WEBVTT
1
00:00:11.360 --> 00:00:15.900
All right. Quick question before we start: show of hands,
2
00:00:16.140 --> 00:00:20.620
who here has ever sent a Stablecoin transaction? Okay,
3
00:00:20.800 --> 00:00:25.240
pretty good. And then who here has used an agent of any type?
4
00:00:26.700 --> 00:00:30.660
Okay, so a little bit more agent-heavy than stablecoin heavy. So maybe we start,
5
00:00:33.400 --> 00:00:35.760
do stablecoins replace credit cards for agents?
6
00:00:37.440 --> 00:00:39.060
Seems like the apt question to ask me.
7
00:00:40.100 --> 00:00:40.100
Yeah.
8
00:00:40.100 --> 00:00:44.120
So my view is that stablecoins are not a replacement. They're an expansion.
9
00:00:44.900 --> 00:00:48.360
And so it's not really an and or an or
10
00:00:49.860 --> 00:00:54.860
situation. It's more they can both be complementary and expand the ecosystem.
11
00:00:55.400 --> 00:00:58.540
And I think if you take a step back and zoom out,
12
00:00:59.120 --> 00:01:02.020
the internet basically has a whole new set of buyers.
13
00:01:02.180 --> 00:01:05.180
It's a different profile of customer who's showing up and purchasing,
14
00:01:05.340 --> 00:01:06.280
and that's AI agents.
15
00:01:06.920 --> 00:01:10.620
And AI agents will outnumber humans on the internet.
16
00:01:10.680 --> 00:01:12.600
There's literally going to be billions and billions of agents.
17
00:01:13.180 --> 00:01:17.680
And I think that they will actually be the core drivers of the internet economy.
18
00:01:18.460 --> 00:01:21.440
And cards are a fantastic rail,
19
00:01:21.740 --> 00:01:24.600
but they were designed for human initiated payments.
20
00:01:25.200 --> 00:01:27.080
And there's a lot of bright spots around cards.
21
00:01:27.140 --> 00:01:30.000
They offer unparalleled consumer trust and protection,
22
00:01:30.980 --> 00:01:33.880
broad merchant acceptance, broad consumer adoption,
23
00:01:34.260 --> 00:01:37.400
and they perform all of those roles very well.
24
00:01:37.920 --> 00:01:42.000
And I think that agents in many ways will accelerate the adoption of cards as
25
00:01:42.060 --> 00:01:44.960
well. But with a new buyer profile,
26
00:01:45.080 --> 00:01:48.740
there's net new payment flows and agents will pay for things in different ways
27
00:01:48.780 --> 00:01:52.000
than humans do. So things like machine-to-machine payments,
28
00:01:52.100 --> 00:01:55.660
programmatic pay-per-use payments or streaming payments,
29
00:01:56.100 --> 00:02:00.260
those are things that card rails weren't natively designed to handle.
30
00:02:00.980 --> 00:02:04.040
And stablecoins are programmable,
31
00:02:04.960 --> 00:02:07.120
always on global money.
32
00:02:07.620 --> 00:02:12.460
And so the architecture is natively designed to meet the needs
33
00:02:12.600 --> 00:02:13.260
of agents.
34
00:02:13.260 --> 00:02:18.040
And so I think that in many ways they're well-suited to address a lot of those
35
00:02:18.100 --> 00:02:22.700
net new payment flows that you'll start to see emerge from
36
00:02:22.880 --> 00:02:26.900
agents as they become a larger share of the internet economy.
37
00:02:27.900 --> 00:02:30.920
I bet a bunch of people would've expected to say stablecoins would replace
38
00:02:30.980 --> 00:02:34.900
agents or cards. So I'm curious, do you actually believe what you just said?
39
00:02:35.420 --> 00:02:37.000
I could be a "maxi" and say yes, but no,
40
00:02:37.480 --> 00:02:41.740
I do genuinely believe that it's not going to be that stablecoins rule the
41
00:02:41.800 --> 00:02:43.360
world, and it's an end-all, be-all.
42
00:02:43.540 --> 00:02:46.280
I think that we'll live in a multirail system,
43
00:02:46.500 --> 00:02:49.960
and the payments ecosystem will just get hyper,
44
00:02:50.280 --> 00:02:55.240
hyper efficient and rails that meet certain
45
00:02:55.380 --> 00:02:59.300
use cases and certain payment types will succeed and they'll thrive.
46
00:02:59.580 --> 00:03:04.540
So example would be, if an agent wants to make a very high-order value purchase,
47
00:03:04.880 --> 00:03:07.360
card rails are great for that because you have chargebacks,
48
00:03:07.420 --> 00:03:08.340
you have fraud protection,
49
00:03:08.440 --> 00:03:11.980
you have layers of consumer protection that today stablecoins don't provide.
50
00:03:12.540 --> 00:03:14.900
But for machine-to-machine payments, or,
51
00:03:15.140 --> 00:03:19.540
like I was mentioning for paper use payments that will emerge,
52
00:03:19.640 --> 00:03:23.360
I think that agents will defer default to stablecoins over time.
