Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Let's dive into it today. You've heard of it. It's
probably brand new on your radar. It's called x four
O two and it has exploded onto the payment rail scene.
We're going to break it down for you and let's
just jump into it today. We brought one of the
senior engineers over from Coinbase to break this down, Eric Repel,
who's coming in from Coinbase, the developer platform.
Speaker 2 (00:20):
Thanks for jumping in. Great to meet you, Paul.
Speaker 1 (00:23):
All right, So a couple of things I want to
get into first. For a lot of people, they may
not completely understand what we're dealing with right now around
the explosion of x fo O two and what it
means for payment tech AI agents, how all this applies.
Give us a rundown of what this is in simple terms.
Speaker 3 (00:45):
Yeah, so X four two is an open standard for
Internet native payments, and so X four two really is
the most boring piece of tech you will probably ever encounter.
But I think people don't realize that the Internet is
entirely based on these standards that everyone agrees upon to
use HTMLHDP, CSS, JavaScript, These aren't controversial. There actually isn't
(01:09):
a standard for payments that everyone just agrees is the
one you should use. And so what x RO two
is is like an open, neutral standard for communicating between
two computers how to perform a payment. But what it
unlocks is what's so exciting about it. Because now if
you can move money in the same way you move data,
and everyone speaks the same language of how to move
(01:31):
that money, you get like these crazy unlocks of agents
being able to pay other agents, programmatic payments from different
computers to different computers, and that unlocks micro transactions and
all sorts of different use cases.
Speaker 1 (01:44):
This is kind of almost a residual of tim berners
Lee in terms of the technology that you're talking about
in terms of use case. Is it only been unlocked
because of the fact that we have digital money now
in the format of something like a stable coin.
Speaker 3 (02:01):
Yeah, that's actually exactly right. So X four oh two
is a nod to the four H two status code,
which was enshrined in I believe it was nineteen ninety
seven under the HDP one point one spec And I'll read,
I'll tell you the entire spec of the four H
two status code. Four H two. The entire spec of
it is payment required. Four H two is reserved for
(02:24):
future use, and there's been a lot of people. That's it.
That's the entirety of the spec There's nothing else, and
so there's a lot of people who have tried this
over the years. Market Jurason spent a non trivial amount
of time trying to get payments to work on the
Internet in the nineties, back in the Mosaic days, back
in the Netscape days. Brian Armstrong tried to get payments
to work with bitcoin native to the Internet in twenty fifteen.
(02:47):
Bology Shunivasen tried at twenty sixteen. There is definitely not
a new idea, But I think the reason as to
why now is one, the technology has gotten way better
and you now have stable coins and the mechanisms of
payment that are like better, faster, cheaper, subtle, instantly, And
then I think even that may not be enough of
an incumbent for us. The second one is that you
(03:07):
now need programmatic payments because we have AI entering the
ecosystem and being a new economic actor on the Internet.
Speaker 1 (03:16):
Well, and I think what's interesting about this is it's
all converging at the same time, and it's what's also
somewhat eloquent is it's using technology from the nineteen nineties
to be able to execute on what could be the
next generation of technology around AI. I'm just looking at
X four h two X scan and this was kind
of the explosion that occurred right there. Coinbased, of course,
(03:39):
leading the way we were trying to get the payai
guys on. But that's a good explanation on it. But
why did we see the explosion over the last couple
of days. What was the catalyst that caused this?
Speaker 3 (03:52):
I think really the catalyst was Twitter noticed, and it
just it became front and center of everybody's mind very
quickly that oh, X four h two is a thing,
and then it's extremely easy to adopt. X four two
is a standard. You spend a lot of effort making
it so it can you can accept payments with X
four two in less than twenty minutes. You ended up
(04:12):
seeing a huge explosion both in transaction volume for a
bunch of different use cases, some crypto related, some not
crypto related. And then I think the number of servers
that you can pay I think at almost one hundred
X at least at least twenty five X that have
to go look at the stats and the number of
sellers more than one hundred x and so you kind
of had this moment of a big, huge amount of
(04:34):
attention and crypto Twitter kind of discovering it, and then
people actually taking the time to think, oh, maybe I
should build something on this, and then a bunch of
new use cases popping up.
