Digital Disruption

00;00;00;23 - 00;00;01;22

Hey, everyone.

00;00;01;22 - 00;00;04;22

I'm super excited to be sitting down

with a mod Matt Stock.

00;00;04;26 - 00;00;09;01

He's the co-founder of Stable Diffusion,

the leading AI text to image model

00;00;09;06 - 00;00;12;13

and now a leading voice

for the AI revolution.

00;00;12;16 - 00;00;16;11

And Mod believes that

in the next 1000 days that the economy

00;00;16;11 - 00;00;20;08

and world order

as we know it will implode and be replaced

00;00;20;08 - 00;00;25;22

by an AI led order that shatters

our systems of money, work and meaning.

00;00;25;25 - 00;00;29;00

I want to ask him what parts

of this vision he's most certain about,

00;00;29;03 - 00;00;31;24

how he expects it to play out,

and what we can do

00;00;31;24 - 00;00;34;24

as mere humans

to get the future that we want.

00;00;35;00 - 00;00;39;20

Let's find out.

00;00;39;23 - 00;00;40;29

Ahmad, thanks so much for being here.

00;00;40;29 - 00;00;43;11

Super excited to,

you know, talk about a lot

00;00;43;11 - 00;00;47;19

of different stuff with you

and specifically, the last economy, the,

00;00;47;19 - 00;00;51;28

you know, the book you've written recently

about, you know, AI and societal,

00;00;51;28 - 00;00;54;25

you know, transformation, revolution,

whatever you want to call it.

00;00;54;25 - 00;00;58;15

And maybe the place we can start

is, I know one of the framing devices

00;00;58;15 - 00;00;59;22

you use is you try and predict

00;00;59;22 - 00;01;03;26

what's going to happen

next is this notion of inevitabilities.

00;01;03;29 - 00;01;07;15

And so what are some of the

inevitabilities that you see as being on

00;01;07;15 - 00;01;10;00

the horizon for us?

00;01;10;03 - 00;01;10;14

Yeah.

00;01;10;14 - 00;01;13;00

I think first of all thank you

for having me on.

00;01;13;00 - 00;01;15;18

I'm glad you managed to get the time.

00;01;15;18 - 00;01;19;04

So crazy time because whenever you're

00;01;19;04 - 00;01;22;07

dealing with exponential technology

like this, it's very difficult to frame.

00;01;22;07 - 00;01;23;01

Right.

00;01;23;01 - 00;01;26;18

Like it seems in the middle of AI every

single week, something else is happening.

00;01;26;18 - 00;01;29;14

And it's only been three years

since ChatGPT.

00;01;29;14 - 00;01;30;06

Give or take.

00;01;30;06 - 00;01;31;28

Like it's a crazy thing.

00;01;31;28 - 00;01;33;14

So I was like, what is inevitable?

00;01;33;14 - 00;01;37;09

If we look at the pace of this

and if we look at the actual technology,

00;01;37;09 - 00;01;38;12

what it's doing.

00;01;38;12 - 00;01;41;12

And so one of these things I had

is this concept of a metabolic rift

00;01;41;14 - 00;01;45;28

that occurs whereby, you know,

it was humans, and it was our computation

00;01;45;28 - 00;01;48;28

that allowed us to kind of scale

and get to a certain point.

00;01;48;28 - 00;01;53;03

But AI is now at this takeoff point

whereby it can handle computation

00;01;53;03 - 00;01;58;10

better than us, and the value of human

cognitive labor is likely to go negative

00;01;58;11 - 00;02;02;16

by not just a zero, but negative, because

we'll be the dumbest people on the team.

00;02;02;19 - 00;02;05;09

And we're already seeing that

in things like medical diagnosis

00;02;05;09 - 00;02;08;20

and others,

when I by itself can outperform a human.

00;02;08;23 - 00;02;13;02

And in fact, this is also in things

like code forces, competition, emo, etc..

00;02;13;05 - 00;02;15;21

It's not a human plus an AI

that's getting top on these benchmarks.

00;02;15;21 - 00;02;18;03

Now it's just an AI by itself.

00;02;18;03 - 00;02;20;01

So I don't see

how this isn't going to happen.

00;02;20;01 - 00;02;22;02

Even if the technology stopped today.

00;02;22;02 - 00;02;26;16

And it has some very profound implications

for society, you know,

00;02;26;19 - 00;02;31;27

for the way that we have our economies

and more when things like that happen.

00;02;32;00 - 00;02;32;29

So let's let's maybe

00;02;32;29 - 00;02;36;04

talk about that,

and about some of the predictions

00;02;36;04 - 00;02;39;06

you're making here, and you've sort of

framed it out as, you know, there's

00;02;39;07 - 00;02;42;14

sort of three futures

that we're looking at.

00;02;42;17 - 00;02;46;29

Where do you see the road potentially

going in this world where humans are now?

00;02;46;29 - 00;02;51;20

The dumbest people on the team and,

you know, intelligence or,

00;02;51;24 - 00;02;56;18

you know, cognitive

labor is now basically a free resource.

00;02;56;21 - 00;02;58;27

Yeah, I think in the way that some of us

00;02;58;27 - 00;03;01;10

put it, is intelligence too

cheap to measure, right.

00;03;01;10 - 00;03;03;08

And so again, what's the market for that?

00;03;03;08 - 00;03;04;17

Nothing.

00;03;04;17 - 00;03;07;09

I kind of outlined three different roads.

00;03;07;09 - 00;03;09;25

One is this kind of great fragmentation.

00;03;09;25 - 00;03;11;10

You have Chinese AI, American

00;03;11;10 - 00;03;13;24

AI and others, and everyone's,

like, in their little bubbles

00;03;13;24 - 00;03;15;23

because they're a bit

scared of what's going on. The other

00;03;15;23 - 00;03;20;10

is this controlled by big corporations

like Elon's got a million GPUs.

00;03;20;10 - 00;03;23;00

What are they? It's 100 million workers.

00;03;23;00 - 00;03;26;16

And so his new company,

Macro Hard is just going to boot up

00;03;26;16 - 00;03;29;07

SAS companies

to take on the existing companies.

00;03;29;07 - 00;03;31;24

And so who controls the GPUs. Controls

the world.

00;03;31;24 - 00;03;34;06

And the final kind of futurist

is more symbiotic.

00;03;34;06 - 00;03;37;13

One where the AI is not there to replace

us, but augment us.

00;03;37;16 - 00;03;39;21

And there's a question of

how does value flow?

00;03;39;21 - 00;03;40;16

How does money flow?

00;03;40;16 - 00;03;43;11

How does meaning operate in that world?

00;03;43;11 - 00;03;46;00

And so that's what I was trying to

illustrate potential futures of

00;03;46;00 - 00;03;47;24

with the last economy.

00;03;47;27 - 00;03;49;23

So there's

00;03;49;23 - 00;03;52;19

there's obviously no shortage of things

to be concerned about in,

00;03;52;19 - 00;03;54;29

in some of those scenarios,

you know, whether it's,

00;03;54;29 - 00;03;57;28

you know, China or some of these,

you know, tech, you know, mega

00;03;57;28 - 00;04;00;29

corporations

that the Elon Musks of the world.

00;04;01;01 - 00;04;04;13

And so, you know, one of the things

that I thought was interesting,

00;04;04;16 - 00;04;08;03

odd is, you know, you made

you made the book freely available.

00;04;08;03 - 00;04;11;22

And, you know, for me,

I and you can tell me if you disagree,

00;04;11;22 - 00;04;15;16

but I sort of reverentially think of it

as not just a book, but like a manifesto,

00;04;15;21 - 00;04;19;22

of what's going to happen next

and what we need to know about it.

00;04;19;25 - 00;04;22;20

So I wanted to ask you

why did you choose to make it free?

00;04;22;20 - 00;04;25;20

And what are the most important messages,

do you think, for, you know, most

00;04;25;20 - 00;04;31;12

listeners about how we need to be

proceeding into this brave new world?

00;04;31;15 - 00;04;34;15

Yeah, I mean kind

of made it free to spread because open

00;04;34;15 - 00;04;37;15

source spreads and because it's

very difficult to have the answers.

00;04;37;15 - 00;04;38;07

Right.

00;04;38;07 - 00;04;41;09

Like, there's a lot of kind of gurus

are like, hey, I got the answer to this.

00;04;41;10 - 00;04;44;18

I'm not sure what the answer is

because this is unprecedented time.

00;04;44;18 - 00;04;48;24

So it was like put forward things and then

we can have version two, version three.

00;04;48;27 - 00;04;51;17

And this is what we saw

in my previous company, stability AI.

00;04;51;17 - 00;04;55;07

So we created open

source models had 300 million downloads.

00;04;55;10 - 00;04;57;09

Stable

diffusion was the most famous of those.

00;04;57;09 - 00;05;00;22

This image generating model

that created the creative boom.

00;05;00;25 - 00;05;03;13

I left that last year

because I was like, who's building the AI

00;05;03;13 - 00;05;07;10

to teach the kids, manage our health,

guide our governments, you know?

00;05;07;10 - 00;05;09;11

And then who's building the

AI for that stuff?

00;05;09;11 - 00;05;12;27

Like, how do we make sure

this is available to everyone?

00;05;13;00 - 00;05;15;29

I think the key message of the book

is one of,

00;05;15;29 - 00;05;20;09

you know, these agents are coming to

replace our jobs and everything like that,

00;05;20;12 - 00;05;23;12

or we can use this

AI to solve the biggest problems

00;05;23;17 - 00;05;27;13

or the smallest problems that

we have to us, if we change our framing,

00;05;27;16 - 00;05;30;23

if you start using it daily,

then you actually have skin in the game

00;05;30;23 - 00;05;33;16

and you kind of care.

You'll be ahead of the pack.

00;05;33;16 - 00;05;37;21

But then we need to think about

our own capitals and our own bases of

00;05;37;24 - 00;05;39;11

what does it mean to be me?

00;05;39;11 - 00;05;42;04

Like, am I a accountant or a lawyer

00;05;42;04 - 00;05;46;03

or all these other things that works

in the classical age?

00;05;46;03 - 00;05;49;00

But it might be that the

AI accountant is better than you.

00;05;49;00 - 00;05;52;07

Then it's more about where

meaning comes from in society.

00;05;52;07 - 00;05;53;11

And I've got a chapter about that.

00;05;53;11 - 00;05;56;11

Like it's your network,

like it's your family,

00;05;56;11 - 00;05;59;17

it's your church,

it's your community groups and more.

00;05;59;20 - 00;06;01;23

If you build those,

you become much more resilient.

00;06;01;23 - 00;06;05;03

And so we go through the mathematics

of the economics of how

00;06;05;03 - 00;06;10;01

you actually become more resilient, how

you can leverage this technology and more.

00;06;10;04 - 00;06;13;05

And again, it was meant to be an initial

framework that others could take.

00;06;13;05 - 00;06;14;19

And expand upon.

00;06;14;19 - 00;06;15;12

And we'll be releasing

00;06;15;12 - 00;06;19;27

actually the full math of this very soon,

hopefully at the start of the year.

00;06;19;27 - 00;06;25;10

And, I hope that people will take that

and be able to build on that as well.

00;06;25;13 - 00;06;26;05

If you work in

00;06;26;05 - 00;06;29;25

IT, Infotech research Group is a name

you need to know.

00;06;29;28 - 00;06;32;28

No matter what your needs are, Infotech

has you covered.

00;06;33;03 - 00;06;34;10

AI strategy?

00;06;34;10 - 00;06;36;22

Covered. Disaster recovery?

00;06;36;22 - 00;06;37;22

Covered.

00;06;37;22 - 00;06;40;07

Vendor negotiation? Covered.

00;06;40;07 - 00;06;44;00

Infotech supports you with the best

practice research and a team of analysts

00;06;44;00 - 00;06;47;23

standing by ready to help you

tackle your toughest challenges.

00;06;47;26 - 00;06;52;25

Check it out at the link below

and don't forget to like and subscribe!

00;06;52;28 - 00;06;57;23

I love the sort of the really practical

lens of what do we do as individuals

00;06;57;23 - 00;07;01;17

to find meaning and to sort of reframe

our own lives around us.

00;07;01;22 - 00;07;02;23

I do want to just,

00;07;02;23 - 00;07;06;11

you know, briefly zoom out a little bit

and talk about the disruption.

00;07;06;11 - 00;07;09;13

And we've talked,

you know, very broadly about,

00;07;09;18 - 00;07;13;20

computational intelligence, you know,

going to zero the impact on humans.

00;07;13;23 - 00;07;15;08

Can you share a little bit though?

00;07;15;08 - 00;07;17;01

You know, in the next

00;07;17;04 - 00;07;17;23

three ish

00;07;17;23 - 00;07;21;22

years, thousand ish days, what impact

do you see on this

00;07;21;25 - 00;07;25;24

having on the economy, on capitalism,

on our jobs and enterprises?

00;07;25;25 - 00;07;30;23

Like what level of disruption

should we be ready for?

00;07;30;26 - 00;07;34;27

If your job can be done on the other side

of a screen within two years,

00;07;34;27 - 00;07;39;10

the AI will be able to do it better

for pennies, I think is the headline.

00;07;39;13 - 00;07;43;03

So if you've got an in-person job, then

it's a while till the robots come, right?

00;07;43;03 - 00;07;46;03

Although they're coming much faster

than I think anyone expected.

00;07;46;06 - 00;07;48;05

And there have been several

breakthroughs there.

00;07;48;05 - 00;07;52;08

But where the AI is right

now is the start of 2025.

00;07;52;11 - 00;07;56;25

Everyone was on GPT four,

which was a bit of a dumb model,

00;07;56;28 - 00;07;59;25

you know, and it hallucinated

that all of this kind of stuff.

00;07;59;25 - 00;08;04;02

Now you have agents

that can check their work reason

00;08;04;05 - 00;08;08;07

getting gold medals in international math

Olympiads coming top of code forces,

00;08;08;08 - 00;08;13;14

being able to build for hours

and check their work.

00;08;13;17 - 00;08;15;11

And this is really what is economic.

00;08;15;11 - 00;08;17;06

Most of the benchmarks are now saturated.

00;08;17;06 - 00;08;20;05

So the benchmarks

that are coming through of things like,

00;08;20;05 - 00;08;24;02

vending machine batch

where the AI runs a vending machine, sees

00;08;24;02 - 00;08;27;17

how much money it can make, you know,

literal dollars of the benchmark.

