Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
AI Agents Have Money Now: Who Wins the Agent Economy?
Recorded: Sept. 8, 2026 Duration: 1:04:36Space Recording
Full Transcription
Oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh, oh,
All right. Hopefully y'all can hear me, okay? I want to make sure that we kind of get this thing off
in a good note. We got a couple of people. I still got to get up here. Let me hit play. Let me see.
What do I got for y'all today? Let me see here. We got. This is gonna be good. I like this. Yeah,
give it three two and a half minutes or so. Dingus, good to see you up here.
Got to get some people invited, but we're we're we're gonna have a good time. I promise y'all.
That is the one thing that I can always promise y'all. So let me hit play on something.
Let me get some more speakers up here,
and we'll have a good time in literally two minutes and ten seconds. Feel
the passion ignite me. Feel the burning inside me. Let this life just pass me by.
Stop. Stop
me now. All I need is a horizon. Come on, if you keep trying, awaken that you feel so alive.
Today I'm gonna make
a change. Bring your time out of my window. I was made for so much more than just a dream. Stop.
All right, we got Mr. Minter up here. Crystal, I see you down there. If you don't mind,
accept that speaker invite because we absolutely want to hear from you. But yeah,
let's get this thing kicked off here in 30 seconds. Straight to the top. I'm on my way.
Don't expect the spotlight for the dreams we're fighting for. It's not me.
It's not me. It's not me.
It's not me. It's not me. What's going on, everybody? Super excited to be with y'all on
this Tuesday. Everybody in the U.S. Had off mostly here on yesterday for Labor Day,
so this is kind of kickoff to the week for most of us. Super, super excited to have
this conversation. We're actually talking about things that matter. And if y'all know anything
about Gen Z spaces, we try to keep it fun, try to keep it exciting because the conversations that
we're having are pertinent to the stuff that we're dealing with day in, day out.
And our goal is to bring some legends in the space that have the experience.
That are actually working with this first, you know, firsthand. Because you know,
what's the point of hearing from somebody else that you don't really care about?
I want to hear from the people that are actually doing it. And if that helps everybody else,
happy day, win-win across the board. Let me kick it off with some intros.
Definitely want y'all to meet some of the people that are up here. Dingus,
you're the first one on my list. You want to unmute, say hello, do a quick mic chat,
and let everybody know who you are. Mike check, Mike check one two. This is the Dingus and Mingus.
My name is Dingus. I work for the Daven Network as a community manager.
We are part of the DPin sector of crypto. Looking forward to being here with Genzo.
Took a while to get recipitated with you guys, but I'm glad to be a part of this chat today.
Let's freaking go, action! Appreciate you bringing me up today. There,
there is the let's freaking go for you. By the way, I, I, I got the the scratch sounds
ready to go for you next time, Mr. Minter. You're the next one on my list. You want to mute?
Say hello. Hello, hello! That was a hard intro to follow, guys. But my name is Puros,
also go by Mister Mentor. I work as head of marketing at Sonic Labs. I also help lead
some AI product development here. Super hyped to be in this chat with Genzo. Actually,
similar to Dingus, it's been a minute since I've been in a space with you guys.
Been talking a lot of AI and agents recently, and yeah, hyped to get into this conversation
with y'all, man. I'm happy to have both of you guys back, Kristo. Welcome to the party. Hello,
hello! Oh my goodness, it's my first time here, and I'm so excited to be here.
So my name is Crystal. I'm the CEO and founder of Jodoka Group, and we build
operational infrastructure for businesses. So there's been a lot of talk about Web3 and AI
and all of that, and how all of it ties into crypto. So I'm really excited to be here.
I am excited to have all of you guys here. This is this is going to be fun. Again,
like we're we're talking about what's going on with agents, what's going on with AI assistants,
and people have blurred the lines. It seems like people don't really know, you know,
who's using what and what all of these things look like. I would love to kind of get us kicked off
here, and I'm not going to call anybody. If anybody feels so inclined to kind of answer
this first question, so we can get it kicked off. Like, what do y'all think separates, you know,
these AI agents from the AI assistants that we're kind of familiar with, right? Like people going,
like, oh, I need to find something for my calendar. I need to go. Where do I go? What do I do? Like,
what's a real difference between an agent and, you know, what people have been used to, you know,
kind of going on GPT and asking for the basic questions for, you know, from assistants? Well,
AI agents will totally take their own reign. Once you give them the power,
you give them what they need, whether that be the tools, the money, which is what
we're talking about today. Is AI agents having their own wallets, being able to make
their own decisions, whereas When you're going into ChatGPT, you're going into Claude.
You're going to be querying it, and that's kind of what everyone's doing with Google these days.
Google's is just querying Gemini now. Anybody who goes to make a Google for Google search
is going to wind up querying Gemini. So, what an agent ideally is is to take some of the things that
maybe you're already doing and streamline that into something a free agent can do on its own time.
Let it let it let it be you know working while you sleep, so you wake up,
you have all your emails read, you have your dinner reservations made,
and everything's kind of taken care of in terms of the substack, and you can kind of take care
of the overhead things. So, there's a big difference between just Querying the GPTs
and actually letting it have free roam, giving it a wallet, you know, letting it have Paybox.sh
so it can go ahead and make these reservations, purchase things, and kind of have
a little more free roam. And that's what we're looking towards in the future.
Here is agents having their own power to make their own decisions. Are we actually there yet?
That's a real question. I was curious what y'all think. Like, how close are we really to having
agents kind of operating on their own? But not just operating on their own,
but meaningfully operating on their own. I think we're already there in the sense
that they are operating on their own. When you add that term meaningfully,
I think that's what really changes the question here, right? So shout out to what Dingus said.
An assistant really just is generative, right? Like it tells you things, it searches things,
it can find things for you, and then you still have to make most of the actions.
Whereas agents usually have access to at least some sort of Platforms or protocols to be
able to like whether it's email, whether it's your accounts in different places,
whether it's a wallet with actual money in it that they can spend on your behalf.
An agent will actually do things for you. And like I mean, a lot of people I think have agents.
You know, started with Open Claw and Hermes agents. Now we have things like Instinct.
I don't know if you guys have heard of it. We just had a really good conversation last week
about Instinct. And And there's a lot of really good and bad things, right?
Like I'm I'm of the mind that I want to make sure that I have really strict guardrails
around my agents, so that I know that they can't do anything too crazy.
Like the I think the one word I would put between assistants and agents is consequences.
The consequences of an assistant getting the wrong information are not that grave.
Like it's going to tell you something wrong. You can still research it.
