Google launches AI money

Duration

24:00

Captions

1

Language

EN

Published

Sep 17, 2025

Description

The latest AI News. Learn about LLMs, Gen AI and get ready for the rollout of AGI. Wes Roth covers the latest happenings in the world of OpenAI, Google, Anthropic, NVIDIA and Open Source AI. ______________________________________________ My Links 🔗 ➡️ Twitter: https://x.com/WesRothMoney ➡️ AI Newsletter: https://natural20.beehiiv.com/subscribe Want to work with me? Brand, sponsorship & business inquiries: wesroth@smoothmedia.co Check out my AI Podcast where me and Dylan interview AI experts: https://www.youtube.com/@Wes-Dylan ______________________________________________ #ai #openai #llm

Captions (1)

00:00

So, Google just partnered up with

00:02

Coinbase to create a sort of a virtual

00:05

economy for AI agents. So, a lot of

00:08

people have speculated that this was

00:10

coming, that AI agents will need their

00:12

own currency, their own way to pay for

00:15

services so that they're able to acquire

00:17

resources and go out there and get stuff

00:20

done. You've probably heard this idea

00:22

talked about before, but it's here. And

00:25

I for one got caught off guard just a

00:27

little bit because I didn't expect it to

00:29

be here this fast. So Brian Armstrong is

00:32

the co-founder and CEO at Coinbase. Now

00:36

Coinbase is a crypto trading platform.

00:39

So why is it partnering up with Google?

00:42

Well, just a few days ago, Google

00:43

DeepMind published this big paper. What

00:46

is this big paper about? It's about

00:50

virtual agent economies. Here's the key

00:53

part. The rapid adoption of autonomous

00:56

AI agents is giving rise to a new

00:59

economic layer where agents transact and

01:02

coordinate at scales and speeds beyond

01:05

direct human oversight. So again, not a

01:07

new idea, but it just feels like it's

01:10

here a little bit ahead of schedule. And

01:13

notice for the keywords, they do mention

01:15

blockchain. Now, a lot of people are

01:17

probably hesitant to mess around with

01:20

crypto and the blockchain just because

01:22

of the sheer amount of scams and various

01:25

shenanigans and various shenanigans that

01:27

got pulled with crypto and cryptoreated

01:30

projects. But a lot of that had to do

01:32

with just the possibility of speculating

01:34

on prices going up and down. This, as

01:37

far as you can tell, is completely

01:40

separated from that. So, we're not

01:42

talking about any tokens that you can

01:44

bet on. We're just talking about some

01:46

sort of a payment system that can be

01:49

carried out away from the dollar, away

01:52

from, you know, needing Visa and the

01:54

credit cards companies to transact those

01:56

things. Something that can be done very

01:58

fast, done automatically by AI agents

02:00

and something that can be very, very

02:02

transparent. So, none of this is for you

02:04

to buy anything. None of this is going

02:06

to the moon, nothing like that. But

02:08

let's take a look at what it actually

02:10

is. So, this is Coinbase. They're saying

02:12

Google Agentic Payments Protocol. So, I

02:14

guess they're calling it X42.

02:16

So, if you're not aware of the

02:18

background, there's a lot of these

02:19

protocols that are being built out right

02:21

now for various AI systems to interact

02:25

with one another. MCP was the model

02:28

context protocol that was an open-source

02:31

protocol by Anthropic and just allows

02:34

the various large language models and

02:36

other AI systems to interact with other

02:39

things whatever those things may be

02:41

where it's banking or services online

02:44

services or other software. It just

02:46

helps them to kind of plug in and

02:48

communicate to each other and to get

02:50

stuff done. Then we had a much less

02:53

known one, but one that's sort of going

02:56

to be getting more and more visible.

02:58

That's the A2A.

03:00

That's Google's agent 2 agent protocol.

03:04

So, we've covered this in a video from a

03:07

while back when it got announced. So,

03:09

we'll come back to that. But just really

03:11

fast, Google has been quietly building

03:14

out an entire Aentic marketplace. at

03:16

least they briefly announced it at one

03:19

of their conferences at Google IO either

03:22

earlier this year or late last year and

03:24

it it seemed like a big deal but it was

03:27

there was not too much to find out about

03:29

it. So it's something that it seems to

03:30

be like they're working in the

03:31

background but I wasn't able to get

03:33

access to it or anything like that. And

03:35

so there's kind of that agentic

03:36

marketplace. We'll get back to that in

03:37

just a second. There's agent to agent.

