Welcome to another episode of Cloud

Unplugged.

So we have some topics obviously inspired

by the news today.

I'll kind of cover them in order that

we're going to kind of discuss them.

We've got the OpenAI's financial leak.

So they were going to IPO and then

there was a load of their finances that

have managed to get leak prior to that.

SpaceX buying Cursor for sixty bills.

Quite epic amount of money.

What's happening with cloud capacity at

Microsoft?

Because they're running GitHub in Amazon.

And also a little bit of a flashback,

Commodore is back.

New Linux flip phone, basically.

So let's get into OpenAI and the

financials.

I don't know if you've seen this, Salman.

um that they were kind of leaked i

don't know who leaked them but it was

a guy called ed citron who's a journalist

he's kind of got his own podcast show

but um they were basically given to him

somehow i don't know how um and there

was obviously the audits of the financials

there weren't there to be met i think

there were there were the audits for the

um for the ipo um

it showed interesting numbers of like

basically money moving back and forth um

as in like some of them revenue they

made was kind of under revenue from

investors um so that was a bit peculiar

um they are spending an insane amount of

money um i think the operating cost of

like twenty

point two billion um or twenty point nine

two billion or something um is what their

losses are at the moment but they're going

to be spending more as well even though

they're making a loss they're going to

spend more on that loss for more capex

expenditure on data centers

um and then they've like increased their

uh marketing spend by five times and

that's like up to five point seven three

billion pounds on marketing um that

they're kind of spending at the moment

which is kind of a little bit wild

anyway what do you think to those numbers

Wow.

Are these numbers even real, John?

What is this?

Is this a hallucination?

These are the real numbers.

These are the real numbers.

That kind of indicate, obviously,

we spoke about in the last podcast episode

about all the hype,

but we don't mean necessarily the

technology hype because the technology is

sound,

but obviously the valuation hype and I

guess the mind share that they're all

trying to get of everybody.

by just talking about mythos and all the

fables, it's there, it's not there.

And so it's like a constant AI washing

almost constantly in the media.

But whilst all that's happening and all

the hypes there,

obviously the valuation is inflated whilst

all that's happening.

And then this is kind of the truth

in the end,

the numbers get leaked and then you really

see, well,

the demand isn't really there to warrant

that valuation.

I think sometimes we just have to be

a little bit careful with the numbers that

show.

But again, it's a leak,

so we don't know how accurate the leak

is.

Within the same leak,

one of the numbers that came out was

that the company is getting more efficient

at

hosting losses.

So in twenty twenty four,

they were they spent two dollars thirty

seven cents for every dollar of revenue.

So, you know, there's a lot of losses.

But in twenty twenty five,

they're down to one dollar sixty for every

dollar of revenue.

And twenty twenty six, the numbers again,

you could choose if you want to believe

it's not a dollar twenty for every dollar

spent.

But that's within a dollar twenty now.

All companies, John,

that we know when they start,

they operate at a loss.

Like Uber, when they started,

they operated at a loss.

So much so that even now,

especially with these LLM companies,

we talked about GPT for taking costing

around hundred million to just train and

make it available.

so yes the numbers are out yes they

show that the companies are operating at

loss but so do most tech companies

alphabet meta in the beginning they do

operate a loss now the the the reason

why they got leaked again if they got

leaked perhaps just so you and i can

talk about it on this podcast john

or others can talk about on this podcast,

just to get their name out even more.

Even people ask the question,

is this company of any value,

of the valuation that's been given?

Is good marketing for the company itself?

So I don't know if seeing these numbers,

the headline numbers that are presented,

does it mean anything for us?

Perhaps not.

Do you think this is good marketing?

I think it's good marketing.

Really?

You think bad numbers are good marketing?

Because people will ask a question.

What is Anthropic doing?

What is Mistral doing?

What are the other companies doing?

so again is this bad i don't necessarily

think this is bad is it spending money

in the company it's the same history in

in the history we've seen all companies

lose money in the beginning okay how much

are you losing is a lot in this

case of course i don't know what the

comparison is to to uber i don't

necessarily think this is too much of an

issue

I don't think it's too much of an

issue, but I think what it showed, though,

is the money movement.

So,

nine percent of the revenue came from

Microsoft and SoftBank,

which are obviously investors in the

company.

So,

you've got money moving between other

companies that are also investors back

into the company that they're invested in

that then inflates, obviously,

the amount of revenue.