53
00:03:24.100 --> 00:03:26.200
Well, maybe we should introduce ourselves. I'm Dan Romero.
54
00:03:26.500 --> 00:03:27.760
I lead go to market for Tempo,
55
00:03:27.940 --> 00:03:31.920
which is the blockchain that Stripe and Paradigm incubated. And Shan?
56
00:03:32.980 --> 00:03:34.700
Shan Aggarwal, the chief business officer at Coinbase.
57
00:03:34.880 --> 00:03:38.780
I actually had the pleasure of, I'd say, hiring you at Coinbase back in the day.
58
00:03:39.840 --> 00:03:44.340
And one thing it's kind of worth noting is that Shan was in the room when you
59
00:03:44.360 --> 00:03:48.840
guys originally did USDC at Coinbase. So I mean,
60
00:03:48.900 --> 00:03:53.320
maybe stablecoins weren't necessarily even a thing when we were thinking about
61
00:03:53.340 --> 00:03:54.100
it back then.
62
00:03:54.100 --> 00:03:56.780
Yeah. I mean, back then, frankly, when we launched it,
63
00:03:57.140 --> 00:04:00.660
I was blown away by the number of people who asked me a question of, well,
64
00:04:00.800 --> 00:04:04.140
why do I need a programmable digital dollar? What's the point of it?
65
00:04:04.400 --> 00:04:05.240
What's it good for?
66
00:04:05.860 --> 00:04:10.800
And I think that these technologies and the growth of the agentic
67
00:04:10.860 --> 00:04:12.520
economy and stablecoins now,
68
00:04:12.980 --> 00:04:17.360
you can very clearly see how programmability and open
69
00:04:17.580 --> 00:04:22.380
ecosystems work very symbiotically in a way that wasn't so apparent back
70
00:04:22.440 --> 00:04:26.360
then. And so when we started USDC in 2018,
71
00:04:27.160 --> 00:04:31.200
the first use case that actually saw adoption was not payments-oriented.
72
00:04:31.800 --> 00:04:33.980
It was supporting crypto capital markets.
73
00:04:34.440 --> 00:04:38.300
There's a lot of onchain protocols and there's a lot of exchanges around the
74
00:04:38.320 --> 00:04:42.640
world that don't necessarily have seamless access to fiat rails.
75
00:04:43.040 --> 00:04:47.260
And so you can run those entire exchanges on stablecoin rails and it looks a lot
76
00:04:47.300 --> 00:04:52.020
like a dollar.
And I think you're starting to see similar things now happen with
77
00:04:52.060 --> 00:04:56.280
payments where you can have a virtual account that looks like a US dollar bank
78
00:04:56.320 --> 00:04:56.860
account,
79
00:04:56.860 --> 00:05:01.800
but it can be issued to anyone anywhere and it's inherently
80
00:05:01.840 --> 00:05:02.640
global.
81
00:05:02.640 --> 00:05:07.080
And so it's interesting to now see this second use case expand and expand very
82
00:05:07.140 --> 00:05:07.973
quickly.
83
00:05:08.480 --> 00:05:10.200
And maybe to touch upon that,
84
00:05:10.420 --> 00:05:14.880
why do you think it's shifted more towards payments and given that it started at
85
00:05:14.920 --> 00:05:18.760
capital markets and now we're on stage talking about agentic payments?
86
00:05:18.960 --> 00:05:21.500
Yeah. Well, it depends on what metric you look at.
87
00:05:21.560 --> 00:05:24.840
If you look at stablecoin supply versus stablecoin transactions,
88
00:05:25.280 --> 00:05:26.180
for stablecoin supply,
89
00:05:26.620 --> 00:05:29.840
still the predominant use case for stablecoins is in crypto capital markets.
90
00:05:29.940 --> 00:05:33.720
The vast majority of supply supports crypto trading,
91
00:05:34.340 --> 00:05:37.100
borrow/lend protocols, and credit markets and things of that nature.
92
00:05:37.920 --> 00:05:40.640
But we've started to see significant pickup on the payment side.
93
00:05:40.700 --> 00:05:45.120
And I think a lot of that has really been sparked by the infrastructure pieces
94
00:05:45.200 --> 00:05:49.960
coming together in a way where a lot of the early adopters saw stablecoins
95
00:05:50.160 --> 00:05:51.620
as a lower cost,
96
00:05:51.720 --> 00:05:55.440
more efficient way to just move money across borders and even within
97
00:05:56.740 --> 00:06:01.360
their own company. And we still see significant demand from corporates and
98
00:06:02.660 --> 00:06:04.480
enterprises for those use cases,
99
00:06:04.960 --> 00:06:07.740
but where we're seeing green shoots-and I say "green shoots" because it's still
100
00:06:07.760 --> 00:06:11.580
like very, very early days, the infrastructure pieces are there,
101
00:06:11.640 --> 00:06:16.270
but the ecosystem is pretty thin still-is with agents that
102
00:06:17.190 --> 00:06:21.370
sort of organically looked to stablecoins as a payment rail
103
00:06:21.810 --> 00:06:26.730
because of the inherent elements of programmability and the fact
104
00:06:26.750 --> 00:06:27.810
that they're always on,
105
00:06:29.510 --> 00:06:32.750
gives them a lot of flexibility for specific use cases that we're starting to
106
00:06:32.750 --> 00:06:32.900
see.
