Hello everybody, how's everyone doing today?
Doing great, doing great, how are you?
Doing good, happy to have you on the space here with us.
I see that Tom, who'll be moderating the space, he was just kicked out from the stage.
This happens on Twitter as we know it, so we're going to give him a few more minutes
Oh, I think he's back now.
Tom, can you hear us okay?
Yep, I got you, hello, hello.
Hello everyone, welcome to another XSpaces.
I'm thrilled to have everybody here to talk about the future for AI DApps, super exciting.
I think we're waiting for a few more speakers to join, but we're almost ready to get started,
Let's see here, got a good morning today, I'm excited about this, this is a topic I don't
know much about, and so I'm super excited to learn, I think everyone's focused on the
LLMs and kind of the obvious big applications, but the AI DApps is obviously the next generation
of everything, so super excited to dig into that more.
Let's see here, who else do we have?
Well and just as people are joining, just before we kick it off and introduce and how about speakers
introduce themselves, just a second, I mean some of you guys obviously, poor Julian's
got to hear this for the fifth time, but Fluence is a cloudless computing platform, so that
means we are basically decentralized serverless computing, and we think the reason why we care
about AI is we think that there is a real potential for decentralized AI to be really
important, and we've had some spaces in the past, we've talked about that, and this
space isn't really about decentralized AI, it may bleed into some of the things depending
on what the use cases for some of these DApps are, but we think that even if it's not
decentralized, even having provable data and had models that were trained provably on
certain data sets will be very important even in centralized AI, and then also being
able to run models provably and audibly will also be important, so that's kind of
our interest in the AI topic, and so maybe some of that will come out with some of these
applications we're going to talk about, but with that I figured we might as well get started,
we're 5.05, so don't want to keep the audience waiting too long, so with that, who, any
of the speakers ready to kick it off first, who's interested to introduce themselves and
kind of talk about their project a little bit?
Yeah, we can go ahead and kick it off, my name is Travis, I do BD and communications over at
Vayner, we are a new L1, we just went into testnet last week, and a new L1 for the billions
in entertainment, we're doing a lot with gaming, with AI, and also with traditional
blockchain systems such as RWA, DeFi, and things of the sort, so this will be a really
interesting conversation, we're excited for it, I saw you guys have a Galaxy campaign going,
which is awesome, I just got joined on that, we've got one as well, so thanks so much for inviting
us. Great, well I mean just to understand a little bit more, how does your L1 intersect
with AI and with DApps, you mentioned some, you know, obviously some vertical use cases
for an L1, but can you dig into a little bit how that's specifically relevant to AI and to
some of these applications? Yeah, so we're working with, and this will be announced
more closer to GDC and at the event there, but we're working with an AI company for
some AI integrations in different DApps and also in some releases that we're doing that will
integrate AI in some fun new cool ways, different things such as some systems kind of stop at the
endless scrolling through social media and stuff and kind of condense your stuff, giving you some
AI personal assistant type programs and platforms to help consolidate your news, your social feeds,
events and relative things that you're interested in so that you don't have to do the, you know,
hour long of scrolling through Twitter or Facebook or Instagram or things of those sorts,
and also doing some things with AI as far as digital assistant, or I'm sorry, digital twins,
so people could upload their kind of digital identity, digital twin that can be put into
different games, platforms and things of the sort, and it can be used for companies and
businesses as well as like a virtual customer, customer service associate, people have questions,
people want to find something, have, you know, questions about a product, the LLM and AI
systems can help assist you in those kind of fashion fashions. And so is it fair to say that
this is your you and the applications you're working with are using AI to enhance and streamline
kind of existing type of applications to make them more relevant and more user friendly?
Yes, yes, that's a real good way to put it on, you know, that angle that we're going with,
and it's, you know, kind of, it's going to do a real great job at kind of
pushing out the noise in a way, I guess you could say, and helping you just receive that kind of
info or that feed that you're really relating to or really interested in at the moment.
We know that subjects and, you know, kind of brain past shift, where today I may be real
interested in the, you know, NHL and what's going on with that and everything else where next week
I may be more interested in the NCAA basketball tournament and want to see more towards that.
So it does a lot with kind of consolidating those interests and helping put those more
forefront for you, as well as with kind of the digital assistance where, you know,
if I'm in a new area, say I've traveled to San Francisco and, you know, I'm into restaurants or
I'm into gaming events or sporting events or whatever else that may be, it can, you know,
give me different kind of activities or things that are going on relevant to that in that area.
And then with those digital twins, someone like a, you know, celebrity or maybe a large head of
the company that AI can scrape data from everything that they're doing, their mannerisms,
how, you know, stuff they're talking about, things to that to help convey that message to people.
Great, great, great. Fair enough. Now I got a good hand of it. So we've got a whole bunch
of other speakers and sort of on topics similar to this. So who's ready to go next?
I'll go. Can you hear me? Yep. Gotcha. Yeah. My name is Cody. I'm from layer one X. I'm the
chief experience officer. If you're unfamiliar with what layer one X is, we're a next gen
layer one that is fully decentralized looking to transform the web three experience. And we're
doing this through our flagship proprietary technology that focuses on bridge list interoperability.