03:40

So Google Cloud's agent 2 agent A2A

03:43

protocol AI agents across multiple

03:45

platforms can already communicate and

03:47

interact with each other to complete

03:49

tasks. So when they introduced this, if

03:51

I recall correctly, Google gave an

03:53

example of I think something that that

03:55

already happened like an example of how

03:56

these agents communicate. So let's say

03:58

you have an HR department, right? So

04:00

you're hiring people for your company.

04:03

You have your little HR agent and it's

04:06

tasked with doing certain things in

04:08

order to hire those people. In this

04:09

case, let's say we needed a background

04:11

check. That was this was actually the

04:12

exact example that Google used. So your

04:15

little HR agent would go out there on

04:17

the Agentic marketplace and try to find

04:20

other AI agents that provided background

04:22

check services. So it did a bunch of

04:24

research. should interact with a lot of

04:26

the other agents and found out that

04:28

this, you know, agent X, let's say, was

04:31

the best background agent checker for

04:35

USbased people and then agent Y was the

04:38

best for international background

04:40

checks. So, let's say you had a number

04:42

of applicants, some of them local, some

04:44

of them international. So, it's

04:46

interacted with agent X, right? So, it

04:48

sent them the information about the

04:50

applicants and they said, I need a full

04:51

background check. agent X. So that's

04:53

from a different company that owns Agent

04:55

X, the background check company, right?

04:57

So it, you know, did the background

04:58

check and it send it back to our HR

05:00

agent and then also the HR agent gave

05:03

the um international applicants to agent

05:06

Y. Agent Y did a background of them and

05:08

gave it back to our HR agent. Right? So

05:10

all that happens seamlessly without

05:12

human interaction. Those agents do the

05:15

research, get information, make sure

05:18

that everything is what they need.

05:20

Right? So the background check meets the

05:22

specifications that we need and they

05:25

complete the task. So now our little HR

05:28

agent was able to get the background

05:30

checks on our applicants. Question is

05:32

how does the payment take shape, right?

05:35

Does our little agent pull out their

05:37

Visa rewards credit card and and then

05:39

scan it or whatever? Well, maybe that

05:41

could have worked, but there's tons of

05:43

problems with that. Number one, I mean

05:45

the fees for using Visa and stuff like

05:48

that. We might need there to be tons of

05:50

microtransactions,

05:52

right? A few cents or maybe even

05:54

fractions of a cents that these agents

05:56

can exchange quickly for small

05:58

microservices, etc., right? Like if

06:00

we're able to create something that has

06:02

no friction, that's super fast, doesn't

06:03

have huge fees. That would be a massive

06:07

unlock of course. So they continue here

06:09

now with the addition of X42

06:12

with an extension of Google's newest

06:15

agentic payments protocol AP2. These

06:18

agents can do something entirely new,

06:20

handle payments. As one of the first

06:23

extensions to AP2 and the only stable

06:26

coin facilitator, X42 enables agents to

06:29

monetize their own services, pay other

06:31

agents or handle micro payments

06:33

automatically on behalf of users. A

06:35

stable coin is just basically means that

06:38

it's backed by let's say a dollar,

06:40

right? So if I wanted to give you a

06:41

dollar and you needed to give me a

06:43

dollar, right? We we would have to go

06:45

through let's say the credit card

06:46

company, I'd pay you a dollar. They take

06:48

some insane fee, right? So basically

06:50

every single time we move that money

06:52

around, somebody gets paid. That's great

06:54

for Visa and Mastercard and all those

06:56

companies, but tons of issues and costs

06:59

and slowdowns arise with that. But if we

07:02

are able to take that dollar, turn it

07:03

into a stable coin. So basically

07:06

something that's worth a dollar, but

07:08

something that we can split into a

07:09

million pieces and send quickly

07:11

everywhere in whatever fractions that we

07:13

want, that would just make things a lot

07:15

easier. And that's what a stable coin

07:16

is. So what does this mean? That means

07:18

that certain tasks which previously

07:20

required manual oversight like paying

07:22

for data crawls, services or microtasks

07:25

can now be executed seamlessly on behalf

07:28

of users by the agents themselves. They

07:30

do have a demo here. Here's this demo

07:33

that they're talking about using X42 and

07:36

Google's A2A to buy a new refrigerator.