107
00:06:32.900 --> 00:06:36.790
And maybe, why now with agents and payments?
108
00:06:37.210 --> 00:06:39.530
And you've talked about the intents becoming the interface.
109
00:06:39.770 --> 00:06:41.370
I'm curious if you can expand upon that a bit.
110
00:06:41.950 --> 00:06:42.250
Yeah.
111
00:06:42.250 --> 00:06:47.230
So I think the way that humans interact with the internet is completely
112
00:06:47.290 --> 00:06:51.730
changing. So the flow today really involves
113
00:06:52.370 --> 00:06:56.790
UIs and a common flow where you browse, select,
114
00:06:56.990 --> 00:07:00.590
and then check out. You go to a website, you browse a bunch of options,
115
00:07:00.810 --> 00:07:03.450
you pick a thing, you put something in your cart and then you check out.
116
00:07:04.370 --> 00:07:06.350
And going forward,
117
00:07:06.510 --> 00:07:11.430
intents are the new interface where you express an intent through a
118
00:07:11.450 --> 00:07:14.550
terminal and then an agent just figures it out for you.
119
00:07:15.490 --> 00:07:19.630
And the interesting thing about that is it sort of collapses that three-step
120
00:07:19.690 --> 00:07:23.530
transaction flow into one. Instead of browsing, selecting and checking out,
121
00:07:23.910 --> 00:07:28.530
you instruct through an intent, the agent, who can then fulfill that task.
122
00:07:28.810 --> 00:07:32.390
And from a payments perspective, the interesting thing-and we know this,
123
00:07:32.770 --> 00:07:37.470
it's a common concept within the crypto space-is you can sort of embed the
124
00:07:37.510 --> 00:07:42.250
payment in the intent resolution.
So instead of having the checkout be the
125
00:07:42.650 --> 00:07:47.290
final step, the agent can say, "It's a conditional payment.
126
00:07:48.330 --> 00:07:52.010
I will transact if these conditions are met and therefore the payment is made in
127
00:07:52.050 --> 00:07:55.490
real time." And so it's all bundled into one transaction.
128
00:07:56.090 --> 00:07:59.670
And I think that's just a very different way for how the internet is sort of
129
00:07:59.810 --> 00:08:04.610
architected today where today it's primarily a visual marketplace
130
00:08:04.730 --> 00:08:05.870
for humans, but in the future,
131
00:08:06.680 --> 00:08:11.680
it's a programmable marketplace for agents and stablecoins are
132
00:08:12.160 --> 00:08:16.800
very well-suited for that inherent need for programmability.
133
00:08:17.040 --> 00:08:22.000
And so that's why I think that for agent-native transactions and
134
00:08:22.120 --> 00:08:24.560
these new payment flows that will get created,
135
00:08:25.030 --> 00:08:27.280
stablecoins are the native payment rail.
136
00:08:28.120 --> 00:08:31.400
Rather than retrofitting card rails to work for agents,
137
00:08:31.710 --> 00:08:36.120
stablecoins were built with these elements from the very beginning.
138
00:08:36.880 --> 00:08:40.240
Yeah. I mean, I think from the Tempo perspective, one of the demos,
139
00:08:40.360 --> 00:08:44.320
maybe you were watching the keynote when Will kind of walked through the example
140
00:08:44.480 --> 00:08:49.450
of an agent discovering the fact that the API endpoint that his
141
00:08:49.470 --> 00:08:54.360
demo was doing accepted payments over crypto
142
00:08:54.600 --> 00:08:57.880
and then the ability to actually do it at a very, very granular and small level,
143
00:08:59.320 --> 00:09:02.200
just not possible with other payment methods. At the same time,
144
00:09:02.600 --> 00:09:07.280
you can also accept other payment methods through programmable and
145
00:09:07.560 --> 00:09:10.880
protocol-based like x402.
146
00:09:11.200 --> 00:09:14.080
Yeah. And I don't think that's a bad thing. To your first question,
147
00:09:14.880 --> 00:09:17.760
it's growing the pie because those types of payments,
148
00:09:18.000 --> 00:09:20.520
they weren't previously happening. So now they are happening,
149
00:09:20.640 --> 00:09:23.610
and I think a lot of them will be facilitated by stablecoins,
150
00:09:24.090 --> 00:09:27.910
but many transactions will continue to run on card rails as well.
151
00:09:28.590 --> 00:09:29.423
Do you think that there's
152
00:09:31.190 --> 00:09:35.570
any specific differences between how an agent is kind of
153
00:09:35.810 --> 00:09:38.830
choosing, or kind of flowing through, with payments versus humans?
154
00:09:39.730 --> 00:09:40.710
We talked a little bit about this.