Yes, we understand that that term is very cliche in this industry. And so we've rebranded it to
X talk because X talk solves the need for using risky bridges. Likewise, X talk not only allows
you to move assets, but also data and logic freely across EVM and non EVM chains, like I said,
bridges. Our ethos pretty much is to build a united web three where all blockchains and projects
come together and the flow of data is pretty much effortlessly occurs that way, you know,
almost like currency, right? And, you know, we're today we're going to be doing for data what
Bitcoin did for payments. And so love the AI talk, excited to kind of jump in on some discussions
with it. And so thanks for having us on the day. Fantastic. Thank you, Cody. Who's up next?
Yeah, great stuff. I don't want to go next. So I yeah, I'm Karen, I'm the head of community
for Madman first. Madman first launched in 2021. And yes, since then, we've, we've built a really
good community together and, you know, put a lot of time and a lot of money and resources
into building the first platform and also the first game. Yeah, for Madman first,
that's been fully tested now as well by by our community. And, and so yeah, we haven't actually
gone ahead and launched that yet. Basically, this was all built just before the year the last bear
sort of set in. So yeah, there was a big change in the meta there for web free gaming. And so yeah,
we've pulled back and we're now working on developing another game. And yeah, we're using
AI to develop that game as well. So yeah, sort of interesting journey from your 2021 to where we
are now. So yeah, really looking forward to getting into this and yeah, really interesting
topic today. So yeah, cheers for having me. Great, Karen. One question for you. You said you're using
AI to develop the game. Is AI helping you build the game or is it helping are you integrating AI
into the game? No, we use it that's helping us to build it. So yeah, we have, yeah, really
interesting. You know, actually, the, the one of the biggest changes that we've seen, especially in
gaming is just how much cheaper it is now to actually build games. That shift has happened
almost instantly. So it's actually kind of scary. But yeah, yeah, looking forward to getting
more into that, you know, and discussing that a little bit further. Great. Thank you, Karen.
Who's next? I'll go next. Hi, everyone. My name is Alejandro. I'm the Senior Community Manager
at VicRit. We are a data science platform. It's basically web two, we've been around since
2017, made up of a whole bunch of data scientists, both of the founders are Japanese company was
founded in Tokyo. And we have decided to finally go web three, because we see the need to have our
community of 35,000 data scientists strong, we see the need to have a token that we're going
to be launching soon at the end of the month for an IDO, to basically allow us to move our
data science competitions forward, outside of tri-fi outside of wire transfer payments when,
let's say bounties are completed. And later on this year as well. And we're also going to be
launching our AI marketplace, which is essentially the decentralized version of what OpenAI is doing.
And so we are not building any dApps in particular, we will, I mean, of course,
be doing through competitions, but we have the data scientists to facilitate and work with partner
web three AI companies to host these competitions and use our collective brainpower in our community
to build these dApps. So, you know, see is that sort of a labor resource pool and also as a
marketplace. And they're really excited to be here to talk and hear from everyone else. This
seems to be a very good space. Well, listen, super interested by that. How can you spend a
second on the decentralized OpenAI? That is a big thing to bite off. Yeah, it really is. So
right now, like with our current, the current competitions, it's like I said, it's built in a
very web two style that's easy for data scientists to engage in. Like we have ongoing competitions,
we have ones coming out this month. And that's very important for these data scientists that
may not be web three natives, like ourselves. However, we also do see the need for this, like
I said, the token, the financial wrapper to allow for these permissionless payments, these, you know,
borderless transactions. So in essence, the token itself is going to be like the financial wrapper
that allows for these, these AIs that are being built in these competitions, in these private
clouds, to be able, of course, to, you know, pay out these data scientists and to use them
and to allow these data scientists to upload and sell them on the upcoming API marketplace.
Ideally, as well, what we're looking to do is rather than using private cloud
stabilities, we're also looking to work with companies like gaming, which is a decentralized
GPU company, and others out there to allow for the facilitation of using these clouds that are
decentralized, using these decentralized GPU power, rather than relying on something like AWS.
So it's really quite a multifaceted, you know, how to say this, this intersection that we're
looking at, but we see what open AI is doing with their centralized version of,
let's say, the Google or Apple Play Store, or iOS store. And obviously, we don't like it because,
you know, there's people in AI space calling out for decentralization. And we have web three have
been like that since the beginning. So it's like a perfect marriage.
Terrific, great to hear how those run decentralized. Love to hear we'll get into that a little bit later.
Who's next? We got at least three more speakers, maybe four interviews themselves.