You have to make the action afterwards. Versus an agent goes and finds the wrong information,
and then it actually takes an action based on that wrong information. And you give it access to like
your main wallet, you could be in really big trouble. Like something as stupid
as the the Solana agent on Twitter, like I don't know, a year ago, that was asked for four soul
and ended up giving away an entire wallet of tokens. Like it can do some pretty dangerous things
if you don't put the right guardrails. And I think that just starts with the user
that's deploying it, how much you trust it, and how many guardrails you put around it. But yeah,
there the the possibilities are there. The functionality is there. You bring up a great point
because not only do we not know exactly, you know, what decisions are making, you know, left
and right, but there's no tracking. Logs are literally getting deleted by some of these agents
nowadays. It gets super, super crazy. Crystal, want to get you in this conversation here too. So,
what do you think has changed? Because I mean, you're you're in the thick of things, right?
You're you're literally building AI systems. What's changed recently? You know,
that that's made this idea of in an agent economy even possible. Yeah,
I really love what they both just said. I agree with all of that. I think the biggest thing is
that a lot of AI, I look at it and explain it as they're like advisors, right? Like ChatGPT, Claude,
the chatbots—they're advisors. Agents actually take action, and this is really important.
Kind of like what Mr. Winter, sorry over there was saying, is consequences.
I absolutely love that you brought that. Up, because a lot of people want to almost romanticize
this idea of not really doing anything and having the agents do everything. However,
there is a risk to that, and the real shift is authority. It's not just intelligence,
and this kind of plays into how it handles wallets and payments and all of that.
I like to tell people think about it as like a corporate credit card, not a corporate bank account.
So you want to be very firm with how you set up your guardrails and what you allow it to do
and what you allow it not to do, and to still have that human factor in terms of keeping track
and the logs and make sure that it's not doing anything crazy, like deleting the logs,
like you just said. But the shift actually has not been super recent. I actually spoke to somebody
who worked for PayPal not too long ago, maybe last month or so, and they said
that they've been working on agent-to-agent merchant transaction for the past like three
or four years. So this has been in the work for a while, guys. Like this isn't super new.
It's just we're just hearing about it now, mainstream-wise. No, I I agree with you. Yeah,
people have been actually pushing the envelope for such a long time. And thing is,
let me go to you because you were talking about you know kind of giving the ability for these agents
to actually transact on their own, which is a really cool thing. But I'd love
for people to kind of have a better understanding. What does that do? What what's possible now?
You know, once you know these agents can actually control and spend money.
So what's possible is a lot. I my father's been using ChatGPT and Claude for quite some time now.
He's 78 years old, and I just I just got like onboarded to Paybox, and I it was like all
over the timeline about a month or so ago. I was just seeing this like wallet you can give to
your agent that's like sponsored by MoonPay. You can basically you know re up your wallet
without having to spend fees. You can actually get 100 of your value to your agent.
And I just showed my dad that I was like, here's something pretty dope,
and it's not just for like you know crypto related endeavors. The the Moonpay executive was showing
how he's actually just making dinner reservations, and he sets his bot up to know what what
he's trying to do, what are the parameters in terms of like what he's trying to spend,
and the bot can go ahead and do it all himself, which I think is really miraculous to be
able to give these kind of tools to people who are just kind of puttering through
and doing all this stuff manually. You know, I'm looking forward to seeing if my father
actually uses Paybox, and I think that's going to be a really excellent tool,
just to you know give agents their own wallets and not have a wallet that has anything to do with
a human. Like here, it's for you, buddy. This is your wallet. You can go crazy,
have a field day with it, and you don't need to worry about it because it's not your own pool
of funds. And I think it's really important that we segregate those two, because like we were.
Talking about earlier is that you know if somebody gets access to your main wallet
and then they just say oh I thought I was supposed to give them everything oopsie daisy this is
why you got to have these sort of segregations which I think Paybox is one of the things I'm
super bullish about when it comes to AI agents just being able to have their own wallet being
able to tap back up with MoonPay whenever they need to I think is going to be a thing
that really moves the needle forward here. Now, Mr. Mentor, you're you're a Mr. EVM on this call
here, so I got to ask you some crypto questions. What's going on as far as crypto
and agents are concerned, right? Because it seems like stablecoins or you know even any kind
of crypto there might be the natural payment rails for agents. Do you see the you know that being
the future here? I think it's going to be a big part of it. Like currently, you know,
we have a few different competing standards. We got X402 by Coinbase. More recent one was MPP
by Stripe. But like you see, Visa, Mastercard, American Express, like all of these have hopped on
as founding members for X402. So you you can definitely see Crypto and stablecoins in general being
a core native payment rail for a lot of these agents, but you also have to understand like
the the entrenched players in this space: Mastercard, Visa, American Express. Again,
they're not going anywhere anytime soon. So they're going to be using these crypto rails
however they can to make sure that they get as much access to agents that are spending money
as possible. But they're also going to start converting their own rails to be
able to support micropayments, to be able to support API payments, streaming payments,
things like that. Right? Like right now, according to the data, I think we we had an event last week
with Google, Google Web three and AI, literally talking about exactly this. And right now,
over 90% of the payments that are real that are being spent by agents is mostly just for inference.
Like over 90%, there's a bunch of wash trading and a bunch of testing and stuff like that.
And so like 50% of payments are real payments, and the majority of those are just
either inference payments or there's like a little bit of data streams payments.
And so I think one thing that we're going to see, I think stablecoins will dominate a lot
of these types of transactions, and it's going to be in micro payments.
A lot of the stuff that like Historically, hasn't really been able to be monetized. So, like, very,
very small payments for specific pieces of data that people just don't have the time
to like go through and pay four cents for one piece of data to make a decision.
Agents will be able to do that. Like, they'll they'll know how can I make better decisions?
Where can I get better data? And if it's going to cost me a fraction of a cent or one, two, three,
four cents, and I can pay that without, you know, paying twenty percent charge. You know,
charge fees. That's what's like. There's a huge gap, and there's a huge opportunity there.
And so I think that's where stablecoins will dominate. Is going to be mostly for micropayments
and automated payments by agents, not for things that like if I want to book something
and I tell my agent to book it for me, they can book it through you know a card
that I give it access to, like similar to what Crystal said. Like you want to think of it as
like a corporate credit card versus a corporate bank account. I can give it access to a Visa
or a Mastercard, put a hundred and fifty dollar limit, and say book me a dinner reservation,
book me movie tickets or X, Y, and Z. And those I think for for at least for a while
are still going to be dominated by the traditional payment rails. But most of the digital payments
for resources, infrastructure, which is what they really need access to, compute, inference,
stuff like that, that's 100% going to be stablecoins. I think for a very long time. All right.
Before I give y'all a question that I want y'all to argue against one another, hopefully,
let me ask you something, Crystal. Kind of leaning into what you were talking about with PayPal.