So, I think it's more things like that.

And then...

all of the r d money which you

kind of expect that was kind of spent

and i think then they've invested in

microsoft obviously for their expenditure

for the compute and things like that for

open ai and that money's then kind of

returning back so it's just

It's just oddity of the flow of where

all the capital is going back and forth.

And then the revenue is kind of part

of that capital, which is not normal.

Obviously,

revenue you'd expect to be based on the

things you are selling,

not selling them to the companies that are

already invested with you.

But yeah.

But if they have invested in you,

but if they're selling a service to

Microsoft, for example, Microsoft is,

you know,

you can go into Azure and start using

open AI models.

And company as an organization will also

use the open AI models, right?

If the revenue is that,

if that's where the revenue is coming

from, I think that's legit.

No problem whatsoever, right?

Because they're selling a service.

I don't know about SoftBank, John.

Okay, interesting.

I think so.

Maybe SoftBank is using their models.

Maybe they are using.

I have no idea.

I knew Sam has got to you,

hasn't it?

I just knew it.

I saw you.

I just knew this was happening.

I told you, John.

It's just what I'd like everyone to know,

that before we even started this podcast,

someone had to move a life-size cutout of

Sam from behind that he had,

that he had to take out of shot

of the video before we started this

podcast.

So, I mean,

weather is independent and there isn't

a... I'm not really sure, but...

But John, just humor me for a minute.

We've seen this money move around between

the same companies a lot, right?

NVIDIA investing in OpenAI,

OpenAI investing in...

This money is just moving around.

At the end of the day,

who's going to win?

Who's going to win at the end of

the day?

Well,

we all know it's going to be Anthropic,

right?

Yeah, I mean,

OpenAI is also going to get its share.

Anthropic is also going to get its share.

NVIDIA is also going to get its share.

Because the issue at the moment is, yes,

LLM technology is improving.

It doesn't need as much compute that it

needed in twenty twenty four to be able

to train and to be able to do

inference.

The chips are now getting designed

specifically for training and inference

for these LLM models.

So it's definitely going to get a bit

more.

Yes, it still costs a lot of money,

but the people who are going to lose

out is

that the reason why this cost is because

this training takes a lot of money.

But then all these companies are moving

LMS to usage based fee.

Now, instead of giving you credits,

I think what's going to happen is the

end users will end up paying a lot

more for these services because the

companies realize, okay,

Operating at a loss,

we need to adjust our prices to reflect

how much it actually costs.

So people who will lose out as well.

I don't necessarily think that means

losing out.

It's the same model every company uses.

Uber, when they came in,

they were much cheaper.

They used to give more money to the

drivers, but now they don't.

I think the difference in this, though,

is just the amount of money

we're talking about like this isn't like

you know five hundred grand here or like

a couple of million over there we're

talking like billions like billions that

most companies aren't even valued at this

amount so most you could buy an insane

number of companies for that amount of

money that's kind of going that they're

talking about here which is

all kind of fictitious to a degree.

So, you know,

it's not like it's evidenced hugely on the

capital expenditure.

So that's the bit that's a little bit

warped because if you look at it through

a market lens, its own market,

it's all quite, you know, on the up.

The taxi market or the car hiring market,

less so,

or the delivery market or delivery market,

probably less comparatively,

less so because it's more stabilised.

But I do know what you mean.

Obviously, companies that start

have to obviously make a lot of

expenditure to kind of grow their foothold

in those markets.

But they're spending five billion on

marketing, five billion on marketing.

That's the cost of most companies.

And you're spending five billion on

marketing in a market to then get how

much are you going to get back from

five billion expenditure?

What's the ROI on five billion on your

know token usage like what are you

actually going to get for that what's the

what's the business model that allows you

to know you can if i spend five

billion on marketing i'm gonna get how

much back as an roi on that marketing

Yeah, this is hard, right?

Because five billion in marketing seems

quite a lot.

But then these companies are taking out

Super Bowl ads and videos every day and

whatever's coming up.

But yeah,

five billion in marketing definitely

sounds a lot.

But, John,

these companies are in a different

ballpark now.

We're talking about companies with a lot

of backing.

So them spending five billion is like

other companies spending five hundred K.

So

I don't know.

AI has got to me, John.

The numbers don't seem real,

but I'm in for a ride.

Because it can't be quantified at the

moment because it's so new,

we don't know how people are going to

make their money back,

what the services are going to be.