07:39

So, looks like this is Coinbase. They're

07:41

using some sort of an agent development

07:42

kit. So, here's two agents talking

07:46

together. Looks like our agent or the

07:49

human is talking to a Lowe's

07:51

merchant agent and they're trying to buy

07:54

a refrigerator. They get several options

07:57

for a refrigerator and it looks like

07:59

once they confirm which refrigerator

08:01

they want, they're saying, "All right,

08:02

I'll process that right away with our

08:04

secure X42 payment technology." So, it

08:07

gets paid a certain amount of USDC

08:10

units. So, that that's that stable coin,

08:12

I assume. And so, this allows you to do

08:15

various purchases, you know, without

08:17

leaving the agent window, without the

08:19

chatbot window, you know, you don't have

08:21

to bring in your credit card. And also I

08:23

mean eventually of course this could be

08:24

100% automated with just the agents

08:27

making those decisions doing various

08:29

micro payments. Maybe you're having some

08:31

threshold at which you need to approve

08:33

the transaction right but like if it

08:35

needs you know 75 cents to get some

08:38

computes or whatever to do something

08:40

like I don't want to have to approve

08:41

that. Just go ahead and get it up to a

08:43

certain amount. So why this matters?

08:46

This is actually kind of a big deal

08:49

because it's going to unlock a lot of

08:52

things we couldn't do before. So, for

08:53

example, I mean, this solves the

08:56

challenge of agents needing to interact

08:58

with each other. We're setting the stage

09:00

for agentic commerce. So, they're saying

09:02

here for developers, this introduces low

09:04

friction pay-per-use payments, ideal for

09:07

microp payments, per crawl fees, and

09:09

other agentdriven scenarios. Now, this

09:12

is huge. Just yesterday we were talking

09:15

about a situation where a number of

09:17

companies are for example suing Google

09:20

because their AI search overviews they

09:23

claim sort of takes away from their

09:25

traffic. So if they're doing

09:27

investigative journalism and Google you

09:29

know crawls their website and just kind

09:31

of summarize them for example that might

09:33

take away from their traffic. Lots of AI

09:36

companies might crawl Reddit to use that

09:38

training data for training their own

09:40

models. oftentimes artists are not very

09:43

happy with AI models, you know, using

09:45

their artwork to train the AI models. We

09:49

currently don't have any good solutions

09:51

for dealing with that, right? Even if

09:53

you wanted to pay like it, how do you

09:55

negotiate those payments? You would have

09:57

to talk to the company or talk to the

09:59

artist. You there's tons of headaches

10:02

involved and it just does not work at

10:04

scale. It doesn't work fast. Right now,

10:06

whenever I do deep research with Chad

10:08

GBT, it might go to 20 to 30 different

10:11

sites. It gets information from those

10:13

sites and organizes into a report for

10:15

me. And I love it. I can't complain. But

10:19

the reality is all those sites, all that

10:21

effort that they've been putting into

10:23

building and having that information. I

10:25

was able to get around whatever

10:28

monetization schemes they had, right?

10:30

So, I don't see their ads. I don't see

10:32

their pop-ups. And again, while I'm very

10:34

happy with that, I mean, at scale, that

10:36

does have negative impacts, right?