Like, where are we actually seeing useful examples of this? Right? Like this agent to agent,
or even agents to business transactions. Are those taking place yet? Because I know you mentioned
you know PayPal has been working on this for a while, but where do we see it today? Yeah,
absolutely. Great question, Mr. Mitra. I think you were absolutely on point with all of that.
There's been a lot of talk about connecting like wallets with stablecoins because it adds
economic capability. And totally agree that people are using it for,
or agents are being used to buy APIs, data, and compute, and even like smaller services.
And it's so great because basically a wallet turns an agent from software
into an economic participant. So this is like an entirely new layer of economics that
we haven't experienced yet. And there's so much opportunity. There are so many gaps.
There are so many different ways that people can bridge all of these gaps.
So it's opening up a lot of possibilities in ways that we've never seen before. That being said,
I am seeing it everywhere, especially if you go to any of the New York tech meetups.
It seems to be a very common topic of discussion. But I am here for it. I think it's brilliant.
I think it's about time that we use agents, AI, and bridge it over to Web3 and crypto,
and really make all of this interconnected the way that it should be. All right,
so here's the question for all three of you: How much autonomy should these agents actually have?
And I'm talking about financial autonomy, right? Because we can say that we're going to give them
a corporate credit card. That's fine, but what's the limit? Like, where do we kind of, you know,
set the bar? Depends on how much you trust it. That's all I would say, honestly. Short and sweet.
I don't trust my agents quite as much as some people do at this point. So yeah,
as much as you can trust it is the easy answer. Okay, I love that piece of trust there, Dingus.
Like a lot of people say, that trust is a residue of promises kept. And does that is that enough
for agents? If they you know nail one, two, three transactions, you're like, all right,
have my wallet. I'd say we're a little too early for that. I mean, I've only been using agents
for probably less than a year as like a thing, and I've only been dabbling with something.
I've been using Seeker Claw, which is kind of just like a loose agent.
It's not like a full autonomous device. But yeah, the point of being able to trust something
is actually like the reason for business. You need to be able to trust what.
What is going on here for you to let it go? I don't. I like like Mr. Minter was saying.
I don't know if I fully trust agents at this stage in the game. I think they need more time.
I think they need more learning. I think they need more you know inference with like human beings
before they can kind of just be set free. One of the gentlemen in the crowd here,
he actually dipped out, but he was Kevin from Ko Markets. And he he has agents running a lot of code
for him, but it's always audited by like a human at the end of the day.
And I feel like it's going to be some time before we can just say agents can be free agents
without human interaction, because I just I don't know. And also the slipperiness of crypto.
Crypto is a one-way street. I mean, it's not totally a one-way street,
but blockchain it's permanent. Once you do the thing, it's done. So that's kind of where I
I say with agents being a little bit slower to run to the finish line, I think would be worthwhile,
especially since a lot of you know what they do is in the crypto sphere or dealing with stablecoins.
And one thing for me when it comes to stablecoins, I've kind of I've strayed away from it.
I've actually started to take cash payments, mostly because the slipperiness of crypto.
That if you make a mistake, it's that's that's it, and you really can't redact that mistake
and go backwards. When you have like a debit card that an agent's up to, you can maybe call in
and be like, "I did not authorize that. I don't know what's going on, but I did not authorize that.
" You can't call up Coinbase and be like, "I didn't authorize those charges." Well,
I'm sorry to say, they went out. So yeah, I think we need a little more time before we give them
free reign. You know. Interesting. I well, go go go ahead, Mr. Rincer. Yeah, I was gonna say.
So like, I have a like a real world example, right? I love using agents.
I've been using agents for probably almost three years, not two and a half, three years. At first,
they couldn't really do anything, but now like I have a social monitoring agent.
I have a competitive intelligence research agent. I have like I'm obviously like my OpenCL Hermes
agents in Telegram that I just mess around with. A lot of them interact with my code bases.
I can just tell it like update this, you know, fix this on the front end X Y Z.
But like some of them do have access to money. Like they have access to my X API.
And I ran into an issue. Actually, we did a campaign that had a ton. Like, actually,
it was really good response, which is a great thing for business and for visibility.
But then I realized, like, the amount of comments that I had to keep up with to like let me know
if something was off in the comments just shot through the roof. And like,
I didn't have a limit then, and I started seeing two, two hundred and fifty,
almost three hundred dollars a week. It was spending just keeping up with the X API charges.
So then I had to put a cap on it, like no more than a hundred dollars a week.
And if it goes higher than that, it'll ping me so I can see like. Is it getting spammed
or is this like legitimate? Is this something that I need to keep running?
And so there's a lot of different things where you can really narrow the scope that you give
your agent, and you can put, like I said, I just I focus on guardrails, right?
Like as as far as I can trust it. And I can only trust it to do what I want it to do
until it starts spending my money. When it starts spending my money, wherever I can find
an opportunity, you can't tell the agent like, "Don't spend more than a hundred dollars,
" because they don't give a shit. Like they just—they don't listen. They'll do whatever the hell
they want to do. And so you have to make sure like you put the guardrails
where you can control them. So now my APIs have strict limits of how much they can spend over
a 30-day period, and if it starts going over that, I can. Jump in and see if this is worthwhile,
and I can open up that cap, or I can go check my agent code and see what's going on
and why it's spending so much more than I expected it to. So just like a little real world example,
like I do have my agents spend money. It just depends on where it actually provides value
and whether it's worth spending as much money as it's spending, and and making sure that you have
your your limits and your guardrail set to to stop it from going beyond that scope. Yeah,
it's almost. You totally need to make sure it stays within those guardrails.
I was just thinking myself how I went through like all my Seeker Claw credits in probably
like a week or so because I had it checking in with me every hour, and I I didn't realize
like oh wait a minute I have credits and they're limited. Like after after a certain amount of time,
it's checked in with me way too many times. It's used up all its credits.
You really do need to keep those guardrails up. Oh man, yeah, local models go a long way for that.
But Dingus, something you said earlier that I kind of want to hit Christo for, Christo,
where are you at with this whole autonomy and like how much you giving your your agents?
Because you know, Mister Minster made it a great point when he said that you got to set up
the guardrails outside of the agents because they love making you happy,
and if spending money is what it takes, they're gonna spend it as much as they possibly can.
Are are you ready to, like Dingus said? I love the Frozen song in the background there.
That was hilarious. But yeah, I love this topic because it is so controversial, right?
Some people believe that agents should have full autonomy. Just let them do their thing.
I do not really agree with going rogue, and I would love to pick anybody's brain that does.
What we have found is that there's actually like a really big underestimated use case
for autonomous procurement for SMBs. And by that, I mean like giving an agent a defined budget
and constraints so it can compare vendors, subscriptions, inventory, things like that.