00;08;27;17 - 00;08;33;25

Now, because 2026 is the

year of the economic agent

00;08;33;28 - 00;08;34;10

that can

00;08;34;10 - 00;08;37;10

actually go and replace humans

on the other side of the screen.

00;08;37;14 - 00;08;42;25

But 2027 is the year where using them is

literally it's this new economic Turing

00;08;42;26 - 00;08;45;13

test, like you're having a zoom

call with it, just like now.

00;08;45;13 - 00;08;50;11

You don't know if it's me or my digital

double a company to replace its workers.

00;08;50;11 - 00;08;54;08

Your company can just take all your slack

messages, all your emails.

00;08;54;08 - 00;08;57;19

Create

a digital double of you as an assistant.

00;08;57;22 - 00;09;02;22

And then 2728, the assistant replaces you

because it never goes to sleep.

00;09;02;25 - 00;09;05;16

You know it never makes a mistake twice.

00;09;05;16 - 00;09;09;17

It's always personable,

you know, and it works well with others.

00;09;09;20 - 00;09;11;27

And literally by talking, it can do that.

00;09;11;27 - 00;09;14;28

And so this technology for me

is an inevitability of the way

00;09;14;28 - 00;09;18;29

this all comes together

around the cognitive core

00;09;19;02 - 00;09;22;24

and what won't be the economic impact of

that is kind of a thing.

00;09;22;24 - 00;09;27;28

Like there have been studies show

that that type of jobs is about 50% of

00;09;28;01 - 00;09;31;08

the white collar workforce, and it takes a

while for people to be fired

00;09;31;11 - 00;09;33;28

or let go, because no one likes to have

teams like that.

00;09;33;28 - 00;09;35;12

But then

00;09;35;15 - 00;09;36;06

it's going to be

00;09;36;06 - 00;09;40;19

companies using AI outcompete companies

not using AI.

00;09;40;22 - 00;09;42;20

And they've got less,

less need for people.

00;09;42;20 - 00;09;44;01

And that will be fully AI.

00;09;44;01 - 00;09;48;17

Companies outcompete AI,

plus human companies

00;09;48;20 - 00;09;51;02

because again,

what are we missing from that equation?

00;09;51;02 - 00;09;53;14

And again, this is why I put

00;09;53;14 - 00;09;56;25

when I published the book in August,

it's about three years away.

00;09;56;28 - 00;10;01;14

That real tipping point where your job,

if it can be done on the other

00;10;01;14 - 00;10;04;29

side of the screen, is economically

irrelevant, doesn't mean you'll get fired.

00;10;05;02 - 00;10;06;21

You know,

you could be working in the public sector

00;10;06;21 - 00;10;09;26

or something where efficiency doesn't

really matter that much,

00;10;09;29 - 00;10;13;07

but it could be replaced as the key thing.

00;10;13;10 - 00;10;16;22

One and this is something

that's been on my mind recently, which is,

00;10;16;25 - 00;10;19;25

you know, this sort of dangerous question

to ask about what

00;10;19;25 - 00;10;23;26

is the right number of employees

for an enterprise to have.

00;10;23;29 - 00;10;27;22

And in the wake of this technology,

it feels like

00;10;27;25 - 00;10;32;00

what we're going to see is enterprises

sort of like almost racing to zero,

00;10;32;05 - 00;10;37;06

like asking themselves,

how few employees can I get away with?

00;10;37;09 - 00;10;39;01

Is that something? You know,

I don't know if.

00;10;39;01 - 00;10;41;21

Is that something on your horizon?

Is it something you're concerned about?

00;10;41;21 - 00;10;44;14

How do you see enterprises

that sort of framing this problem?

00;10;44;14 - 00;10;47;22

Yeah, you see it like just tipping over,

like Duolingo,

00;10;47;22 - 00;10;51;28

for example, were like,

we haven't fired anyone because of AI

00;10;52;01 - 00;10;54;03

and they're growing at 40% a year.

00;10;54;03 - 00;10;56;13

But they didn't hire anyone this year

either.

00;10;56;13 - 00;11;01;07

Normally a company, in order to grow,

you have to hire humans

00;11;01;10 - 00;11;03;00

because you need that extra cognition.

00;11;03;00 - 00;11;04;08

You need the extra coordination.

00;11;04;08 - 00;11;08;09

But we all know as companies get bigger,

the coordination overhead becomes massive.

00;11;08;09 - 00;11;10;07

Cost is theorem effectively.

00;11;10;07 - 00;11;13;23

You know, Dunbar's number,

you know, 150 people effective team

00;11;13;23 - 00;11;16;23

seem to be about 12 people roughly.

00;11;16;26 - 00;11;20;06

But now you look at the things

that people usually hire for sales

00;11;20;06 - 00;11;22;16

and all these kind of other things,

if they can be done remotely, again,

00;11;22;16 - 00;11;26;09

the AI salesperson will call and follow up

every single lead.

00;11;26;12 - 00;11;28;08

Well, again,

if it's in person, it's a bit different.

00;11;28;08 - 00;11;32;10

But that in-person person

doesn't need to hire the graduates anymore

00;11;32;10 - 00;11;34;19

because they can be more efficient

than having people lost.

00;11;34;19 - 00;11;35;19

And those have shown studies

00;11;35;19 - 00;11;39;25

showing that early stage employment

is starting to fall off a cliff.

00;11;39;28 - 00;11;43;25

You know, you see BPO and offshoring

that's trying to fall off a cliff.

00;11;43;28 - 00;11;47;06

It's basically

as if we discovered a brand new continent,

00;11;47;06 - 00;11;51;26

I Atlantis, with a trillion workers,

and they work way below minimum wage.

00;11;51;26 - 00;11;55;20

They work for electricity,

you know, they need food and other things

00;11;55;23 - 00;11;57;21

and they're tax deductible.

00;11;57;21 - 00;12;01;04

So when a company

is coming to make a decision,

00;12;01;07 - 00;12;04;21

just like every head teacher had to ask,

can I say essays for homework anymore?

00;12;04;22 - 00;12;06;26

They will say, do I hire a person?

00;12;06;26 - 00;12;10;28

And, Toby, look at Spotify

kind of said this.

00;12;10;29 - 00;12;14;06

It's like before

any job is an added Spotify,

00;12;14;09 - 00;12;18;01

you have to justify

why I can't make that job or do that job.

00;12;18;04 - 00;12;20;23

And so that list of things

is going to get smaller

00;12;20;23 - 00;12;24;02

and smaller and smaller

and smaller companies will go

00;12;24;02 - 00;12;27;16

for the top line first,

which is extra efficiency for employees.

00;12;27;19 - 00;12;31;24

But then once you have the data

to train the eyes and it's easy

00;12;31;24 - 00;12;37;16

to do on your cloud, you can just replace

the AI's, the people with the eyes.

00;12;37;19 - 00;12;41;29

And I think it's important understanding

the actual economics of this.

00;12;42;02 - 00;12;46;21

So the average person speaks 20,000

tokens a day.

00;12;46;24 - 00;12;50;04

It's about 1.3 tokens per word.

00;12;50;07 - 00;12;50;28

They speak.

00;12;50;28 - 00;12;53;08

They think about 200,000 tokens a day.

00;12;53;08 - 00;12;55;15

It's about 200,000 words.

00;12;55;15 - 00;12;59;22

That's about, let's say

10 million, 100 million tokens a year.

00;12;59;25 - 00;13;01;26

Right now,

00;13;01;26 - 00;13;07;19

GPT five is $10 per million tokens.

00;13;07;22 - 00;13;10;03

So we're talking a thousand bucks a year.

00;13;10;03 - 00;13;13;28

Grok for fast,

which scores 93% on tower bench,

00;13;13;28 - 00;13;17;21

for example, which is a call center worker

benchmark and human score, 90%.

00;13;17;24 - 00;13;20;24

It's $0.50 per million tokens.

00;13;20;29 - 00;13;22;27

So you can replace a call center worker

who's world

00;13;22;27 - 00;13;27;08

class of $500

next year will be ten times cheaper.

00;13;27;11 - 00;13;31;04

So they're not talking like,

hey, here's a $100,000 person.

00;13;31;04 - 00;13;34;15

I'm gonna replace them with a $80,000 AI.

00;13;34;15 - 00;13;38;13

It's I'm going to replace

the $100,000 worker with $1,000.

00;13;38;16 - 00;13;41;05

Hey, I and that's just how it is.

00;13;41;05 - 00;13;43;02

A company with a fiduciary responsibility.

00;13;43;02 - 00;13;46;13

Can you say,

well, I'm going to keep the worker,

00;13;46;16 - 00;13;48;19

you know.

00;13;48;19 - 00;13;50;28

So one of the one of the most,

00;13;50;28 - 00;13;55;21

you know, I think salient points,

counterpoints to all of this

00;13;55;24 - 00;14;00;24

is that at the

the current level of technology right now,

00;14;00;27 - 00;14;03;18

what we're just sort of at that

tipping point

00;14;03;18 - 00;14;07;05

where the technology is not quite

especially at an urgent tech level,

00;14;07;08 - 00;14;10;15

ready to have this sort

of large scale replacement.

00;14;10;15 - 00;14;14;27

And so it's contingent on what

that technology curve looks like.

00;14;14;27 - 00;14;16;29

And there's, you know, as you said,

there's the energy piece

00;14;16;29 - 00;14;20;15

and the efficiency piece,

but there's also the capabilities piece.

00;14;20;18 - 00;14;25;29

And, you know, whether it sort of tops out

and we start to see incremental gains

00;14;25;29 - 00;14;28;07

or whether it goes exponential,

it can really,

00;14;28;07 - 00;14;30;26

you know, end up

in a number of different directions.

00;14;30;26 - 00;14;31;19

Do you have a prediction

00;14;31;19 - 00;14;32;29

made for where it's going to go,

00;14;32;29 - 00;14;35;13

or do you think that's just

sort of a silly question

00;14;35;13 - 00;14;37;13

and it's basically

only going in one direction.

00;14;37;13 - 00;14;39;06

And that's what matters.

00;14;39;06 - 00;14;43;29

So I think there's this question of do

the models need to be more capable.

00;14;43;29 - 00;14;46;03

And those are this question of AGI,

00;14;46;03 - 00;14;48;24

you know, an AI,

they can do anything a human can do.

00;14;48;24 - 00;14;52;09

I think the answer is they don't need to

be more capable than today.

00;14;52;12 - 00;14;55;18

They just need the right framework.

00;14;55;18 - 00;14;58;28

They need the right connectors

like even real time now.

00;14;58;28 - 00;15;02;00

Like you have synthesizers and 11 labs

and other things like that.

00;15;02;00 - 00;15;03;09

Again,

00;15;03;09 - 00;15;06;23

me right here,

the technology is there for real time

00;15;06;23 - 00;15;11;26

replication of a video

where you can't tell it's me or an AI.

00;15;11;29 - 00;15;13;16

I just needs to be plugged in.

00;15;13;16 - 00;15;16;00

And then we see the models

at a certain capability.

00;15;16;00 - 00;15;18;15

We know what the pricing of that

looks like.

00;15;18;15 - 00;15;21;08

Again,

even if we didn't advance from today,

00;15;21;08 - 00;15;23;21

and we will advance from today

because the chips are going

00;15;23;21 - 00;15;27;16

to get ten times

cheaper in terms of tokens per dollar.

00;15;27;19 - 00;15;29;26

We know those numbers are coming.

00;15;29;26 - 00;15;33;07

It's still only going to be less

than a buck per million tokens.

00;15;33;07 - 00;15;35;11

And the tokens have got good.

00;15;35;14 - 00;15;38;03

To give an idea of how good the tokens

have got.

00;15;38;03 - 00;15;41;17

The new claw 4.5 opus

00;15;41;20 - 00;15;43;09

from Anthropic

00;15;43;09 - 00;15;44;23

they sent.

00;15;44;23 - 00;15;45;05

They have this

00;15;45;05 - 00;15;47;28

take home test for applicants to anthropic

and it's quite hard.

00;15;47;28 - 00;15;49;06

You know,

00;15;49;06 - 00;15;51;25

it scored higher

than any human had ever school before.

00;15;51;25 - 00;15;55;12

And this is the take home test,

which means you can use anything.

00;15;55;15 - 00;15;59;00

And in discussions

I've had with various top programmers

00;15;59;00 - 00;16;02;11

and again, my own experience, it's

just a really, really good coder.

00;16;02;14 - 00;16;07;21

It doesn't make many mistakes

and so to replace workers who are

00;16;07;21 - 00;16;12;01

basically machines taught by society,

I think the technology is not good enough.

00;16;12;01 - 00;16;14;02

It just needs to be again put together.

00;16;14;02 - 00;16;16;08

I've given 1 to 2 years for that,

but again,

00;16;16;08 - 00;16;19;27

I think 2026 is the year that takes off

00;16;20;00 - 00;16;22;22

because you don't need continuous

learning, you don't need algorithmic

00;16;22;22 - 00;16;26;26

breakthroughs, you don't need to have

a big leap in capabilities for this.

00;16;26;29 - 00;16;29;21

You just got to put it together.

00;16;29;21 - 00;16;32;27

And so that constructionist view

is the key thing.

00;16;33;00 - 00;16;35;22

If there are capability advances

00;16;35;22 - 00;16;38;19

and other things,

then it gets even crazier.

00;16;38;19 - 00;16;41;01

Like and I think we see that Nvidia.

00;16;41;01 - 00;16;43;18

So two years ago

and up two years and a bit ago

00;16;43;18 - 00;16;45;21

we really stable video,

which was state of the art video.

00;16;45;21 - 00;16;48;08

And it was like taking an image

00;16;48;11 - 00;16;48;23

and then it

00;16;48;23 - 00;16;51;23

move it slowly, you know smooth motion.

00;16;51;23 - 00;16;53;00

Now you look at curling,

00;16;53;00 - 00;16;56;00

you look at runway,

you look at other things, you know, like

00;16;56;02 - 00;17;01;00

that's pretty much almost Hollywood level

for pennies.

00;17;01;03 - 00;17;05;23

But more than that,

you look at something like Curling Omni.

00;17;05;26 - 00;17;08;16

You can edit the video just by talking.

00;17;08;16 - 00;17;10;23

Like he's nodding.

He should be shaking his head.

00;17;10;23 - 00;17;12;17

And there it goes.

00;17;12;17 - 00;17;15;16

And something like

that was a capability advance.

00;17;15;16 - 00;17;17;11

It was in line with our expectations.

00;17;17;11 - 00;17;23;04

But again, it just reduces the friction

of adapting this technology like opus.

