Hello,

welcome to another episode of Cloud

Unplugged and we're going to be talking

today about

AI and whether it is coming for everyone's

jobs,

whether people need to reskill and rethink

their career paths,

what it means to young people maybe that

are finishing university and trying to get

employed,

and whether or not the AI disruption is

real or whether people are justifying

layoffs on AI just because it looks

better.

Salmon.

How are you doing and what do you

think?

I'm doing well, John.

What do I think?

Well, what I think,

let's look at some of the data.

Let's see what the data is telling us

because, John,

we love a bunch of data and so

does AI.

AI does also love data.

So the headline numbers, John,

are that technology has led all the

industries with roughly a hundred and

fifty thousand job

cuts throughout july up from this year

from this year through to july almost

hundred fifty thousand this is a report

that came out by challenger gray and

christmas and the companies you know all

the big companies that we heard about

layoffs in the beginning of this year and

the end of last year

And all the companies have cited AI as

a reason for layoffs for consecutive

months.

We've seen that happen because these

companies have been saying that using AI,

we can be a bit more productive and

there's been a bunch of job losses from

the big tech companies.

And we've all seen a bit of a

trend where the companies have gone back

on what they were saying and rehired some

of these people.

So we'll come to that later.

So about one hundred and fifty thousand

from there.

And there's some other feeds as well of

news there.

I mean,

it's between one hundred and fifty to two

hundred thousand job losses,

losses roughly,

if you also include like month of August.

So that's the stats that are coming out

like a high level.

Is that global?

Yeah, it's global.

This global job is global.

And the main thing here is that AI

has been cited as a reason number one

for job losses.

So that's like a high-level thing.

But they're saying that another thing that

perhaps we need to look into is that

Stanford's Digital Economy Lab has found

that the employment for young people,

people between the age of

almost twenty percent below uh where would

we would where it would be without ai

so yeah think about a couple of years

ago three years ago four years ago that's

that's the two numbers that we are of

course there's a bunch of other numbers as

well about the skills gap and the training

that's been provided but this is high

level number are you surprised john by

hearing these numbers um i i'm i don't

know about the the layoffs um because i'm

not sure i've ever

you know, as in,

if we're going to trend it over time,

um, cause a lot of factors for layoffs,

isn't there,

then a lot of growth and bubbles,

you know,

it's very hard to kind of predict whether

there was a,

a COVID kind of hype spike.

And then like the dust settles,

the economy starts to kind of crash a

bit, um, and dips,

and then everyone starts to laying people

off because, you know,

things haven't done as well,

or the demands reduced or whatever's going

on and whether that's AI or not,

I'm not really sure.

Um,

I don't think I'm too surprised.

I think a lot of the layoffs are

from massive companies, aren't they?

So like Google, Amazon, Meta,

when they lay people off,

it's high numbers.

So it's not that everyone's losing money.

jobs all over the place because this is

like a sprawl of businesses across the

world that are all laying everyone off and

everyone's been impacted and feeling it.

It feels isolated to the mega big four

or big five organisations.

But interestingly,

I was speaking to somebody who worked for

Meta and their job had significantly

changed for recruitment.

And they were saying that they let a

lot of people go.

They've also not been hiring as much in

general,

which obviously won't be unique to Meta.

So the demand has kind of come down

a bit.

So obviously then the demand on the talent

team is reduced.

but also they've automated a lot using AI

for the talent team.

So actually they don't need as many junior

people because they've got quite a lot of

tooling and automation and they've made it

quite slick and efficient that actually

somebody that's very senior can use all

the right tools and then maybe become

I guess,

a little bit less strategic and a little

bit more of an executor with the tools.

Whereas before, maybe career path wise,

you would have been thinking, all right,

I'm going to become more strategic.

I'm going to have people underneath me and

I'll drive the strategy with these people.

And so that's a bit of a,

I guess,

a little bit of a shift on the

standard archetypal pattern of career

progression and how a business runs and

the hierarchy and people under you and the

kind of structures is now shifting.

So that's a little bit different.

um but yeah not too surprised but it

is still quite high numbers yeah so so

you think you don't really fully buy that

the cuts are because of ai but maybe

perhaps because over hiring by these

companies for over covert time and then

cutting that down but it's interesting you

mentioned about meta and having like

senior people's being executors and not

having people underneath them

do you think there will ever come a

time where there'll be a gap in the

market where like you have the senior

people who replaces the senior people or

we'll get to a point where the tools

will be so good that you don't need

anybody to execute specific

roles that are doing these bits?

Do you think what's going to happen?

Because right now,

I know you gave an example of Meta.

I'm not sure if this is happening

everywhere,

but you don't need junior people to have

the career progression.

But what will happen in the ten,

fifteen years time, John?

I don't know.

I think we've already got this problem

with the boomers.

If you look at the boomers,

they're still all working.