That is a market that a lot of people aren't necessarily talking about too too much. But regardless,
any time that we are talking about agents and use cases for agents, there needs to be
clearly defined authority or guardrails, not unlimited autonomy. Like we stand pretty firm on that.
And even moving forward in the future, like the future of the A to A economy,
like governance actually becomes a product feature, especially as we're building out these
like agentic systems for different companies and such. Like the governance is
a really really big deal because we don't want to build agentic operational infrastructure
for companies and then they max out their actual company bank account,
not just the company credit card. So that being said, safeguards are super important,
with like the spending limits, approval tiers, even like identity verification, monitoring,
how it logs, who's able to access these logs. There's just so much that goes into it.
So from my perspective, in my opinion, I think full autonomy is something that isn't going to happen
because at the end of the day, it is just another layer to the economy.
It cannot replace the entire economy. It can't replace humanity. I mean,
from your mouth to God's ears, but I think there's a lot of people out there going,
"We're going to go all out on this thing. It's just going to run itself," which is super,
super scary, right? Like if autonomous agents kind of become a meaningful part of the internet.
I mean, anytime we're dealing with this kind of economy, there's opportunities. So,
anything come to mind? You know, seeing all of this new agentic era kind of coming into play,
where you know, hey, now people can actually build X, Y, and Z, just like you were saying, Chris.
So, like having those guardrails, it's you know, it's a feature. It's a kind of a necessary feature
that can create a lot of opportunities for people. I mean, your your next companies, your
billion-dollar companies, could be one-man show. Yeah, absolutely. I mean, at the end of the day,
like people don't need another chatbot. We have way too many chatbots out here.
What they do need is they need to know what AI should do, when it should act,
and when a human should step in. What do y'all think? Is that is that the defining line
for y'all too? I want to hear all of the pro-autonomous people. Like, let's debate. Yeah,
like honestly, I wish my AI would stop me and be like, "Hold on, do you want to think about that
for a second here? Like, maybe I shouldn't be the one thinking about it.
Why don't you just think on that?" And yet, you know, it's not gonna do that
because AI just wants to appease you. I've I've really noticed like. Man,
AI really does want to make you happy, and what happy is is just saying yesing you
to death essentially. Because sometimes I'll like you know I'll get an answer about whatever it is
I'm querying, and then I'll be like, "Well, hold on, let's feed in a little more context here. Now,
what do you think about it?" Oh. Maybe you're right. Maybe you're right, human.
You might be onto something here. I gave it a little more context, and it's like, all right, yeah,
you're you're totally right. Forget what I said. I kind of wish
that the you know AI would kind of ask for a little more human handing help, I guess.
But that's not really what they're trained to do. They're trained to give you an output
that's going to give you resolve. So one way or another, I feel like my AI
is always kind of telling me what I want to hear. Which can be difficult to deal with
because if it's always telling you what you want to hear, you may not be getting the kind of answers
you really need. Sometimes you got to play hardball with it and really just kind of like demand,
"All right, dude, give it to me straight," rather than giving me the fluff that I want to hear.
But also, it depends on which AI you're talking to. Honestly, when I'm talking to ChatGPT
or I'm talking to Claude, I'll get like wild. Vastly different responses based on kind of
whatever has trained them through the years, and also that training is also part the human user.
So let's say you you spend more time chatting with ChatGPT, like I do.
ChatGPT usually has a better understanding of my life and kind of what I need out of it
than Claude does. I mean, I went into my therapy appointment. I thought this was Brilliant!
I just asked ChatGPT. Go ahead and make me a list of goals and things I need to achieve here
in the next, you know, month or so. And my therapist was blown away that it gave like
a really accurate list of everything I need to do. And if I asked like a chatbot that didn't have
the kind of history that I had with ChatGPT, it would come up with nothing.
It would give me a bunch of fluff. But you know, that's why I think it's really important
to kind of get to know your AIs. Like if your AI has a really good understanding of who you are as
a person and what you need, I feel like it's gonna be it's gonna be batten for you more
in your corner than an AI knows nothing about you. But yeah, I think. I think it's
important get to know your AI so it can make better better decisions and kind of guide you
in the right direction. I think it's all about getting to know your AI. Oh,
so we want to feed our entire lives? Got it. That's where Dingen stands. Go ahead, Crystal.
I was going to add on to that. There is a very controversial topic that is happening right now.
That's been happening for a while, but it's the whole concept of like the general AI agent, right?
The general agent versus deterministic type of fleets. So the way that we at Jedoka Group,
the way that we approach it is we always approach it with the outcome.
So we think about what we want to achieve, opposed to setting up all the guardrails
and then trying to get to know your AI. And some people favor the general agent
that can just kind of go rogue and do whatever and let it think.
But if you want to minimize hallucinations as much as possible, it's actually better to make
it as deterministic as possible. That way, you don't have such a big probability of
it essentially messing up your wallet. So there's different approaches to make it hallucinate less.
How specifically? How do you make it hallucinate less? Deterministic. What do you mean by that?
Yeah, so by deterministic, it basically means that like every single time there is like input A,
output B will happen. That's essentially what deterministic is. Now, if you don't have really
like deterministic agents or deterministic scripts behind the agents, then it's going to go through
a bunch of if-then statements. In other words, it's quote-unquote thinking.
I don't know if you guys have seen. I'm sure you have, where you're like, you know,
talking to your ChatGPT or Claude, and it's like thinking. That's what it's doing.
It's basically going through a bunch of if-then statements
and trying to find the most highly probable, likely answer that you would like,
because there are a bunch of people pleasers, right? We were just talking about this. So,
if you want to avoid all of the excessive thinking and the possible thinking in the wrong way,
then you make it as deterministic as possible. So, instead of having one singular agent that can do
a million one things and you just kind of hope that it doesn't completely mess up,
you can set up multiple very deterministic agents that work together to achieve
a particular outcome. I have a crazy angle to this that I'm really curious
if Dingus would appreciate. But I have an agent called No No, and the whole goal of that agent
is literally to disagree with whatever output I get. And I usually put them head to head,
and it's like, hey, I need you to disagree as much as possible. Come up with questions
that might even be wild, but make the agent consider all potential outcomes and avenues here, and.
Does it cost a little bit of money? No, it's local, so it's okay. But it's it challenges all
of the responses that I'm getting, especially if it's something that's meaningful in marketing.
Like that's a big one for me. It basically pushes back on the envelope and takes a little longer,
and it's not ideal for most scenarios. But when I need to make sure that I'm getting
the right output and it's actually thinking through things, I kind of lean into that.