155
00:09:40.830 --> 00:09:45.550
Yeah. So the first is agents are rational and
156
00:09:45.990 --> 00:09:49.450
ruthless optimizers. So we all have our friends,
157
00:09:49.510 --> 00:09:51.450
those friends who are like overoptimizers,
158
00:09:51.610 --> 00:09:55.030
and there's kind of a negative connotation when a human is an overoptimizer.
159
00:09:56.810 --> 00:10:01.590
And humans tend to be loyal. We express brand loyalty, we value convenience,
160
00:10:02.070 --> 00:10:04.270
but agents just see through all of that.
161
00:10:04.750 --> 00:10:07.130
And it's purely based on what is the lowest cost,
162
00:10:07.550 --> 00:10:12.550
most efficient service to provide or complete whatever task
163
00:10:12.590 --> 00:10:16.090
that I need to complete, and then payment rail that I can leverage.
164
00:10:16.170 --> 00:10:20.790
And so I think the implication of that is that switching costs are relatively
165
00:10:20.870 --> 00:10:25.810
low because there's no embedded loyalty for subjective elements like,
166
00:10:26.410 --> 00:10:30.150
"I like this brand more than that brand and I'm willing to pay a premium for
167
00:10:30.210 --> 00:10:31.150
it." Historically,
168
00:10:31.210 --> 00:10:35.650
that has been a very important lock-in for human-based transactions.
169
00:10:36.890 --> 00:10:41.750
The second is agents don't really have identity and that's obviously
170
00:10:41.810 --> 00:10:44.730
very important when you're moving money.
There's no bank account,
171
00:10:44.810 --> 00:10:48.170
there's no home address, there's no POA or anything like that,
172
00:10:49.370 --> 00:10:50.310
and that's a key difference.
173
00:10:50.430 --> 00:10:55.150
And then I think the third key difference is that agents
174
00:10:55.670 --> 00:10:58.730
transact at machine speed and at machine scale.
175
00:10:59.010 --> 00:11:03.610
So it would break a human's brain to make a thousand transactions in a day.
176
00:11:03.890 --> 00:11:06.790
You could do it, but how would you audit all of it?
177
00:11:06.850 --> 00:11:08.330
How would you keep track of all of it?
178
00:11:08.390 --> 00:11:11.610
You probably make 10 or maybe a hundred transactions max,
179
00:11:12.130 --> 00:11:15.250
but agents can make thousands or millions of transactions a day.
180
00:11:15.750 --> 00:11:20.750
And so I think that has implications for the cost structure of the ultimate
181
00:11:21.690 --> 00:11:26.570
underlying rails and where blockchains are very well-suited to serve.
182
00:11:27.090 --> 00:11:27.510
Yeah.
183
00:11:27.510 --> 00:11:30.870
One frame I always think about is if someone comes out with a brand new credit
184
00:11:30.890 --> 00:11:35.570
card that maybe is doing 4% cash back in Bitcoin, that seems interesting,
185
00:11:35.650 --> 00:11:37.750
but that's like a lot of work to go get that new card,
186
00:11:37.930 --> 00:11:39.950
switch all of your subscriptions over to that.
187
00:11:40.410 --> 00:11:42.290
Where's an agent that work could just happen.
188
00:11:42.550 --> 00:11:42.970
Yeah.
189
00:11:42.970 --> 00:11:46.830
And so kind of that rational, ruthless optimization of, "Oh,
190
00:11:46.970 --> 00:11:50.670
if I get an additional 0.1% cash back here,
191
00:11:50.730 --> 00:11:51.810
I'm going to use this payment rail.".
192
00:11:51.870 --> 00:11:56.590
Yeah, agents will make you a better optimizer in your life.
193
00:11:56.910 --> 00:11:59.910
It's like when you're dealing with all of these credit cards that have these
194
00:12:00.350 --> 00:12:04.550
embedded bundles and these schemes that no one likes around-"You can get this
195
00:12:04.610 --> 00:12:09.390
benefit in this quarter"-an agent can help navigate all of that for you
196
00:12:09.730 --> 00:12:14.450
and make you a better card user so you can optimize your reward spend in your
197
00:12:14.490 --> 00:12:15.270
example.
198
00:12:15.270 --> 00:12:19.350
So let's say agents take over and are doing a large percentage of transactions.
199
00:12:19.690 --> 00:12:23.330
Are we actually in a multirail world in that case or is it that they're going to
200
00:12:23.350 --> 00:12:24.550
all optimize for the same thing?
201
00:12:25.250 --> 00:12:28.370
I think we'll be in a multirail world because I think different rails will be
202
00:12:28.470 --> 00:12:30.030
best-suited for different use cases.
203
00:12:30.410 --> 00:12:35.370
And I think that that optimization function that agents go through
204
00:12:35.810 --> 00:12:38.870
will lead them to whichever rail is the best-suited.
205
00:12:39.090 --> 00:12:42.610
So I gave the example around card rails. Card rails,
206
00:12:43.030 --> 00:12:46.510
there's a lot of good reasons for why you would transact with a card and a lot
207
00:12:46.550 --> 00:12:49.310
of different transaction types for where that will be well-suited.