Obviously, money has been made.

They are making money.

They said they've made €

Oh, no, they had a ten billion target,

but I think they made thirteen billion.

So obviously they are making money,

and that's a very high revenue.

But it does feel a little bit on

the hype comparative to kind of other

markets,

but a little bit mental to me that

that amount of capital has just been spent

with all the other things kind of going

on.

But I think, John,

the hype is there because the market is

way bigger than it was for Uber,

let's just say, right?

Let's take that as an example.

True.

And the market is way bigger than it

was for cloud.

Yes, we have companies, you know, okay,

we can talk about the cloud era as

well when it all became cloud native.

Of course, everybody needed to use,

not everybody,

but companies who wanted to host the

services, they will go tend towards cloud,

right?

So you had all these companies pop up.

But with these models, John,

Everybody can use a model.

Anybody who has access to a device can

use a model.

I know people in healthcare,

they use this model for various reasons,

right?

And I know people who are not even

like barbers,

they're using these models to come up with

some designs and whatnot.

Barbara or barbers?

A person.

Barbers.

Barbers.

All the barbers.

All of the barbers.

out there today, if your name is Barbara,

you are using the model.

And people are spinning up their own

websites.

They're doing their own design for the

logos and running their businesses.

So I think the market is way bigger

for this than it was for anything I

have seen myself.

Right.

I'm just talking about my time.

I hate to say this, John,

but you're right.

Air has got to me.

has got to me but i think that's

why the numbers are too big that's why

they want to spend that five billion

because they think they can get a bigger

share of the market and the competition is

tough against open ai because you're up

against anthropic who really is leading

the market right anthropic is leading the

market and hence why you know people say

why do they models got banned i'm not

saying anything john but yeah maybe

there's got something to do with it

Yeah,

I don't want you going in an anthropic

whilst you're a big advocate for open AI.

I'm not an advocate for open AI.

I'm not an advocate for entropic.

No, we get it.

We know where your alliances lie.

But yeah, I don't know.

The B to C aspect of it all,

I don't think most people are building

websites and doing all these things.

I think they're uploading pictures and

then being like, can you...

I don't know,

put long hair on me or can you

do this thing?

So the cost, though, of all those tokens,

the general consumer B to C at the

moment that you're saying it's a bit like

the Uber,

how sustainable that's going to be on the

capex.

I don't even know whether you'll even make

a return on that cost.

How much are you going to price those

plans at based on the cost of the

capex to get your money back?

What's the volume of

consumption because the average person

that's using at the moment it's affordable

but when it doesn't become affordable

obviously people can't use it and won't be

able to have the money to use it

also let's be honest they won't be able

to use it because they don't have a

job i'm joking because they've all been

taken by ai so um that was being

facetious jesus uh but yeah i still i

still think it's i still think it's a

little bit out there i don't i feel

it's definitely yet there's a market for

it

yes, it's a little bit more universal,

but the same as all websites need cloud

computing to host, you know,

isn't that much different for like new

services needing models to do features.

I don't think it's that like radically

different necessarily, or like,

some market that is that differential

because obviously to run things Netflix to

get those services that you've got

delivery or require internet or require

cloud you know you wouldn't say that those

markets are like eight trillion dollars

big and all these other things you know

so it's seems a little bit inflated but

No, but that's market for you, right,

John?

It runs on sentiment.

The sentiment behind this technology is

very bullish.

That's the point.

And that's why all these numbers seem

conflated.

But what are these numbers?

Are they even real?

I don't know, John.

I'm just here for the ride.

Well, we know where your sentiment lies.

That's for sure.

So we move on to the SpaceX.

purchase of Cursor as well,

which was also a bit of a curveball,

but quite smart, I think,

really by Elon in kind of trying to

get into the code area and to be

kind of more of a competitor to OpenAI

and Anthropic and others.

So it does kind of make sense.

But sixty billion, obviously, of stock,

basic stock to acquire Cursor.

Cursor being

Obviously,

they're for helping developers write code.

They have their own model.

They extended the IDE for VS Code.

What did you think about that acquisition?

Well, first of all, John,

let's start with this.

SpaceX IPO-ed last week, week before.

SpaceX, one trillion, right?

The IPO,

the share price started at one sixty,

one hundred and sixty dollars,

and it shot up to two hundred within

a day, right?

Two hundred dollars.

