If we could go back, they always say,

you know, if you could go back in time,

Warren Buffett would joke, we'd kill the

Wright brothers because airlines have

been such a terrible investment. I think

our biggest regret about this era of

tech is going to be what we let happen

to our children here.

Welcome to PropG on AI. This is a

special series where we are joined by

Greg Scho, the CEO of Section to tackle

your questions on what AI means for

work, business, and the future. So

disclosure, I'm the founder of Section.

Section is a firm that helps uh deploy

AI to the workforce of companies. All

right, Greg, let's get into it. Our

first question comes from user Brosa

8020 on Threads. They say, "It seems AI

is the only thing giving the S&P 500 its

gains over the past year or two. How

much longer can this last and when will

the vast majority of the remaining

companies start to show signs of life?"

Greg, what are your thoughts about the

concentration

uh of a small number of stocks

specifically? It feels like AI is

driving the S&P and to that extent kind

of valuations of the global markets.

What what are your thoughts?

Yeah, I think we've got a couple more

years to find out how this is going to

end. Meaning they are uh these

valuations have been driven by an

assumption that the independent AI

companies and I would specifically focus

on open AI and anthropic can actually

generate at least a hope of profitable

revenue. They've got revenue now. Can

they show even a hope of profitable

revenue in a couple three years? I kind

of feel these companies got those two in

particular right the whole we're we're

heading to AGI give us you know

bucketloads of cash pitch has worked.

could just both those companies just

closed new rounds of capital and I feel

like they can do it one or two more

times which feels which seems like two

to three more years of runway at which

point the whole thing starts to collapse

under its own weight and only the guys

with the picks and shovels actually made

money so the data centers the chips the

energy uh but we've got to get to a

place where you know we can see some

revenue that that might have some margin

I agree it's pretty striking you have

there's just four big tech stocks Nvidia

Microsoft Meta and Broadcom um that

account for 60% of the S&P's index total

return so far this year. So the majority

of the index's returns are driven by

just four companies. And the S&P is up

2.6% this year just based on just based

on Nvidia.

And in addition, it's it's not just the

S&P. The S&P is becoming the global

market. The Magnificent 7 make up over

20% of the MSCI all country index. So

they're not only driving and dictating

the returns in the S&P, they're driving

and dictating the returns of just the

global market. Nvidia, Microsoft, and

Apple each carry a heavier weight than

all of China's stock market combined.

Think about that. So those three

companies, their movements are more

influential and important than the

entire Chinese stock market. And the

premium here is pretty substantial. The

S&P 500, I think, trades at about 23 and

a half forward earnings. And if you take

away those four megga cap stocks, the

multiple falls to 19

uh.4.

And if someone were to tell us uh and

I'm curious if what you think of this,

Greg, that in 6 months we know that

there'd been a crash in the markets that

there had been a global sell-off

and we were asked to guess what it would

be. It would be fun to catastrophize a

civil war or an outbreak of war. Article

five triggered because of attack drones

coming from Russia in in Romania and

Poland. I think the most obvious

explanation would be to be one to be

kind of more boring and that is a big or

a series of big corporations,

traditional corporations announced in

their earnings calls that they were

dramatically scaling back their

investments in AI. That their

investments had not shown the ROI they'd

hoped for. you'd see Nvidia crash and

then you just see this ripple effect all

the way through the uh S&P uh AI supply

chain and people just don't realize how

dramatic it is. Most recently, you had a

38% one-day gain in Oracle which made

Larry Ellison the richest person in the

world. He increased his wealth by about

$120 billion

in one day. And uh this was driven off

of an announcement that OpenAI was

committing to spend, get this, $300

billion

to rent compute from Oracle over 5

years. So when OpenAI looks at its

consumer behavior in its business, it is

so confident in its business that it is

committed to a contract that will

obligate it to pay $60 billion a year

for the next 5 years off a company

that's currently doing 10 billion a

year. So the expectation of this market

is just so incredible that any any kind

of flaw I mean this isn't priced to

perfection. It's priced to insanely

perfect. Yeah, I agree Scott. I don't

think investors realize how fragile

these businesses are. I mean just look

at the last couple weeks, right?

Anthropic settled one class action

publisher lawsuit and had to use 10% of

the capital they just raised. 1.5

billion out of a$1 13 billion round they

had to pay out just to settle one

lawsuit. Right. XAI just laid off 500

people uh this past week uh because they

realized the way they were training

their AI models didn't work as they had

planned. U you know Zox pays 14.5

billion to get one CEO Alexander Wang

from scale to help him recruit other AI

researchers. You know the these

companies

have faced unprecedented challenges.