Hello, my name is Julian. I am the co founder of Masma. A little background about how I got
started in the space. I got started during the Dogecoin era. My previous experience or career was
in web two products, focusing on user experience and growth. When trying to actually buy Dogecoin
during my during that time, I couldn't do it. Because I've never seen it before, right? You
kind of see the main nets, you're like what the F is that. And then I had to settle for another
doggy coin on a centralized exchange. But so along the way, I wanted to actually make,
or I guess along my journey to make web three grandparents certified. I met my co founder
me tree, who has 14 years of experience and software engineering. He's built mass
market unicorn companies like as a staff and principal engineers and companies like Spotify
and Autodesk. So yeah, we basically identify that there's not an easy way to use web three without
hours of learning and mistakes. So basically, we solve web three user experience by providing a
web three everything app that replaces your wallet, your social media, and any type of
decentralized exchange or centralized exchange into one app, right, which solves web three
and forwarding. Because there's no learning curves anymore, right? We make it very similar
to any other applications you've ever used before, by using utilizing cutting edge technologies like
account abstraction to move the complexities and crypto jargon, etc. So to date, we have
bootstrapped a proprietary product, which has, I would probably say, I think most people would say
that the most superior on chain synergy never before seen web three. So we're coming after,
you know, all your wallets, we're coming after all the exchanges or DEXs, we have, we have quite
a few 1000 people on our wait list, so definitely come up and sign up for that. We have a very
highly anticipated feature, which provides a distribution model, actually, for any founders,
or if you're interested to find tokens in a very transparent manner. You can imagine actually,
if you're actually using x right now, you see a hashtag, let me use you see and take a dollar
sign about Pepe from your, you know, your, your influencers, and you can literally just click that
and I'll bring you to like a coin market cap like page, right with claim base exchange
experience, just one click, you got cash in your platform on asthma, you want to buy any single
token in the world, just from the social page, you can do that. And that's one of the many other
features that we kind of 10x the user experience for you will also be, it's not officialized yet,
but will be actually most likely part of the few select other group of members be the first
insured self custody platform on the market. And yeah, so you want to think about us work a
claim based metamask PayPal demo meets claim market app at Twitter. And we will be launching a token
as well, probably in the coming up, probably not the TGE, but we will have a utility.
We're in the works right now building tokenomics, because we do want to take
or capitalize on this bull.
Well, listen, Julie, you've been in a couple of these spaces, that is the most comprehensive
description you've given of your project so far. So we are still still people can be in the space
multiple times. They're still learning from you. So I appreciate it. Yeah, yeah, coming together,
every single space. Fair enough, fair enough. It's also the first time I've heard of someone
referred to the Dogecoin era. So I didn't know that wasn't era. But now now I know, I guess.
Oh, for sure. There was no such thing as a dog era until the almost posted about Doge.
Fair enough, fair enough. Great. Well, who is up next?
I'll jump in here. Big ask after Julian's great speech there, I guess, but might as well give
it a go. And I'm Jeff, I'm the founder of Amplify. So Amplify is DeFi yield aggregator that
is basically using AI and ML to power automation decentralization and safely optimize yields
across our protocol. Basically, you know, our research shows the 74% of current real yield
aggregators use a manual strategy management, which means that the user has to manually manage
their own strategies. And if they want to move from one pool to another, they must do that manually
and pay gas every single time. Other aggregators, but 90% of other aggregators only optimize for a
single asset. So what does that mean? Like it will optimize for USDT on Ethereum or USDC on
Ethereum or DAI on Ethereum if we're choosing stablecoins, for example, but it doesn't optimize
multiple assets simultaneously. And only 26% of aggregators that we looked at actually
auto rebalance people's portfolios. So obviously DeFi yields are dynamic. So there's large
fluctuations on a day to day basis and only 26% of aggregators actually rebalance this to
optimize yields over time. So we use our technology basically to do real time automation.
So the idea is that somebody comes in, deposits their assets in our pool, chooses a growth
strategy, and then it will automate the entire process of earning yields for that user.
We bucket assets together to create multi asset optimization. So using the example I gave,
we will actually bucket together USDC, USDT and DAI into single pools. And then we'll distribute
those assets across multiple pools across multiple protocols simultaneously to distribute
the risk, first of all, but also to get a more balanced APY over time. And then we obviously
also use the technology to auto rebalance. So if we see, for example, some pools are dropping
an APY, then the system will basically auto rebalance and increase the sort of the
distribution in pools that have a higher APY. And we do that on an ongoing basis to ensure
that basically that we keep it sort of higher and stable over time that if you were to manually
do it yourself. Terrific. That's a real value add, no doubt. It saves a lot of stress from
someone trying to manage themselves and also just takes out the, you can easily automate
and take out the need of a PM, which is terrific. Great. Who is next?
I can go. My name is Kevin Yunai and I'm the founder and CEO of a project called rwa.inc.
And we are a real world asset tokenization and investment ecosystem.
So we fractionalize real world assets and list them on various launch paths and exchanges.
So I'm very excited to be here today and send my warm greetings from Bangkok where
I'm here with a company called Gave for their events. Terrific. Well, welcome. Welcome. Welcome
from Bangkok. It's probably a night over there, right? So one night in Bangkok. All right.
Exactly. It literally is one night in Bangkok because I'm flying back tomorrow.
Fair enough. I think we've got almost everybody, but not everybody who has not introduced themselves.
I'm thinking, um, Kelano, have you introduced yourself yet?
I haven't actually. Hello, everyone. Yeah. In the beginning, yeah. Yeah. Co-founder of
Lewis together with Tom. So we built in decentralized cloudless platform, basically like
the decentralized computing infrastructure that is, uh, you know, web tree native.
And, uh, we believe that it's going to be big resource for AI and, uh, you know,
incredibly neutral infrastructure can really help AI, like to train AI models, to run AI models,
um, you know, with, with more central persistence, with less dependency on, on
centralized entities, uh, at much cheaper price. So we are here at infrastructure side.
Um, you know, and excited to speak with everyone about AI use cases.
Great. I guess, um, you know, I'll kick it off with a question for people.
You know, when we think about AI use cases, you know, we've talked a little, you know,
we've a little bit of AI and building, building some of the games, right? Which is a use case.