They were at the right time at the

right place,

and they're holding on to loads of

property.

they're still like in quote senior roles

and then all this you know their children

are living with them and they can't afford

to buy a house and so there's been

all this like generational kind of

opportunity that's changed anyway right i

mean that's apparent today i know this is

obviously supposed to be about ai and jobs

but i guess the point being that there's

already

differences per generation on like how

things function through to historic

reasons so I guess you're already seeing

those shifts so I'm not sure in terms

of what does a business do in terms

of skills and capability like how do you

acquire the knowledge to become senior and

if you haven't done the job from the

bottom to begin with and then you know

are those people in

limited supply but if it is if the

demand is reducing anyway then you only

need a certain amount of supply so it

probably equals itself out i suppose maybe

there is less of them and maybe the

that makes you more scarce a resource you

know over time in the market um but

i think people will always have people

join like i think economically

having graduates is cost effective for a

lot of companies.

I think you still want to have those,

whether they're reducing in number,

probably.

but you still will have them because,

you know, businesses have margins,

you know, there's run costs to a company,

even your product business,

you've still got OPEX costs.

If you can get, you know,

cheaper resources doing things,

that's obviously better for the bottom

line.

So I think that's always going to be

true.

That will never change that bit.

So there'll always be, I think,

graduate schemes and opportunity just is

probably shrinking a little bit rather

than changing.

What do you think?

Yeah, because I think, well,

you are up here, John,

are hiring graduates.

So shout out to Appia for hiring

graduates.

I just see, John,

that I think some areas are impacted more

than others.

Like, for example,

if you look at customer service or data

entry,

That's kind of where I know this company,

which used to have like forty paralegals

to check the lawyers paperwork and

whatnot.

Now they have got these agents, AI agents,

and they only require four agents instead

of forty paralegals to check that work.

So some of the bits that could be

automated are getting automated.

It definitely is having an impact,

but then at the same time,

now the role is shifting.

You're actually now becoming, you know,

you're checking the output of the AI agent

rather than being a full paralegal.

Yes, you're still a paralegal.

Of course,

you still wear the same hat because you

need to have that knowledge of being able

to say, like, this is correct,

this is not,

because at the end of the day,

still predicting next token, right?

I know you hate it when I say

LLM is predicting next token.

You still need to be able to figure

it out.

But yeah,

there's definitely cuts for sure.

And yeah, it's impacting.

But yeah, you're also right.

You will need

we saw this from from meta themselves and

even uber they mentioned that they got rid

of a bunch of people and sorry meta

got rid of a bunch of people then

they hired some people back and uber

finishing all their ai credit because

don't forget

the tokens that people have to use are

also not cheap.

So it's not like you're getting it for

free, right?

Yes, you might get rid of some people,

but you still have to pay for these

subscriptions.

You still have to pay for these tokens.

Unless you use the Chinese models,

then you're trading something else.

But they're much cheaper than the ones

that we use day to day.

But these are Uber burnt all their twenty

twenty six budget the tokens in April for

the whole year.

And the CEO said they didn't see anything

beneficial come out of it.

So, yeah, it's a little bit weird.

But if you think about it,

because I was saying that it's like

Sixteen percent is it in the data,

so sixteen percent of sixteen to twenty

four year olds can't find work compared to

five point one percent nationally.

But that's probably always been triggered

out of any experience.

So it's going to be hard to find

work without any experience.

That's just a given.

But I think it's kind of increased.

And they were saying that then the job

postings for graduates is down seven

percent year on year.

So it's not that that isn't new,

obviously, because of AI.

I think it's just been happening anyway.

So there's some things that are just going

to happen.

There's some things that will always have

been true.

And then we're in this weird place where

no one really knows

What are all the new jobs that are

going to emerge once?

Because everything kind of moves forward

and something new will happen, right?

So at the moment,

we're all trying to figure out the usage

of AI.

People are adopting it.

It's not foolproof.

You can't really trust it.

You need a lot of assurances around the

quality of its work.

Someone's still going to steer it.

um so that will probably go on for

quite a long time yeah yeah then there'll

be just new things you know there'll be

new markets that will probably pop up

around ai to start to fill the gaps

there'll be new businesses that will start

where someone's like oh my god a bit

like clouds you know before the cloud

security

You know,

cloud security wasn't a massive thing.

Cloud comes up,

they don't really do very much around

security.

Everyone's like,

I don't really know what's going on with

my cloud.

I don't know if I'm secure or not.

We've still got all this regulation.

We've got loads of these assets in the

cloud.

And all of a sudden the massive market

starts and you get into this big,

security market new companies whiz all the

others kind of start and you know it

becomes like a serious market and they

start employing people and then people

start to move into that space and and

on it will go so i think it

tends to roll forward rather than like

cause disruption uh fully forever if you

see what i mean

Yeah,

and don't forget the Kubernetes engineers,

the role that came out.

Sorry, what's Kubernetes?

Yeah, I don't know, John.

I don't know.

Is that a new model?

Is that a new model by China?

Is that a new Chinese model?

No,

it's a model that existed before any of

the chat GPT models.

Right, okay.

The Kubernetes model is like the best

model.

That's the real deal.

The best model that there is.

That's the best model.

The interesting thing here, John,

is that Stanford's digital economy lab

looked at the payroll data in USA.

That's how they figured out the numbers

that you give like, sixteen,

nineteen percent for twenty two to twenty

five years old.