10:39

Because now we're disentivizing people

10:41

from building, putting that information

10:43

together, from journalists going out

10:45

there doing investigative journalism,

10:47

and then writing about those things,

10:49

like we're cutting them off from their

10:51

ability to pay for that content to to

10:54

support those efforts, etc. There's tons

10:56

of publications where I wouldn't mind

10:59

paying them to read their stuff if it's

11:01

high quality enough, but it introduces

11:03

various headaches. One is it's usually

11:05

going to be some monthly subscription

11:06

that it's going to be hard to get out

11:08

of. You got to put in your credit card

11:10

details. You got to deal with a whole

11:12

bunch of nonsense. Imagine having an AI

11:15

agent that can take care of all of that

11:17

while keeping everybody satisfied. Let's

11:19

say there's some, as they put here, per

11:21

crawl fee. That could be a few cents. It

11:23

could be a fraction of a cent. So my

11:25

agent goes through, checks everything

11:27

out, goes, "Oh, I think this would be a

11:29

good thing to include in the report. I'm

11:31

authorized to spend a few pennies to

11:34

acquire that information or whatever

11:35

that number may be." And if the

11:38

publication agrees, that thing gets, you

11:40

know, paid. I get the information and

11:42

gets put into the report that I need,

11:44

everyone's happy. Same thing if I wanted

11:46

to use somebody's artwork, for example,

11:48

on some product or whatever, right? This

11:51

could make it a lot easier to license

11:53

those things. Instead of me sitting

11:55

there going through their terms and

11:56

conditions or whatever, the agent's

11:57

like, I want to put your artwork on this

11:59

thing. This is my budget. Yes or no? And

12:02

they handle all the nonsense needed to

12:05

get that thing to happen. Cheap, fast

12:08

micro payments let all of those things

12:10

happen. It allows all those use cases to

12:12

become reality. And as they say here,

12:14

this marks a big increase in agent

12:16

autonomy that will only accelerate as

12:18

agent discovery expands through tools

12:21

like the X42 Bazaar. I'm always going to

12:24

have a hard time for some reason saying

12:25

that X42. It almost seems like it spells

12:28

something out, but it doesn't. So, they

12:29

have the X42 Bazaar. We've already

12:32

enabled agents to automatically discover

12:33

and interact with any service that's

12:35

added to the ecosystem. So, by the way,

12:37

here's the X42 Bazaar, right? So this is

12:40

self-improving AI agents kind of going

12:42

out there and being able to interact and

12:44

conduct various financial transactions.

12:47

So they have one of the first projects

12:49

is this or maybe it's pronounced pre a

12:52

stock price API allowing agents to

12:54

create up-to-date financial report and

12:56

various image and video generation

12:58

endpoints. Right? So instead of signing

13:01

up for some SAS, some software as a

13:03

service for $15 a month, you can just,

13:06

you know, if you need one report, you

13:07

can negotiate with this company like,

13:10

okay, for 25 cents, generate this

13:12

specific report or something more

13:15

complicated for $5. The point is your

13:17

agent should be able to maybe negotiate

13:19

that price for you knowing your budget,

13:21

your needs, etc. Really gets the wheels

13:24

moving, doesn't it? I mean, there's a

13:26

lot of possibility here. They give you a

13:29

great sort of example of how this would

13:30

work, right? So maybe you want to have

13:33

some outdoor event. So help me find the

13:35

best one. Book it for me. Right? So the

13:37

agent could pay less than a penny to get

13:39

the weather for next month. So this is

13:42

the thing that I'm kind of talking about

13:43

because there's a lot of these little

13:44

services that you might need once and it

13:47

doesn't make sense to sign up for it.

13:49

You don't want to see their ads, right?

13:50

So the the agent is going to be able to

13:52

pull that information without you going

13:54

to your website watching all the stupid

13:55

ads. If they charged you a dollar to

13:57

check the weather, you'd be like, "No,

13:59

that's insane." But if it was a fraction

14:01

of a penny, I I certainly wouldn't mind

14:03

it. If that request at the end of the

14:06

day cost me 5 cents, I feel like I'd be

14:09

okay with that. Meaning the the actual

14:11

completion of, you know, planning and

14:13

booking the entire event with this

14:15

fraction of a penny thing being as one

14:17

of the line items on there. And we've

14:19

talked about this before, but this idea

14:21

of building autonomous businesses,

14:24

right? So this idea of you being the

14:26

only sort of human overseeing a bunch of

14:28

AI agents that are going out there on

14:31

your behalf doing various things sort of

14:33

saying in the near future there might be

14:34

a self-driving taxi that owns itself and

14:37

starts to run its own business. It

14:39

starts simple with the taxi having

14:41

access to a wallet and being able to

14:42

give people rides and using its earnings

14:45

to pay for its own maintenance and

14:46

operations. It uses this payment method

14:49

to learn how to create a website for

14:51

itself, find people to hire if it needs

14:53

cleaning, or post on social media

14:55

advertising and services. Soon, we could

14:57

have entire businesses and storefronts

14:59

operated by AI. Now, again, like we've

15:02

we've been talking about this for a

15:05

while now, and this is indeed where

15:07

things are going. We're still very

15:10

early, but as you can see, serious

15:12

companies, I mean, a lot of the Google's

15:14

behind a lot of this. They're building

15:16

out the infrastructures. They're

15:18

thinking ahead and going, "Okay, this is

15:20

happening. We know it's happening." And

15:22

they're trying to sort of skate to where

15:24

the puck is going by building out the

15:27

infrastructure on which this whole thing

15:30

will likely run. So, I think we're going

15:31

to need to do a separate video about the

15:34

X42 bazaar because that is the discovery

15:39

layer, right? So, it's the discovery

15:41

layer for that whole ecosystem. So

15:43

notice we already have at least two sort

15:46

of marketplaces of this nature. Google

15:49

and now Coinbase. Okay, we have to read

15:52

this little note that they have here cuz

15:54

it's hilarious. They're saying the X42

15:57

Bazaar is in early development. While

15:59

our vision is to build the Google for

16:01

agentic endpoints. We're currently more

16:03

like Yahoo search functional but

16:06

evolving. Features and APIs may change

16:08

as we gather feedback and expand

16:10

possibilities. So like we want to be

16:12

like Google eventually. That's the goal.