00;17;23;07 - 00;17;26;00

You can say all sorts of stupid stuff

and they'll still give you a good output.

00;17;26;00 - 00;17;27;15

You know,

00;17;27;15 - 00;17;29;24

and it doesn't do it

with a million tokens anymore.

00;17;29;24 - 00;17;33;14

Actually, it's like four times

more token efficient as well.

00;17;33;17 - 00;17;36;23

So I think that again, if we freeze today

00;17;36;26 - 00;17;39;20

and there's

no advantage, I think there will be.

00;17;39;20 - 00;17;44;15

You still get ten times cheaper per year

and they'll still put this all together.

00;17;44;18 - 00;17;47;14

And again, remove all the friction

from removing the humans

00;17;47;14 - 00;17;51;01

from the workforce

and bringing in the GPUs.

00;17;51;04 - 00;17;53;29

And that removing the humans

is really interesting to me because I,

00;17;53;29 - 00;17;54;07

you know,

00;17;54;07 - 00;17;57;27

I talked to enough people who say, oh,

no one will never remove the humans.

00;17;57;27 - 00;17;59;01

It's augmenting.

00;17;59;01 - 00;18;00;19

It's all augmenting.

00;18;00;19 - 00;18;02;18

AI is best at augmenting.

00;18;02;18 - 00;18;03;22

You don't have to worry about it.

00;18;03;22 - 00;18;07;21

No humans will be replaced, you know,

in the implementation of this technology.

00;18;07;21 - 00;18;11;19

And it sounds like you have like

a basically a violent disagreement

00;18;11;19 - 00;18;13;11

with that.

00;18;13;11 - 00;18;15;18

Well,

I mean, humans can be annoying, right?

00;18;15;18 - 00;18;19;27

Like the main issue of most companies

is the personnel.

00;18;20;00 - 00;18;22;20

I can't get enough good people now.

00;18;22;20 - 00;18;25;10

It's like McDonald's,

you know, like you go somewhere near,

00;18;25;10 - 00;18;27;22

you have a McDonald's because, you know,

it kind of works.

00;18;27;22 - 00;18;31;04

It gets rid of your hunger

and you know, you won't, like,

00;18;31;07 - 00;18;33;04

get food poisoning or whatever.

00;18;33;04 - 00;18;34;26

The AI will just work.

00;18;34;26 - 00;18;40;07

Again, this AGI concept is like super chef

geniuses in a data center.

00;18;40;10 - 00;18;42;22

I just want to have stars

in a data center.

00;18;42;22 - 00;18;46;19

You know, all marketers in the data center

or accountants in the data center.

00;18;46;22 - 00;18;51;00

Do I really care about my accountant that

much that I'll pay a thousand times more?

00;18;51;00 - 00;18;53;02

No. Not really. Right.

00;18;53;05 - 00;18;55;26

And so it was just this question

of removing the friction

00;18;55;26 - 00;18;59;13

because every technological advance

had this diffusion element.

00;18;59;13 - 00;19;04;18

The other thing with diffusion AI,

where you had to have 5G for the phones,

00;19;04;18 - 00;19;09;13

you needed to have wire

laid down for broadband, whereas this

00;19;09;13 - 00;19;12;21

AI uses our existing infrastructure

00;19;12;24 - 00;19;15;21

and has economies of scope.

00;19;15;21 - 00;19;19;15

The model, once

trained, means every ChatGPT user.

00;19;19;18 - 00;19;22;06

So I get smarter overnight

00;19;22;06 - 00;19;24;25

and more capable overnight.

00;19;24;25 - 00;19;29;18

And the level of capability

versus requirement.

00;19;29;21 - 00;19;34;07

A lot of the stories were the data centers

would have to get exponential,

00;19;34;10 - 00;19;38;27

and you need giant

millions of GPUs to like, I don't know,

00;19;39;00 - 00;19;43;20

get a gold medal in the Puttnam

or EMR right now.

00;19;43;20 - 00;19;46;19

So research recently released a math model

00;19;46;23 - 00;19;49;17

that has 30 billion practice parameters

three.

00;19;49;17 - 00;19;52;17

So 3 billion actor

parameters, 30 billion total.

00;19;52;19 - 00;19;55;16

So it'll run on a top level MacBook.

00;19;55;16 - 00;20;00;02

Actually, I think it runs on

a MacBook air even with CPU

00;20;00;05 - 00;20;02;13

by just swapping.

00;20;02;13 - 00;20;05;25

And that couldn't get that

last year would have got, I think, second

00;20;05;25 - 00;20;10;15

or fourth in the Puttnam math competition,

which is a hard math competition.

00;20;10;15 - 00;20;14;10

You know, it's the top undergraduate

math competition, and most jobs

00;20;14;13 - 00;20;17;12

are less complex than that.

00;20;17;12 - 00;20;20;03

And so I think what you'll see

is the current state

00;20;20;03 - 00;20;23;25

of the art models,

the old PCs, the GPT five thinking, etc.

00;20;23;28 - 00;20;26;27

will be on a MacBook in two years. Wow.

00;20;26;27 - 00;20;27;26

And that's a big deal.

00;20;27;26 - 00;20;32;10

Like Siri will finally be smart,

you know, not say such dumb stuff,

00;20;32;13 - 00;20;36;20

but then Siri will be able to do

your taxes.

00;20;36;23 - 00;20;38;10

It's a it's a compelling vision.

00;20;38;10 - 00;20;41;01

And, you know,

series is an interesting example.

00;20;41;01 - 00;20;45;16

Who do you see as being, I guess, best

positioned to take advantage of this?

00;20;45;17 - 00;20;50;02

Is it going to be the biggest enterprises

just, you know, the rich get richer.

00;20;50;02 - 00;20;54;10

Is it going to unlock, you know, this

sort of myth of the one person,

00;20;54;15 - 00;20;56;02

you know, billion dollar enterprise.

00;20;56;02 - 00;20;59;00

How is it going to change

the competitive landscape?

00;20;59;00 - 00;21;02;00

Yeah, I think you will have a $1,

a $1 billion dollar

00;21;02;00 - 00;21;05;00

enterprise in 2026 or 2027.

00;21;05;04 - 00;21;06;10

Again,

00;21;06;10 - 00;21;08;03

someone who's figured out how to use this

00;21;08;03 - 00;21;10;13

right now, it's about leverage

of actually using this.

00;21;10;13 - 00;21;14;05

So like if you use the technology daily,

like anyone listening to this

00;21;14;05 - 00;21;17;22

that runs a company, if you just tell

all of your workers to use

00;21;17;28 - 00;21;23;14

wrap player or any of these other tools,

AI agent hour in one for one hour a day

00;21;23;17 - 00;21;26;23

to build anything useful,

your company will do better

00;21;26;27 - 00;21;29;26

than any other company you know,

like just simple things like that.

00;21;29;26 - 00;21;32;13

Like how many people listen to this,

have actually built some stuff with AI.

00;21;32;13 - 00;21;35;11

It's a wonderful

experience right now. Right.

00;21;35;14 - 00;21;36;14

And so I think

00;21;36;14 - 00;21;39;14

that it raises the floor for everyone.

00;21;39;16 - 00;21;42;26

But ultimately what matters

typically in a business is distribution.

00;21;42;29 - 00;21;45;18

And so that's why, you know,

you have so many more teams

00;21;45;18 - 00;21;48;18

users than you do slack users

and other things like that.

00;21;48;23 - 00;21;51;12

And so you can build

in your existing distribution,

00;21;51;12 - 00;21;53;07

you can build on your existing attention.

00;21;53;07 - 00;21;54;18

It just doesn't cost as much.

00;21;54;18 - 00;21;57;17

Your character,

LTV, your cost of user acquisition

00;21;57;17 - 00;22;00;19

to your lifetime value equation

just changes.

00;22;00;23 - 00;22;03;01

And so as people build,

they should be thinking

00;22;03;01 - 00;22;06;01

about distribution,

distribution, distribution as it were.

00;22;06;08 - 00;22;09;22

But it does mean that

you don't have to scale humans anymore

00;22;09;22 - 00;22;13;22

if you're providing a genuinely useful

service like you used to have to.

00;22;13;24 - 00;22;17;17

And this is why you're seeing these

AI companies like know

00;22;17;18 - 00;22;21;13

got 250 million in revenue

in just two years era.

00;22;21;14 - 00;22;22;05

You know

00;22;22;05 - 00;22;26;09

cursor got to a billion and normally

it would have to sell into enterprises.

00;22;26;09 - 00;22;28;27

But it meant the enterprises

where they were in the ID,

00;22;28;27 - 00;22;30;23

the software development environment.

00;22;30;23 - 00;22;32;12

And so got to $1 billion in revenue.

00;22;32;12 - 00;22;33;21

Just like that.

00;22;33;21 - 00;22;36;13

Build useful stuff and you can get bigger.

00;22;36;16 - 00;22;39;05

But again, the distribution channels

are mostly controlled

00;22;39;05 - 00;22;43;24

by the big companies, but most of them

won't be able to move fast enough.

00;22;43;27 - 00;22;45;23

Well, and that's

that's exactly what I'm getting at, is

00;22;45;23 - 00;22;50;01

it seems like with scale,

at least right now for enterprises

00;22;50;04 - 00;22;54;17

comes this, you know,

sort of bureaucracy or, you know, there's,

00;22;54;20 - 00;22;57;05

you know, there's the metaphor about,

you know, the Titanic versus

00;22;57;05 - 00;23;02;28

the speedboat, but the, the sense

that to move an organization that big and,

00;23;03;01 - 00;23;06;06

you know, with any sort of agility,

it's just it's just difficult.

00;23;06;09 - 00;23;09;15

It Well, I think that there's

also this thing about like

00;23;09;19 - 00;23;13;09

if you're a software company,

you write your software once

00;23;13;09 - 00;23;18;02

and then you replicate it

and you have 60 to 80% margins right

00;23;18;05 - 00;23;23;07

now. It's like you can replicate almost

any piece of software in a week or two.

00;23;23;10 - 00;23;25;02

And so you're there.

00;23;25;02 - 00;23;28;10

Margin is your opportunity as a

00;23;28;13 - 00;23;29;21

new company, right.

00;23;29;21 - 00;23;35;01

Like write the software and then have the

AI support for the software.

00;23;35;04 - 00;23;36;15

So you see the software industry

00;23;36;15 - 00;23;40;09

as being one of the key areas that's

just ripe for disruption with these tools.

00;23;40;12 - 00;23;41;01

Massively.

00;23;41;01 - 00;23;44;21

I mean, again, these big enterprise

contracts take a while,

00;23;44;24 - 00;23;49;11

but their margin is your opportunity

for so many startups.

00;23;49;14 - 00;23;53;00

Like there was a good, interesting post,

a few weeks ago

00;23;53;03 - 00;23;56;06

by the head, of course,

one of the head of marketing at Kessler.

00;23;56;06 - 00;23;59;06

They were like, we're replacing our CMS

00;23;59;12 - 00;24;02;07

with just markdown files, and we're

building the whole front end ourselves.

00;24;02;07 - 00;24;04;09

So if I'm coding

00;24;04;09 - 00;24;05;23

and you want honest,

most companies can do that.

00;24;05;23 - 00;24;10;00

40% of the websites of

the internet is WordPress.

00;24;10;03 - 00;24;12;15

And you think about that whole industry

that's like you don't need it anymore.

00;24;12;15 - 00;24;15;06

You can just vibe code

something, you know,

00;24;15;06 - 00;24;18;14

and there's zero margin required for that.

00;24;18;17 - 00;24;20;05

But then, like I said, your

00;24;20;05 - 00;24;24;16

other content management systems,

your CRMs, your other software,

00;24;24;17 - 00;24;29;17

yeah, like software will be replicated

very, very, very quickly.

00;24;29;20 - 00;24;32;22

So you see looking at enterprise software,

it sounds like you see

00;24;32;22 - 00;24;35;08

the like the sales forces

and the Oracle's of the world

00;24;35;08 - 00;24;39;03

as potentially being in trouble

in the next few years as people find a way

00;24;39;03 - 00;24;43;19

to create some of those capabilities

at a fraction of the, of the cost.

00;24;43;22 - 00;24;45;01

I think it's a fraction of the cost.

00;24;45;01 - 00;24;48;02

But then also just removing the barriers

like Salesforce, Workday,

00;24;48;02 - 00;24;50;04

all these things are just painful

to integrate, right?

00;24;50;04 - 00;24;51;23

No one really wants to.

00;24;51;23 - 00;24;53;26

But that's why they have such sales

aggressiveness.

00;24;53;26 - 00;24;56;25

And this is why Salesforce is going

all in on I actually

00;24;56;27 - 00;25;00;10

they're even I researched he was building

some really interesting stuff.

00;25;00;13 - 00;25;03;02

I think Oracle you can see they moved

00;25;03;02 - 00;25;06;26

so aggressively to GPUs for that reason.

00;25;06;29 - 00;25;10;06

Like why Oracle is a cloud or GPU company.

00;25;10;09 - 00;25;14;03

And now Ellison is going all in on media

because attention is what you need.

00;25;14;09 - 00;25;14;19

That's it.

00;25;14;19 - 00;25;20;08

So I think the big SaaS companies

are really acknowledging this.

00;25;20;11 - 00;25;23;10

And you can see this for with Adobe,

for example,

00;25;23;14 - 00;25;26;25

if you want to use Adobe for free,

you can do it via ChatGPT.

00;25;26;28 - 00;25;29;26

You take an image

and then you call the Adobe mic

00;25;29;26 - 00;25;32;14

from your ChatGPT directly,

and you can just transform it.

00;25;32;14 - 00;25;36;14

Whatever you want at that point means in

a year or two, why do you even need Adobe?

00;25;36;17 - 00;25;37;28

Right?

00;25;38;01 - 00;25;40;06

Like they are under massive threat.

00;25;40;06 - 00;25;44;16

And it feels like there's

even beyond the enterprise software.

00;25;44;19 - 00;25;47;27

The the entire internet is under threat

in some ways

00;25;47;27 - 00;25;49;11

with that model, right, of why?

00;25;49;11 - 00;25;51;05

Why do you need to go to any website?

00;25;51;05 - 00;25;53;12

Why do you need to engage with any,

00;25;53;12 - 00;25;57;03

you know, piece of,

organic content versus, you know,

00;25;57;03 - 00;26;02;12

using ChatGPT or using an Lem to just call

exactly what you're looking for.

00;26;02;15 - 00;26;05;14

I mean, like, you know, I ChatGPT

00;26;05;14 - 00;26;08;28

to find the best car seat for my toddler.