208
00:12:49.370 --> 00:12:50.610
So for ecommerce,
209
00:12:50.690 --> 00:12:55.450
I think that that will continue to be the primary or default modality
210
00:12:55.530 --> 00:12:59.650
that agents use as well. But for stablecoins,
211
00:12:59.710 --> 00:13:03.830
we think more about these net new use cases and consumption patterns that are
212
00:13:03.850 --> 00:13:06.630
starting to emerge. And like I said,
213
00:13:06.870 --> 00:13:08.870
it's very much the experimentation phase,
214
00:13:08.950 --> 00:13:12.130
but that's kind of what we're starting to observe and see.
215
00:13:12.190 --> 00:13:15.350
So an example of this, to make it a little bit more concrete is:
216
00:13:16.910 --> 00:13:21.750
I have an agent and I want it to be my financial analyst and say,
217
00:13:21.870 --> 00:13:22.190
"Hey,
218
00:13:22.190 --> 00:13:26.710
can you go build a report and put together a perspective on should I invest in
219
00:13:26.810 --> 00:13:30.950
Coinbase stock?" And the agent might need to consume data from Bloomberg,
220
00:13:31.010 --> 00:13:31.843
from Morningstar,
221
00:13:31.930 --> 00:13:36.250
from all of these different proprietary data sources that have paywalls.
222
00:13:36.950 --> 00:13:39.110
And if you have a proprietary data source,
223
00:13:40.410 --> 00:13:42.810
you want to scale demand as much as possible.
224
00:13:43.050 --> 00:13:47.830
And so now a lot of these sites are starting to opt in by
225
00:13:48.990 --> 00:13:50.270
becoming agent-accessible,
226
00:13:50.750 --> 00:13:54.570
and that allows an agent to pay cents for specific points that it might want to
227
00:13:54.670 --> 00:13:55.810
access within the dataset,
228
00:13:56.150 --> 00:13:59.210
produce a very robust report that gives me a lot of utility.
229
00:13:59.690 --> 00:14:03.930
And so that's one example that we see both in the traditional finance world and
230
00:14:03.950 --> 00:14:05.370
then also in the crypto-specific space.
231
00:14:05.950 --> 00:14:09.530
But some people might say that micropayments never have worked,
232
00:14:10.230 --> 00:14:11.063
too much friction.
233
00:14:11.590 --> 00:14:15.390
So do you think just because agents are hyper-rational,
234
00:14:15.490 --> 00:14:18.270
willing to deal with the schlep and the pain, that that actually does work?
235
00:14:18.630 --> 00:14:22.350
I think they didn't work because the rails and the infrastructure didn't allow
236
00:14:22.410 --> 00:14:23.243
for them to work.
237
00:14:24.090 --> 00:14:27.050
I think there's a narrative out there.
238
00:14:28.310 --> 00:14:32.330
Obviously Tempo is working on MPP, Coinbase is working in x402.
239
00:14:33.850 --> 00:14:36.830
It's kind of this head-to-head agentic protocol.
240
00:14:38.010 --> 00:14:41.590
Do you think it's actually the case where it's kind of like this battle
241
00:14:41.630 --> 00:14:45.010
VHS/Betamax or is it a kind of a multiprotocol world?
242
00:14:45.570 --> 00:14:46.403
What do you think?
243
00:14:46.640 --> 00:14:47.473
Well,
244
00:14:51.250 --> 00:14:54.550
I think agents, to your point, they're polyglots, right?
245
00:14:54.730 --> 00:14:58.210
So you can ask it to write a programming-type script and have it switch over to
246
00:14:58.250 --> 00:15:02.670
Python. So in the world where there are multiple protocols, they can speak both.
247
00:15:03.050 --> 00:15:04.670
And I think in most cases,
248
00:15:05.430 --> 00:15:06.910
if you just kind of look at the history of technology,
249
00:15:07.330 --> 00:15:11.070
there are very few protocols or standards that is the only kind of winner.
250
00:15:11.550 --> 00:15:15.410
I do think the web and SMTP for email are two maybe rare examples,
251
00:15:15.470 --> 00:15:18.350
but I think generally agents don't care.
252
00:15:18.470 --> 00:15:20.090
They just kind of want to get the job done, right?
253
00:15:20.390 --> 00:15:21.990
So if you hit an API endpoint and says, "Hey,
254
00:15:22.120 --> 00:15:26.710
I speak protocol X or protocol Y," they're just going to get the job done,
255
00:15:26.770 --> 00:15:30.970
assuming a stablecoin is interoperable with either protocol, right?
256
00:15:31.350 --> 00:15:34.130
And similarly, maybe use a different chain here or a different chain there.
257
00:15:34.690 --> 00:15:37.030
And in the case that it doesn't accept stablecoins,
258
00:15:37.090 --> 00:15:39.710
they can issue a virtual card and solve it that way.
259
00:15:40.070 --> 00:15:42.550
Yeah. I mean, we see it pretty similarly.