16:13

Right now we're kind of like Yahoo. So

16:16

work in progress. All right. So also

16:18

keep in mind that few days before this

16:21

got released, Google DeepMind published

16:23

their virtual agent economies. So this

16:26

is the Google research paper that they

16:28

published. By the way, quick side

16:30

tangent. There's this article about it

16:32

saying Google researchers warn of

16:33

looming AI run economies. These sandbox

16:37

economies could emerge that magnify

16:39

inequality, monopolize resources, and

16:41

create systemic market risks. Google

16:43

researchers warn. Risks include systemic

16:46

crashes, monopolization, and widening

16:48

inequality. And while those things are

16:50

sort of mentioned as potential risks,

16:52

that not not really what the paper was

16:54

about. The paper was really about this

16:57

idea that hey, we we do need to create a

16:59

separate sandbox economy and this is

17:03

what that would look like. In other

17:04

words, the point wasn't all of this is

17:07

coming and this is in our future. The

17:08

point was, hey, just to avoid issues

17:10

like this, we need to have this

17:12

solution. It needs to be separate so

17:14

that the agents and AI, they can kind of

17:17

do their own thing off to the side

17:18

without interacting with the regular

17:21

economy. That's why they're building all

17:23

this stuff. But so just FYI, if you see

17:26

scary headlines, understand that that's

17:29

not what was in the paper necessarily.

17:31

They said, "Yeah, these are risks.

17:33

Therefore, we should do this. If you're

17:35

interested, here's the point. They're

17:36

saying, "Our current trajectory points

17:37

us towards, you know, this big shift

17:40

that's coming, presenting us with

17:42

opportunities of awesomeness as well as

17:45

significant challenges, which they list

17:48

here. I'm just concerned that all the

17:50

newspaper articles will just focus on

17:53

this part and just ignore the rest of

17:55

the paper." And the whole point is this.

17:57

We discuss a number of possible designs

17:59

that may lead to safely steerable AI

18:02

agent markets. To boil down the point of

18:05

that paper into one sentence is that

18:07

these AI agent markets, they're coming.

18:09

They're coming in one way or another.

18:12

Accept that as fact. And Google saying

18:14

we need to figure out the rules of

18:15

engagement like how these marketplaces

18:18

will function. And they propose some

18:20

things that they think will make it run

18:22

better. This is part of the effort I

18:24

think of coordinating with Coinbase. But

18:26

my point is focus on the first portion

18:28

of that statement. These AI

18:31

marketplaces, they're coming. Assume

18:33

they're going to be here maybe a year

18:34

from now, maybe 5 years from now.

18:36

Knowing that, how do you approach things

18:40

differently moving forward? This is

18:42

going to be a massive change in society,

18:45

in the economy. Are you setting this one

18:47

out? And if you are, that's totally

18:49

fine. No shame in it. Just making sure

18:51

no one is, you know, a year, five years

18:53

from now going, "Shoot, I wish I knew

18:56

this was happening because I could have

18:57

been in the right place to prepare for

18:59

it." Don't be that person. We knew this

19:01

was coming and we're seeing the first

19:04

sort of examples of this being showcased

19:07

here. You're going to watch it being

19:09

developed alive by very capable tech

19:12

companies. Very soon we're going to be

19:14

hearing stories are on fully autonomous

19:17

businesses being run by people with AI

19:19

agents. And just like people made

19:21

millions in the dot era by creating

19:24

websites and online businesses, this is

19:26

like the next sort of iteration, the

19:28

next evolution of that. By the way, in

19:30

my experience with a lot of these online

19:32

businesses as the whole thing was

19:34

starting off, the people that are on

19:36

there in the beginning tend to do

19:38

extremely well because a lot of these

19:40

marketplaces or whatever they are,

19:42

right? So, Google's search engine, if

19:45

you were one of the first people to

19:46

figure out how to do SEO, you're under

19:48

you were doing great. Same thing with

19:50

YouTube and Twitter and Amazon and

19:52

Facebook ads and just about everything.