And now it's settled down to like one

hundred and eighty to one hundred and

eighty five.

So before we even talk about Cursor,

the IPO for SpaceX,

was a success because it didn't go down

below the IPO price.

And SpaceX doesn't just include SpaceX

itself.

It also includes Starlink.

It also includes XAI, John.

So now some of these moves are,

of course,

being done to increase their footprint.

Because we have a player, Anthropic,

in development LLMs,

and we also have OpenAI.

But really, there's no other.

So because now SpaceX have that capital,

I think it's good for them.

Well, that's what's happening.

They've taken a bet on the billion dollars

that you said, right?

Now, Cursor, as we know,

is a developer platform that people can

use to use any models they like.

And I think, John, at the moment,

Cursor spends a bunch of their revenue to

Anthropic.

Is that correct?

Like Anthropic do charge some of that

because they use the Anthropic models?

Claude, is that right?

Oh, I didn't actually know that.

I don't know.

Is that right?

I didn't know that bit.

I know they have their own model called

Composer or that was like what they built.

Yeah, you can also use it.

So I think the idea here is that

SpaceX is going to I don't know if

they have the boarded yet or is it

still in process of buying?

I thought they had bought it.

Yeah, but I assume they had bought it.

But yeah,

I reckon it's to take down OpenAI,

I think.

I reckon he's going to be a gun

for Sam Altman.

Yeah, so...

I think that's a fair point because Elon

Musk and Sam Altman,

they started OpenAI and there was a

fallout and Elon Musk left OpenAI and

there's been a bit of back and forth.

But I think then we probably will see

that Grok

models will be competition for the models

by OpenAI and Anthropic.

Because right now, Grok models exist,

but they don't have a massive user base.

People who are using X use those models.

Developers don't use those models.

Where do developers live?

Well, they live in Cursor at the moment,

right?

And that's one of the toolings people use

for these models.

So perhaps it is to capture that developer

market just so they can.

I think you're probably right.

They're gunning for OpenAI and Anthropic,

not just OpenAI, both of them,

because this is a huge market.

So they did quite well with the SpaceX

IPO, but they probably will do way better.

Now it's just marketplace.

You're in the IPO.

Now you have to keep making moves that

the market is going to like and the

sentiment is going to increase.

It will get better.

So I think this is a reasonable move

for SpaceX to make.

Now you're going to tell me that Elon

has got me too.

SpaceX has got me too.

I'm just talking about from market

sentiment, right?

I'm not saying this is right or this

is wrong.

I'm just talking about a market sentiment.

If you're running this company,

you want to increase your market share and

you want to increase your market cap,

what would you do?

And this is where the hot topic is.

But you don't think though, I mean,

obviously for him,

it didn't really mean very much.

I think it was like three percent equity

essentially of SpaceX.

That's kind of what he's given up to

then acquire Cursor.

So obviously it hasn't really cost

anything necessarily per se.

But Claude Code, I think, has like,

forty-six percent of the market.

I think Cursor had nineteen percent.

It was still growing because I think,

like you said, the market was massive.

But I have a funny feeling, though.

I mean,

I knew a lot of people that started

on Cursor who then moved over to Claude

Code and a lot of people that were

using Cursor,

but actually just using the Claude Code

terminal,

just using it as an IDE and then

stopped paying for Cursor because they

weren't really using their model anymore.

And I imagine that trajectory will

probably increase.

And even like today,

a lot of the customers we work with,

all I've ever heard is call code more

than cursor.

I don't hear anybody saying, hey,

how can I roll out cursor safely?

And what's the process for that?

It seems to be because Anthropix more of

general model,

not just a development dev tooling

centric.

Obviously,

it makes more sense to have an agnostic

approach to everything.

development for your business um you know

for your genetic services for you know you

don't so i i don't know it felt

like an odd thing unless he's got a

desires to to make the same his model

which is obviously the thing is it's

obvious to use that skill to get you

know um xai's model to be also a

universal model for everything as well and

help train it on lots of things i

presume

But John,

you're forgetting some things that we will

never have the information to decide to

make our own opinion, right?

Yes, Claude is all the rage today,

but is it going to be all the

rage tomorrow?

We don't know.

Here we go.

Anti-anthropic every time.

No, I'm just talking about this, right?

Which is why everybody needs to use

OpenAI.

God's sake, everyone.

No, but remember, John,

last week what happened?