260
00:15:44.090 --> 00:15:48.550
So we built x402 last year just because we saw a gap in the market where
261
00:15:49.010 --> 00:15:53.890
there wasn't a HTTP-native payment protocol that was built
262
00:15:53.930 --> 00:15:56.050
for machines. And so we felt like, hey,
263
00:15:56.110 --> 00:15:59.350
if machine transactions are going to increase, then there should be a way for
264
00:16:00.930 --> 00:16:03.070
machines to fulfill that payment request.
265
00:16:03.130 --> 00:16:07.490
And so that was the sort of genesis behind x42. But we fundamentally,
266
00:16:08.150 --> 00:16:09.450
we don't believe it should be proprietary,
267
00:16:09.570 --> 00:16:12.610
which is why we open sourced it with Stripe and Cloudflare and a number of
268
00:16:12.730 --> 00:16:13.563
others.
269
00:16:14.110 --> 00:16:18.370
And so if another protocol emerges that fits the agent's use case,
270
00:16:18.470 --> 00:16:21.950
then that's great. That's totally fine. I think, from our perspective,
271
00:16:22.090 --> 00:16:27.090
we want to ensure that we try to avoid walled gardens and
272
00:16:27.130 --> 00:16:32.130
fragmentation that potentially limits the utility and ability for agents
273
00:16:32.190 --> 00:16:33.310
to transact on the internet.
274
00:16:34.010 --> 00:16:35.430
If I'm being totally intellectually honest,
275
00:16:35.490 --> 00:16:40.030
when I scroll Twitter and I see a bunch of the kind of agentic protocols and
276
00:16:40.070 --> 00:16:43.210
some of the charts, there's a lot of chart crime on crypto, Twitter,
277
00:16:43.290 --> 00:16:46.630
especially where they put big volumes and then it's like, "Okay, well,
278
00:16:46.750 --> 00:16:49.570
what's actually behind that volume?" Maybe if we just step back,
279
00:16:50.110 --> 00:16:53.610
what do you think is holding back the takeoff for agentic payments? I mean,
280
00:16:53.930 --> 00:16:57.930
you saw some of the charts that Patrick was showing in terms of just things kind
281
00:16:57.950 --> 00:16:59.950
of going vertical from the singularity standpoint,
282
00:17:00.210 --> 00:17:02.570
but I don't feel like we've hit that for agentic payments. I'm curious,
283
00:17:03.170 --> 00:17:04.110
what's the missing ingredient?
284
00:17:04.310 --> 00:17:06.570
Yeah, we've definitely not hit that for agentic payments.
285
00:17:07.450 --> 00:17:12.330
I think it's like inning one or maybe even the warm-ups in the bullpen or
286
00:17:12.390 --> 00:17:15.570
something like that. In terms of what's missing,
287
00:17:16.150 --> 00:17:18.310
I think there's a few components that come to mind.
288
00:17:18.370 --> 00:17:21.530
So first is we need broad payment acceptance.
289
00:17:21.870 --> 00:17:23.470
Just like when cards launched,
290
00:17:23.810 --> 00:17:26.430
cards had minimal utility if no merchants accepted cards.
291
00:17:27.070 --> 00:17:30.510
And so we need merchants, PSPs, et cetera,
292
00:17:30.630 --> 00:17:35.490
to make their services agent accessible. And so x42, MPP,
293
00:17:35.750 --> 00:17:39.190
these are protocols that can help those businesses do that.
294
00:17:39.350 --> 00:17:43.530
So I think that's one. The second is, I mentioned earlier:
295
00:17:43.570 --> 00:17:44.710
the identity component,
296
00:17:44.770 --> 00:17:49.010
but I do think that that is actually quite fundamental because if you think
297
00:17:49.030 --> 00:17:51.990
about an agent that comes and requests to
298
00:17:53.790 --> 00:17:57.330
make a payment or make a purchase, there's no way right now to validate,
299
00:17:57.450 --> 00:18:01.070
is this like a legitimate transaction that that agent is doing on behalf of
300
00:18:02.670 --> 00:18:03.690
the human that it's working for,
301
00:18:03.910 --> 00:18:08.810
or is it just going rogue?
And so I do think that we'll need more
302
00:18:08.930 --> 00:18:13.290
portable and flexible identity layers for agents that are
303
00:18:13.510 --> 00:18:15.830
consumable on the acceptance side.
304
00:18:17.230 --> 00:18:21.410
And then the third is I think the entire compliance layer and trust layer is
305
00:18:21.770 --> 00:18:26.610
very much like uncharted territories. And so we talked a little bit,
306
00:18:26.670 --> 00:18:31.350
that's where cards really shine. And cards have invested a tremendous amount.
307
00:18:31.410 --> 00:18:35.780
And I think that's been a huge driver of the success that cards have seen in
308
00:18:35.820 --> 00:18:38.300
terms of achieving broad scale consumer adoption.