I'm getting, it looks like most of these products are not being built using AI. Um,
where do we think, you know, AI is most useful? Like, how do you, where in the applications,
what, what uses is it most useful? It sounds like we've seen everything from,
you know, trading from automating a yield process to making your feed better. There's,
there's a whole number of different things, but I guess if you abstract it out a little bit,
how do you guys think about where either the easiest or most value add use cases of AI are
now or are going to be in the next couple of years with regard to the kind of consumer facing
applications? Yeah, I think it's quite an interesting, uh, question. Actually, I mean,
like, I mean, personally from, uh, yeah, from my standpoint as a community lead and also,
you know, working with the dev teams as well, um, you know, it's helped across the, in the entire
board, uh, you know, everything from marketing content creation to development. Um, you know,
it, it, it's become very, very apparent very quickly. I mean, we've, we've been using it now,
uh, you know, since probably late 2022, 23, uh, early 23. So yeah, we've been making really good
use of it. Uh, you know, uh, mad metaverse for quite some time now where the best value
added though, probably out of, you know, everything, at least from, uh, you know, our standpoint,
um, it's definitely being able to, um, automate and also make more efficient. A lot of the processes,
uh, you know, so generally now for myself, uh, you know, it's very, very easy for me to be
the social media manager, the business developer, the community manager, uh, you know, all at the same
time. Uh, you think there's, there's so many different tools now available. Um, and those
tools are only getting better as well over time. So, but I think one of the most interesting bits
or I think where it's going to get very interesting, I think it was Vayner, I think we're already
talking about it or what they're generally trying to work on there, which is the, the digital
assistant. Uh, you know, I think that's probably where, uh, you know, there's going to be a lot of
movements there, uh, you know, in the future because we already had those anyway. Right. So,
uh, and it's kind of surprising to see that Siri and Google and some of those others aren't already
using it with their, you know, with their, uh, assistant systems essentially. So, uh, yeah,
I think that's where it's at. Uh, you know, for us, and that's been probably, like I said,
one of the biggest value arts, uh, you know, so far is that efficiency.
I'll piggyback off of that. Um, from, uh, being a UI UI UI UX expert, uh, for over two decades,
uh, I'll come at it from a user experience standpoint and design, um, for us over at
layer one X, you know, we're, we're focused on the users controlling and maintaining their data
and where I can see UI becoming a good experience, uh, for the user is, um, enhancing, um, their
overall performance of these different applications that are being built on top of a layer one.
And, uh, what I mean by that is, is, uh, you know, take a game, for an example, uh, if,
if, uh, user, uh, the way we're building it out, the user will basically be able to control and
maintain their game gamer data. Um, so that as they enter into similar like, uh, like, uh, uh,
games that they play, I'll just use the first person shooter cut type games. Um,
that data could be AI could help, um, translate that data over into the games existing, um, data
structure in the, in the sense that, um, you know, if you've ever played a game, um, you know,
sometimes you got to start out, they teach you, they educate you on how to, how the system works,
how their ecosystem works, um, the different levers, the different, uh, utilities, you know,
things like that. And then they, they gradually get harder and harder and harder. And with AI,
it could definitely speed that process up or even eliminate it if it's a very similar game. Um,
and basically getting the, uh, player up to speed at their level very, very quickly. So
there's that component. The other component is, uh, with what we're kind of building with the
L1X app is, is that, um, we're putting the user in control of how they want to experience web three,
meaning that they get to dictate when, why, and how they consume the data. And so, um, by building,
having projects build widgets that they can basically put onto their own kind of OS system
that they get to design. That's another way that AI can come in and help kind of enhance that
experiences is becoming more of a predictive, not just predictive analytics, but just a predictive
thing. Um, I'm the huge fan of, you know, getting the right content to the right people at the right
time type of mentality when we design. And that's basically what we're focusing on and,
and allowing, uh, projects that build on top of us to tap into with AI.
Excellent. Super cool. Um, Julian.
Yeah. I mean, also the piggyback off of, uh, your guys talk about efficiencies, um,
at the startup level. Um, and there's also ties back also, obviously with, uh, um, the
sent a bit from, um, from Sam that, uh, obviously it's gonna be like, you know,
single person or very small team, uh, billion dollar startups, because it's reducing the
efficient, uh, it's improving the efficiencies of the humans. Right. Um, as a bootstrap founder,
I'm noticing that like, and I'm sure all of us here who work in, uh, in, in, in, in this
type of organizations, like the human, uh, um, um, I guess, uh, bottleneck is like,
we need to have meetings with other humans and then communicate, right? Like what I'm seeing now,
especially with how the AIs are now working together in terms of like, you know, um,
opening is a marketplace. And then the way they like, you're able to hook stuff and add
context to buttons on the backend for, um, um, you know, the, the software to the AI
knows exactly what to do. Like I'm kind of envisioning, like how that's going to play out
is I'm going to be in a meeting, right? My marketing meeting, my tech meeting or whatever,
right. I'm instill the time with the agents on me, talking to my co-founder, we're going to
be talking about, and the same thing that you're prompting on chat to BT, except the
prompting that's going to be happening is like an active, like how we're building the company,
like, Hey, retreat, we need to do this, this, this, and this. And at the same time,
like there's an AI actually listening to us fabricating this, thinking about this.