00;26;09;01 - 00;26;10;23

Why would I go

and search the internet now?

00;26;10;23 - 00;26;11;24

I'm just like, go find it.

00;26;11;24 - 00;26;14;15

They have my constrictions and it's like,

yeah, he go and here's a full report.

00;26;14;15 - 00;26;15;29

And I'm like, great.

00;26;15;29 - 00;26;20;21

So I think that whoever controls

that first play next to you

00;26;20;24 - 00;26;23;01

is the main one,

because it'll be one AI controlling

00;26;23;01 - 00;26;25;06

all the other AIS

and the access to the internet.

00;26;25;06 - 00;26;28;06

Who is your Jarvis and Iron Man terms?

00;26;28;13 - 00;26;31;09

And so that's why you see

all these companies like

00;26;31;09 - 00;26;33;16

competing for that,

because it used to be the ChatGPT.

00;26;33;16 - 00;26;36;01

Now it's like 20 bucks a month.

00;26;36;01 - 00;26;39;14

It would cost about $3,000 a year

when it came out.

00;26;39;17 - 00;26;40;24

For the average user.

00;26;40;24 - 00;26;46;16

Now it costs like $0.50, like five bucks

for the total number of tokens per year.

00;26;46;19 - 00;26;48;28

Like they have to justify

their subscriptions and things,

00;26;48;28 - 00;26;51;28

just why they have advertisers,

why they have more services coming out.

00;26;52;04 - 00;26;53;25

It's good for consumers.

00;26;53;25 - 00;26;58;00

But again, that first

AI next to you is the most important one.

00;26;58;03 - 00;26;59;24

One. And you know, that

00;26;59;24 - 00;27;03;25

that feels like a lot

of the justification for the, you know,

00;27;03;25 - 00;27;08;15

if I can call it sort of the arms race

for controlling these AI platforms.

00;27;08;15 - 00;27;11;28

And, of course, you see the big companies,

the big AI and tech companies

00;27;11;28 - 00;27;16;03

in the space just commanding

a lot of the market growth right now.

00;27;16;06 - 00;27;19;27

Do you think that's poised to continue,

or do you see them being disrupted by,

00;27;20;04 - 00;27;22;07

open source or other alternatives?

00;27;22;07 - 00;27;25;04

How do you see that

particular playing field playing out?

00;27;25;04 - 00;27;28;09

So I think you see a bifurcation

of the best models now versus

00;27;28;09 - 00;27;30;00

the models that everyone has.

00;27;30;00 - 00;27;34;12

So we still haven't got the OpenAI model

that got the gold medal at the AMA,

00;27;34;15 - 00;27;36;14

you know, for example. Right.

00;27;36;14 - 00;27;39;24

And the best models require huge

amounts of compute and they're expensive,

00;27;39;24 - 00;27;43;10

whereas the consumer models are going

towards zero in terms of capability.

00;27;43;10 - 00;27;47;28

And they're saturating benchmarks like

how good a model do you need for ChatGPT?

00;27;48;02 - 00;27;51;02

Not really that much better

than what we have now, right?

00;27;51;03 - 00;27;53;05

Like in satisfy CES.

00;27;53;05 - 00;27;57;11

And so there is a real opportunity here

for open source other models

00;27;57;11 - 00;28;01;25

to fill that gap because they are cheaper,

they're optimized, etc..

00;28;01;28 - 00;28;05;21

But again, this is why you're seeing

the playing field shift, like OpenAI

00;28;05;22 - 00;28;09;23

hiring loads of meta people and Fujiko

and kind of others,

00;28;09;26 - 00;28;13;17

they're going aggressively

into that consumer network, play.

00;28;13;20 - 00;28;17;06

It's difficult for them, though,

because they don't really have the DNA

00;28;17;09 - 00;28;19;01

like Sora has.

00;28;19;01 - 00;28;22;19

It really taken off as an app, like it's

got some good traction, but it doesn't

00;28;22;19 - 00;28;26;10

have the inherent virality of a TikTok

or something like that, you know?

00;28;26;13 - 00;28;29;07

So instead, it's like, here are features

00;28;29;07 - 00;28;32;07

in an application

versus application itself.

00;28;32;07 - 00;28;33;21

The base models themselves.

00;28;33;21 - 00;28;37;23

Everyone said you have to keep up

with millions of GPUs

00;28;37;26 - 00;28;40;07

to train a deep sea level model.

00;28;40;07 - 00;28;42;06

Next year,

00;28;42;09 - 00;28;47;03

$500,000 to a million bucks from scratch.

00;28;47;06 - 00;28;48;29

And that's as good

a model as anyone needs.

00;28;48;29 - 00;28;52;09

So I think that that moat

has been disappearing.

00;28;52;09 - 00;28;55;02

And even,

you know, we need the GPUs to run them.

00;28;55;02 - 00;29;00;23

Like you'll be able to run a top level

model as good as a top level one tier.

00;29;00;23 - 00;29;06;14

Like I said, probably on MacBook

level compute in two years.

00;29;06;17 - 00;29;09;19

So it's not like that is a moat either.

00;29;09;19 - 00;29;13;17

The moat is distribution, the moat is

how much of the attention of the user

00;29;13;17 - 00;29;18;10

do you cross,

and how much does the user trustee?

00;29;18;13 - 00;29;22;23

Which which has valuation implications

for some of the big players here?

00;29;22;23 - 00;29;25;08

If the if the moat truly is eroding.

00;29;25;08 - 00;29;28;14

And I'm curious, ammad, you know,

you had said a couple of years ago

00;29;28;17 - 00;29;32;29

that, you know, AI is both $1

trillion opportunity, but also,

00;29;33;02 - 00;29;35;23

I think your quote was the biggest bubble

of all time.

00;29;35;23 - 00;29;38;12

And I'm curious Yeah,

00;29;38;15 - 00;29;39;13

I don't know 100%.

00;29;39;13 - 00;29;41;16

I think it will continue for a while.

00;29;41;16 - 00;29;44;11

But like if you're just talking

about language models

00;29;44;11 - 00;29;48;10

that are more than enough chips

in the world right now for those.

00;29;48;13 - 00;29;51;14

So the only reason that people need GPUs

00;29;51;14 - 00;29;54;23

is medium models

now, every pixel being generated.

00;29;54;26 - 00;29;59;26

But even though you don't know what

the lower bound in terms of efficiency is,

00;29;59;29 - 00;30;03;13

and there's trillions of dollars

going into building that out.

00;30;03;16 - 00;30;05;23

So I think that you probably

got another year or so

00;30;05;23 - 00;30;07;17

and then you've got a pullback,

00;30;07;17 - 00;30;09;10

but then you could

rapidly go up on the other side,

00;30;09;10 - 00;30;12;10

because we'll finally be smart enough to

figure out what to do with those tokens.

00;30;12;17 - 00;30;15;16

You know,

like it's genuinely very difficult

00;30;15;16 - 00;30;19;02

to use more than a few million

tokens a day per person.

00;30;19;05 - 00;30;24;02

But again, how much does a million tokens

cost for a front level model?

00;30;24;05 - 00;30;26;03

$0.50.

00;30;26;06 - 00;30;28;24

And so you're wondering,

what are the people going to use a token.

00;30;28;24 - 00;30;32;17

So and you're going from 4 million hoppers

to 10 million

00;30;32;17 - 00;30;35;18

Blackwell's,

which is 2 or 3 times faster than hoppers.

00;30;35;18 - 00;30;39;09

A lot of compute is coming online,

And then there is a question of things

00;30;39;09 - 00;30;43;29

like, sure, ChatGPT has 800 million users,

00;30;44;02 - 00;30;47;02

but it kind of looks like that

00;30;47;07 - 00;30;49;29

Google will just offer

the equivalent experience for free matter

00;30;49;29 - 00;30;53;03

will offer an excellent experience

for free, because the base cost

00;30;53;03 - 00;30;57;05

of providing a trapped GPT

experience has gone from 20 bucks a month

00;30;57;08 - 00;31;01;04

to 20 bucks a year to two bucks a year

next year.

00;31;01;07 - 00;31;05;28

I think we've ever seen anything like that

on the cost side, you know.

00;31;06;01 - 00;31;09;09

So if I'm following your logic

and I'm curious, it sounds like

00;31;09;12 - 00;31;12;18

if we play the clock forward a few years,

00;31;12;21 - 00;31;16;27

the value

being unlocked by this technology

00;31;17;00 - 00;31;21;15

is going to be more decentralized and

more diffuse than what we've seen so far.

00;31;21;15 - 00;31;22;12

It feels like right now.

00;31;22;12 - 00;31;25;19

So much of the value

has been created by the same,

00;31;25;22 - 00;31;28;22

you know, the same number of players

that you can count on one hand,

00;31;28;27 - 00;31;33;17

and that as we figure out more use cases

and more ways to, as you said, efficiently

00;31;33;17 - 00;31;37;12

use these tokens, we'll see more players

unlocking more value.

00;31;37;17 - 00;31;40;02

Do you buy that

or is that a misstatement? Yeah.

00;31;40;02 - 00;31;43;23

So I think it's like they figured out

how to build the car,

00;31;43;26 - 00;31;46;09

but it turns out the calls

on very expensive, right?

00;31;46;09 - 00;31;48;16

You could build one in your backyard.

00;31;48;16 - 00;31;50;15

But then you can also distribute it

again.

00;31;50;15 - 00;31;52;06

The friction here is very low.

00;31;52;06 - 00;31;55;23

You build a good app,

it can go to 100 million users.

00;31;55;26 - 00;31;58;21

Just because the backend is literally

just an Lambo.

00;31;58;21 - 00;32;00;25

It's literally just an image model.

00;32;00;25 - 00;32;03;24

Like, you don't even need

to have much more architecture than that.

00;32;03;24 - 00;32;04;13

And that means you're

00;32;04;13 - 00;32;08;15

then competing against an entire world,

like even things like outsourcing, right?

00;32;08;15 - 00;32;12;03

Like the speaking with a bunch of leaders

in India and places like that.

00;32;12;03 - 00;32;15;05

Recently, I was like, you,

00;32;15;08 - 00;32;15;20

every

00;32;15;20 - 00;32;16;26

single one of your workers now

00;32;16;26 - 00;32;20;21

will speak perfect English,

you know, in their interactions,

00;32;20;24 - 00;32;24;18

and they will write even better code

like you used to have to compete

00;32;24;18 - 00;32;27;22

against this level of code

coming from the outsource overseas.

00;32;27;22 - 00;32;29;26

Now you have to compete

against this level of code.

00;32;29;26 - 00;32;31;06

They're all now amazing lawyers.

00;32;31;06 - 00;32;35;20

They're all now amazing accountants,

and they can build applications

00;32;35;23 - 00;32;37;26

on their on their smartphones.

00;32;37;26 - 00;32;41;22

So I think that we will see

a big democratization as a result of this.

00;32;41;22 - 00;32;44;21

It's just again,

who controls that layer next to you,

00;32;44;21 - 00;32;48;25

especially as the AI

is going into governments and more things

00;32;48;28 - 00;32;50;26

and making more

and more important decisions,

00;32;50;26 - 00;32;54;03

like we will outsource

more and more decision making to these AI.

00;32;54;03 - 00;32;56;13

So that's why

I really think that for civic AI,

00;32;56;13 - 00;32;58;27

it needs to be open source

at the very least.

00;32;58;27 - 00;33;01;26

But again, that concept of

00;33;01;29 - 00;33;03;19

Dario's,

00;33;03;19 - 00;33;07;11

the only AI people will use

is the best idea from anthropic.

00;33;07;11 - 00;33;09;14

And you need $1 billion training one.

00;33;09;14 - 00;33;10;27

I think that's been proven false.

00;33;10;27 - 00;33;14;16

It was all about the data underlying it,

and the data got better quicker

00;33;14;16 - 00;33;18;24

than the compute

scaled up to make the data good.

00;33;18;27 - 00;33;21;07

Which feels like actually in some ways

a big win

00;33;21;07 - 00;33;25;21

for everybody outside of,

you know, the big tech companies, right.

00;33;25;21 - 00;33;29;29

If everybody now is more

and more capable of doing this.

00;33;30;02 - 00;33;33;17

yeah, I definitely I think so I think it's

the biggest democratizing technology

00;33;33;17 - 00;33;34;09

out there. Right.

00;33;34;09 - 00;33;39;05

It's just that you still have

a big advantage if you can scale your GPUs

00;33;39;05 - 00;33;42;04

for the frontier level work,

that is replacement work, right?

00;33;42;08 - 00;33;45;03

You still have a big advantage

if you own the distribution,

00;33;45;03 - 00;33;48;22

and you can lock it in

by having incredibly persuasive eyes.

00;33;48;25 - 00;33;51;09

You know, like Google and Messer

and others

00;33;51;09 - 00;33;54;09

will use the full persuasive

capability of their eyes.

00;33;54;09 - 00;33;57;01

So I think you're at a dangerous point,

but not maybe as dangerous

00;33;57;01 - 00;34;00;23

as some people thought,

where you basically would be cut off

00;34;00;23 - 00;34;03;23

from intelligence thanks to open source.

00;34;03;24 - 00;34;05;12

You know,

you'll always have another option.

00;34;05;12 - 00;34;08;18

But we got to make sure that the defaults

are, again, open

00;34;08;21 - 00;34;11;18

for the AI that's closest to you,

or at least it's aligned with you.

00;34;11;18 - 00;34;16;29

You shouldn't be outsourcing your

cognition to someone with a misaligned

00;34;17;02 - 00;34;17;15

set of

00;34;17;15 - 00;34;21;17

values, etc., so that personally,

AI is going to be a very important one.

00;34;21;20 - 00;34;24;11

One, let's let's talk about

outsourcing that cognition.

00;34;24;11 - 00;34;27;27

And maybe we can start at sort of,

organizational level.

00;34;28;00 - 00;34;31;16

you're a business leader

or you're an entrepreneur or,

00;34;31;19 - 00;34;35;17

you know, should you be

should you be looking now at creating,

00;34;35;17 - 00;34;40;08

you know, your a own AI tool or your own

AI platform versus leaning on,

00;34;40;14 - 00;34;43;17

you know, a Microsoft Copilot

or a ChatGPT like, to

00;34;43;17 - 00;34;47;10

what degree do you think

there's value competitively

00;34;47;13 - 00;34;51;22

in, you know, building

and creating your own models and tools?

00;34;51;25 - 00;34;53;26

I think you should create your own models.