309
00:18:38.940 --> 00:18:41.280
And for the most part today,
310
00:18:41.340 --> 00:18:46.120
stablecoin transactions are irreversible and that puts us a bit of a
311
00:18:46.440 --> 00:18:47.490
ceiling in terms of,
312
00:18:47.880 --> 00:18:50.640
am I really going to be comfortable making this multithousand dollar purchase
313
00:18:50.700 --> 00:18:52.520
using stablecoins? I think back to
314
00:18:54.660 --> 00:18:55.730
when mobile phones came out,
315
00:18:56.800 --> 00:19:00.460
an interesting stat that I always noticed is there was a significant delta
316
00:19:00.520 --> 00:19:03.880
between the average order value on desktop versus mobile.
People felt
317
00:19:03.900 --> 00:19:06.460
comfortable spending more money on desktop than they did on mobile.
318
00:19:06.990 --> 00:19:09.060
And why was that? It's a little bit of consumer behavior,
319
00:19:09.210 --> 00:19:14.120
I'm spending a lot of money, but that builds up over time.
320
00:19:14.180 --> 00:19:17.640
And so I think that entire trust layer is very much uncharted territory that
321
00:19:18.360 --> 00:19:19.193
needs to get solved.
322
00:19:19.490 --> 00:19:20.800
So if we rebuild that trust layer,
323
00:19:20.860 --> 00:19:24.280
are we just rebuilding the cost that exists in the existing system?
324
00:19:24.760 --> 00:19:25.593
Well, if you can do it,
325
00:19:25.860 --> 00:19:28.640
I think if you can do it onchain with smart contracts in a
326
00:19:30.400 --> 00:19:33.020
more open software-forward way,
327
00:19:33.300 --> 00:19:37.420
then my hope is that the cost will be significantly reduced.
328
00:19:37.780 --> 00:19:39.440
And then it's not just a cost thing,
329
00:19:39.580 --> 00:19:43.470
but it's also making sure that these things are apples-to-apples.
330
00:19:43.640 --> 00:19:46.140
They should be natively developed together, payments and identity
331
00:19:47.860 --> 00:19:52.820
rather than retrofitting legacy identity systems for a new agent-led
332
00:19:52.840 --> 00:19:53.600
economy.
333
00:19:53.600 --> 00:19:54.940
Yeah. From my perspective,
334
00:19:55.000 --> 00:19:58.620
I think payments actually feel like at least the plumbing feels solved.
335
00:19:59.140 --> 00:20:01.720
It's the identity piece that you mentioned. It's like, okay, well,
336
00:20:01.940 --> 00:20:03.440
in order to pay for this service,
337
00:20:03.500 --> 00:20:05.040
you actually need to go sign up with an email address,
338
00:20:05.160 --> 00:20:09.560
and then come back because we don't want anonymous agents paying for the APIs.
339
00:20:09.740 --> 00:20:12.880
And then I think the second component is how do you actually fund the agent,
340
00:20:13.520 --> 00:20:17.620
like the classic problem, which it always comes back to. And it's like, okay,
341
00:20:17.700 --> 00:20:21.200
moving from fiat into something like stablecoins, there's a fraud vector,
342
00:20:21.580 --> 00:20:22.740
and things like that as well as complete.
343
00:20:22.800 --> 00:20:23.633
Yes.
344
00:20:23.780 --> 00:20:28.260
So a bunch of folks in the audience, probably CEOs, developers.
345
00:20:28.840 --> 00:20:32.680
If you had a company right now, and let's say you really only accept cards,
346
00:20:33.840 --> 00:20:36.840
how would you be thinking about agentic payments? Is this a P0,
347
00:20:37.940 --> 00:20:39.690
have the whole team work on it tomorrow or...?
348
00:20:40.620 --> 00:20:43.800
Yeah. Well, it's hard to comment on P0,
349
00:20:43.800 --> 00:20:46.020
P1 without knowing all your priorities.
350
00:20:46.360 --> 00:20:48.780
I think it's something that you should start to experiment with for sure right
351
00:20:48.840 --> 00:20:53.240
now. And I think the two things that I would think about are: one,
352
00:20:53.340 --> 00:20:57.220
if you believe that agents will be a core driver of the internet economy,
353
00:20:57.460 --> 00:20:59.360
like they will be core purchasers, core...
354
00:20:59.800 --> 00:21:02.520
I think about them effectively as a new customer segment.
355
00:21:03.100 --> 00:21:07.280
It's a new customer segment that didn't exist previously that is coming to your
356
00:21:07.340 --> 00:21:10.540
door and is going to grow very significantly.
357
00:21:11.080 --> 00:21:15.360
And so is my product accessible by this new customer segment?
358
00:21:16.540 --> 00:21:18.860
If you're purely relying on a UI,
359
00:21:19.900 --> 00:21:22.960
then your product is effectively invisible to an agent.
360
00:21:23.160 --> 00:21:25.740
And so I would think about, again,
361
00:21:25.840 --> 00:21:30.020
how can I build my product so that it is natively agent accessible
362
00:21:30.500 --> 00:21:35.160
because I do believe that they will become important actors in the economy.