And then boom, the moment I finished my meeting, right? Marketing plan, uh, already a marketing
plan hasn't put out into my slack, right? And as well, not only that there's a content, uh,
strategy plans, there's a, there's infographics being put, there's ads being put. And at the same
time on the tech side, like on my tech channels, it's like, okay, well, um, um, you got, uh, um,
your, your UI UX already designed up, which it does a really amazing job and just all based on
the conversations that we just had in that meeting simultaneously happening. Um, and then
like that moves into your first MVP, right? Um, and then like the next meeting you're having is
just actually just going through like, okay, well, so this specific part on this tech, uh,
UI needs to be changed a little bit. Boom. It's still, and it's still being, uh, modified. And
at the same time, the marketing and the tech are talking to each other, right? Like, because
obviously you're going to need your, your, maybe your product demos are going to change
inside of your, um, your, uh, you know, and whatever your documentation is, um, your marketing
is going to change based on that as well. The strategy based on that. I think that is actually
how it's going to play out. Um, and that's going to be pretty crazy. Like if someone puts
that together, man, like I don't, I'm not sure if that's like, you know, a side project for
you, Tom, but like that is going to make a million startups a day. Yeah. I mean, listen,
that is not a side project for me. I promise you that. I don't know. There may be some people
that's a side project for, but that seems like some effort. And that would be, but I mean,
I see that as you say is basically inevitable. I just, a question of timing. Um, and I,
what the one thing I'm not close enough to know is whether that is, you know, six months away,
18 months away or three years away, but it's probably not much more than that is my guess.
Would you agree? Um, I mean the technology is right here, right? You can make that MVP,
right? I mean, you know, in, in the, uh, if you play around with the AI, it's just basically
you're, you're putting a bunch of solutions together. Um, you're just basically hooking,
like, you know, you create the con, like you would just build the, the systems of,
okay, the context of this, this, this pulls from here that you can build that MVP. I think
now today, I, I, if I were to go hardcore at it, you could build that, um, that MVP out race,
you know, I mean, I think it was, it was still work. You would get a pretty good, uh, like,
because the thing about when you're in a startup is like, you have a conversation. And then the
thing is like humans, we got things going on in our lives. Like, you know, I'm not going to like
do the task right after the meeting, like, you know what I mean? And like, I want shit done.
When I'm at the front, like you want shit done, like this. Um, and so you can move and do very
things that are quickly and I don't have to waste time in the human aspect, which is like, I mean,
it's just, we're humans, right? It's, it's normal. Um, but like there's lots of people on
the startup stage that would probably do such thing and it wouldn't be expensive. Um,
and when it helped, it would help them actually move a little bit farther. Um,
and that's just MVP one, right? Just, you know, and then you'd go up to be two, three, four,
five. And then like, as everyone else evolves the technology, you're just, it's just like web
three, like I'm using the Uniswap, um, liquidity pools, like they're innovating into the three
and I'll assume before like the same thing with Sora or like, you know, all these solutions,
like we're just going to pull it together. Someone can just build a Frankenstein product
that is basically just the arbitrator for all of that into one package.
Hey, Julian, could you imagine if somebody did a Zapier and automate IO over here on web three?
I mean, that would go nuts. Zapier and what? Automate IO, you know, those kind of platforms
just for non code. Yeah. But in the web three space, it would go nuts.
What do you have? What type of use cases are you seeing?
I mean, you could just basically do that Frankenstein build together thing that you
were just talking about, you know, uh, that's basically what Zapier is in the,
in web two space for the non coders. But I'm just saying, you know, with the non coders over here,
it would be a great way to kind of implement that Frankenstein workflow, logic flow,
kind of thing, if then type statements happen, right?
Yeah, I'm already seeing a few startups now that implement the intent AI, which is, for example,
is a project called parallel finance. They use a kind of like, I mean, for the construction,
you can you know, kind of make everything one click, right? But what they've done
is you prompt, you can like kind of prompted like, Hey, I want to put X, Y, there's a few of
these projects kind of doing this now. I haven't really followed up on them. Because some people
are kind of scared giving away their like, depends on the business model, I guess, but some of them
asks for access, like the AI, they actually have full access to your C phrase. And some people
are a little bit scared about that. So, you know, but it's 100% like possible, the AI builds
the whole intent structure for what you want to do. I have $5 right now. Where should I put this
shit coin? What's the best training shit coin right now? And then like, it'll put it, it'll find it
script through the web three, put it there, or like, you know, I want to, you know, whatever you
want to do, actually, it's a little possible. Now, it just depends on how people in web three
is a little bit different. Because like, you know, it's like money and like the scams and
hacks, and like, you kind of want to feel in charge and power for that. So, you know,
so I've got quite a good example of this, actually. Yeah, the first the first game and platform that
we built, you know, we're talking we're talking seven digits, you know, it cost us to develop that
this new game that we're building now is bigger and probably flashier as well than the first
game we built. And it's more like a five digit, you know, price tag this time around. So like,
and I'm not saying it's like a Frankenstein, you know, put together like what you say.
But one of the things that sort of really helped us with this as well is because
we're going to launch the game with a project called BEGA Arcade. And they have a whole
platform set up there for games to come on have instant on and off ramps like for
like with fear as well. So, you know, there's no need to this non-constodial and custodial wallet
access as well. So users can just come straight on logging with their Google account, throw some
money down with their debit cards and you know, start playing and they're using the blockchain
then without even realizing that they're doing so. And then, you know, once they also want to
pull the money off, they can pull it off in into fear as well. But the options there
obviously to do it in crypto, probably better percentages as well there. So, you know, that's
one of the systems that that's helping us to be able to launch with that, not just without,
you know, not only with the ease of access, you know, removing all of those boundaries,
but also, you're having that user base there to start playing the game from the word go as well.