They have no idea what they're doing.

And I think you're right. What we're

seeing is super companies are being

built potentially if OpenAI, Anthropic,

and others can actually get to $20

billion in revenue with a fraction of

the headcount that Google can with

revenue per employee metrics that are

more like 3 to 5 million per employee.

These are literally super companies in

today's context and traditional metrics.

And I think what the market's saying is

we want to back super companies only.

And so I think that you know if this

thing all works out I think we'll see

this concentration potentially getting

even narrower even you know even even

tighter meaning the rest of corporate

America has to figure out how to become

a super company in some way using AI or

they won't be able to attract any

capital or any investor interest.

Support for the show comes from

Neutrifall. If you're dealing with hair

issues like thinning or shedding, it's

easy to feel stuck on what to try next.

There are so many products out there and

it's hard to know which ones are

actually going to do anything. Neutrifll

is a number one dermatologist

recommended hair growth supplement brand

trusted by over 1 and a2 million people.

You can feel great about what you're

putting into your body since Neutrifll

hair growth supplements are backed by

peer-reviewed studies and NSF content

certified, the gold standard in

thirdparty certification for

supplements. You can purchase online and

there's no prescription required.

Automated deliveries and free shipping

keep you on track. Plus, with a

Neutrifll subscription, you can save up

to 20% and a Headspace meditation

membership is included. See thicker,

stronger, faster growing hair with less

shedding in just three to six months

with Neutrifll for a limited time.

Neutrfil is offering our listeners $10

off your first month subscription and

free shipping when you go to

neutrfil.com and enter the promo code

prop. Find out why Neutrifol is the

bestselling hair growth supplement brand

at neutrfil.com spelled nutrfol.com.

Promo code propg. That's neutrfull.com

promo code propg.

[Music]

All right, question number two comes

from Reddit. Hawaiian bri. I think I can

give it to Reddit. The people are come

up with very interesting titles.

Anyways, so Hawaiian Bry asks, "Scott,

you said AI won't take your job. Someone

who knows how to you use AI will take

your job. What are the key areas in

which a competitive employer candidate

really needs to demonstrate their AI

competency today? Personally, I'm in the

legal space and use AI for research,

summarizing law regulations, and

prioritizing my inbox. But a lot of my

work is still very manual. Plus, I don't

want to be one of those people who cite

AI hallucinations in my work. So, I

spent a lot of time vetting and revising

AI output, which often feels like I'm

doing all the work myself. Anyways,

Greg, thoughts on AI and more

specifically AI in the legal profession.

First of all, I think we all have to be

you've got to show that you're really

good with AI and you've got to be a

driver of your AIS and and not what I

call a passenger. So, if you're using

AI, you've got to really be in the 1% of

people that know how to use it. And what

that means is you're able to steer your

favorite AI model to get the kind of

outputs you like. And that includes

managing for hallucinations. And we can

all complain about hallucinations, but

the reality is a lot of them are

manageable and and can be eliminated or

at least reduced to make yourself more

productive. So what I mean by steering

AI, I mean that you need to be able to

provide context with your AI. You need

to be able to know how to upload your

own documents of data. You need to be

able to ask AI to adopt a persona when

it's working with you. You need to be

able to uh work with reasoning models

and multi-step

uh AI tasks with your AI. again to get a

successful outcome. So when we interview

people, we're really looking right away

for their ability to steer AI. The

second thing we're looking for is do

they have their favorite use cases? And

it sounds like this listener hasn't yet

found the use case that generates a

decent return in terms of their efforts.

But you got to be able to to talk about

three to five use cases, whether it's

personally at home or but certainly at

work where you've been able to get AI,

you've been able to coax and steer AI to

get the outcome you want to make

yourself more productive and a little

bit smarter. And we all have those. When

I talk to candidates that that really

have very superficial use cases, that

tells me they're not spending enough

time with AI and they're not really

figuring this out. And the last thing we

look for is hacks, Scott. Everybody

who's a really active AI user has a

couple three personal hacks, things they

figured out. Might be talking to AI,

using advanced voice mode, or, you know,

whatever it is. We're just they've been

able to sort of find a workaround or a

way to sort of uh, you know, get get

more from AI. So we've just got to be in

that 1% of people that are using AI kind

of like power users and that that's the

for us that's the base case for a

candidate.