363
00:21:36.460 --> 00:21:39.700
The second is, I think,
364
00:21:39.700 --> 00:21:41.900
thinking keenly about pricing models,
365
00:21:42.100 --> 00:21:46.820
we live in a world right now where there's a heavy preference towards bundled
366
00:21:46.880 --> 00:21:47.713
subscriptions,
367
00:21:48.060 --> 00:21:52.360
and bundling is like a very core strategy for a lot of large-scale businesses,
368
00:21:53.160 --> 00:21:57.420
but I think that's fundamentally sort of at odds with agents that are ruthless
369
00:21:57.540 --> 00:22:01.820
optimizers because you think about what's the calculus on,
370
00:22:01.920 --> 00:22:06.340
"Am I going to pay for this large, heavy bundle when I only need to consume one,
371
00:22:06.400 --> 00:22:08.120
two, three, four, five services,
372
00:22:08.240 --> 00:22:12.260
or am I going to look for an alternative?" And so I think starting to think
373
00:22:12.460 --> 00:22:17.240
early about different types of pay-per-use payment models will be really
374
00:22:17.280 --> 00:22:18.113
important as well.
375
00:22:18.380 --> 00:22:20.680
And just to maybe go back to Coinbase,
376
00:22:21.300 --> 00:22:23.020
as you guys think about the business model of Coinbase,
377
00:22:23.180 --> 00:22:24.560
is that changing any of your thinking?
378
00:22:24.820 --> 00:22:25.880
Yeah, definitely. I mean,
379
00:22:26.480 --> 00:22:29.500
we've built out a new product line that's called a Coinbase Developer Platform
380
00:22:29.540 --> 00:22:30.880
that's purely API-first.
381
00:22:31.780 --> 00:22:36.480
And then we're thinking about these elements for how agents ultimately end up
382
00:22:36.600 --> 00:22:39.420
accessing, for example, our crypto exchange. Historically,
383
00:22:40.320 --> 00:22:43.700
every customer that came to the crypto exchange had to complete KYC and set up
384
00:22:43.720 --> 00:22:47.000
an account and prefund the account, but that's going to look very different.
385
00:22:47.880 --> 00:22:51.220
If you believe that demand for trading is not going to go away,
386
00:22:51.380 --> 00:22:56.000
but the people or the agents that are driving the trading activity changes,
387
00:22:56.320 --> 00:23:01.060
then we need to adapt our business to ensure that we reduce friction as much as
388
00:23:01.100 --> 00:23:03.800
possible and make that as seamless as we can.
389
00:23:03.860 --> 00:23:07.580
All right. I have two final questions. So the first would be:
390
00:23:07.900 --> 00:23:11.420
what is something you strongly believe will be the same 10 years from now?
391
00:23:11.840 --> 00:23:12.380
The same?
392
00:23:12.380 --> 00:23:16.500
Yeah. With all these changes, right? What do you think is the same?
393
00:23:17.800 --> 00:23:22.020
People are still going to use cash for offline payments. That's not changing.
394
00:23:22.080 --> 00:23:23.260
Yeah, that's definitely probably true.
395
00:23:23.440 --> 00:23:24.960
Yeah.
396
00:23:26.100 --> 00:23:28.520
And then I guess the final question is:
397
00:23:29.380 --> 00:23:34.140
what's a version of the future that maybe you're thinking about or you've heard
398
00:23:34.180 --> 00:23:37.900
that you actually think is likely to happen as it relates to agents?
399
00:23:39.000 --> 00:23:42.760
So one that I've heard, which is I noodle on, but I think
400
00:23:44.300 --> 00:23:47.440
we were talking about the UX layer for the internet being designed for humans
401
00:23:47.500 --> 00:23:48.320
natively.
402
00:23:48.320 --> 00:23:52.380
And the business model that has been applied on top of that has been an
403
00:23:52.480 --> 00:23:56.100
ads-based business model where the largest companies in the world monetize on
404
00:23:56.140 --> 00:23:56.973
attention.
405
00:23:57.220 --> 00:24:02.160
But if the internet is no longer driven by humans engaging
406
00:24:02.200 --> 00:24:05.160
on the internet, but driven by intents and agents,
407
00:24:05.680 --> 00:24:08.020
then does that ad space business model hold?
408
00:24:08.660 --> 00:24:11.240
Do agents care about advertising,
409
00:24:11.660 --> 00:24:14.640
and does that influence their purchasing decisions or their direction?
410
00:24:15.200 --> 00:24:19.020
Facebook and Google's entire business is based on directing flow of traffic
411
00:24:19.520 --> 00:24:24.060
through ads. And I think that in 10 years,
412
00:24:24.080 --> 00:24:26.840
that that will look very, very different. And again,
413
00:24:26.920 --> 00:24:31.880
I think that it will favor more granular- and programmable-type payments that
414
00:24:32.680 --> 00:24:37.460
allow agents to optimize the usage of different services. And again,
415
00:24:37.700 --> 00:24:40.580
I think that's one of the ways that stablecoins is well-suited to serve.
416
00:24:41.380 --> 00:24:42.880
Well, Shan, thank you very much.