The new model came out, Fable.

Everybody lost their mind.

I lost my mind.

I was like, wow, this thing is amazing.

Next morning I woke up, bam.

models not available in your country

because you're not an american citizen

right so why was that banned i don't

know john but also at the same time

it doesn't the tokens for fable were

really expensive that model is super

expensive if you come out as cursor

feat featuring xai and your token

consumption because token consumption is a

problem now right things are getting

really expensive and if you can somehow

say you know what this model is going

to cost maybe it's going to be even

a bit cheaper than what anthropic model is

perhaps you can get some people in your

you can get some of the market share

So all I'm saying, John,

is that I'm not anti-any company.

I'm not anti-anthropic.

But what's happening today is what not

necessarily is going to happen tomorrow.

So I think it's a good gamble,

in my opinion.

I'm calling it a gamble,

but it's a business decision because

SpaceX valuation is two trillion.

And as you said, what,

sixty billions or roughly three percent of

the equity slice.

imagine if you spend sixty billion you end

up getting two hundred billion why not why

not because you know you've got the space

side covered you've got you already have

XAI which is like the X and the

grok perhaps increasing that but you know

that's it'll be interesting to see what

happens in future do you think you will

manage to achieve building the data

centers out in space and then running

those models out in space and then using

Starlink to

then obviously stream back down and do you

reckon that's the yeah so so even in

their own ipo papers they said there's a

lot of challenges around running this

right so when the papers came out there's

a lot a lot of challenges but he's

it was a lot of challenges with spacex

really to be fair like everyone yeah you

know it's like to be able to do

that and to land back down and be

so precise and you know all those

challenges and make it that cheap and um

Yeah, I think technically, no, it's true.

It's true.

But the problem is, well,

it's not a problem,

but there's a slight difference because,

yes, making it reusable was amazing.

But I think there's like radiation is a

problem as well up in space.

But I think the reason why SpaceX was

successful is because SpaceX needed that

technology to send the satellites out

there to land people on the moon, right?

But data centers,

there's still a lot of land that needs

to be explored.

You know,

we talked about you can run your data

centers in the ocean.

And I know, John,

that you want to swim next to a

data center in a sea.

I know that's one of your bucket list

items.

So I don't know if they will focus

on running data centers in space because

you want to run some stuff on Earth

first.

When you run out of space, then perhaps.

Yeah, I mean,

that's the way to look at it,

isn't it?

Like, let's really...

the best way to align everybody to

building data centers in space or just

space in general is to ruin us first.

If we can get everybody

aligned on that if we can ruin it

first then we all want to go to

space we're like all on board we're all

aligned on that vision we're like yeah i

mean who wants to live here now it's

totally destroyed i at least we then share

the common vision of like how do we

support spacex and getting us to space now

that we've uh you know destroyed earth i

suppose it becomes a common maybe we can

get as a hotel there as well so

you're like you're all begging elon like

get us off this planet

Get me a seat on SpaceX.

Get me off this planet.

We'll get you into your flying Tesla.

You're not wrong, John.

Amazon came out last week.

Well, actually,

they came out last week and they said

their data centers use about two point

five billion gallon liter of water.

Billion gallons liter of water.

That's two point five.

And same with the same with I think

Google had similar numbers and I mean,

a little bit less or whatnot.

So you're right.

We are ruining the earth.

I think they are saying that they want

to become by twenty thirty water positive.

So as in like they don't waste the

water more than they use.

But yeah, this is becoming the efficiency.

This is how we're already ruining it,

right?

We're already ruining it.

Like body positive.

You become water positive.

Water positive, John.

yeah and what about this uh what about

this microsoft um capacity issue where

it's actually having to run github in

amazon what what have you got to say

about that so john we're talking about

ruining the earth right uh we need

capacity because we're making so many data

centers we're not finished ruining the

earth itself so we started moving into

water with data centers in water hopefully

coming soon and then maybe space as well

but here's the problem

So GitHub,

our favorite source control solution,

one of our favorite source control

solution.

In twenty eighteen,

Microsoft bought it and they said by

twenty twenty seven will fully migrate it

to Azure.

Right.

Because, you know,

they were running some stuff in AWS

before.

Yeah.

Yeah.

They still got it.

Yeah, they've got it.

Yeah.

So here is.

Everything's related, John.

So they were running this stuff.

They were fine.

They were on track to move on from

AWS to Azure.