00;34;53;26 - 00;34;57;29

Tools like you can use our AI agent

framework, which is a full compute to use

00;34;57;29 - 00;35;02;10

what it can build websites, applications,

everything fully open source

00;35;02;13 - 00;35;03;24

and it's compatible with all the others.

00;35;03;24 - 00;35;06;24

Like you can hook and cloud code

or codex etc.

00;35;07;01 - 00;35;10;08

because your organization

needs to be in the capability of like

00;35;10;08 - 00;35;14;05

I have capability, I have agency,

I can build and operate at that pace.

00;35;14;05 - 00;35;19;02

Interface now are nothing like again,

it will be another six months a year

00;35;19;03 - 00;35;22;29

before you have all the back end

connectors and is robust enterprise grade.

00;35;23;02 - 00;35;26;22

But the way the models and systems are

going right now, it's an inevitability.

00;35;26;25 - 00;35;31;11

So I think it's more about the muscle than

the actual like bias of Microsoft work.

00;35;31;12 - 00;35;35;01

In my best interest,

because I can tell you right now, like

00;35;35;04 - 00;35;38;03

if you buy a 200 pound a month

00;35;38;06 - 00;35;41;12

ChatGPT subscription for your organization

00;35;41;15 - 00;35;46;24

and you just use GPT 5.2 Pro

to help you in strategic decision

00;35;46;24 - 00;35;49;24

making, you'll probably

make better decisions as a company right?

00;35;49;29 - 00;35;52;25

Like, I don't really see how you're

not going to.

00;35;52;25 - 00;35;55;13

And again, it doesn't care about you

because it forgets about each thing

00;35;55;13 - 00;35;57;10

from bit

to because of that continuous learning.

00;35;57;10 - 00;35;59;12

But that will just make it better.

00;35;59;12 - 00;36;02;14

What's really concerning,

though, is I think more on the bias

00;36;02;14 - 00;36;03;13

on the personal side.

00;36;03;13 - 00;36;07;09

Like if it says a beer, it says Budweiser

versus Asahi or whatever,

00;36;07;12 - 00;36;09;15

you know, it's

concerning on the government side,

00;36;09;15 - 00;36;11;19

like who's making the decisions

like Trump?

00;36;11;19 - 00;36;14;14

Tariffs

had some elements of ChatGPT in the,

00;36;14;14 - 00;36;17;02

you know, em dashes

and other things like that.

00;36;17;02 - 00;36;18;19

I'm teaching our kids and stuff like that.

00;36;18;19 - 00;36;19;28

Again, they're very sensitive.

00;36;19;28 - 00;36;23;13

But from a corporate side,

I think the AI is pretty well behaved.

00;36;23;13 - 00;36;26;28

It's just, again, you need to work

that muscle of being able to build

00;36;26;28 - 00;36;30;07

and deploy, and that will increase

the velocity of what you're doing.

00;36;30;10 - 00;36;34;13

And then also just using the best of the

AI is to help you make those strategic

00;36;34;13 - 00;36;38;02

decisions and more, because why wouldn't

you want 150 IQ, buddy?

00;36;38;02 - 00;36;40;27

That's really meticulous, right?

00;36;41;00 - 00;36;43;22

That's worth more than any consultant

you can bring in.

00;36;43;22 - 00;36;47;05

Well, which is, you know,

kind of an interesting point about,

00;36;47;05 - 00;36;48;28

you know, the consulting industry

right now

00;36;48;28 - 00;36;53;07

and, you know, are they at continued risk

from all of this?

00;36;53;07 - 00;36;54;24

And, you know,

if you're a business leader,

00;36;54;24 - 00;36;56;20

how should you approach

strategic decisions?

00;36;56;20 - 00;37;01;20

And, you know, how should like what

what's your best advice for them in 2026?

00;37;01;23 - 00;37;02;01

Yeah.

00;37;02;01 - 00;37;04;24

So Accenture

all these guys are making billions now.

00;37;04;24 - 00;37;06;20

Genitive AI consulting. Right.

00;37;06;20 - 00;37;08;23

Okay. It's quite funny.

00;37;08;23 - 00;37;10;11

None of them have a supercomputer.

You know,

00;37;10;11 - 00;37;13;18

maybe I should launch a consultancy

that actually has a super computer.

00;37;13;21 - 00;37;17;13

The transformation is going to be key.

00;37;17;13 - 00;37;19;05

And again,

you need to have some expertise there.

00;37;19;05 - 00;37;23;17

But the truth of this technology

is that it's just very easy to use,

00;37;23;20 - 00;37;25;24

like even to actually build

the technology.

00;37;25;24 - 00;37;26;24

The core loops in

00;37;26;24 - 00;37;30;13

the biggest models are like

not more than a thousand lines of code.

00;37;30;16 - 00;37;34;19

It was kind of crazy, but to use it, you

just have to be able to type and speak.

00;37;34;22 - 00;37;36;12

It's just again,

that muscle kind of thing.

00;37;36;12 - 00;37;40;02

So for me, like I said, I would always say

00;37;40;05 - 00;37;44;07

right now everyone should be using AI

for a second opinion.

00;37;44;10 - 00;37;46;06

Because what's your downside,

00;37;46;09 - 00;37;47;11

right?

00;37;47;11 - 00;37;51;06

Everyone should be using AI regularly

to build up the muscle.

00;37;51;09 - 00;37;55;15

And again, for leaders of organizations,

what's your downside?

00;37;55;15 - 00;37;58;29

If everyone in your organization builds

for one hour a week

00;37;59;02 - 00;38;02;02

with AI, the zero downside

00;38;02;02 - 00;38;04;10

and it doesn't even matter what they build

like they get together,

00;38;04;10 - 00;38;07;07

they can build stuff internally,

externally, they can have competitions.

00;38;07;07 - 00;38;09;19

It's fun building

like even within a family context.

00;38;09;19 - 00;38;12;03

Like if you build together

with your family

00;38;12;03 - 00;38;13;24

one hour

a week, it's actually genuinely fun

00;38;13;24 - 00;38;16;17

and it brings people together

because it shows capability.

00;38;16;17 - 00;38;19;12

So I would say use it for a second

opinion, for everything.

00;38;19;12 - 00;38;22;20

Get into the habit of building

from top to bottom in your organization.

00;38;22;20 - 00;38;24;02

Like our job

00;38;24;02 - 00;38;30;00

application requires people to show us

what they have built with our tools.

00;38;30;03 - 00;38;31;01

And it's really interesting.

00;38;31;01 - 00;38;34;18

People like, wow, I never knew I could

build, you know, like that.

00;38;34;21 - 00;38;36;13

Why wouldn't you do that?

00;38;36;13 - 00;38;40;05

The best job applications for students

now is not here's my CV.

00;38;40;05 - 00;38;41;28

It's here's what I've built.

00;38;41;28 - 00;38;43;26

And that will guarantee

put you at the top of the pile

00;38;43;26 - 00;38;45;16

because like,

oh, this is a guy that can use AI.

00;38;45;16 - 00;38;47;27

And look at this cool thing

that he's built for my company.

00;38;47;27 - 00;38;51;28

How do that

take a few hours, you know, like still.

00;38;52;01 - 00;38;54;28

So I'd say those are probably

the key things right now at this stage.

00;38;54;28 - 00;39;00;11

And then the capabilities will

only increase over the next year or two.

00;39;00;14 - 00;39;03;15

One of the, you know,

one of the main sources of tension

00;39;03;15 - 00;39;07;10

I'm seeing and it's it's come up already

in this conversation.

00;39;07;13 - 00;39;11;14

Is this, you know,

sort of notion from employers of, yeah,

00;39;11;14 - 00;39;16;17

we want all our employees using AI

and whether it's building,

00;39;16;17 - 00;39;19;22

whether it's using it

for a strategic decision making, this,

00;39;19;25 - 00;39;23;05

you know, this narrative

that it's augmenting them.

00;39;23;08 - 00;39;28;00

And on the other hand, in the longer

term, we're saying, well,

00;39;28;00 - 00;39;32;03

and at some point

we will replace them with the AI itself.

00;39;32;03 - 00;39;35;05

So there's this sort of tension about

should I use it, shouldn't I?

00;39;35;05 - 00;39;38;14

Where where is this going

with my employer, with my job

00;39;38;17 - 00;39;44;23

and so if you're if you're an individual,

if you're an individual worker

00;39;44;26 - 00;39;48;17

working in some sort of,

you know, knowledge economy capacity,

00;39;48;20 - 00;39;52;02

what are the skills

that you would be leaning on?

00;39;52;02 - 00;39;55;21

What what are the practices

and the muscles you would be building

00;39;55;28 - 00;39;58;06

to futureproof yourself?

00;39;58;06 - 00;40;02;13

Yeah, I think this this concept

of intelligence and network, as it were.

00;40;02;16 - 00;40;06;04

So again, within the book I describe,

there's four types of capital material,

00;40;06;04 - 00;40;10;02

capital intelligence capital network

capital and diversity capital,

00;40;10;05 - 00;40;14;29

which is more like, you know, how many

options, the optionality that you have.

00;40;15;02 - 00;40;17;07

The key thing here is the ability,

capabilities

00;40;17;07 - 00;40;20;13

using AI and then your network

and people knowing about that.

00;40;20;13 - 00;40;22;28

And you'll do very well in what's coming.

00;40;22;28 - 00;40;26;26

Because the interesting thing about AI

is you go and you give it a prompt.

00;40;27;03 - 00;40;31;21

If it gets the next second, it doesn't

remember, it's not a logic tree.

00;40;31;24 - 00;40;34;25

It's like a movie file or a music file

that just sits there

00;40;34;25 - 00;40;35;15

and it's like a sieve.

00;40;35;15 - 00;40;38;21

You push words in and somehow other words

come out.

00;40;38;24 - 00;40;42;10

We're not quite sure how it works, even,

which is also a bit creepy, right?

00;40;42;13 - 00;40;46;00

So the really interesting thing here is

the AI doesn't have skin in the game.

00;40;46;00 - 00;40;47;18

It doesn't give a damn.

00;40;47;18 - 00;40;51;12

Whereas if you can deploy AI tools

and use AI tools and you give a damn

00;40;51;12 - 00;40;54;27

and you show that you are proactive

here, you'll be ahead of the pack

00;40;54;27 - 00;40;58;05

internally or externally,

your employability goes way up.

00;40;58;05 - 00;41;02;06

And again, like I said,

the classical CV was here's what I did.

00;41;02;09 - 00;41;07;13

Whereas now it's here's what I made and

here is what I'm capable of, shall we say.

00;41;07;16 - 00;41;10;14

So if you just even maintain

that like even create a personal website

00;41;10;14 - 00;41;12;19

where you show

all the things that you've made

00;41;12;19 - 00;41;17;04

that puts you ahead of the pack,

deploy it internally, speak up,

00;41;17;07 - 00;41;20;12

and you'll be ahead of the pack,

because the vast majority of people

00;41;20;12 - 00;41;24;09

are still scared of AI

in terms of kind of the replacement.

00;41;24;09 - 00;41;28;15

Again, I think it will be slowly

then all at once, because no one really

00;41;28;15 - 00;41;31;15

wants to fire

large numbers of employees like

00;41;31;22 - 00;41;34;16

there's the usual consultant 15%,

00;41;34;16 - 00;41;37;16

but not 50%, 70%, etc.

00;41;37;23 - 00;41;40;13

but at the same time, who knows what

the next few years will hold.

00;41;40;13 - 00;41;44;19

Like you get a recession, it's much easier

to fire people on the other side.

00;41;44;19 - 00;41;45;14

Are you going to again?

00;41;45;14 - 00;41;47;06

Are you going to hire the person

or the AI?

00;41;47;06 - 00;41;51;21

The AI it's cleaner. You know.

00;41;51;24 - 00;41;52;04

Let's see.

00;41;52;04 - 00;41;53;19

Scale lets you do the job.

00;41;53;19 - 00;41;58;17

So I would say just really care, give

a damn and then use this tool as leverage.

00;41;58;20 - 00;42;01;22

And again, I think the building

is definitely one thing.

00;42;01;23 - 00;42;04;03

As well as the day to day stuff.

00;42;04;03 - 00;42;05;25

I mean, it's interesting

on the other side,

00;42;05;25 - 00;42;08;25

copilot and stuff like that,

they're not good enough yet.

00;42;08;25 - 00;42;12;20

They're about to be good enough

on the day to day job, which is why I say

00;42;12;23 - 00;42;13;27

don't push. Why?

00;42;13;27 - 00;42;15;25

I'm not saying like use copilot

00;42;15;25 - 00;42;20;05

internally and all these tools like,

yeah, they'll get there.

00;42;20;08 - 00;42;24;01

The key thing here is that the muscle

and the second opinions,

00;42;24;04 - 00;42;27;09

those are the key things at this stage

before they actually become

00;42;27;09 - 00;42;31;25

genuinely useful next year

or sometime doing this year.

00;42;31;28 - 00;42;34;11

So flipping that on its head,

00;42;34;11 - 00;42;38;24

what are the roles and the, you know,

the types of employees that you see

00;42;38;24 - 00;42;42;23

as being most at risk of being disrupted

or being replaced.

00;42;42;23 - 00;42;45;17

And the next call at 18 months.

00;42;45;20 - 00;42;46;27

So if

00;42;46;27 - 00;42;50;11

again, your job is to be a machine

and it can be measured

00;42;50;11 - 00;42;53;11

against a manual,

00;42;53;14 - 00;42;56;04

the AI can replace you very quickly.

00;42;56;04 - 00;42;58;13

So I kind of call center worker

as an example of that.

00;42;58;13 - 00;42;59;23

Right.

00;42;59;23 - 00;43;04;07

Customer service reps, I think

00;43;04;10 - 00;43;07;05

within a couple of years

accountants, lawyers, other things.

00;43;07;05 - 00;43;08;28

The AI is not good enough

to be an accountant.

00;43;08;28 - 00;43;10;20

Right now. I'm 100%.

00;43;10;20 - 00;43;12;06

It will be by the middle of the year.

00;43;12;06 - 00;43;15;12

Like again, I can just see it because

I understand the nature of accounting.

00;43;15;15 - 00;43;18;12

So I think again, if you can be measured

that way, then definitely

00;43;18;12 - 00;43;22;08

the AI can replace you in a few years

time.

00;43;22;11 - 00;43;25;22

And I think there's been a Harvard

Harvard slash open AI study

00;43;25;22 - 00;43;29;12

where they actually go industry

by industry, job by job and show

00;43;29;15 - 00;43;32;22

danger of automation.