So, you know, you're saying earlier about all the tools being available now for these
and I wouldn't mind, you know, I've been hearing this a lot when you do air maze and all that,
a lot of air maze that I go on as well. People are saying, Oh, well, what's the text here? And I'm
thinking the technology is here. Like, we're already implementing a lot of this technology
that you're talking about that you can't wait for. And, you know, people don't realize that
a lot of the accuracy is here and that, you know, they're sort of sorely getting left behind
at this rate as well. So, yeah, this is why it's always good to keep up with everything that's
going on in this space, right? Especially now, because we're getting saturated. But I think
that was a question actually, I want to ask, yeah, Tom, everyone, but anyone else that's actually
creating an AI at the moment, because obviously, as I said, we're only really using it, we're not
creating an AI, because well, to be perfectly honest, we believe the Mac is very oversaturated
at the moment. But yeah, we wanted to see, you know, what are your guys opinions on that in terms
of like the amount of AI projects and, you know, apps, also dApps in general coming out at the
moment. You know, the number of web free projects are also jumping on, you know, this bandwagon as
well. You know, do you think it's still a long way to go yet? Or is it already, you know,
majorly oversaturated? Well, I mean, I guess what I'd say to that I don't have the exact answer for you,
but I'll phrase it this way, is that AI is hot. And so everyone is putting AI in names, just like
people put .com in names a long time ago, or more recently, I put blockchain in names. And so
they're really what gets interesting. The reason I asked the question earlier about
where the use and strengths are is because they're places where it can be very useful
in places where it's less useful. And so I don't have a good sense as to who is actually
implementing AI in really either novel or super high value added ways versus people just putting
it in because listen, you could you could there's certain some levels of AI you could have
articulated, you know, a year or two ago in certain types of applications in terms of,
you know, whatever sentence completion or all kinds of, you know, predictive, you know,
movie searching on Netflix, right? That's some level of that, right? So, you know, it's just
a question of degree, and I'm figuring out where within this next generation is really going.
And so I guess the short answer is, I see a little more and more people using it. But what
is actually being used in a proper, really, you know, differentiated value add way. And that's
much less clear to me. But Andre, you may have a comment on this.
Yeah, um, no, I think it's very important to realize that, you know, for example,
we look at Nvidia, you know, the start as a GPU company, obviously, you know, back in the 90s,
and they really started getting into AI around the 2010s. And a lot of these enterprise companies,
including big grid, you know, they have we have this experience like goes back now going, you know,
seven, eight years. And it was still very much to retail, like, okay, I'm going to think about the
movies like, you know, the with all the robots, and then all of a sudden, chat GPT comes out in
late 2022. And as you're saying, is the buzzword is popular, and especially with this bull market
going on right now. There are a lot of projects and things that are oversaturated. My personal
take is that a lot of the web3 projects are being built out there. And in AI are going to serve
retail and people that are really already in the space. But when you look at like corporations
such as SoftBank, or for example, government entities like NASA, which we've worked with,
they probably just want to take the regular traditional route of working on platforms like
Kaggle or hugging face, or even ones like ours that are like a mix of web2 and web3,
because they, they either don't want to have all that information out there to be completely open,
there has to be some level privacy for them to remain competitive. So that that's my take,
I think the real winners, although with web3 AI space, when the bear market comes, of course,
will show their their selves. But yeah, there's so many chat pods out there, for example,
that is just getting quite ridiculous.
I'd like to kind of jump in on that with what both of you guys have said.
In the web2 space, I've had the opportunity to work for a couple of different smart device
companies, designing their apps and infrastructure that way. And, you know, let's be honest,
smart devices are only as good as as the data that they have to back it with, right. And so in a
sense, a lot of these smart devices are actually pretty dumb, unless you can connect it to you
together to other devices and grow the data set that these smart devices could actually tap into,
right. And I look at AI kind of the same way. Unless you have a great data set to pull from or
tap into a kind of a non custodial type of thing, you know, like chat GBT and all these guys that
are that are tapping into the same pool, it's kind of it's it kind of raises the question of
ethical responsibility when using AI in in that sense. I think Alejandro, you know,
you were the one that just mentioned, you know, about having data out there for everybody to tap
into, you know, and if we look at the ethos of what web3 is all about is decentralization and
user own data and this and that. It kind of it kind of makes it interesting because it's like,
do you do you tap into a centralized data point for your AI to be built on to run off of or do you
kind of respect the privacy of a user? And I think that that's an ethical question as
builders that we're going to have to at some point come to the rationalization and answer
that for ourselves of what we're willing to do. That's sort of the entire side of it's getting
actually quite extreme now to this point as well, right. And there's actually a couple of good
examples really that apply to this because you've already got Twitter obviously that building their
AI up, which obviously that will be running entirely. Yeah, that poll will be Twitter as a
whole, you know. And then obviously you got Reddit who recently announced that they're going
to be selling, you know, their data from their servers to an AI company to help out build,
you know, their LMM. And so, you know, the ethical side of it, you know, everyone was complaining about
Facebook, etc., and all these other social media companies selling people's data or using people's
data. And to be perfectly honest, at that time, you know, yes, that data can definitely be used
to profit. But, you know, it wasn't really, it can be used dangerously, you know, and like,
that's not true either. But to the most part, you know, it was relatively harmless, I believe,
in comparison to where we're getting at now, where our entire like digital profiles on some of
these platforms that we've been using for a lot of years, you know, that data is now going to be
used to, you know, to train some of these AIs. So I do wonder how that's going to go and where
that's going to really fit in. You know, people are complaining about now about, you know,
where they pull their data from, etc. And I think in cases like this, you know, it can really,
really start to cause some issues, especially when you consider how many companies also use
Reddit, you know, on a regular basis to, you know, to promote themselves or, you know,
to have their communities there. So yeah, I do wonder where that's going to be heading,
you know, in the very, very near future. Well, here's a question I have for you guys. Like,
if you look at the internet, it clearly rewards scale. And that's why you've got his global,
I guess we got Amazon and Google, Netflix, and, you know, really, you know, benefits,
because of the basically global nature of the bigger you get the more you can spend on
marketing and customer acquisition, which makes you more profitable, which means you can spend
more on customer acquisition. And you know, that flywheel is hard to supplant. And I guess
the question is, is AI level that playing field for small companies? Or is it the same thing all
over again, where pick your use case, whether it is an app that helps you read Twitter faster,
whether it's one that, you know, picks your, you know, maximizes yield for a portfolio,
do we end up in a world with, with just a couple of AI resources for each one of those,
like we do for our, you know, applications or internet things now? Or does it democratize the
level that the buried entry is so low, the fragments out to just a large number of different uses?