But your favorite company,

Anthropic and OpenAI,

came into the picture.

And they told developers, oh,

in twenty twenty five,

you committed one billion commits of code.

That's cool,

but let's have a bit more slop.

Right.

Not just a bit more,

fourteen times more slop in our

repositories than we had before.

I feel like you're aiming this at me

and my PRs or something.

It feels really personal.

The way you directed that felt really

Like an attack on me, actually,

a little bit there.

Have you been reviewing my code or

something?

I've been reviewing your code, right?

I've been reviewing your code.

So the code commits within one year have

gone from one billion to fourteen billion

commits.

So the scale is massively increased.

And there were a lot of issues.

You might be aware of GitHub going down

because of the scale.

It was kind of not predicted that it's

going to be at that scale.

So basically what's happening is at this

scale,

Microsoft do need extra capacity to run

this.

And hence why they're like, okay,

you know what?

We do need to use,

maybe stick with a bit of AWS because

we're getting all these outages.

We're running out of capacity.

So this is a bit of a problem

that this is being produced because of all

these bots, commits that are coming,

not just developer-assisted commits,

but also agents committing code

themselves.

So this is a bit of a problem

for Microsoft because GitHub is losing a

bit of market share.

And the point here is that this is

an architecture lesson,

that even your own cloud isn't infinitely

elastic on demand.

And multi-cloud is sometimes, sometimes,

reasonable solution and that's what that's

what microsoft is saying themselves oh

maybe we need to be multi-cloud i mean

but that scale is it's completely bonkers

right eighteen fourteen billion commits

and a bunch of other things are happening

so this is uh so that's what's happening

john all these things are related with

each other and perhaps you're right maybe

we do need those

satellite data centers they're flying

around do you reckon github because of the

cost to them on hosting now will put

the price up of github surely they must

must do because the amount of like the

cost to like what's the margin on all

of that now they're going especially if

they're bursting into amazon they've got

to pay amazon for all that compute it'd

be interesting to see if they put the

prices up

Prices up for GitHub Enterprise.

Prices are for Copilot.

Yeah,

I think there is going to be an

adjustment, John, for sure.

Is it going to go up?

Is it going to go down?

No idea.

But inflation is also kicking in as

always.

So the price will only go up,

I guess.

But yeah,

this is a bit of an issue.

But the thing is that this is the

reflection point.

This,

we've seen AI infrastructure on demand is

really increasing a lot.

But we do need a bunch of that

scale.

It's not just for training the models,

John, right?

It's not just for those tools that are

going to give you all that slop.

But it's also for when you produce all

that code, where you host it,

where it runs.

So the demand is only going to go

up, John.

I think we should change and go into

data center technology, John.

I think that's what we need to do.

We need to get into chips.

Chips?

Not data chips, obviously, you know,

like GPU chips.

We need to get into chips and data

centers.

You're right.

That's where we need to do.

Oh,

we can get into the other technology that,

who was it,

NVIDIA had where you can strap it to

your house, basically.

Just like a mini data center strapped to

the side of your house.

That's...

That's the way forward.

Yeah.

Like a solar paneled data center on the

roof of your house or something.

Maybe.

Everything is compute.

But what about this Commodore?

Commodore is back.

Not the Commodore sixty four.

This is the Commodore flip phone.

Flip phone.

That has come back.

Yeah.

And running Android.

Are you going to buy one?

I'm going to buy one.

Why not?

It's a Linux powered flip phone.

It's going to cost around five hundred

dollars.

So it prevents.

This is the this is the kicker.

It prevents social media app installation.

John,

I know it's not going to work for

you.

Because you won't be able to install

TikTok.

I'm out.

You're out.

Yeah, I'm out.

But this sounds really good.

It's got a forty-eight megapixel camera.

We've now become tech reviewers now, John,

right?

So that's what we're talking about.

But the point here is that with UK

government deciding that we need to ban

social media for under-sixteens,

I don't know what your take on that

is.

I agree with it.

Plus, I'm not under-sixteen,

so it's okay for me.

Do you agree with it?

I'm not surprised.

I agree with it.

I agree with it.

I agree with the fact that generations

have grown up without social media.

And I think that was probably a positive

thing.

But I don't necessarily agree with the cop

out of not forcing companies to legislate

against protecting people on the platform.

It feels like the people penalized.

for the issues is now the children rather

than making the adjustments in the right

places to those platforms to begin with.