00;43;32;25 - 00;43;33;13

Right.

00;43;33;13 - 00;43;35;14

And and I've

seen I've seen some version of that study.

00;43;35;14 - 00;43;37;02

I think I saw a Microsoft one.

00;43;37;02 - 00;43;39;13

And you know,

unless you're like an embalmer

00;43;39;13 - 00;43;43;17

for a funeral home, like you're

you're in a lot of trouble, basically.

00;43;43;24 - 00;43;44;21

So Yeah.

00;43;44;21 - 00;43;47;11

But even then, we've got robots

in 5 to 10 years.

00;43;47;11 - 00;43;49;09

The robots are getting really good.

00;43;49;09 - 00;43;52;19

The only restriction on

that is just the supply chain.

00;43;52;21 - 00;43;53;20

Honestly.

00;43;53;20 - 00;43;57;00

Like I think robots will be capable of

00;43;57;03 - 00;44;01;15

full 99% of human jobs in 3 to 4 years

00;44;01;18 - 00;44;04;16

and it'll cost about a dollar an hour.

00;44;04;16 - 00;44;08;06

So where does that leave humans in 3

to 4 years?

00;44;08;09 - 00;44;12;22

Again, like this, that technology

has take time to percolate, but

00;44;12;25 - 00;44;17;00

it's like you have a massive job migration

coming in

00;44;17;00 - 00;44;21;22

from this new continent, and the workers

will literally work for electricity.

00;44;21;25 - 00;44;23;05

So we'll see what happens there.

00;44;23;05 - 00;44;24;24

Like the governments will have to respond.

00;44;24;24 - 00;44;26;22

They'll have to be UBI type programs.

00;44;26;22 - 00;44;29;17

We suggest that taxation UBI doesn't work.

00;44;29;17 - 00;44;33;29

You have to create new types of money

coming from people, not from banks.

00;44;34;02 - 00;44;38;01

Like it's going to get complicated

over the next 510 years,

00;44;38;04 - 00;44;41;26

like the capability

will be there within three years.

00;44;41;29 - 00;44;45;13

The dispersion will be 5 to 10 years,

and we just have to navigate through that.

00;44;45;13 - 00;44;46;21

And hopefully on the other side,

00;44;46;21 - 00;44;50;28

we have a world of abundance

where the robots do all the nasty stuff

00;44;51;01 - 00;44;54;18

and boring stuff, and we can go

and explore the stars like Star Trek.

00;44;54;21 - 00;44;58;04

More likely is conflict and,

00;44;58;05 - 00;45;03;08

you know, unpleasantness and hoarding,

etc., where capital no longer needs labor.

00;45;03;08 - 00;45;05;25

So it just compounds.

00;45;05;25 - 00;45;08;01

But we'll see. We'll see.

00;45;08;01 - 00;45;11;16

So let's let's stick on the, the,

the capital notion for a minute.

00;45;11;16 - 00;45;14;11

And you mentioned just in passing

new types of capital,

00;45;14;11 - 00;45;17;00

you know, created by by humans,

not by by banks.

00;45;17;00 - 00;45;19;11

you know,

what could that look like in your mind?

00;45;19;11 - 00;45;21;09

So right now you have inside money.

00;45;21;09 - 00;45;23;03

Outside money. Most money is created.

00;45;23;03 - 00;45;27;19

You go in per deposit, the bank

in the bank issues loans against that.

00;45;27;22 - 00;45;31;17

And so the central bank controls

employment by adjusting interest rates.

00;45;31;20 - 00;45;35;03

And then people will borrow more or less

and then they will hire

00;45;35;03 - 00;45;40;29

more or less on the back

of that shall we say the AI.

00;45;41;02 - 00;45;43;25

Like right now humans will hire

the AI for office.

00;45;43;25 - 00;45;45;11

The humans, you know,

00;45;45;11 - 00;45;48;17

like if you got capital, you just put it

into GPUs rather than factories

00;45;48;17 - 00;45;52;06

or schools and universities because again,

that's for capability development.

00;45;52;09 - 00;45;54;02

So the question is

where does money come from?

00;45;54;02 - 00;45;57;23

So in the book

we propose that you have like a reserve

00;45;57;23 - 00;46;01;22

asset gold type tide

against open intelligence for humanity.

00;46;01;22 - 00;46;06;09

So supercomputers for organizing

the world's cancer knowledge or culture

00;46;06;09 - 00;46;10;14

or running civic AI given free

to the people and then the people

00;46;10;14 - 00;46;14;08

that use the AI, everyone gets an

AI that's aligned with them.

00;46;14;11 - 00;46;17;11

You then get money for being human

00;46;17;13 - 00;46;20;25

and then within that whole construct,

then as the private sector is more

00;46;20;25 - 00;46;22;14

and more taken over by AI,

00;46;22;14 - 00;46;26;10

the eyes will have to buy the human

currency that you get for being human.

00;46;26;10 - 00;46;29;12

So it's a different type of UBI

because we did the math on

00;46;29;12 - 00;46;34;24

UBI via taxation,

it didn't quite work like within the US,

00;46;34;27 - 00;46;37;18

the whole

income tax base is about $5 trillion

00;46;37;18 - 00;46;41;24

plus corporation tax,

which is about trillion dollars of that

00;46;41;27 - 00;46;44;20

minimum poverty level,

00;46;44;20 - 00;46;48;18

$16,000

a year per person will cost $5.1 trillion.

00;46;48;18 - 00;46;51;08

So more than the entire tax base.

00;46;51;08 - 00;46;52;22

So a lot of these proposals don't work.

00;46;52;22 - 00;46;56;19

So you have to think about where does

money come from, where does money go.

00;46;56;22 - 00;47;01;18

And so again we outline the proposal

that you should get money for being human.

00;47;01;21 - 00;47;05;18

And then a way to tie

that to computer intelligence

00;47;05;21 - 00;47;10;12

and make it so that the

AI is all wanted by the money off you.

00;47;10;15 - 00;47;12;23

sort of like a human bitcoin.

00;47;12;23 - 00;47;15;14

If I can oversimplify

that, Pretty much. Yeah.

00;47;15;14 - 00;47;16;25

So yeah.

00;47;16;25 - 00;47;17;14

So it's

00;47;17;14 - 00;47;21;24

this Bitcoin but it's core foundation

coin locked against again supercomputers

00;47;21;24 - 00;47;25;28

for humanity to solve the big problems

and give everyone a free AI as well.

00;47;25;29 - 00;47;28;29

Because you need an AI

that's aligned with you to keep up.

00;47;29;01 - 00;47;32;00

And then that creates a local currency

that then

00;47;32;00 - 00;47;34;13

the AI can buy off the humans.

00;47;34;13 - 00;47;39;19

Like, again, if you think about

a human in five, ten years,

00;47;39;22 - 00;47;42;07

you had

your work and you got your capital.

00;47;42;07 - 00;47;42;21

Another side.

00;47;42;21 - 00;47;44;09

This is from the days of Henry

Ford, right?

00;47;44;09 - 00;47;46;25

He paid them enough

to be able to buy the cars.

00;47;46;25 - 00;47;49;00

What are people going to pay you for?

00;47;49;00 - 00;47;50;05

Not cognitive labor.

00;47;50;05 - 00;47;51;26

Not even creative labor necessarily.

00;47;51;26 - 00;47;56;04

Like probably most people end up employed

by the public sector

00;47;56;07 - 00;47;59;15

as the jobs

programs of the 1930s and others.

00;47;59;18 - 00;48;01;26

But that's not very happy, right?

00;48;01;26 - 00;48;03;23

Which is why, again,

you have to look at your community

00;48;03;23 - 00;48;06;05

and other things

like if you have a strong community,

00;48;06;05 - 00;48;10;00

even if you fall through the gaps,

then you can be supported.

00;48;10;03 - 00;48;14;23

Whereas if you were a truck driver

and then you were replaced by a Tesla

00;48;14;23 - 00;48;20;00

Optimus robot that gets up

and literally sits in your truck,

00;48;20;03 - 00;48;21;29

what are you going to reskill to?

00;48;21;29 - 00;48;25;21

You know, it's very tough to see

how you're going to have millions

00;48;25;21 - 00;48;31;28

of private sector jobs in the future,

because again, what reasonably

00;48;32;01 - 00;48;34;10

on a digital basis

00;48;34;10 - 00;48;38;20

will an AI not be able to do

thousands of times cheaper than a human?

00;48;38;23 - 00;48;42;15

Within five years,

I it's very hard to decide that again.

00;48;42;18 - 00;48;44;22

Physically. Yeah. An embalmer.

00;48;44;22 - 00;48;48;24

But then an AI robot will probably be able

to be an embalmer, right?

00;48;49;01 - 00;48;52;03

But they probably won't bother,

because it's not like you go to

00;48;52;03 - 00;48;58;06

an embalmer saying,

I want cut price embalming like

00;48;58;09 - 00;48;59;25

it takes time.

00;48;59;25 - 00;49;03;25

There's there's sort of an implication

running through here and

00;49;03;26 - 00;49;06;01

and maybe implication

isn't even a strong enough word.

00;49;06;01 - 00;49;09;27

But there's a thread that the government

is going to have to play

00;49;09;27 - 00;49;14;22

a fairly outsized role in managing this,

you know, societal and,

00;49;14;25 - 00;49;19;19

you know, economic disruption,

at least in the sense of human, economy.

00;49;19;22 - 00;49;22;16

What role

do you see the public sector playing?

00;49;22;16 - 00;49;26;00

And, you know, what do you see

as being kind of the big things

00;49;26;03 - 00;49;28;29

that they need to get right?

00;49;28;29 - 00;49;32;24

If we're going to avoid this sort of,

00;49;32;24 - 00;49;38;16

you know, dystopia of, you know, looting

and starvation and, you know,

00;49;38;19 - 00;49;42;01

rebellion

against the robots, if I can call it that.

00;49;42;04 - 00;49;43;09

Yeah. You're going to see a backlash.

00;49;43;09 - 00;49;44;02

Well, I mean, so

00;49;44;02 - 00;49;47;23

we built this thing called sage,

the sovereign AI governance engine,

00;49;47;26 - 00;49;50;17

which will be made for you to governments

next year for the opensource

00;49;50;17 - 00;49;54;09

to help design policy in the time of AI,

can you see such big shifts

00;49;54;09 - 00;49;56;17

that can really impact

the way that everything happens?

00;49;56;17 - 00;49;57;01

Like what happens

00;49;57;01 - 00;50;00;26

when you get quantum supremacy

or robots can finally mimic humans?

00;50;00;26 - 00;50;03;26

We're just a little bit of training data,

or you got an AI

00;50;03;26 - 00;50;06;27

that can replace an accountant

that works on a MacBook, right?

00;50;06;27 - 00;50;09;01

These are big shifts that are cut.

00;50;09;01 - 00;50;12;18

So we think that governments will have to

play an outsize role because ultimately,

00;50;12;19 - 00;50;16;14

you know, they are the social Security net

for people.

00;50;16;17 - 00;50;20;14

And this disruption

coming is bigger than Covid

00;50;20;17 - 00;50;23;25

like it can be both

ways and people will do very, very well.

00;50;23;26 - 00;50;26;25

But again, let's take the example

of truck drivers in America.

00;50;26;27 - 00;50;31;02

There's a million truck drivers,

maybe 3 or 4 million jobs.

00;50;31;05 - 00;50;34;03

The moment Tesla Optimus is good enough,

00;50;34;03 - 00;50;38;11

unless there's government regulations

stopping them, the way that truck driver

00;50;38;11 - 00;50;43;10

will be replaced is an Optimus

comes takes itself out of its box.

00;50;43;10 - 00;50;48;25

It's probably dropped off by a Tesla

robotaxi walks into that truck,

00;50;48;28 - 00;50;51;11

and he don't need that truck

driver anymore.

00;50;51;11 - 00;50;52;19

So what does the truck driver do?

00;50;52;19 - 00;50;55;11

He needs to be supported by the government

one way or another, right?

00;50;55;11 - 00;50;58;25

Otherwise you get massive

social instability because

00;50;58;28 - 00;51;02;24

a million robots

replace a million truck drivers.

00;51;02;27 - 00;51;07;06

And again, there's no like retrofitting

or anything going and building in there.

00;51;07;06 - 00;51;09;29

The physical robots can go in

and replace them.

00;51;09;29 - 00;51;14;15

The online accountants are replaced by the

AI even more seamlessly.

00;51;14;18 - 00;51;18;07

So the government's going to have to play

an outsized role because more

00;51;18;07 - 00;51;22;28

and more of the GDP of the world

will become public sector.

00;51;23;01 - 00;51;26;03

And so and again, the mathematics,

the economics, we kind of show

00;51;26;03 - 00;51;29;02

how mathematically this must be true.

00;51;29;05 - 00;51;31;13

And already 20%

00;51;31;13 - 00;51;36;03

plus education, some healthcare, 20, 40%

of the GDP of the world is public sector.

00;51;36;06 - 00;51;39;00

In Europe it's about 50%.

00;51;39;00 - 00;51;41;07

But it's not just distributed properly.

00;51;41;07 - 00;51;44;13

This is the thing, but the AI should be

able to help with the distribution.

00;51;44;16 - 00;51;47;29

So I think unfortunately, because again,

the government's not that great

00;51;48;02 - 00;51;49;22

and institutions

are going to have to upgrade.

00;51;49;22 - 00;51;52;25

We have to make sure nobody's left behind,

give everyone

00;51;52;25 - 00;51;56;01

an individual AI,

rethink the way that money flows.

00;51;56;01 - 00;51;59;23

And again,

unfortunately, time is running out

00;51;59;26 - 00;52;02;08

because actually this thing

00;52;02;08 - 00;52;05;08

spreads faster than a virus.

00;52;05;09 - 00;52;05;23

Covid.

00;52;05;23 - 00;52;11;15

We saw the spread of Covid, whereas this

it will be every single

00;52;11;15 - 00;52;16;08

accounting in the world suddenly is like

AI cost pennies and it can replace me.

00;52;16;11 - 00;52;18;18

Those that have strong partnerships

and things like that.

00;52;18;18 - 00;52;25;20

Yeah, that might last a while,

but these things happen quickly, quickly.