Speaker 2 (04:43):
I'm looking at diagram coming over from X.
Speaker 1 (04:45):
This is kind of the interesting point right there, before
taking a online payment and how it works with an
AI agent versus what this would do for it. Talk
to me about that, because this starts to change the
very nature of how payments can be done, especially when
we're talking about AI agents being involved. This changes everything
(05:06):
for companies like Visa master, anybody that's doing online payments
right now, which has been using old technologies as thirty
years old. How does this help the modern day financial
system as it is today? The day has to just
adopt this, and they're you know, they're immediately caught up
to where we are now.
Speaker 3 (05:25):
I think one of the things that makes X four
two really powerful is that it's not coupled to the
underlying payment network, and so you know, quin basically think
stable coins are the best way to move money, but
there's nothing that would stop Visa MasterCard, or you know,
any payment mechanism even ach from being expressible and usable
via X four O two. The amount that you can
(05:46):
pay may be different, the like limits may be different.
But as a all X four two is a is
a way of expressing how a payment should be conducted,
and you could you could easily say that payment could
be conducted on the Visa network or any other payment
network in the world.
Speaker 2 (06:00):
Okay, and what.
Speaker 3 (06:01):
State, And then you marry x roy two with stable
coins and you get subscent transactions, and you get these
micro payments, and you get all these payments that suttle
in real time. But X row two you can easily
imagine a situation where you have an AI agent that
is paying like one cent per request to do better
research for you, to like research where you want to
go on vacation, and then you do like a last
(06:22):
mile larger purchase using a credit card to buy the
actual plane ticket.
Speaker 1 (06:26):
Well, that's interesting because obviously that's kind of the new
granular method of how AI agents can really execute in
terms of drawing in data. We're starting to see it
now on token based systems around AI, especially as we
see a lot of these lms kind of explode in
terms of their just the narrative of what they have
to look for in terms of the syllabus. When you
(06:49):
look at this statement right here, This came in from
A sixteen Z. This was an interesting component where he
said crypto may be the only viable payment rail for
AI agents and showing x.
Speaker 2 (07:01):
RO two as a possibility.
Speaker 1 (07:04):
First of all, is that do you think that is
going to be the case for AI agents or do
you think there might be other technology that could eventually
adapt to the speed and the demand for AI.
Speaker 3 (07:16):
I think, as of today, stable coins on modern blockchains
are the fastest, cheapest, quickest to settle payment mechanism that
exists on the planet. In the future, maybe these payment
networks adapt their technology or transpiration from blockchains and kind
of modernize and reduce things like fees and so on.
But right now I think I agree with that statement. Yeah,
(07:38):
I think that stable coins are the best way for
AA agents to transact.
Speaker 1 (07:43):
Yeah, and it's moving quickly, Eric, when you look at
what this could potentially unlock, you know, as far as industries, obviously,
we saw the coinbased partnership with City here recently, just yesterday,
there's a lot converging on the finance side of it.
But could this also unlock other areas of digital use
(08:05):
cases out there?
Speaker 3 (08:08):
I definitely think so. If you think about it, you've
probably never made a purchase on the Internet that's been
under a dollar. Yeah, because one, there's a lot of friction.
You have to enter a credit card and you need
to it's just a bad ux. But to the these
networks haven't really been designed for those low dollar values.
And so you can imagine a world where instead of
(08:30):
needing to pay for subscriptions, you can just pay for
the content that you or your agents are actually accessing.
And you could pay it ten cents, or you could
pay one cent, you could pay a dollar. And we've
already seen a product that Coinbas shipped a couple of
weeks ago that flew under the radar is something that
we call payment MCP and what it does it gives
(08:51):
Claude or any MCP compatible AI as a stable coin
wallet and now it can make autonomous purchases on your behalf.
And what we've seen is you can use that to
compose to get more out of these AI agents because
now they can access more data, more context, and perform
more tasks than they were able to before.