So I learned something from you um and

that is and I'm going to hold this up

what you called uh you suggested that

everyone have a second screen. I don't

know if you remember saying this, but I

took it to heart. And now, wherever I

am, I have my screen where I do my

regular work. And then I bring with me a

second screen. And up on the second

screen, if anyone could see it, is is um

is Claude. And I have Claude and Chat

GPT. Those are the two I use. I need to

get better at using the rest. I keep

seeing those Tik Toks saying, "If you're

only using Chat GPT, you're

No, you don't. You don't, Scott. You You

I think this is one of the mistakes a

lot of people are making. They're sort

of getting overwhelmed and with too much

AI and and trying too many tools. Pick

one or two at most, but really one GPT

or Claude or maybe Gemini if you love

Google and just get really great at one

AI.

That's interesting. I use them for

different things, but my point is I've

taken your advice and what I do is I use

it for everything now or I test it and I

use it as a thought partner is the way I

would describe it. When I first started

using it, I thought, "Oh, this is going

to be great. I'll just have I'll just

give it a very long prompt and have it

write my next book and boom, I'll

collect my advance. Doesn't work that

way. It's very anodine. But what it is

really good for is when I write my

newsletter on Friday's No Mercy and

Malice. If I need to trim it by 200

words, I say, "Trim this post by 200

words," and then highlight which words

you trimmed. Or I'll say, "I don't like

my ending. It's too emotional and angry.

How do I tone it down? Which words?" And

they'll come back with a couple ideas.

And I I do compare uh Claude against

Chad GBT to see what they're coming at.

What I've noticed recently is they're

they're becoming the same AI that

they're reverse engineering each other

and coming back with the exact same

answers. I I'm literally using it for

almost everything. We're in a

negotiation right now about taking PropG

Markets,

editing it. We do two and a half hours a

week of I don't even call it podcast of

our content. editing it down to 40 uh 8

minutes and then running it on Saturday

or Sunday morning on this network and

doing a rev share because they need, you

know, the problem with cable news is

that it's declining so fast that they

need to dramatically decrease the cost

of the means of production. So I went to

someone who runs one of the bigger

networks and I said, "Give us an hour

and it's zero incremental cost to you.

We have all of it already done and we'll

do a revshare." Okay, the rev share. So,

we're talking to them, right? I go on to

chat GPT and Claude and say, "This is

the deal. This is the show. This is the

network based on viewing patterns, ad

rates, ad revenues, and previous deals

like this. What should the rev share

split be?" And they both came back with

the exact same answer and a bunch of

rational reason in a range. And it was

just so helpful. And in this instance,

you know, human instinct kicks in. I

don't mind going at the low end of the

range because I want my partners to make

money and I just, you know, quite

frankly, I'm excited about the

opportunity to have a new channel. But

before I just would have had absolutely

no idea where to start. This gentleman

in the legal field, I think law is just

going to get gutted and it's going to be

a metaphor for America. What do I mean

by that? Everything's being optimized

for the top 1% or the top 10%.

And I am now using uh Chat GPT to drop

legal contracts. I'm just going to spend

less money. And I keep the one place I

keep hearing corporations saying the

lowhanging fruit. I talked to one of the

largest private equity firms in the

world and the guy who runs their one of

their their biggest divisions and he

said, "We've been tasked with cutting

our legal expenses by 30 to 50 million

this year." And he said, "We're going to

exceed it using AI." Any closing

thoughts here? Yeah, Scott, I want to go

back to what you said about that the AI

monitor or screen. I think that all of

us need a reminder to use AI. We're not

AI native. We're not in we're not in

school now. The kids in school now are

AI AI native and when they come out and

go to work, they won't even think about

should I use AI for any task. They'll

use AI for every task. I think for the

rest of us, you know, the monitor was

the the sort of hack I I started in

January this year. Before that, I had a

post-it note. You know, talk about loi,

right? I just had a yellow post-it note

on my monitor saying ask AI just to

remind myself to, you know, kind of

consult AI. I also now use that. I use

the GPT5 widget on my phone. So, I'd

recommend this for others as well. Put

either Claude or GPT5, whatever your

favorite AI is. I use GPT5 widget right

on my on my home screen. So, it's always

there. And I basically have turned off

Google as much as I can in my life. It's

really important that you we all get to

something like a hundred conversations a

day with AI. That's what power knowledge

workers are doing today and in some

cases hundreds. And I'm not talking

about talking to your AI companion. I'm

talking about, you know, when you're at

work, lots of conversations for all

kinds of questions with AI. You really

got to you got to kind of jump over from

Google into this world as fast as you

can.

