
dogmas, but you still need to have enough quality.
Because the process is like moving your hair
from your back to the front,
whereas there is more hair coming from each particle
in the back.
So if you are like stage seven,
like you're confident you can do your hair transplant completely, but when you're like stage 7, you can't do it completely.
But when you're like stage 3 and going slowly into stage 4, you can still have a lot of
different ways to solve your problems.
Wow.
So we've talked a lot about hair, but you mentioned you're planning to do something
with skin.
And is there anything else you're planning to expand into?
Yeah, definitely.
And what devices would you use for that?
Yeah, for us, our goal is to build the total state of wellness, AI agent for everyone,
and especially for farm health.
Because when we started the research, one thing we realized is that most of the demontology
code database, they are trained on male data, and especially Caucasian male data.
So that will lead to a lot of deficiency
when it comes to something like disease analysis
or remote treatment.
So when we are actually providing this whole like
DPAI, DSI solution to like hair loss issue, acne,
eczema issue or later on early stage skin cancer or even we can do like
teeth checking, eye checking, gynecological checking at home and you can
have like the whole process affordable, accessible
and actually be rewarded from the same input.
Awesome. I guess Fiona, thank you. That was great.
Thank you so much. Yeah, I really enjoyed it.
Wow, it was a very vulnerable moment.
Waiting for that score, you know.
You should have one and check every...
Oh, I don't get to keep that? Yeah, pass the Indian founder and bring the Italian flag. Hi. The Italian guy will be the highest and lowest score.
You can send in the last one. Yeah, just Lily Pod.
Again, so diverse. The range of...
Every time we do an episode, I just want to get everybody... They we do an episode I just want to get everybody.
They're all great.
I just want to get everybody to win.
Yeah.
All five of them into one.
Hello.
Hello.
Drone, hair inspection, copy trading.
That informs your buying strategy.
Hey, how's it going?
Pretty good. Nice to meet you. That informs your buying strategy. Hey, how's it going?
Pretty good, having a great day.
Talking to all the other content pitches.
Yeah, it's been fun.
What's your name?
My name's Jay.
Jay, ah, Jay, great, great.
Nice to meet you.
Nice to meet you as well.
From Lilypad, right?
Yep, from Lilypad Network.
What's Lilypad?
Sure, so Lilypad is a network in order for you
to deploy AI models and bringing a mic
a little closer so everybody can hear you better.
Sounds good. So yeah, it's a network for you to deploy AI models and then monetize them
and make them accessible to other people. So if you're trying to interact with LilyPad,
you can interact with our CLI or our API. And what you can do is interact with open
source models. So if you've heard of like Beepseq or Mistral
or any one of those models,
if you were the creator of it,
you can just bring it off the network, deploy it,
and it becomes a module that other people can use.
So if I'm, let's say building an application
and I want an LLM within my application,
instead of going to, let's say, OpenAI
where I'm paying for their API service,
I'm limited by the credits, the rate limits,
maybe opening my own GPU warehouse.
It's a little bit out of my reach.
I have to worry about the electricity, the security,
the GPUs themselves, which are millions of dollars
to fundraise and actually deploy.
So instead of worrying about all of that,
we run the models on our distributed compute network.
So all you have to do is you request a job,
you want to interact with the LLM,
send the job over to our network. It pairs you with to do is you request a job, you want to interact with the LLM, send the job over to our network, it kind of pairs you with the best compute that's available,
it processes that job and sends it back. So whenever you're calling for one of these requests,
you're paying with a Lillipad token, and then that token is getting split up between the person
providing the GPU for you, the person who created the model, and the network itself.
So the problem that we're seeing is that a lot of the people
who are researchers deploying these models,
they're kind of, they're just PhDs,
they don't have the funds to do this.
And if you look at their research, it's actually drawing up.
A lot of the grants that they use for their research
is drawing up the GPUs they don't have access to.
So they're coming to us and we found that
that's our like biggest user base.
It's a bunch of scientists working on these AI models.
So one of the first applications we've seen built on it was from Stanford. We
have a machine learning scientist there that's used our protocol in order to
deploy AI oncologists. So the oncologist is trained on a bunch of proprietary
scientific research data, some stuff they want to host locally. They can still
interact with the LLM, still ask it questions, and a scientist can use that
as its advisor.