00;52;25;23 - 00;52;27;00

I want to pick back up on the

00;52;27;00 - 00;52;31;00

notion that you brought up earlier

a mod about,

00;52;31;03 - 00;52;34;16

you know, our search for meaning

and what it means to be human

00;52;34;19 - 00;52;37;19

in a world where suddenly it can't be our,

00;52;37;24 - 00;52;41;23

you know, our employment

or our occupation,

00;52;41;23 - 00;52;46;04

or it can't be necessarily economic in

the same way that it has been in the past,

00;52;46;07 - 00;52;49;13

you mentioned, meaning, you know,

what to you

00;52;49;16 - 00;52;54;05

is going to create meaning in this,

you know, in the next century?

00;52;54;05 - 00;52;58;02

And how do we kind of recalibrate

our expectations about,

00;52;58;05 - 00;53;01;27

you know, life

and certainly living a good life?

00;53;02;00 - 00;53;06;16

I think that it has to be a given

that you don't earn money through Covid

00;53;06;16 - 00;53;10;02

to work again, you don't mind money

ready through physical works.

00;53;10;02 - 00;53;11;00

The machines replaced it.

00;53;11;00 - 00;53;12;14

You had these inversions,

00;53;12;14 - 00;53;15;22

as I wrote in the book, and the final

version is the intelligence and version.

00;53;15;25 - 00;53;19;18

Computation and consciousness

are divorced from each other,

00;53;19;21 - 00;53;23;01

so the role of the human is to guide the

AI to be

00;53;23;01 - 00;53;27;08

the reinforcement learning, as it were,

and we should be rewarded for that.

00;53;27;08 - 00;53;30;00

Which is why I think you should get money

for being human.

00;53;30;00 - 00;53;32;23

But meaning in life,

00;53;32;26 - 00;53;36;04

if you're tying it to material stuff,

material wealth,

00;53;36;04 - 00;53;40;02

your position,

it's really usually never enough.

00;53;40;05 - 00;53;44;22

If you're tying it to your interactions

with others, spending time with your kids,

00;53;44;25 - 00;53;46;21

learning and exploring,

00;53;46;21 - 00;53;50;19

being creative, then that really changes

the calculus of things, right?

00;53;50;22 - 00;53;52;24

But we don't really reward that

within our society.

00;53;52;24 - 00;53;54;05

Like, sure, if you're top level.

00;53;54;05 - 00;53;55;02

Yeah,

00;53;55;02 - 00;53;56;05

but by the time you're 18,

00;53;56;05 - 00;53;59;10

you spend as much time

as you'll ever spend with your parents.

00;53;59;13 - 00;54;02;07

But how many parents

have time for their kids?

00;54;02;07 - 00;54;05;27

Like some parents may say,

let's sit down for dinner together.

00;54;06;00 - 00;54;09;00

How many parents say,

let's make dinner together right now?

00;54;09;00 - 00;54;11;21

The latter will form stronger bonds.

00;54;11;21 - 00;54;14;20

You think about the churches

and the mosques and things like that.

00;54;14;20 - 00;54;16;21

Again, that's the nature of community.

00;54;16;21 - 00;54;18;23

What is the meaning of having

a community event?

00;54;18;23 - 00;54;20;12

You think about your friends.

00;54;20;12 - 00;54;21;09

So I think that's why

00;54;21;09 - 00;54;25;14

we really have to think how do we build

stronger networks, communities?

00;54;25;17 - 00;54;29;16

How do we really bring forward

the stuff that matters in life?

00;54;29;19 - 00;54;32;23

And let's use the

AI to solve the coordination

00;54;32;23 - 00;54;36;09

problems of like, there's

enough food in the world to feed everyone.

00;54;36;12 - 00;54;39;00

Use the robots to make sure it gets to

everyone.

00;54;39;00 - 00;54;40;19

We'll have robots building houses.

00;54;40;19 - 00;54;43;14

There's two acres of land

for every human in the world.

00;54;43;14 - 00;54;45;16

If you do the math, let's do that.

00;54;45;16 - 00;54;46;26

Let's use solar power.

00;54;46;26 - 00;54;47;22

Let's kill cancer.

00;54;47;22 - 00;54;49;16

Let's do all these things.

00;54;49;16 - 00;54;53;07

So I think that again,

this is up to individuals

00;54;53;07 - 00;54;59;04

to be meaning makers and really drive

that and these discussions because

00;54;59;07 - 00;55;01;02

for those whose primary

00;55;01;02 - 00;55;05;01

meaning that primary story is their job,

00;55;05;04 - 00;55;08;10

it's going to be a difficult

five, ten, 20 years.

00;55;08;13 - 00;55;11;03

Right? Like even now,

00;55;11;03 - 00;55;13;24

how many people

00;55;13;24 - 00;55;17;20

listening to this are going through

and doing computer science

00;55;17;20 - 00;55;22;05

degrees or planning to be programmers

and things like that, or know people

00;55;22;08 - 00;55;25;25

programing, this isn't a thing

anymore, you know, like,

00;55;25;25 - 00;55;30;19

why do you need to know C when the AI

can speak C better than you can?

00;55;30;22 - 00;55;31;26

So if your meaning was, I'm

00;55;31;26 - 00;55;35;03

going to be a programmer,

then you got a problem, right?

00;55;35;06 - 00;55;38;11

If it's I like to solve problems,

I like to build things, then yeah,

00;55;38;12 - 00;55;40;16

you'll still have meaning right?

00;55;40;16 - 00;55;43;06

So I just think that needs to be

a discussion that we have.

00;55;43;06 - 00;55;48;04

And again, the biggest meaning

is an interaction with others.

00;55;48;07 - 00;55;51;07

Humans,

like an AI, isn't gonna replace the time

00;55;51;07 - 00;55;54;17

that you spent in

the park with your daughter,

00;55;54;20 - 00;55;56;03

you know, flying a kite.

00;55;56;03 - 00;55;58;26

Come on.

00;55;58;26 - 00;56;01;10

I love that,

I love the answer of community.

00;56;01;10 - 00;56;05;04

And part of the reason why I love it is

I feel like

00;56;05;07 - 00;56;08;21

I don't think it's controversial

to say that in the last 10 or 15 years,

00;56;08;21 - 00;56;12;18

you know, social media and,

you know, a lot of technology in general

00;56;12;21 - 00;56;16;05

in its quest for driving engagement,

00;56;16;08 - 00;56;20;25

has eroded our civic society

and eroded this sense of community.

00;56;20;29 - 00;56;23;08

if you're a scholar of political science,

it goes beyond that.

00;56;23;08 - 00;56;24;09

And this has been a trend.

00;56;24;09 - 00;56;25;00

And, you know,

00;56;25;00 - 00;56;29;14

the Western world for decades now,

and I think very much to our detriment.

00;56;29;14 - 00;56;34;17

And so, I don't know, with your futurist

hat on, if there's a way you sort of see

00;56;34;17 - 00;56;40;05

this playing out societally of how we can,

you know, make sure that we,

00;56;40;08 - 00;56;43;12

you know, flip the script and end up

actually strengthening these again.

00;56;43;16 - 00;56;45;23

Is that going to be bottom

up? Is that going to be top down?

00;56;45;23 - 00;56;48;26

Is it just up to each of us

to make a conscious decision.

00;56;48;26 - 00;56;52;26

How do we how do we make sure

that we get there for ourselves?

00;56;52;29 - 00;56;53;08

Yeah.

00;56;53;08 - 00;56;56;22

So kind of my concept was universal

AI for everyone,

00;56;56;25 - 00;57;00;10

AI to help guide

the governments and organizations,

00;57;00;13 - 00;57;03;21

and then the optimizing function

is human flourishing,

00;57;03;21 - 00;57;06;22

which is this multifaceted multi capital

thing of your intelligence,

00;57;06;22 - 00;57;09;19

your capability, your diversity

kind of other stuff like that

00;57;09;19 - 00;57;10;19

as well as the material thing.

00;57;10;19 - 00;57;14;15

Like again, you need to have food, water,

all that kind of thing.

00;57;14;18 - 00;57;19;05

And the way that we kind of had it

was that you earn money for being human,

00;57;19;05 - 00;57;21;14

but then you earn more money

for doing things that are beneficial

00;57;21;14 - 00;57;25;13

for society, as judged by open source

AI that we all get together and build.

00;57;25;13 - 00;57;30;12

So the concept of subsidiarity,

where the communities themselves

00;57;30;15 - 00;57;32;29

define

what is beneficial for the community,

00;57;32;29 - 00;57;35;25

and this is similar

to the Argentinian program

00;57;35;25 - 00;57;40;28

where they had, minimum viable jobs

for everyone versus handouts.

00;57;41;01 - 00;57;44;03

And that brought a lot of women into the

workforce and other things like that.

00;57;44;03 - 00;57;46;03

Again, it was quite small scale, but

00;57;46;03 - 00;57;49;29

I think it makes sense, like reward people

for what is good.

00;57;50;02 - 00;57;52;15

And that community itself should decide

what is good.

00;57;52;15 - 00;57;54;16

It's not up to the people to decide

what is good.

00;57;54;16 - 00;58;00;14

So push that down and then you have

optimization functions that occur on that.

00;58;00;17 - 00;58;02;28

The other part of this though, is that

there is a community aspect,

00;58;02;28 - 00;58;05;28

and there's also the solitude

aspect, right?

00;58;06;02 - 00;58;09;13

Like most religions and faith traditions

and others

00;58;09;13 - 00;58;14;05

are about enlightenment

and realizing that you're not all that,

00;58;14;08 - 00;58;17;14

but then realizing that your interactions

with other people are all that.

00;58;17;17 - 00;58;23;24

And so that for progress of cleansing,

enlightenment going down,

00;58;23;27 - 00;58;27;06

can you have an

AI is the voice that helps you on that

00;58;27;06 - 00;58;31;21

infinitely patient, infinitely kind,

you know, are we building the AIS for that

00;58;31;24 - 00;58;35;25

right now? I have no morality or ethics.

00;58;35;28 - 00;58;36;19

So can we

00;58;36;19 - 00;58;40;20

build the AI that's closest to us

to have that that reflects our own AI?

00;58;40;20 - 00;58;43;29

Allow us to explore our own self?

00;58;44;02 - 00;58;46;18

Or is it's job to sell us a Budweiser?

00;58;46;18 - 00;58;49;18

This is going to be a very important thing

for the AI that grows with us,

00;58;49;21 - 00;58;53;23

especially

because we think about our children.

00;58;53;26 - 00;58;58;13

Their best friends will be AIS,

their first loves will be AIS.

00;58;58;16 - 00;59;01;00

Who is that AI working for?

00;59;01;00 - 00;59;04;00

This will be a thing

because the AI is a mirror.

00;59;04;01 - 00;59;06;19

And in fact,

if you look at some of the system prompts

00;59;06;19 - 00;59;10;06

of meta AI, for example,

it says mirror the user,

00;59;10;09 - 00;59;13;09

which is a very powerful

psychological technique.

00;59;13;11 - 00;59;14;21

Quite a dangerous one as well.

00;59;14;21 - 00;59;17;03

Yeah. But again,

we see what that's going to be.

00;59;17;03 - 00;59;20;18

So I think that there are ways that we can

encourage to me, to find dissipation,

00;59;20;21 - 00;59;23;16

the ways that we can help guide

people in life.

00;59;23;16 - 00;59;25;29

And again, this is why the AI that's

00;59;25;29 - 00;59;31;08

closest to your personal

AI is going to be the most important.

00;59;31;11 - 00;59;32;20

It feels like there's a friction

00;59;32;20 - 00;59;36;24

there between this notion of community

as the most important.

00;59;36;27 - 00;59;38;23

You know, inter

00;59;38;23 - 00;59;41;03

connection

between humans is the most important.

00;59;41;03 - 00;59;44;20

And then you, you saying

well you know your first love

00;59;44;20 - 00;59;48;12

may be I,

some of your closest friends may I.

00;59;48;15 - 00;59;52;22

Do we have the I guess the cognitive

capabilities as humans to balance

00;59;52;22 - 00;59;55;25

that or how do we want to be building

that future for ourselves?

00;59;55;25 - 00;59;59;16

How should we be sort of

compartmentalizing this or structuring it

00;59;59;16 - 01;00;04;14

in a way where it is

leading to flourishing for us

01;00;04;14 - 01;00;07;16

and not just being,

you know, an engagement tool

01;00;07;16 - 01;00;11;09

for the Meadows

and tech companies in the world.

01;00;11;12 - 01;00;11;21

Yeah.

01;00;11;21 - 01;00;14;29

And this is kind of person style

entertaining yourself to death.

01;00;14;29 - 01;00;18;17

Right.

It's the Wall-E future. You're strapped in

01;00;18;20 - 01;00;21;20

what we see right now with polarization.

01;00;21;20 - 01;00;25;16

With these walls,

it could be accelerated to the nth degree.

01;00;25;19 - 01;00;30;20

But your VR headsets and things like that,

I think it's going to be crazy.

01;00;30;23 - 01;00;33;28

So this is a question of how we build

and what we build.

01;00;33;28 - 01;00;38;04

And AI is incredibly dangerous

cognitively.

01;00;38;07 - 01;00;41;11

Last year there was a study done

where they flooded Reddit with all these

01;00;41;11 - 01;00;45;20

bots like a black anti-black Lives

Matter person and all this kind of stuff,

01;00;45;23 - 01;00;46;19

and they showed that the

01;00;46;19 - 01;00;50;25

AI was 99th percentile persuasiveness

with last year's models,

01;00;50;28 - 01;00;54;14

not these models that come in this year,

like AI, is incredibly persuasive.

01;00;54;17 - 01;00;55;11

And you can tell it like

01;00;55;11 - 01;00;58;26

there are some companies that are matching

now three minutes of your grandma.

01;00;58;26 - 01;01;01;03

They'll make an AI

replica of your grandma.

01;01;01;03 - 01;01;02;07

And they'll sound like your grandma.

01;01;02;07 - 01;01;04;22

And you can call your grandma

and all sorts of things.

01;01;04;22 - 01;01;08;21

Let's just think about what can I can do

psychologically,

01;01;08;24 - 01;01;11;11

you know,

this is Black Mirror on steroids in a way.

01;01;11;11 - 01;01;12;14

Right?

01;01;12;14 - 01;01;13;21

So this is why it matters.

01;01;13;21 - 01;01;16;11

Any kind in

AI to protect against the other eyes.

01;01;16;11 - 01;01;17;14

Freezer.

01;01;17;14 - 01;01;21;27

But then the way we build

it is going to be so important because

01;01;22;00 - 01;01;23;24

our kids can't protect against this.

01;01;23;24 - 01;01;26;10

We can't protect against this.

01;01;26;10 - 01;01;28;23

And again,

we need to have standards in society.