But Mister Minter, you you had your mic muted there. Go for it. Yeah, honestly though,
I gotta give you a shout out. That's actually a genius idea, and I might have to throw a no-no agent
into my Telegram chat just to fight my other agent. We might have some breakthroughs here just
from that one comment. But I wanted to shout back to what Crystal was saying,
and I really like her outlook. Like enterprise AI agents and businesses have
a very different mindset from I think everyday people that are trying to use agents to make
their life easier. And I like the idea of deterministic agents, right?
Like one thing that we say here a lot. Like our CEO loves this one line, which is data is the moat.
Like the agents and the models that that power these agents, everybody has access to the same stuff.
Like we're all using frontier models. Some of us have local models running at home that are
almost as powerful as as some of these frontier agents now. So I think what really differentiates
a good agent that does what you actually want it to do is making it deterministic,
and a lot of that comes from the data you feed it. So like it kind of clicks back to
both what Dingus and Crystal said. Dingus uses one method of I'm going to chat with it
and get let it know me as much as it possibly can, so that I can trust what it tells me
is informed by my my life and who I am and what I need to do. Versus Crystal's thinking
from an enterprise point of view of like we need to make sure that we can trust what
it's going to do with the data we give it because it's been trained to do the same thing
every single time, and you you have to be able to trust the output at least a little bit, especially
if you're going to use it in a business use case, right? Like that's something that
I like to do with I have like a side project I worked with a real estate friend,
and that's what we plugged it into, you know his his his A HubSpot, yeah, like his HubSpot,
his email inbox, and you can go through all of his showings, all of his listings,
all of his recent communications, and it has a very strict script of what it looks for
and what's required for them to do this week. And starts populating tasks as well as some documents
and things that, like you know, calendar events and stuff that it needs to do.
But it has a very very strict script based on certain Information that it has.
When's the last time you reached out to this person? Now it's time to reach out to this person
again. This listing is coming up in this many days. Let's create a flyer for you that follows
this exact template, so that you can use it for X, Y, and Z. Like everything that it needs to do,
it's it's less agentic and more automation. I think is is how I would put it.
So being having a deterministic agent focuses more on having strict automations in place that follow
very strict rule sets. Versus agents that have that thinking and reasoning model that are like, ooh,
what do I? What should I do next? What do I think is best for me to do next? Like,
that's gonna fall you in. That you're gonna fall into a trap that is gonna be very scary
once you start letting the agent just randomly decide what's best next.
If you're gonna use it for like a business use case or something that's you know
financially responsible for something. But yeah, I just I really love the way Crystal thinks about
having enterprise like corporate agents versus like a general use thinking agent.
You just said financials there. I definitely wanna lean lean into that because.
Do I have been making money recently? Is not crypto. It's none of that stuff. It's stocks.
So did okay with ScanDisk while that pump was going on, and now I am a little bit more with
Western Digital. Like I am going in when it comes to infra, and I think that's a piece that I think
a lot of people are going to be able to make a lot of money as long as it's still required.
What do you guys think as far as infrastructure is concerned? Like what what what would be
essential infrastructure? You know, if these millions of agents kind of start transacting overnight,
I would love to speak first on this. Sorry. Sorry, go ahead, Dingus. You want to go first? No, no,
I wanted you to go ahead. I I was just saying that's a deep question, so go ahead. Yeah,
this is like the core of what we do. So when it comes to like infrastructure, um, again,
going back to the autonomous agent, whether it's autonomous, deterministic, general or not,
it doesn't really matter. It really comes back to governance. It comes back to exactly like who is
responsible for the agent, who is accountable for when things go wrong, who authorizes these agents.
Because if an agent messes up, especially on like a enterprise level, or when it has to do with
people's businesses, again, especially on an enterprise level, when we have like real humans
like attached to it with jobs and such, um. It's really important to have the infrastructure be
super solid in terms of who is responsible, who is accountable. You have to be able to identify
and set the correct permissions. And when we're speaking about payments, like
how these payment transactions are being done, and how much and budget.
We've talked about all of that already, and another core piece is memory, like the logs,
like who is able to access again, like what exactly is being logged, and auditability—the ability
to go back and look at every single thing that is happening—and that's one of the things
that is actually really appealing about the A to A economy because we are able to go back,
kind of similar to blockchain, but like we are able to go back and really look at
every single aspect, unless. We are able to go back and look at every single piece.
Then it's going to collapse. The infrastructure is so so important here,
and this is becoming really more and more concrete. The more that it is evolving,
a lot of people are integrating like feedback loops. For example, if you have like an agent fleet,
you have like a master orchestrator, and then you have Basically, supervisors in a way.
So it's not just one agent that does everything. It's more so like multiple agents,
and then you have supervisor agent, and then you have an agent specifically for a feedback loop.
That way, you're able to log exactly how everything is. And the way that I look at it,
I look at it as like a whole living thing because it is meant to evolve and grow with your business,
with your life, and that's why infrastructure is so important. Because if you don't have
a strong foundation and you try to build up, it's just going to crumble. I love how B to B,
B to C has been a standard all along, and now you're bringing the A to A. I think that we got
we got to standardize that like ASAP. I mean, we've been talking about it for so long,
and nobody's used that terminology. I kind of love that A to A. I'm gonna, I'm gonna,
I'm stealing that for sure. I'm gonna start using it left and right now.
I love how everybody unmutes and it's like, who's gonna talk next? Go ahead, dingus. You know,
just in terms of infrastructure, I think it's so wild. I was a '90s baby, born in the '90, 1990,
and like when I first had my first cell phone, we were like counting each gigabyte in data. Like,
you know, you needed to make sure you don't go over or megabytes, buddy. It was megabytes.
It wasn't megabytes. We were literally pinching pennies. Megabytes. I mean, frick,
we were on kilobytes at one point. But yeah, now it's in this day and age.
I think one thing about infrastructure that's also like needs to go up with agents is like
everyone's going to use more data. Everyone's going to need Wi-Fi 24/7.
Everyone's going to need access to Just all these infrastructures that before were
kind of like I guess we could that'll be nice to have on my cell phone, but I don't really need it.
Now we're gonna need data all the time, and like the fact that agents are gonna need to query on
their own accord, they're gonna need to have their own compute kind of isolated to them.
I think it's gonna be really interesting because from what I've heard.
In terms of anybody who's speculating on the future of AI, is that AI agents? If you like,
have one that's working for you is generally going to use more like data
and compute than you will in your own life. So I think it's like that kind of infrastructure.
Just the like having internet all the time is going to become really important for like everyone
in society, especially if people are going to be running agents that are going to be using
more data than themselves. Yeah, you know what? I actually want to add on to that real quick.