Speaker 1 (09:13):
Yeah, there's the MCP wallet right there, just showing Anthropics
Claude integrating it. So this is integrated now. And you said,
this only takes about twenty minutes for somebody to roll
this out, which is an example of kind of the
explosion that we saw in terms of merchants that have
already jumped on this. What would be the big area
(09:34):
that you think would be a huge uptick on this?
Would it be gaming? Could it be e commerce? Where
do you think the growth will be?
Speaker 3 (09:43):
I think the growth is going to be in areas
that we don't even see coming because the Internet is
really not composable right now. Yeah, everything kind of exists
in these silos, and AI has the potential to make
it so you can combine multiple sources of data, multiple
services into singular workflows in a much more, much more
(10:06):
robust way than has previously be impossible. And so I'll
give you some concrete ones. It's like you can imagine
your AI agent going out and paying for a bunch
of data on a site like LinkedIn and giving you
paying a dollar or something to access a bunch of
profiles of candidates, and then going and searching the web
for all of their web presence and then giving you
an analysis based on their social media and their web
(10:28):
presence of do they think this candidate is going to
be a good culture fit. And that's like a novel
use case that you can just kind of do because oh, now,
if you can pay to access that data, you can
compose it with other capabilities these ais have.
Speaker 2 (10:44):
Well.
Speaker 1 (10:44):
And also I think you look at the potential here
around subscription models that are existing in the media space
right now, because that is a I would say it's
a huge barrier right now in terms of content. I
know a lot of the publishers talk about this bit.
I go to publisher conferences all the time, and they're
always getting in on, you know, how much they're going
(11:05):
to embrace AI.
Speaker 2 (11:06):
The problem that's always brought up is micro.
Speaker 1 (11:09):
Payments of being able to just say, okay, they only
want a few stories out of my website, but I
only have you know, a monthly or an annual program.
So being able to do this now would be pretty
significant around this. What about the server growth expectations? Does
this change.
Speaker 2 (11:26):
Anything right now? I'm just looking at the marketplace.
Speaker 1 (11:28):
Active merchants have exploded new registrations are moving fourteen thousand
new merchants. What is this going to look like in
just in a few months.
Speaker 3 (11:39):
I mean, I think if we can keep writing this,
you'll we cross a million dollars in volume two days ago,
and then I believe it, day later, we crossed two
million dollars and so okay, if the trend continues, and
I hope it does, like we'll probably be seeing a
billion dollars of commerce through X four h two per
week in six months.
Speaker 2 (11:58):
So this is exponential.
Speaker 1 (12:00):
I mean, this is a major onlock for how quickly
do you think retailers, finance, media companies, banks will will
pick this up?
Speaker 3 (12:10):
I think traditional companies can take a while. And companies
are starting to get more and more comfortable with stable coins. Yeah,
and so I think that I think progress is only
going to continue to accelerate. Will there be a dip
at some point of course? Right, like the these things
always kind of have a bit of a trend, and
you know that X four two is really there's a
(12:31):
lot of attention and excitement about it right now, and
inevitably people are going to it's going to become mundane
and then we'll see, like the real utility as more
and more people come online. You might have caught about
a month. A month ago, we announced that we're moving
X four two into a foundation, and cloud Flair and
coinbase will be the two first members of the foundation.
(12:51):
An example of that is cloud flair is going to
be using X four two to enable pay per crawl
there on any website hosted on claud flair, which is okay,
twenty percent of the Internet, and so now if you
host your content on cloud flair, you can pretty easily
enable paper crawl and accept payment VX for it too,
which will.
Speaker 1 (13:12):
I wonder if that would translate to a lot of
the hosting platforms that are hosting for I think Squarespace
is on cloud Flare, which is an example of e
commerce platforms that could launch forward pretty quickly. They're doing
payments now, if they got into micropayments, this would be
interesting for Squarespace, I think to maybe roll something out
(13:32):
like this, and they're already touching I think, what is
it a few hundred thousand sites a week in terms
of growth, so huge opportunity in terms of where this
might go around paywall and subscriptions. Let's talk about the
UX because this is another area that I see as
kind of an inhibitor. Right now, at least on Web three,
(13:54):
most of the time it's been wallets. You've got these
pop ups, you know, this Web three integration, which has
been very clunky. It's very clunky. Does any of that
change with this kind of integration.