Great. Thanks, Greg. We'll be right back

after a quick break.

Fox Creative.

This is advertiser content from

Betterment.

My name is Daniel Egan and I am the VP

of behavioral finance and investing at

Betterment.

My job at Betterment is to think about

how we design financial advice and

investing services to help people be

successful at managing their money.

There's a lot of research that's been

done on how investors make financial

decisions. So, it's my job to take those

learnings and help us help our

customers. So, how do we actually

implement behavioral insights into our

products? The first idea was making the

core function of it focusing on trying

to achieve goals rather than the dollars

or the account type. It's what we call

goals-based investing. So, Betterment

allows you to invest for purposes like

house down payment or kids college

education and you can put multiple

different kinds of accounts inside of

that. But the key thing there is we can

let you name the purpose and critically

we actually let you upload a picture and

I myself have done this my daughter's

education fund. I have a picture of my

daughter there such that when I'm in

there and I'm thinking about what are my

financial priorities I have to look at a

picture of my daughter and there's been

research done showing yes people who

have personalized goals they name them

something very specific and they upload

a custom image of some kind those people

are more likely to have consistent

saving patterns and they're far less

likely to sell out if the market is

getting rough.

Go to betterment.com to learn more.

Investing involves risk performance not

guaranteed. Support

for the show comes from Gruns. They used

to say that an apple a day keeps the

doctor away. Well, that's a nice

thought, but even so, you still won't

get all the nutrients you need that way.

Here's a tip. Add Gruns to the mix.

Gruns isn't a multivitamin, a green

gummy, or a prebiotic. It's all of those

things and then some at a fraction of

the price. And bonus, it tastes great.

All Grun's daily gummy snack packs are

packed with more than 20 vitamins and

minerals made with more than 60

nutrient-dense ingredients and whole

foods. And for a limited time, you can

try their Gruni Smith apple flavor just

in time for fall. It's got the same

snackable, packable, full body benefits

you come to expect, but this time, these

taste like you're walking through an

apple orchard in a cable knit sweater,

warm apple cider in hands. I've tried

Gruns. I find it very convenient and in

general, just super easy to get kind of

that health boost, if you will. Grab

your limited edition Grunie Smith apple

Gruns available only through October.

Stock up because they will sell out. Get

up to 52% off when you go to grun ns.co

and use the code propg.

On to our final question.

Hey Scott, Tim from Long Island, New

York. As a father of four and 5-year-old

girls, I really appreciate the emphasis

you place on being a father. And while

I'm not raising young men myself, I do

realize my duty in providing a positive

example for my daughter so that they may

one day partner with someone and

recognize qualities to look for in a

mate. That being said, my question is

related to what I think is very much a

common theme in your assessment on a

variety of topics, and that is the

generational loss of reasoning and

critical thinking skills. In this coming

age of AI, how can we continue to raise

children that can think critically and

reason on their own?

I'll repeat the last part of the

question. In this coming age of AI, how

can we continue to raise children that

can think critically and reason on their

own? Greg, what are your thoughts?

Well, this is coming from my advice is

coming from someone who who was really

ineffective in terms of intervening with

his kids. So, what I would say is we

cannot outsource this question to our

schools. really that's that's the advice

here which is we do as challenging as it

can be have to intervene I think in this

moment with our kids in that I think

most schools from middle schools up to

colleges are really struggling right now

in terms of how they should talk about

AI how they should use it in the

classroom and so what the road ahead

looks like they'll eventually sort it

out presumably but until that happens I

think we as parents are going to have to

get pretty hands-on with our kids in

terms of how they how they should be

using AI and we're going to have to be

quite explicit around using AI as a

thinking partner, not an outcome

partner. And and kids are going to just

are going to want to use it as an

outcome partner. They're just going to

want to use it to get the damn paper or

assignment done and submitted and get a,

you know, get a good enough grade.

That's the reality. Kids have always

thought that way. So, we're we're just

going to have to talk to them about

things like, "Show me the prompts. Did

you fact check AI? Show me how you

worked with AI to get to get this kind

of outcome that that you're going to

submit." We're going to have to get

hands on here for the next few years.