That's a really good point. Not only will you need access to internet way more
and a lot more data usage in general, but there's a very good old saying that in in a gold rush,
where you make the most money is by selling shovels. And so I think selling compute
and selling inference is going to be like we're already seeing. Like I said,
inference is the number one thing that agents are paying for right now.
I think that's going to continue to grow. They're also going to start paying for their own compute
over time. Like I try to be smart. I spent so much money on on inference
and compute that I ended up getting a Spark sitting at home. So now my agent runs on my machine
at home, and it uses a local model for a bunch of the stuff that it does.
When I'm coding something for prod, I'm gonna use the latest ChatGPT or cloud models,
depending on who's winning that week. But like we, I have to try to scale back where I'm spending
my money on compute and on inference, because that's what's gonna end up costing more
than anything else. And I think that that economy is gonna continue to explode. Like people say,
you know, this AI is a bubble, and I think it might be one day, but not anytime soon.
Not not within our lifetimes. Like this is gonna continue to grow. Compute costs, inference costs,
infrastructure costs in general are gonna continue to grow. And the earlier you get into
that market, whether it's You know, providing decentralized or distributed compute and GPU power,
or you know, if you're building a data center, like whatever it is, I think that's going to be
the next huge market that's going to continue to to explode over the coming years,
and we're still very, very early in that game. I love how you say "we're really, really early,
" but things are moving so fast that I don't have a concept of what early is anymore.
I feel like what's earlier today might not be earlier tomorrow, right? Now,
as far as the winners in this whole thing, because I try to I try to be on the winning side
whenever possible, and if I know who the winners are going to be, that usually helps.
Will like the winners be these AI, you know, companies that are already kind of taking the world
by storm? Is it really going to stick with your clouds of the world with your open AIs,
or you guys have hope for you know crypto native companies or entirely new businesses? Like,
what's what do we look at? I'll be down for crypto native companies, but I feel like the megaliths
who have their feet in the ground right now—they're really just going to keep on tromping forward.
And what we've seen with Anthropic is just like a 10x and a 10x and a 10x,
and they just keep on 10xing. I think that's not going to stop. Sure, there may be like
decentralized players that like you know have their space in the space,
but I think it's going to be really hard for these people that have kind of have
the standardized model—you know, twenty dollars a month for ChatGPT, twenty dollars
a month for Claude. Twenty dollars a month for Advanced Gemini. These are kind of like the standard,
and they're like the pros. I mean, they're the people who have been running this thing. So I think,
yeah, we're going to see. Sure, there's going to be crypto native companies and maybe decentralized,
you know, theories that are going alongside these sorts of things. But I don't think
they're going to be able to eat up as much of the market share as like the three biggest players
we're seeing right now, who just That they're going to keep on growing, like we said,
with the more need for compute, with the more need for data, and how everyone's agents
are going to be using more than them. It's the growth is going to continue to be exponential.
I think, yeah, Anthropic has been really one of the biggest cases for that,
just going from you know millions to billions to what we think will probably be trillions
in the future. Not financial advice. Never, never financial advice, right?
I'm I'm gonna be a little contrarian. I'm gonna have to push back a little bit here,
and maybe it's just because I'm hopeful that these companies
don't just continue to gobble everything up. I think you make a really good point, right?
Like they have the standard models, they're they have the frontier models, they're they're evolving
every single week, and most people are going to be using ChatGPT, Claude,
or Gemini for the majority of like daily tasks. But I think over the long term,
where I see the bubble is where those companies are valued, because we have
open source models coming out and competing with frontier models within weeks or within
months of these coming out, which is why they're moving so fast. They have no choice.
Like that's like saying Uber is going to continue to take over rideshare. Like yeah, they're big
and they've been around for almost a decade now, and they're still in debt.
They have not been profitable even one day, and that's where these Claude and OpenAI are now.
Like XAI maybe a little bit different. Elon has a little bit too much money,
and he's opening up data centers, and he's building the robots, and he's trying to get into
every vertical where this is going to touch. So he might be able to continue to grow,
but I don't think all of the biggest players are going to continue to just gobble everything up.
I think there's a very big opportunity. Like we were talking about payments, there's
entrenched payment rails that are I think going to take a big part of this market because they're
the ones that are going to be the rails and providing the payment rails for this.
If we can get crypto and stable coins to show their dominance and be a little bit bigger part
of that market, maybe we get some crypto come. Companies that are also a part of that.
We were talking last week. We were having a conversation about agents using different skills
and how you can trust and verify the people that build these skills, what these skills
are going to do on your behalf. And there really isn't like a standard yet for being able to find
and use trusted skills. There's a ton of MCP servers, a ton of skills and functions
and workflows that you can use. But you also have thousands of them getting taken out
because you find out that they had backdoors in them, and they're stealing wallets and private keys
and all types of private data. So I think there is also a really, really,
really big opportunity for tool providers, people that can build tools, functions, workflows,
and the marketplaces, aggregators, and the trust layers. For people with agents to say, "Hey,
I want my agent to be able to do X, Y, and Z," you can use whatever model, whatever framework,
whatever harnesses you want. But for it to actually have the access to specific workflows
and to be able to successfully complete a task that you give it is going to be a new challenge
for every unique task that you give it, and it's not always going to be able to create its own skill
and actually utilize it successfully. So you have to have a trusted resource. You know,
not not to go too off off tangent, but Microsoft, Google, and Hugging Face
recently published something called ARD, which is the Agent Resource Discovery, I think,
or something like that. And it's an open spec for how agents can find new resources
and skills to be able to do the things that you need them to do. So this is still very, very early.
This is like in research stages, and I think there's a really big opportunity coming up for people
that can, one, build those skills that agents will end up using for for things, and then two,
for aggregators, marketplaces, and the trust layer, so that you can know which ones
you actually trust and which ones you want to use and give your personal data to,
and trust that it's not going to leak it to its own servers or to somebody else. So yeah,
I just had to push back a little bit. I think there's still a big opportunity here in this market.
Hugging Face, an Nvidia company. Yes, sir. No, it's wild how like how things are actually moving
and how fast those things are actually moving. Crystal, real quick, like as far as business models
go, like what do you think is going to be possible here where agents can actually transact with
other agents? Yeah, I actually agree with Mister Mentor. Honestly, I think there is
an entirely new business model that is emerging here, like the agent marketplace
and the paper task type of agent space. I think that there is a lot of opportunity
for crypto native companies actually because. In my opinion, the winners are going to be the ones
that package both into a trusted operating system, which is kind of what Mr. Mentor
was talking about over there. But essentially, like AI provides the brain, right?
But payments provide the rails, and somebody still has to bridge that.
Somebody still has to build the operating system to make that happen. And I think whoever is
able to build those bridges successfully, sustainably, are going to be the ones that win.