Speaker 3 (14:07):
I think wallet experiences for humans are getting better and
better every day, and wallet experiences for agents are significantly
better than the alternative. Yeah, because there isn't really a
viable alternative, and so giving an agent that you code
a wallet is actually very easy. You just say, hey,
here's you integrate any number of open source SDKs, you
(14:29):
could use CDPs products. It actually is pretty easy to
give your agent a wallet. But at CDP we try
to make that easier and easier every day, and make
funding that wallet easier and easier, and giving your agent
controls and making it safe for your agent to transact
on your behalf using their wallet easier and easier.
Speaker 1 (14:50):
I'm looking at this layout right here where this particular
individual built an AI agent to compete in real time
on auctions. This is reference to NS domains right there,
so if they're in name service. But that's kind of interesting.
That this is interactive. Now, imagine where this could go
(15:10):
around an opportunity like this and back to your point
getting agent agent commerce, because this is a topic we've
covered quite deep on the show. Here is there's an opportunity.
Now it seems like this could be a framework to
help all this become reality.
Speaker 2 (15:25):
What could this unlock when.
Speaker 1 (15:26):
You have kind of this whole AI agent bidding process,
what do you think it can unlock?
Speaker 3 (15:33):
I think it unlocks money moving the same way data
moves on the internet, where there's no friction, money moves
trivially and freely, and you can make payment really naturally,
and you just it's much more budget based and much
more kind of you don't need all these pop ups
and confirms. You just start giving value to the places
(15:55):
that you're receiving value from.
Speaker 1 (15:56):
In line, do you think, Eric, we're what three years
five years away?
Speaker 2 (16:01):
Companies?
Speaker 1 (16:02):
Because the traditional banking system is pretty much the gatekeeper
right now of all money moving online, especially when it
comes to e commerce. The fees, the merchant fees, the
speed problems. Don't even get into cross border settlement. I'm
just talking about dealing with you know, paying for something online,
your Netflix bill, so all of that starts to change dramatically.
Speaker 2 (16:23):
Here.
Speaker 1 (16:23):
Is it a three year five year runway or do
you think it could happen faster with the fact that
we've got so much AI coming online.
Speaker 3 (16:31):
I think that AI has just moved faster than everyone
has projected and predicted. Everything has been faster, and so
it could be six months, it could be two years.
But I wouldn't be surprised if if there was like
multiple billions of dollars, multiple billions of dollars of agentic
transactions happening next year. I think Mark, I think A
(16:53):
sixteen Z and the recent state of crypto projected a
I believe it was a thirty trillion dollar payment market
for agentic commerce by twenty thirty, and so that would
imply like an incredible rate of growth between now and
five years from now.
Speaker 1 (17:10):
Yeah, So I guess the question is could the missing
piece be gaming. We were looking at pay AI's team
kind of what they're doing in terms of unlocking this
using X four O two and being able to get
into what is somewhat micropayments, real time settlement, et cetera.
Do you think that would be the potential gateway for
(17:32):
all of this to maybe take off as the payment ecosystems,
I mean the gaming ecosystems.
Speaker 3 (17:38):
I think gaming is like a really natural fit because
it has so many micro transactions, like so many games,
you're buying loot boxes or skins or other things for
the game, and each one of those is really just
a micro transaction, and right now they need to represent
it as like credits or you know, in game gems
or something like that, because you can't actually just perform
(17:58):
a small payment. And so I think that that definitely
for consumers is like one potential avenue reis. I think
it's going to be utility and composability of different services
and data and so on in order to give AI superpowers.
Speaker 1 (18:16):
Yeah, well, this is something that is starting to I
think catch a lot of people by surprise in terms
of how quickly AI, especially on the Augentic side, because
we've talked about it for I think better part of
a year now now we're actually seeing it. Use case,
I'm looking at the ecosystem market map right here, I'm
looking at chains and protocols. Obvious do you think we'll
(18:38):
see first of all, which blockchains do you see benefiting
most from X four oh two?