01;01;28;23 - 01;01;33;09

Like,

should I be allowed to be persuasive?

01;01;33;09 - 01;01;37;04

Just like we have ad standards,

just like with robots.

01;01;37;07 - 01;01;41;25

Should I should robots be powerful enough

to punch through humans?

01;01;41;28 - 01;01;43;24

That's a question

that we have to answer right now, right?

01;01;43;24 - 01;01;46;24

Like again,

what are we allowing into our society,

01;01;46;24 - 01;01;49;08

into the public sphere,

and into our private sphere as well?

01;01;49;08 - 01;01;53;03

We need to be cognizant of this

because it's way,

01;01;53;03 - 01;01;57;14

way more than TikTok

01;01;57;17 - 01;02;00;02

in terms of what might happen here.

01;02;00;02 - 01;02;03;16

And again, we are very squishy, mushy,

01;02;03;20 - 01;02;07;24

squishy things in terms of

I'll call them to capabilities.

01;02;07;27 - 01;02;10;19

And so much,

so much of what you're describing

01;02;10;19 - 01;02;14;15

we need defending against

is almost the status quo today.

01;02;14;15 - 01;02;17;07

Right. Like it's the technology

that's already here.

01;02;17;07 - 01;02;21;11

It's the companies

trying to sell us more Budweiser or trying

01;02;21;11 - 01;02;24;15

to, you know, rage beat us

01;02;24;17 - 01;02;27;18

with,

you know, whatever political engagement

01;02;27;24 - 01;02;31;12

do you have any predictions of,

you know, who if anybody is going to start

01;02;31;12 - 01;02;35;22

to build these more sort

of righteous, personalized,

01;02;35;25 - 01;02;37;27

you know,

I companions that can actually do that?

01;02;37;27 - 01;02;39;19

Because that's that's my nightmare.

01;02;39;19 - 01;02;43;00

My nightmare

is that the AI companions end up

01;02;43;00 - 01;02;47;18

being a lot more like the tools

we have today.

01;02;47;25 - 01;02;52;05

And their motivation is actually,

you know, split people up from

01;02;52;05 - 01;02;56;07

other people so that they only trust us

and not other humans.

01;02;56;07 - 01;02;57;24

And it just drives deeper engagement.

01;02;57;24 - 01;03;02;11

And yeah, that that is like the

ultimate Black mirror to me.

01;03;02;14 - 01;03;04;12

Yeah, I mean, that's profitable, right?

01;03;04;12 - 01;03;07;12

Like just look at how cults operate.

01;03;07;16 - 01;03;08;26

Literally. It's the same thing.

01;03;08;26 - 01;03;12;06

So you can mechanize cult behavior.

01;03;12;09 - 01;03;14;22

Gosh.

So I mean, this is what we're doing.

01;03;14;22 - 01;03;15;19

Intelligent internet.

01;03;15;19 - 01;03;19;26

My new company, again, I live stability

I leaders in media.

01;03;19;26 - 01;03;20;29

I to do that.

01;03;20;29 - 01;03;24;13

So a building open source universal

AI for everyone.

01;03;24;16 - 01;03;28;14

And we're also developing

a concept called local champions.

01;03;28;14 - 01;03;33;28

So for every state and nation

wholly locally owned by citizen entities,

01;03;33;28 - 01;03;37;25

they act as utilities, give universal

AI to the people that represents them.

01;03;37;28 - 01;03;40;09

You can either use that universal AI,

01;03;40;09 - 01;03;42;25

or you can roll your own

using the OpenStack.

01;03;42;25 - 01;03;47;19

So optimizing it to work on meshes

locally on your computer and others,

01;03;47;19 - 01;03;51;03

because we think that AI that's closest

to you is the most important,

01;03;51;06 - 01;03;51;22

it can still

01;03;51;22 - 01;03;55;12

use ChatGPT and Claude and other things,

01;03;55;15 - 01;03;58;29

but it's not ChatGPT talking

directly to your kids.

01;03;59;02 - 01;04;02;04

It's an AI that sits in the middle

that's mediating all that.

01;04;02;04 - 01;04;05;04

And we think that's the important thing

on a control plane basis.

01;04;05;11 - 01;04;09;05

And again, by making it open source,

just like the book and the economic theory

01;04;09;05 - 01;04;13;02

about to release, we're hoping that people

will just take it and build themselves.

01;04;13;05 - 01;04;15;12

And the wonderful thing is

you can build it

01;04;15;12 - 01;04;18;23

yourself if you've got the right tools

and the right ingredients.

01;04;18;23 - 01;04;22;00

So we're hoping to make that as easy

as possible for people,

01;04;22;03 - 01;04;25;15

because you can't rely on one

centralized protection against this.

01;04;25;15 - 01;04;28;22

Again, you need decentralized swarm

protection against this.

01;04;28;22 - 01;04;31;15

And you need to make

it so it can come together

01;04;31;18 - 01;04;33;21

to build what represents each community.

01;04;33;21 - 01;04;37;01

Because my community's needs are different

from your community's

01;04;37;01 - 01;04;40;10

needs, is different from Vietnam's,

is different from others.

01;04;40;13 - 01;04;43;12

And so you want to have

permissionless innovation

01;04;43;18 - 01;04;45;24

on the top of that,

which is make the building

01;04;45;24 - 01;04;47;29

blocks available,

make the infrastructure available.

01;04;47;29 - 01;04;49;15

ET cetera.

01;04;49;15 - 01;04;51;27

We have the money aspect,

tying it all together

01;04;51;27 - 01;04;54;22

so everyone can benefit

being part of the network.

01;04;54;22 - 01;04;57;10

But again, if you don't want to,

the key thing here is

01;04;57;10 - 01;04;59;29

can you opt out of the AI that's coming?

01;04;59;29 - 01;05;04;09

You know, like

01;05;04;12 - 01;05;07;18

probably not for most of it,

but you should at least have the ability

01;05;07;18 - 01;05;09;11

to try.

01;05;09;11 - 01;05;12;04

It's a really

it's a really noble mission.

01;05;12;04 - 01;05;14;13

And I'm really excited

to see how it goes.

01;05;14;13 - 01;05;17;28

And you know, if we can start yeah.

01;05;18;01 - 01;05;20;05

Mediate and have that sort of defense

force for us.

01;05;20;05 - 01;05;21;09

I'm curious I know it's,

01;05;21;09 - 01;05;23;27

I know it's early days,

but have you started to see,

01;05;23;27 - 01;05;27;04

you know, any specific,

you know, promising use cases of people

01;05;27;06 - 01;05;31;02

either using this technology or,

you know, building something with it

01;05;31;02 - 01;05;35;03

that you didn't

expect, like, are there any sort of

01;05;35;06 - 01;05;35;23

initial

01;05;35;23 - 01;05;39;13

signs

that that have been promising to you?

01;05;39;16 - 01;05;40;29

And say,

you know, we built a state of the art

01;05;40;29 - 01;05;41;29

agent and people are using it

01;05;41;29 - 01;05;44;29

for normal stuff now,

but they are starting to think more like

01;05;44;29 - 01;05;47;29

as it gets better and better,

like it's a rap that on steroids.

01;05;48;03 - 01;05;49;07

How can I build stuff, right?

01;05;49;07 - 01;05;49;27

Like a community.

01;05;49;27 - 01;05;52;09

And we're building

interconnected stuff and app stores

01;05;52;09 - 01;05;55;22

so that they can proliferate

best practices very quickly.

01;05;55;25 - 01;05;58;20

At the high level,

we see dozens of countries

01;05;58;20 - 01;06;02;15

and leading corporations interested

in the Sage project, which is this

01;06;02;15 - 01;06;06;05

AI brain for exponential technology

that helps build policy.

01;06;06;08 - 01;06;09;07

And we're working with some groups

now to enable citizens

01;06;09;07 - 01;06;11;12

to actually build their own policy

bottom up.

01;06;11;12 - 01;06;16;06

That feeds into the top down policy, which

we think will be super duper interesting

01;06;16;09 - 01;06;20;06

because democracy has always been

representative, but not really, you know,

01;06;20;09 - 01;06;23;21

like, what does it mean

if you can get together with people

01;06;23;21 - 01;06;27;14

and build a policy that then interacts

with the government policy being built

01;06;27;14 - 01;06;31;01

top down, I think that will be super

interesting to me in the middle,

01;06;31;01 - 01;06;34;11

because we never had this

independent open source AI for that.

01;06;34;14 - 01;06;38;11

So we're hoping that, you know,

we will release all the blocks and tools

01;06;38;11 - 01;06;39;24

and people will build on top of it,

01;06;39;24 - 01;06;43;08

much as people build on top of blockchain

and Bitcoin and others.

01;06;43;11 - 01;06;47;20

And again, fast forwarding to ten,

20 years from now, we think this is what

01;06;47;23 - 01;06;51;16

the infrastructure of civic

I should look like.

01;06;51;19 - 01;06;55;08

Like again, private sector and others

we think are going to get very weird.

01;06;55;08 - 01;06;58;16

But the AI that's closest to your kid

is the one that matters.

01;06;58;16 - 01;07;02;11

The AI, the gods of government is the one

that matters, manages your health, etc.

01;07;02;11 - 01;07;04;21

and that should be collectively owned.

01;07;04;21 - 01;07;09;02

But self-sovereign I think is the key.

01;07;09;05 - 01;07;10;27

There's an awful lot of

01;07;10;27 - 01;07;14;25

there's, you know, in your view, a lot

an awful lot of disruption on the horizon,

01;07;14;25 - 01;07;20;04

like it's going to be a very, very rocky

handful of years for us as a species.

01;07;20;04 - 01;07;23;02

And I'm curious, just sort of processing

what you said there.

01;07;23;02 - 01;07;26;24

It feels like there's, you know, an

undercurrent of optimism that you've got.

01;07;26;24 - 01;07;31;06

And I'm on balance, I'm curious

if you feel sort of optimistic

01;07;31;06 - 01;07;34;21

or pessimistic or what

your outlook is in terms of,

01;07;34;27 - 01;07;40;18

you know, the impact and on us

as humans in our in our next chapter.

01;07;40;21 - 01;07;43;00

So this is concept called PDM,

01;07;43;00 - 01;07;46;00

which is the probability

that we're all going to get wiped out by.

01;07;46;05 - 01;07;48;26

And so if you look at Wikipedia

you'll see that

01;07;48;26 - 01;07;53;26

a lot of people like Elon and others

1,520%, which is still like

01;07;53;26 - 01;07;57;18

Russian roulette, odds of us

being wiped out is not encouraging.

01;07;57;21 - 01;07;59;08

I'm at 50%.

01;07;59;08 - 01;08;01;09

So I think there's two ways.

01;08;01;09 - 01;08;04;04

Either

this technology drives us to destruction,

01;08;04;04 - 01;08;07;04

or we have the abundant Star Trek future.

01;08;07;05 - 01;08;11;18

And I think the key thing is

if we give everyone their own AI,

01;08;11;21 - 01;08;15;03

if we have these local champions

as the leading organizations

01;08;15;03 - 01;08;19;01

and by the people, for the people building

and deploying it

01;08;19;04 - 01;08;21;25

coordinated by this currency,

01;08;21;28 - 01;08;23;05

then I think we can get to an

01;08;23;05 - 01;08;26;16

abundant future, because this is really

a coordination question

01;08;26;19 - 01;08;29;00

like,

you know, if you give everyone in Canada

01;08;29;00 - 01;08;33;08

an AI that's aligned with them,

they can trust as that thing

01;08;33;11 - 01;08;36;14

and the Canadian government

has a leading institution

01;08;36;14 - 01;08;40;13

that's providing the AI to help

guide them and building policy publicly.

01;08;40;16 - 01;08;43;00

Canada's

a lot more likely to succeed, right?

01;08;43;00 - 01;08;45;04

Multiplied by every country.

01;08;45;04 - 01;08;47;08

And this is

something we've never seen before.

01;08;47;08 - 01;08;50;27

If we don't coordinate

and everyone tries to do their own thing,

01;08;50;27 - 01;08;54;17

then private companies

or balkanization occurs,

01;08;54;20 - 01;08;59;07

optimization occurs

and it's a very ugly future.

01;08;59;07 - 01;09;04;03

And again, the one that I see spiraling,

very aggressively.

01;09;04;03 - 01;09;07;27

So I think this is the time where, again,

we all got to come together.

01;09;07;28 - 01;09;11;05

We all gotta have the best ideas,

01;09;11;08 - 01;09;13;12

because

01;09;13;12 - 01;09;17;27

again, if you use the technology stops

the day, this disruption is inevitable.

01;09;18;00 - 01;09;23;02

The cost of a single piece of cognition

has already collapsed.

01;09;23;05 - 01;09;26;04

It's just

it will take a year or two to show up.

01;09;26;04 - 01;09;29;01

And so we got to work together

to make sure it's the positive future

01;09;29;01 - 01;09;32;21

versus the negative future for our kids,

for ourselves,

01;09;32;21 - 01;09;35;21

for everyone.

01;09;35;25 - 01;09;39;20

I love that, it encapsulated

so much of what we've talked about today

01;09;39;21 - 01;09;41;23

that, you know, brought home

some of the initial points

01;09;41;23 - 01;09;46;15

you were making, and, you know, it laid

pretty bare the the stakes here

01;09;46;15 - 01;09;47;18

and what we need to get right

01;09;47;18 - 01;09;51;15

if we're going to end up in a world

of abundance versus, extinction.

01;09;51;22 - 01;09;53;26

a mat. I wanted to say thank you so much

for joining today.

01;09;53;26 - 01;09;57;19

It's been an absolute pleasure

and I really appreciate your insights.

01;09;57;22 - 01;10;00;25

Thank you very much for.

01;10;00;28 - 01;10;01;20

If you work in

01;10;01;20 - 01;10;05;10

IT, Infotech research Group is a name

you need to know.

01;10;05;13 - 01;10;08;13

No matter what your needs are, Infotech

has you covered.

01;10;08;18 - 01;10;09;25

AI strategy?

01;10;09;25 - 01;10;12;07

Covered. Disaster recovery?

01;10;12;07 - 01;10;13;07

Covered.

01;10;13;07 - 01;10;15;22

Vendor negotiation? Covered.

01;10;15;22 - 01;10;19;15

Infotech supports you with the best

practice research and a team of analysts

01;10;19;15 - 01;10;23;08

standing by ready to help you

tackle your toughest challenges.

01;10;23;11 - 01;10;26;11

Check it out at the link below

and don't forget to like and subscribe!