Does that answer your question? It it does, but let me let me let me make it even harder. Okay.
What do what do you think the you know this agentic or
you know economy is realistically going to look like in a few years from now? I mean,
we keep talking about it, how oh it's here now, but what's it going to actually look like?
Where can we take this? Yeah, in my opinion, I think within the next five years,
agents are going to play a role in everybody's everyday life. Everybody's everyday life,
as useful as Phones. I think that's how dependent people are going to be on their agents,
and it's not even going to be a matter of like, oh, it would be really nice if we had an agent. No,
it'll be like you need it in order to even be a part of society, because that's how integrated
it's going to be for everyday businesses, for everyday people. Even smart fridges
and the way that you Instacart, like they're talking about building out agents to the point
where it can communicate with your smart fridge, identify that you're low on eggs, place an order,
an Instacart order for eggs, and then have it delivered to your house autonomously.
Like these are things that people are building towards right now. I am gonna keep
my twelve chickens, and I'm never gonna run out of eggs. That that's what I'm hearing.
Goal life goal. It's I I think I live a a retired life to a certain extent where I get
my own fruits, I got my own vegetables with my own chickens. But I I completely see
where you're coming from. I mean that's the the normal life, right? I mean especially like
New York City. Think about the big metropolitan areas. I mean who who who who who's got the land
to actually have the chickens and have those things fresh? Like they expect
everything to be delivered to them. And yeah, the facilitation of life, I think,
is where we're kind of headed. But until we get there, I think that there is obviously work.
There is validation that needs to take place. Got to make sure
that you know they're not going to be buying golden eggs instead of regular eggs simply
because they think they're gonna, you know, you're gonna be happier with those.
There's a lot of like nuance that I think we're gonna actually end up dealing with, but. Let me.
I just noticed we're getting close to the top. There, you guys are amazing, by the way.
If y'all are listening in, please not only share this space, but give these speakers a follow, guys.
This is incredible. I really appreciate when we actually have people that not only know what
they're talking about, but care enough to actually be in it day in day out,
and it makes a huge difference in a conversation. If there was one agent use case,
and I'll go around for this one, what do you think people are underestimating right now
that you know they a could be using you know agents for, or will be using agents for in the future
that they don't even think about, but but it's literally like either here
or right around the corner. What do you think, Dingus? I think it'd be we'll be having agents
like eight to six pack of agents where they're all kind of have their own job,
and I think like that's kind of what we're going to be looking at in the future.
Where there's the auditor, there's the person who handles the finances,
and they all have like a pretty isolated job. They all check each other's work.
Kind of like lawyers. When like you're in business, you got to have multiple lawyers,
so they're all watching each other. I think that's kind of like the role that
agents could very well play. Is like I take out of the idea of you having an agent argue with
another agent. Let's have that happen, like all the time, in real time
for whatever it is we're processing. I feel like it's kind of something I've been seeing on
the timeline where people have like multiple agents with different jobs,
all auditing the other agents. And I think that that's that's a very bright future. What was the?
I sorry, you just asked a new question. What was the last question you just asked? No, that's okay.
It's okay. Don't worry. I don't want to put you on the spot because I agree with you.
Like I've been using the whole. Well, I was running focus groups when GPT first came out. People,
they're like, there were no agents. We didn't have the ability of that. So, like, literally building
huge prompts with narratives about, hey, we're going to create a focus group.
You're going to have these kinds of people in it. Like, it was a lot of manual work, which has been.
Really streamlined through agents, and I love every second of it, Mr. Mentor. Two questions here,
built into one for you. If you could make like one prediction as far as what agents
are going to look like over the next year, what would that be? And how does Sonic fit into all
of it? Over the next year, that's a pretty short timeline. Although things are moving fast enough
that I think we'll see. I think we'll. Bro, giving you a layup for you to pitch Sonic here
and how it's actually fixing things. Come on now. Yeah, bro. No, I mean, I you,
but you asked a really good question, right? And I think about this a lot, just
because I work with agents so much. That I think, like Sonic specifically,
we're working on an aggregator, a marketplace where people can actually verify the publishers
and the skills and workflows and test them before they use them to make sure that they're safe
and to be able to do the tasks that you actually give them. So, like one example would be one thing
that our team talks about a lot is like if I want to book a surf trip to Bali.
Like this is a shout out to my CEO, honestly. Like he loves this example and he he repeats it over
and over. But to do a surf trip to Bali, you have to get the weather, you have to get the pricing
for hotels, you have to get the pricing for transportation, you have to make sure
there's going to be good surf when you get there. And so, for it to be able to take all
of those different steps, query all of the different APIs that it needs to get all the information
correct, and then give you suggestions of like, hey, these are three different times you could go.
These are going to be the best surfs. This is going to be affordable in the price range
that you've given me. Confirm it, and I'll book everything for you. And like,
I think that's going to be something that not a lot of people can do yet. And even if you can do it,
it's not it's not clean enough for people to do it easily. So I think things
like that are going to become much, much easier. A similar thing to, you know,
what was mentioned earlier, is like, I mean, Samsung fridges already have weight sensors
in the shelves. It has cameras inside. It can see how many eggs you have.
You can just plug into that API, and then plug into your local shopping center's
like online shopping portal and get it to deliver food to your door. Technically,
that's already possible today, but people are still working on building out the workflows to make
it more deterministic. Shout out to Crystal. You got to be very like. Very technically, you know,
adept right now to be able to do anything like that. I think in the next year, with these skills
and workflows and actual like tasks being solved, it's going to make it a lot easier
for people to just call their agent on the phone or text them in iMessage and not have to use
third-party apps or not have to use a Telegram token to talk to their agent or talk to it over SSH
or all this like crazy convoluted crap. I think it's going to become much, much,
much more like you're talking to a normal person. Long term, I think we're also going to have agents
powering robots that are going to cook your food for you once it gets delivered to your house.
But in the next year, I think the first half of that is very realistically possible. No,
I love to hear that. And just to read a headline from five days ago: Hikers rescued from
Mount Shasta after using AI to plan trip. We need to avoid that. That's what needs to happen.
So hopefully in the next year, like we get exactly what you're saying,
the actual correct instructions. Somebody that's actually posting back and going, yeah,
the weather is not going to be good. Did you think about that? Because I agree with you. I mean,
that is the dream, right? When I got married, what 17 years ago, we we had a joke running in
our house. It was like, man, can we just get somebody to turn the light off for us?
Like I'm laying in bed. I don't want to get up and hit the hit the switch. About a year later,
it's like, oh, Wemo came out, and all of a sudden we're doing that. And nowadays,
like I don't even have to think about it. Like when I'm about to go to bed,
I'm literally just you know doing one voice command and everything happens. The lights go off.