Speaker 2 (18:43):
Right now?
Speaker 3 (18:45):
I think all blockchains that lean into X four two
will see benefit because each blockchain kind of has their
own ecosystem. And one of the one of the designed
decisions that I made very early with X four two
is it's not as zero sum game. You as a
server can accept payments on multiple chains at once, and
you can say, hey, it's not just Base. You can
(19:05):
choose Base and Solana and Polygon and as many as
you care to receive funds on. And so I think
anywhere that has distribution and has people using their blockchain
benefits from adopting X ro two.
Speaker 1 (19:17):
Yeah, And I think the key here is going to
be watching these consumer apps because you guys will see this.
I mean, you probably won't recognize one of these consumer
apps right here, but this is a good example of
the kind of explosion that we could see around what
you're talking about as some of these additional services out
there that really make this a much more composable component.
Speaker 2 (19:35):
Tying into wallets.
Speaker 1 (19:36):
We see you know, Visa tap in their stripe, ACP
Payai guys which we've talked about crossments in their circle. Obviously,
how fast could that grow in terms of wallets and
payments and integrations into x orld two.
Speaker 3 (19:53):
I mean, if you're going to hit thirty trillion in
five years to really yeah, you're going to need everybody. Right,
it's a massively grown py rate, and it's I think
right now, all about proving at use cases, all about
finding novel ways to use this technology. And it's I
think everyone has to be clear at in that it's
we're not going to be able to predict every way
(20:13):
that this is used. If you think back to the
founding of the Internet exactly, I doubt that Tim berners
Lead imagined Uber being created, right.
Speaker 2 (20:21):
It's it's likely not whenever.
Speaker 3 (20:22):
You have a new technology, it's it's the market and
there the builders are going to determine how it's used.
And I can give you a list of things, but
really I'm just excited to see what people do with it.
Speaker 1 (20:33):
It's going to be intriguing and if it does take off,
especially on the Bell curve that we saw in the
early stages of the Internet really taking off, because use
case was really what generated a lot of that Bell
curve in.
Speaker 2 (20:45):
Terms of growth.
Speaker 1 (20:46):
The other thing that is hidden in plain sight is
the fact that this is correlated to money, and if
you look at just the stable coin data right here.
I mean, look at that, guys. This is the stable
coin movement that we're seeing right now, and this is
just getting started. So you unlock micro transactions, you unlock
(21:06):
major transactions, cross border payments, you've pretty much put everything
within this ecosystem. And as you know, Tom Lee talks
about this all the time. Follow the money, which is
mostly sitting right there in ERC twenty, which of course
is u SEC, Teather and some others. But this could
get big pretty quickly. Do you think this is going
(21:27):
to possibly super charge stable coin adoption even more so
in the short term?
Speaker 2 (21:32):
All right?
Speaker 3 (21:33):
I think that's I think that's very likely. Stable Coins
are really the only form of payment where you can
charge a tenth of a cent or a billion dollars
and it's the same thing that happens, and that that
really is like a superpower. And there's no fees right
There's you just pay the gas. There's not some percentage
based fee that you need to cover in order to
move the money. And that is I think a unique
(21:53):
property that makes stable coin like really well suited for
these kind of payments.
Speaker 1 (21:58):
Man, Eric, we're gonna we're gonna be keeping a close
eye on this because I anticipate we're going to see
a lot of new products, a lot of new developments,
even within the products that we've already seen a blockchain
take off. So it's good having you on. Next time,
we got to get you and Dan on because I
want to hear kind of the biz dev side of this.
What's it going to look like, you know, in six months,
a year to where X four two could go. But
(22:19):
thanks for coming in today. We appreciate it.
Speaker 3 (22:22):
Thanks for having me, Paul.
Speaker 1 (22:23):
You bet all right, You guys, make sure and jump
in over on our own group. We'll keep you updated
on things like this, and the way you do that
is click the link down below, and of course follow
me out there on x Appaul Baron. We'll catch you
next time right here on The Paul Baron Show.