What do you guys think? I think that's a, that's an excellent question. And, you know,
from a human behavior, cognitive behavior standpoint, you know, we have a tendency to
find a business model that works. And then we just basically rinse and repeat, you see that
tremendously through even crypto, you know, you see fork after fork after fork, and slapped on
new skin slap on a new logo. And it's the same thing over and over again. So to kind of bring
that to full circle with what you're talking about, I think we're going to see that across
the board with with AI use cases as well as is that somebody's going to find something that works
and everybody's going to try to tap into it just to make some bucks while they can, right?
However, I don't think that that'll be 100% of everybody using AI. I think that the true
innovators are the ones that are constantly looking to not think outside the box, but to expand the
box, even bigger and give it a better scope of what the full capabilities are. And I, I'd like to
think that there's quite a few on these panels, panelists today that are participating in this
discussion. Could be looks like Julian Alexander, Alexandra, we got both of you guys up again.
Go for it. I can tell you guys a funny story about AI. If you have time, I used to actually
do speech recognition and AI back 15 to 16 years ago, we collected all the hospital data,
annotated all the sounds from the dictations and put in all the data in language models back then.
And when you have really worked with AI in practical terms, you find out it's not as
magical as people think, right? It's more a mathematical best guess of some order of words
and that really is what AI is today. It's not as people think, some computer thinking,
a lot of words and things by themselves and then coming with conclusions. It's basically just
mapping of words and sentences. So I think we are still very early in AI in general,
in my opinion, and I'm looking forward to more advanced AI that can do impulsive stuff and really
think for itself, but it's going to take a while. So I just wanted to demystify the AI
from what many people think AI is. That being said, it's an amazing progress with AI and I
think we can use it a lot in the future. And I think really soon we reached a point where
we have so much data in the language models that it will be able to perform any task. We can see
that already. So yeah, I don't know in what context I just said this, but I think it was an
important story to tell. It wasn't exactly on the topic, but it's certainly interesting to be
aware of, no doubt. Alejandro, did you want to jump in here? I want to go to Julian, go first.
Cool. Yeah, I mean to go back onto your comment about the gap between obviously the rich and
the poor, basically, is I think that if you apply it in analogy of efficiency, so for example,
somebody does not have the resources, they're at efficiency like 1%, they're at a 1x.
I think what the AI does is great because now it allows them to move efficiency up to maybe
three, four, five, depending on how trained they are to use those solutions. So now
they've just kind of 4x themselves. Whereas a startup, they have the resources to obviously
be automatically at like 30x or maybe 25x. And that's just because of pure manpower. So if
you're thinking about manpower that way, what's going to be interesting for those companies that
they're actually probably going to then 3x that 30x and then they're at like probably 90x now.
And probably on top of that is that they're going to have specialized AI
prompters, they're going to have specialized analytics. So then, for example,
all of us, every single customer, there's an efficiency rate of how to actually start to
talk to my ideal customer type. There's efficiencies and every single model that we do
plays around with that. So for example, we're just the regular person just utilizing AI
themselves. We're at a 3%, I mean, we're at 3x, 4x. But the thing is, I don't understand how to
leverage that AI to ultimately be able to maybe move myself up to a 4x just because I know how
to really target the customer enough. I really have the information that that they have. Whereas
the startup companies, for example, this is a marketing issue. We create marketing plans for
persona types, right? Like this persona, this persona. We only can do a few. The huge
companies are going to have a million persona types, even one for somebody who just farts,
you know, I don't know, like farts more than the other person, right? And like, you know,
a part of target, this specific person, that AI will be able to, like, their AI is going to be
so sophisticated, because they got people that know how to leverage this, they have additional
resources to really, I know what you do from the moment you wake up, I know what you do when
you go to sleep. And those AI systems are going to be able to more efficiently target those customers.