So we deployed it. You can look on our YouTube. We have a video with scientists that actually work on that.
We're using it for protein research. We have individual developers creating their own applications for,
let's say, like a VS code. We seal a VS code extension that interacts with an LLM.
We saw a Google Chrome extension come out that analyzes the pages that you're reading.
My hope is that one day they just basically take that content
and then read it out for people with disabilities.
So there's tons of different applications.
If you're someone who is a model creator
and you're not sure how to monetize your models, which
we've seen with Stable Fusion, with some of these other LLMs
that have sprung up.
So we allow you to create them, deploy them,
and monetize on our network,
using the Lillipad, essentially Lillipad.
And is the person using Lillipad,
do I choose which model I'm using,
or is it kind of just randomly selected?
If I say any text generated, is it?
Yeah, so as part of whether you're interacting
with the CLI or the API,
that's one of the things that you would have to,
is the model that you have to interact.
Is which model you want to use, got it.
And so there could be multiple different GPUs
that offer these same models.
And it just kind of goes out to like an auction system or?
Yeah, so the GPUs themselves will kind of
choose which model they want to support for various reasons,
like geographic location, censoring,
maybe they don't want to deal with certain businesses.
So their GPUs, they can decide which model they want to support.
Then when a job comes over the network,
it sees which GPUs are supporting this, and then it goes.
It's one randomly or does it pick one randomly or just go in order?
I'm not clear on the exact way it works.
Just for more context on the DevRel side of the business,
I just have to live in New York City.
How big is the team?
Right now we're at about 20, 20.
Yeah, very sweet.
And it sounds like you've got users too, right?
Yeah, so we have scientists creating modules.
We have people deploying applications
and then we have people interacting
with their applications as well.
We see thousands of jobs go through our network daily.
So, so far so good.
What's the roadmap like?
It's the first time I hear the little bit,
so I'm surprised, but at the same time,
yeah, what's it like moving forward?
Yeah, so we launched in a couple of different phases.
We've been working on this for about a year and a half,
the team.
I've been joining more recently.
But the first phase was getting the GPU providers onboarded.
We've already done that.
We're done with that phase.
I'm getting modules onboarded.
Like you see all these AI models
that the public wants access to
and doesn't have access to.
We've done that side of the business.
Now we're going more towards DApp developers, creators,
and we're actually participating in the Filecoin
that's going on right now.
So there's, we held a workshop earlier today
to teach people how to launch DApps and stuff.
So right now our main focus is builders for specific DApps websites of that nature.
But we're also forming partnerships.
We do have partnerships with Dell.
AWS is looking into us.
We also recently got accepted to Fortify, which is an accelerator on Plataeos.
So we're also continuing to go around partnerships on the infrastructure side
and also the model creation side.
So you guys are a little bit farther along than others we've spoken with today. also continuing to grow our partnerships on infrastructure side and also the model creative side.
So you guys are a little bit farther along than others we've spoken with today. So your poke series aid.
This is our second race we're going for. I'm not exactly sure what the class sounds.
Fine. It's, yeah, that's great. Very cool. Is there a token launch coming soon?
So we do have a token launch that's coming soon.
We're planning to launch this quarter.
So we do, if you're a builder and you want to start getting involved,
we have a point system for bounties for particular modules that people are
interested in running on the network that are in the area.
So we have a bounty system for modules. We have some community rewards.
And we just kind of want to participate in spreading the network.
So we have a couple of different point systems that you can participate as we wrap up.
So it's our TG.
You integrate any other GPU deepens?
As of right now, no, we do not.
We are a fairly collaborative group though.
So if there's someone out there that's interested,
I don't see why we can't potentially work with them.
And which system were you eyeing for DG?
As of right now, our test net is on Arbitrum.
So that's where we're testing.
We're having jobs run.
As far as DTG, I can't speak to that.
Is there going to be more on chain than just the token?
So you can process the jobs.
We have been talking about having some of the models
go through ZK proofs and kind of proving
that you're using a particular model, some stuff like that.
So some of the model things are going on.