19:55

initially these companies are super

19:57

duper cool because they just want users

19:59

and they'll do whatever it takes to make

20:00

that happen. They want you on there,

20:02

right? So, in the beginning to kind of

20:04

jump start it to build that network

20:06

effect, which is extremely important.

20:08

It's also extremely difficult to build

20:10

it if you're not the first person to to

20:12

build it like to try to create another

20:14

strong business off of a network effect

20:17

that somebody else already captured is

20:18

extremely difficult. So, don't be

20:20

surprised if Coinbase and Google pour a

20:24

lot of money and effort and marketing

20:25

into building this thing out. What are

20:27

they going to need? They're going to

20:29

need agents that can do stuff. What kind

20:31

of stuff? Whatever stuff you can think

20:34

of. Background checks and financial

20:37

reports and anything. Anything that

20:39

people are searching for, they're going

20:40

to need some agent that's able to do

20:44

that thing for them. That's kind of like

20:45

asking, "What do people search for on

20:48

Google?" The answer is just anything and

20:50

everything. And it will likely be a

20:52

while before the big companies, the slow

20:55

movers will figure out how to use it,

20:57

how to be on there. So, there's going to

20:59

be an AI agent on these platforms that's

21:01

the first one offering legal help or

21:04

offering tax accounting help or help

21:07

with graphic design or copywriting or

21:10

building e-commerce stores or just

21:13

anything. By the way, the agent itself

21:16

doesn't necessarily have to be the one

21:18

building the ecom store. It can just be

21:20

sort of a a sales agent. It should be

21:23

there to interact with other AI agents.

21:26

Give them quotes, right? Negotiating the

21:28

fees, the timelines, kind of getting the

21:31

project specs. So, if there's something

21:34

that you yourself can do well, you can

21:36

have an agent on there just selling your

21:39

services. Then you get an email saying,

21:41

"Here, I've secured a deal for us." Then

21:44

you go to work, you complete that deal,

21:46

you hand it off to the agent, and you're

21:47

done. I mean, yes, that's not

21:49

technically a fully autonomous business.

21:51

But initially, people will just use that

21:54

thing for the sheer novelty factor just

21:56

to find the deals that they're looking

21:58

for, etc. When the various app stores

22:01

launched on on phones, you know, keep in

22:04

mind that there's people that have made

22:06

millions making fart apps. If you don't

22:08

know what that is, it's an app where you

22:10

you push a button and it makes a noise.

22:13

Was this innovative technology never

22:15

before seen a musthave for the modern

22:18

world? No, not really. It was just

22:20

there. It was one of the first apps when

22:22

the thing was just getting launched. It

22:24

was very novel. The app marketplace was

22:27

brand new and a lot of people that were

22:29

first movers made a lot of money.

22:31

Somebody will build something as huge or

22:34

maybe even huger bigger than that, you

22:37

know, for for this for AI agents, might

22:39

be bigger than Google, might be bigger

22:41

than Amazon, might be bigger than any

22:43

other business we have right now because

22:45

you can literally have automation for

22:46

anything you can think of. And these

22:48

companies have tons of resources. They

22:50

can do pretty much anything. The only

22:53

thing that they need, the only thing

22:55

that they can't provide is well that

22:58

network effect. They need people to come

22:59

in and build apps and provide their

23:01

services and put out the AI agents just

23:03

like Google needed people to make

23:05

websites for it to organize. Amazon

23:08

needs sellers on there to sell their

23:09

products. X needs people to post their

23:12

stuff on there. So there's YouTube and

23:13

Instagram and every other content

23:16

platform. That's the game. You're

23:18

witnessing the birth of a brand new sort

23:21

of platform type, a new medium, and

23:25

Google and Coinbase are the first two.

23:27

Let me know if there's any other ones

23:29

that I'm missing, but as far as I can

23:30

tell, these are the first two AI agent

23:33

sort of network effect platforms, or at

23:36

least they're trying to be. Anyways,

23:38

this is obviously big in 5 years. Don't

23:41

tell me that nobody warned you. If you

23:43

made this far, thank you so much for

23:44

watching. Please hit the thumbs up. That

23:46

really helps. Right now, Google is doing

23:48

something weird with the YouTube views.

23:50

Hit thumbs up. That helps to show that

23:52

you are a real person and real human to

23:54

real human. Uh, that thumbs up would go

23:57

a long way. Thank you so much for

23:58

watching.

Video Information

YouTube ID: 8s6nGMcyr7k
Added: Sep 19, 2025
Last Updated: 5 months ago