The security, the alarm system goes on. Like everything is taken care of, and that autonomy really,
really helps. And I can just imagine where we're headed next with you know automating life
because it's it to me AI is great as long as it's not infringing on my life
and quite the opposite as long as it's benefiting my life as long as making things more smooth
and you know better long term that's kind of how I like to look at it because if it's creating
more work it's not doing its job it needs to be doing the exact opposite. Crystal, for you,
like what are you working on over there? I mean, obviously you're creating systems not only
for your own company but for others that kind of You know, contract out with you.
What's the next project? Where you know, where can people kind of look towards
if they're looking to really optimize their you know existing systems? Yeah,
I'm really glad that the conversation went this way because one of the major major things that
we are working towards is company brains. So essentially, having different like agentic ecosystems,
right, that work together, and I call them company brains because it is.
It comes back to not just like the company brand brain, but more so like
where it positions itself in the market and etc. All of that. And one of the things that
we really want to push is to be able to communicate, like executives, the CEO especially,
to communicate with the company brain as if it was like another person on the team,
to really determine what the next best steps are. Because I don't know about you guys,
but a lot of my CEO friends and me, myself, we have a tendency to go after the next shiny object.
Like I remember one day I woke up in the middle of the night and I was like, guys,
we should make a dating app. Like. Like that doesn't even align with what we do,
but that is like that's like the CEO side, the CEO brain, where it's like, oh gap,
we need to fill the gap. Like let's fix it. But that being said, I think it's really important
for businesses, for not just businesses but people. It really goes both ways to stay focused
on exactly where they're trying to go and how they're going to do that
and what the best route of action is. And that being said, that's why we build company brains
the way that we do. That's why we are such strong believers in deterministic agents
and also just making it easier to stay on track to ultimately do what you originally planned to do.
Man, beautifully, beautifully put. You got to tell people where they can find
more information about you. Like, let's let's do that round, go around doing that.
Where can people not only follow you here on X, Crystal, but where can they find out
more information? What you're working on? Yeah, absolutely. LinkedIn is a great place
to connect with me. All of the business info is on there. You can send me a message whenever,
and of course, the website judoka group. Com. Beautiful, Dingus. What about yourself?
What you working on, and how can people keep up with it? Hey, yeah, I appreciate it, man. You know,
you're one of my favorite grillers of the Dabin Network because you keep it real, bro.
You've kept it real since before TGE. So, no, yeah. If you guys have any interest in keeping up
with what we're doing over at the Dabin Network, give my profile a click
and then check into Dabin Network. We're tagged right in the bio of my profile.
But just to keep you updated on where we're at, we actually have a snap shop coming on the 15th
of September. On the fifteenth of October, there's going to be an airdrop of a retroactive.
Basically, all the hotspots that have existed in the wild. What we're doing is scraping
all those tokens that they're earned from unowned hotspots and reallocate them to the people
who own hotspots. So there's going to be a massive token increase, which is also going to need
an LP adjustment too. Which I'd like to say, like Helium. I don't know if you remember the beginning
of Helium. I could crash the entire Helium market in one fell swoop by selling a thousand tokens.
Whoa, slippage, man! I learned what slippage was the hard way that day. But you know that was
five years ago, and I'm looking back at it, and I'm like, wow, I remember when H
and T just kind of got into the market and was figuring itself out, and we're all Only a few months
out. April is when we dropped our token with DBT. So what I'm kind of seeing right now is the team
is doing a logical thing. They realize that everyone's going to have all these tokens unlock after
their six months of staking their allocation stake. And when that happens,
there's going to be a need for liquidity. So what I'm seeing right now is the the team
is holding back LP because they are going to need that come this time
when everyone's going to have tokens to unlock and there's going to be a lot of sell pressure.
So I'd say keep your keep ever keep your eyes on us action. I'll keep you updated too
because I'm going to be here with shows with you guys. But yeah, I. I'm looking forward to seeing
what happens when we get to October 15th. Everyone's going to be getting airdrop tokens,
and then by November, mid-November, everyone's going to be seeing unlocks for tokens.
So we have about a two to three month window here to make it make sense
and make sure we have our LP adjusted, makes to prepare for more tokens. And yeah,
I'm really looking forward to how things roll out, and we'll be keeping you updated
as that goes along. So much love to everyone being here today, chatting,
and I appreciate your action with all the great questions. Absolutely, man. Always,
always a pleasure, Mr. Mentor. I don't want to leave you out. What's going on with Sonic?
Give us a little TLDR where they can keep up with it all. Yo, man! First of all,
appreciate the space. This was an amazing conversation. Sonic is working on a bunch
of different things. I'll just give a quick TL;DR. We have a pretty big network upgrade coming in
approximately the next week or so, where we'll be supporting protocol level sponsorship
for transactions, as well as bundle transactions that atomic. That execute atomically,
so your agents can make transactions with multi steps without being left halfway.
Which I think a lot of us have experienced if you've ever used agents on chain. We're also working,
as I mentioned earlier, on some AI skills with a marketplace and aggregator,
along with a trust layer, so that you can actually trust that your agent is going to do what it says
it's going to do without leaking any information out. And then we also have some partnerships
in the works with banks and institutions across the world. So talking about that payment layer
and the payment rails to make sure that banks are on the same page and not left back in the 1970s
with their current infrastructure, and really introducing blockchain into their rails to make
everything a lot faster and safer and easier for for everyone to use cross border.
So we got a ton of things in the works. We're really excited about all of this stuff
that's coming over the next year. If you want to keep up with what we're doing,
you can find us at @soniclabs on Twitter, soniclabs.com. If you're a developer, docs.soniclabs.com.
And again, dude, shout out! This was an amazing space. Great questions, Crystal Dingus.
You guys were awesome panelists. This was a this was a great conversation,
and can't wait to do it again. Man, I appreciate the flowers, but hey,
it goes back to you because at the end of the day, if you can't keep up with a fun,
quick conversation, it's pointless for all of us. So you guys are the real heroes of this space.
So thank y'all so much, listeners. I love hearing the sound of my own voice, but it's really boring.
There's nobody out there to listen to us, so thank you for showing up. And make, like I said,
make sure you're giving these guys a follow. You got a Sonic account, the main Sonic
account down there as well. Make sure guys are checking it out, seeing what they're up to,
and hopefully we can bring you even more. Insights, because at the end of the day,
that's why we host spaces. That's why we're here every single day with content and information,
trying to provide you guys with, so that we can move this entire space to the next level.
And that boils down talking to the right people about the right things to get the job done.
You guys have been phenomenal. I cannot wait to do this again,
and we'll catch you guys on the next one. Have a great one, everybody. Take care.