I would say the primary use case is basically
requesting the jobs, getting that job sent back.
That's usually what the token queues for. So it's more of an utility token.
Got it. And I guess each node would have their own wallet or identity on chain,
something like that. That's what they're... Yeah.
So they do have a wallet. As right now, during the test net, you basically
providing your wallet, you're paying for the transactions, which is having the
provider check themselves. So they do have a wallet, you're paying for the transaction, which is having the provider to do it themselves.
So they do have a wallet,
but basically whenever you pay for a transaction on-chain,
it would be sent to a smart contract
that would split up the funds.
Yep.
And if your company went down, disappeared,
would I still be able to spend those tokens to,
if that node was still online, to get data from it,
or does it kind of all route through some type of closed system?
Yeah, that's a good question.
All of our transactions would be on-chain.
You would be able to request on-chain.
They're the centralized network that's using our stuff.
If we go down, they should still be up and running.
I would have to consult engineers.
That would be an impressive thing
for any deep in this early.
A lot of them have that aspiration,
demo included, but it's to get to that point.
It's in some people's roadmap, it's not in others.
Interesting design choice.
Good like AI agents use your modules.
So it's a good question
and it depends how you define an AI agent.
A lot of things that people are calling AI agents
are essentially wrappers with a prompt that's just there,
which I've done on the network.
Like I built in that last night.
I actually was testing it with a group
that you just paste in your wallet,
and then it basically roams through all this stuff like that.
So you could deploy agents, bring agents
to market on the platform. We are talking
about working with like, sorry, we are talking about working with like Eliza agents, some of
those other agent frameworks that exist in the process, they may be, you know, compute through us.
So that is on our roadmap. But if you wanted to build an agent from scratch, there's nothing
There's nothing stopping you from doing that today.
stopping you from doing that today. If you don't want to use any of the frameworks.
If you don't want to use any of the frameworks,
you are free to go.
No, like could an agent like use like your modules
to develop something themselves?
I mean, if there's,
it would be a good question to see what they would develop,
but with the inner API, the agent fishes call the API,
get a request back in the process,
whatever they need to do there.
So that's one route to do it.
Maybe the agent can process like generate video well or something like that.
Well, that would be a great piece of case.
Yeah.
Yeah.
So we can have something like a video generation model that's
available on Lillipet.
If your local LLM can't do that, let's say you're using
a call out or something like that.
And yeah, you could just basically interact with the API.
We would have to kind of let your LLM know what our inputs are,
what are the outputs, because they also have to process the outputs
and what that looks like, what are the outputs, because they also have to process the outputs and what that looks like, workplace, and things like that, especially when it's a difference
between images and text.
But yeah, there's no reason why, yeah.
And if you're listening to this, I would love for someone to deploy an AI agent on it, like
you might as well.
Interesting.
So, Rob, in your idea of like putting all of today's projects together as
one aggregate project, I would really pass it.
Yeah, you missed it. So, earlier we talked with folks doing drone detection, hair scanning,
AI agents that can, AI agent like auction market, which probably that one is the most
directly relevant. So, these guys, there's would, you would prompt like what you want. I want tickets
to on a train to go here. I want to arrive by this time and it would kind of put it out
there and whoever can get it for you at the best price, best time would win and execute
that on your behalf.
And I say trading sort of like, yes. yeah, and then the copy trading for
B5 which was also pretty interesting
Struggling to connect
Abias in terms of being able to
Manage a foot for execute real actions and manage the resources to do it
But how do we make that run on a drill on the werewolf testing your hair? Yeah, it's just difficult
I guess you would you would
prompt the
Agent to collect hair samples that would deploy a drone to do it it would get the data and
Run it through your
Models to inform a trading strategy
Strategy
And seeing like put traffic within the store and trying to predict like what is there
You know quarterly report going to look like?
Yeah, sure.
Depending on foot traffic and like car traffic.
So in theory, you could have something like that that has an AI model that's analyzing
classified things that are passing by.
Is this person how much foot traffic?
And that could in theory lead to some sort of thesis on how to invest in that company
depending on how much traffic was getting in that area.