So I mean, in some contexts, I mean, yes, the gap is I guess bridged a little bit,
because both of them are both able to now go onto the playing field and have extra juice for the
customer, right? Because obviously, having something in the market is better than nothing,
right? The question now lies into the end consumer. Are they going to get really sick of
is the mass market going to be much more ignoring the technology now? Because where we are right now
in web two, we're getting mass market on social media, like we all want to delete our social
media, we all want, we don't want ads anymore, etc. So what is the what is that going to look
like for the end user when this gets insane to target the user, like it's going to be basically
an ad built for you? Because like Sora can literally build a video, right? So if we have
a million person times, imagine every single ad that lands to you, just like, Hey, Julian,
I know that you had a hard time on the toilet today. But you know, we want to sell we want
to get we have this this this and like it hits every single need that you've ever had you've
ever thought about in your life. Whereas right now the advertising it's like, you know, oh,
you know, this web three is so different. It's very generalized, right? Whereas like,
the targeted for the marketing side, which for the AI example, it's like,
Hey, I saw you had a hard time with that wallet today. With that trade, this is how
you actually get something much better with this product X, Y and Z.
So Julian, I totally get that. My question is, is that one platform and one tech that is used
to generate that ad across multiple platforms, or is each platform design build their own that does
that? I mean, at the moment, I think it's probably going to be each of its own, right? But
the leveraging aspect here is obviously bigger companies, you're going to be able to hire
much smarter people that know how to leverage these at a much better rate and then create
these specific plants. And whereas the I don't think one startup will be able to do all that
unless it's like, you know, that's, that's more like, that's where the human aspect plays
key here. And you can't really replace the human aspect, because that's generally AI,
right? To actually figure out how to do these specific things in the right context. So I don't
think a one big entity could do that. Because like, that means that that big entity is generally AI,
right? Hold on, this is your question. So I've got it. I've got an answer. Or I've
actually got a hundred per second. Do you want to dive in for a second here?
Yeah, no, thanks, Tom. Um, I mean, what you ask is a very difficult question, because no one has
crystal ball, right? But the power of the markets, the power of the market. So I think when it comes
to like, when you ask your question, if it's gonna be a single product or a single company,
I'm just gonna, you know, state the big elephant in the room, I think open AI personally leads,
you know, the biggest advantage in terms of allowing their customers to pay for a
multi subscription and using their products. And for a lot of retail and consumers, that's all they
want. They don't want to deal with the tech, they don't want to deal with the intricacies of it,
they just want to say, okay, subscription, bump, bump, bump, I can use this for whatever.
On the flip side, though, and that's, and that's also the case for a lot of private
corporations, that they want something that, of course, that is built out by data scientists,
or whatever, that are enterprise, private, built out in the same way that we've seen thus far.
But then, then, of course, to have the web three peers like ourselves, that I think will just
maybe jump around a bunch of different projects and seeing which one works best with the key
that, of course, decentralization is built in, it's on chain. So I, yeah, I really would like
to say that there's going to be a plethora of different products. But again, talk on human
nature. I think people generally like to keep things simple. I know that, you know, yes, we have
Facebook, we have Instagram, we have all different social medias. By the end of the day, I think
people like because they get burnt out, for example, with all those social medias, they choose on just
a handful. So I think open AI personally will be like a leader in that. Maybe there'll be some web
three company that the AI company that really leaves a mark. Hopefully, hopefully, that'll be
big grit. But, but basically, yeah, it's going to come down to like just a few, I think, that will
get to know you as like, like Julie was saying, that persona, right? Because otherwise, people are
going to get overwhelmed, they get too many notifications, it's going to be like having
not just one assistant, but 10 or 20. And no one's going to want that just comes down to the
human experience. Fair enough. Listen, this is super interesting discussion. You know, we,
we had this set up for a hour, which we are just about to hit. And you know, I actually have another
call I got it and I know most of you guys are signed up for an hour. So with that, I just want to,
I just want to note here before before we go and obviously, everyone's welcome to stay on. But
Travis, I know you had your hand up for a while. And I didn't get to you. Do you want to give some
maybe some closing remarks, given, given you were we sort of missed you earlier on?
Yeah, no, I pulled my hand back down, because we kind of moved on from the, from the topic,
but I was just going to add, you know, as a man of very few brain cells, and, you know, developing
and coding kind of department, you know, I'm kind of excited for some AI initiatives into
being able to check code, being able to find vulnerabilities, being able to do some of those
kind of things. And also just overall, you know, glitches that you may have, whether it's a game,
whether it's on your site, whether it's one of those, you know, one in a million times,
if you do this action, then this action, this will will suddenly happen, and be able to, you
know, leverage AI into some of those kind of things. But this has been an awesome discussion.
You know, a lot of people are much more advanced in this topic than I am. So I always learned a
ton in here, and love all the valuable insight, different opinions, and just kind of seeing where
we're going as an industry. So thank you so much for hosting it. It's been a great panel. It's been a
great discussion. Listen, thank you, Travis. I appreciate it. I kind of agree with you. I feel
like we're just getting started here. The hour went fast. So I think we just may have to reload
on this topic in a couple weeks or so and, and dig in a little, a little bit deeper. But I,
thank everybody for joining for, for joining us as speakers really appreciate everybody and all of
your insights. I certainly opened my eyes to some things I hadn't been aware of. But there's a
lot still to figure out. So appreciate it. Hope we can go back and, and dig into this again in a
couple of weeks. Thanks everyone for joining. Thank you. Thank you guys. Appreciate it. Thanks,
guys. Bye. Have a great day.