Anyone who thinks that families have a

good parents. Yeah, I know you and

Cindy. I know you're great parents. I'd

like to think I'm a good parent. Aspire

to be a great parent, but I think

probably I'm just good. And there's just

no getting around it. First off, 90%

90%. Yeah, about 90% of the anxiety,

stress, agit fights in our house have to

do with the phone. I mean this

has if we could go back they always say

you know if you could go back in time

Warren Buffett would joke we'd kill the

Wright brothers because airlines have

been such a terrible investment. I think

our biggest regret about this era of

tech is going to be what we let happen

to our children here. It is just whether

it's addiction to Tik Tok, getting

misinformation and believing is

true they see online because it's a Tik

Tocker they like or their their brains

being rewired to the point where I'm not

even sure my kids are capable of

watching a movie. And that is if I put

them in a seat in a movie theater where

it's like high impact crazy sound and we

get a ton of shitty food they can

consume to keep them all hopped up. they

can sort of do it, but trying to sit

them down for movie night, it's near

impossible because they're so used to

getting dopa hits in 90 secondond cycles

or 20 second cycles. So, I'm sort of now

of the mind that

we need to probably ban phones,

smartphones for kids under the age of

16, and I'm now thinking under the age

of 18. Uh, I just think the downsides

far outweigh the upsides. And there's

two different I mean, some of it's good,

some of it's bad. My youngest is

essentially,

you know, trying to start a bunch of

businesses using AI. He has AI build a

website and he his latest one was he was

going to sell home goods and he came up

with this thing called root home and it

was this bad website and then he asked

AI how do I drive traffic? How do I make

money? And I can look at it and go,

okay, this isn't going to work. But

that sounds like some of your early

e-commerce ventures, Scott.

Yeah, that's exactly right. But I

couldn't be more proud that my

14-year-old, now 15, is using AI to try

and figure out how to build websites and

make money. And I think I want him to be

an AI enabled warrior around this stuff.

And it was like, it's just it's just

fantastic. At the same time, AI is being

weaponized by these social media

platforms to keep him occasionally on

his side like he's in the midst of a

gigantic heroin trip staring at his

phone. Just this bed rot. So I'm of two

minds about it. And that is I think kids

should pretty early learn how to use AI

such that the same way we you know we

learned how to use calculators. At first

they didn't want remember that they

didn't want calculators in school. I

remember in business school they didn't

want us the fir my first year in

business school they didn't want us

using spellch check because they thought

it was unfair and then they then they

got their heads out of their asses and

said of course you should use technology

to have better inputs and better

deliverables. The admissions departments

at universities are struggling with

whether they should let kids use AI or

not. And of course they will. I don't

think there's going to be any way to not

screen it out. Anyway, I'm of two minds

of this. And that is you got to have

competent kids. But if AI is being used

to quite frankly radicalize them, get

them addicted to certain technologies,

we've got to be more mindful there. My

advice to any parent, and I did not do

this, absolutely prohibit social media

and even a smartphone

until probably age 16. I'm with Jonathan

Hyde on this. There's certain upsides,

downsides to that. You won't be able to

communicate as easily with them. Your

thoughts, Greg?

Yeah, we're just back to the question.

Just one last thought on this. Listen, I

used to say to everybody, and I think

this was sort of a mean

Are you mocking me? Are you saying I

wasn't answering the question?

No. No, I was just

Is this part of this? Is this what

happens when you bring your friends on?

I don't know. I was just trying to help

this list around, dude. Uh, listen, I

think that um, we used to say use AI for

your first draft. I used to say this all

the time to my team and and said

everyone I spoke to. And I think

actually I've changed this is a mistake.

I've changed my mind for drafting.

Sorry.

Yeah. Ex. No, I agree. Like I think you

want starting with kids, we all need to

struggle with that first draft. Even

just an outline, just sit with it and

try to think it through yourself and

then frankly you're going to have a much

better conversation with AI anyway. So I

think just that little bit of struggle

doesn't have to be for hours. It just

could be for 101 15 minutes with

whatever the problem is or whatever the

question is. U ask your kids to do that

first and then get to AI next and then

for that for the from there have a good

conversation with AI, learn how to use

it, try to get a good outcome. But we've

all got to struggle I think with the

blank page a little bit more.

That was great. Thanks Greg. Greg

Schauve is the CEO of Section, a company

that helps deploy AI for enterprises. uh

and uh a good friend for 30 years. Uh

Greg, very much appreciate your time

today.

Thank you, Scott. Anyways, if you'd like

to submit a question for next time, you

can send a voice recording to office

hours of profia.com. Again, that's

office hours of profgdia.com or post

your question on the Scott Galloway

subreddit and we just might feature it

in our next episode.

[Music]