So that is a real use case that could be-
Yeah, that's a useful deciding link
if you're gonna put a Whole Foods
in a neighborhood or something like that.
They wanna track these types of things.
Yeah, exactly.
Market valuation based on the average hair score.
Of the team.
Yeah.
It's nice.
And hairs that come into the store.
Nice.
Any other questions for you guys or from you guys
for the event?
What should we look forward to maybe next thing in like,
I don't know, maybe a month?
I mean, you have TG this quarter, right?
Yep.
So I mean, that's exciting to look forward to.
But I'm more excited about the tech side,
especially since I work in DevRel.
So for me, I'm most excited to see what people do, because they're already doing things we
didn't anticipate.
Like the VS Code plugin, that was just mind blowing.
I was like, that's the whole business.
That's Cursor.
That's exactly what they're doing.
This guy did it overnight.
And then we started talking about Google Chrome extension, because I was like, hey, all these
companies that are talking to AI, they're like, can I get an AI chatbot
on my website?
Why not just make a Google Chrome extension that
reads the page, and now you have an AI chatbot
for every page you visit?
We don't need to isolate to one thing.
So these kind of ideas where people
are breaking the molds that they're
used to coming up with different applications,
that's what I'm talking about.
Oh, nice.
Cool.
Can't wait to see them live.
Awesome.
That extension is cool. What is it called?
It's not deployed yet. It's not live.
We're just testing it.
Does it store each page?
You can ask a question about what was
that article I was reading about this.
I'll give you around this.
That's a good distinction because we're doing one-shot inferences.
Right now, it's not holding
a long-term memory of what it's like.
It's more so about this is the task you need to complete,
just complete that task and that's it.
Well, it's on the developer too.
The developer could use Filecoin
for the hackathon you're doing.
They could store some of that stuff there
and then submit that data along with the,
there's probably some TKTLS angle to this too.
Yeah, so one of our teammates
actually did something like that with Filecoin.
And for more context on our team, our CEO was formerly the head of DevRel for Protocol Labs,
so Filecoin, all of us are familiar with them. Our other co-founder for the science side of things
is Science Stanley. He's a machine learning scientist at Stanford School of Medicine. So
that's how we got it to the B-size. Nice. And all those customers that you mentioned, are any of them paying you at the moment?
Well, right now we're on SassNet, so they would be basically paying per inference, and
then we would take a cut of that inference since we're in SassNet.
Right now we're not generating those revenues, we're just making sure that the economics
of the network works and that the actual function works.
Gotcha.
Sweet. That was very cool.
Thank you, Jay.
I appreciate it.
That was great.
Thank you all for inviting me.
That was awesome.
Of course.
See you.
Yeah, we'll have a chat after here.
Yeah.
Hi, guys.
Guys, that was something.
Yeah, so fun.
Wow.
So fun.
Damn, everybody should win.
No, seriously.
They're all great.
Really were.
Yeah. Cool. Well, I guess, Dylan No, seriously. They're all great. Really were. Yeah, cool. Well,
I guess, you know, Dylan, you know, Rob, thank you. That was actually, I enjoyed it. I enjoyed it. Yeah,
those were thoughtful founders. You're pretty good at this, man. You should join us. My own TV show.
Yeah. Night crypto TV show. I would watch it. Yeah, so this is episode 5 for us. Yeah exactly second IRL and we hope to do a lot more of these
Maybe connect some different conferences and just post host get a podcast studio again. Yeah, and
Yeah, and we're doing the Twitter spaces basically every week at this point. So yeah again, we'd love to have you guys
Yeah, you guys can jump on you're approved
Yeah, you fast
Yeah, exactly. I guess we're doing one more episode for
Like round one. We have a couple more projects. Yeah, and then we're gonna start with round two
So we got to start deciding about some of these projects
Yeah, so round two is gonna be a live stream as well
And we're gonna allow the projects to show some slides
Take us through any other material.
Whereas we've only done audio, except for right here.
I'm really scared because I want to choose everybody.
No, seriously.
They're amazing.
They're all amazing.
Sweet.
All right.
Well, yeah.
Thanks, everybody.
Thanks, Claire, from backstage.
Awesome job.
See you all next time.
Good job.