It feels like they kind of got,

I don't know,

it feels like there's going to be

bypassing and they've kind of got,

the people are going to get around it

and

The cultures have shifted, right?

We're just talking about AI.

We're talking about bloody rockets.

We're talking about data centers in space.

The world isn't quite the same.

People are swimming next to data centers.

So everything's kind of altering.

Social media has been a big part of

that for good or for bad.

Probably more bad than good, I would say.

But anyway, that's just my opinion.

They could have adjusted it, though,

and made it more safe somehow and got

them to change it and legislate,

not be like, well, I suppose, you know.

It would be like going on holiday.

Somebody dies from alcohol poisoning.

They're like, you know what?

Just stop kids going on holiday.

Let's just stop all kids going on holiday.

Nobody goes abroad until you're eighteen.

We can't have...

Do you know what I mean?

It's such an extreme...

view to the problem.

Fair enough.

That's because the government legislation,

they've been asking the companies to add

in more guardrails.

They do add guardrails like these

companies,

Instagram and Snapchat or whatever,

that you need to be a certain age

and that before you...

before you can log in,

but it's kind of hard to police that.

How would you stop anybody not like,

I can pick your phone, John,

if I know your passcode and I can

pretend it's you or log in using your

credentials.

People work your way around.

I think

You can argue both ways.

I agree with what you're saying.

But I think a blanket ban is probably

better because there's so much rubbish on

social media nowadays and the pressure

that it puts on the teenagers onto doing

specific things,

following specific trends.

Australia did it, right?

They started that a few months ago.

I think it'll be good to see what

happens.

But I agree.

That's why...

I have respect for making a decision and

actually coming down on a line against

something that is maybe not healthy for

kids and just generally their state of

mind.

I kind of agree with that.

It just feels like the wrong answer to

the problem to me.

But anyway.

John, John, John, let me, let me,

let me stop, finish with this, right?

So UK government mandates that everybody

under sixteen has to go to school.

Agreed?

That's a mandate.

Everybody under under-sixteen has to have

an education.

You could be homeschooled though,

can't you?

You can be homeschooled, correct.

But you have to be schooled.

But that homeschooling is by exception.

So are we saying that when the government

puts in a rule that says everybody has

to go to school, nah, don't do it.

Let parents decide.

Figure out how to do it.

But here's the thing though.

If a child says,

I don't really feel like going to school.

Well,

you have to go to school because the

government says you have to go to school.

Otherwise,

they're going to put you in jail.

So now it's become a lot easier for

parents.

If they say, oh,

can I have a Snapchat account?

Well, actually, you can't.

Did you think people weren't going to

school before that mandate?

Is that what the problem was?

Everybody,

because you'd be walking around.

No one's at school.

Do you remember that?

When you used to walk around?

Yeah.

All the schools were empty.

Do you remember those days?

All the schools were empty.

All the kids were just out on the

streets.

And then the government came along and

were like, you need to be at school.

And then everyone started to go to school.

I mean, that's not what happened.

They didn't need to mandate it.

There wasn't really a problem,

I don't think.

It was like, what was it for?

Like, whatever percentage.

It seems a bit like, yeah.

I mean, bizarre to me,

even mandating going to school.

It seems a little bit silly because...

Yeah,

I don't think that you're a homeschool

kind of guy.

I'm a homeschool kind of person.

I'm a social media.

I believe that people should upskill on

social media.

The best way to get educated is to

just get TikTok and to get Instagram.

And you just scroll and you learn through

scrolling.

And the more you can scroll,

the more you tend to learn.

That's what's got me where I am today,

Salman.

Without social media and scrolling,

I wouldn't be here where I am today.

Welcome to the Diary of a CEO,

where John tells how he got here.

Yeah, I didn't go to school.

It was all social media then.

That's how I upskilled.

But yeah, anyway, Commodore,

you're going to get one.

It looks really cool.

It's a funky screen.

It's like blue in color.

We'll never hear from you ever again once

you've got one, I think.

But it doesn't have social media,

so that's good.

I can't be on TikTok.

I don't know.

Anyway.

Until next time,

I think the next one we're going to

pick a topic and dig into it a

little bit more.

We could have been doing that anyway,

a little bit on these news articles,

but I think we'll probably go a bit

deeper into something more specific going

on in the industry on the next episode.

But yeah,

we shall speak to everybody next time.

See you later.