But what they found for people with

experience,

so we're talking about people with a few

years of experience.

There was no evidence of any job losses

or any disparity between what they had

before and what they had now.

The interesting thing out of all of that

separately came out is that for the class

of twenty twenty six,

the people graduating right now,

that the hiring is still growing,

and then there's actually a shortage in

robotics, automation, generative AI,

AI governance roles.

So it goes back to the point that

you're making.

The roles are shifting.

So yes, there's fewer jobs,

but there's actually a shortage in some of

these roles that are being created.

Like you said, for cloud,

when it came out,

there were probably different new roles,

and now we've got these roles.

young people.

It's not like young people are losing jobs

to AI broadly.

It's specifically entry level workers in

AI exposed roles versus experienced

workers in identical roles.

So basically the thing is experience at

the moment, experience is a thing to have.

It's not the job title.

And junior roles will more often

specifically there because the work was

well defined and repeatable so if you have

a work well defined work and repeatable

you can provide ai can perhaps handle it

and so that's what we're seeing at the

moment so overall yeah sorry go on no

no you you go you go um yeah

i was gonna say like

know paying attention to the i guess the

issue is is that it's quite hard uh

for universities or even the school's

educational systems to kind of keep up

with what's going on because there's a lag

isn't there in those areas they aren't the

industries education industry isn't a

forward-looking industry right it's yeah

academic it's got you know um

what you call it, a syllabus to it,

of educational syllabus that's defined and

static and it doesn't evolve very quickly.

So there's obviously like a serious lag.

But at the same time,

if you're paying attention to the world

and what's going on and you're a bit

more connected, which everybody is,

you know, social media, things like that,

Then we know that chips,

the manufacturing devices, energy,

all these things are all being challenged.

Who can manufacture all of these chips

cheap enough?

What about the semiconductor industry?

Where does all the energy come from?

How can it be efficient and effective for

a lot of the consumption that we now

need based on AI?

so there's like a huge probably deficit

you know on like innovation that's got to

happen in those places there's bound to be

loads of investment that's going to sprout

up um the same way as people are

asking even now for ai skills so i

think that's gonna i don't know however

many percent now i think something like

sixty something percent um of growth of

like people requesting ai skills or

something so

Yeah,

that already shows you that the skill set

deficit of like people aren't very

experienced on something that's new.

And so obviously they need the skills and

then on and on it would go because

what type of experience you need to be

like with the cloud will evolve.

Like you're saying governance.

So then they're like, well,

it's not about how you prompt.

It's about actually the security around AI

and what it means.

Then you specialize.

be like an ai security engineer um or

something um and so they'll be just new

new roles basically i kind of think but

um i guess if you were at university

now and you were you know i don't

maybe university or not university if you

were like getting your grades people are

getting their grades now right a levels

yeah gcscs what would you do if you

were getting your a levels now i don't

know what they'd be what do you reckon

they'd be

What do you reckon your grades would be?

Do you reckon they'd be good?

Oh, yeah.

Yeah,

they'll be good because now I'll smash it.

I'll smash it.

Yeah, you've got nines every day.

You've got A stars.

Yeah.

A stars.

Give me everything.

Nines, everything like that.

I'll smash it.

I think... So where would you go then?

What would you...

If you were like young Salman,

young Salman, young Salman,

what's your move?

I always wanted to be a carpenter.

Because I think that's good.

Carpentry is good, right?

Yeah, carpentry is good.

The important thing, John, here,

before I answer your question,

I'm not trying to sidestep your question,

is that so PwC,

they have this global jobs barometer where

they keep checking job postings throughout

the whole world.

well they saw that the the jobs that

required ai the premium for those jobs in

twenty twenty six jumped to sixty two

percent compared to fifty seven percent

last year right compared to a twenty five

percent the year before twenty four or

twenty five percent the year before so

what that means is if you have any

ai skills

be it you know how to create a

model, train a model, use a model,

doesn't matter what it is.

If you know how to use these AI

tools,

the job that you're applying for that uses

these tools will give you a higher chance

of earning more and landing a job more.

So no matter which field I would pick,

doesn't matter what the field is.

Because most of the fields now have to

use these AI models.

I was reading somebody's blog and they

were like, yeah,

I'm an executive assistant,

but if I don't use AI,

my job is way harder than it used

to be.

So it doesn't matter which area you go

in.

I have my brother-in-law.

He is like an eye surgeon.

use ai all the time so they're using

all these tools to do their job so

doesn't matter where you go you still need

to you need to use these tools it's

like john remember when office came out

and they're like ah yeah you need to

know how to use office before you get

any job like you know microsoft word and

microsoft excel it's becoming that it's a

net job creation excel you've got to excel

in ai now you have to excel yeah

So to answer your question,

I think I would still be in like

some maths field.

I'll still be in some engineering field

because that's what I studied.

That's what I would do.

What about you, John?

What would you,

which area you would go in?

Do you know what?

I am probably not a very good example

of having...

of like i wouldn't know what to do

i copied people uh you know from school

because i didn't know and then i was

like just used to ask people like oh

what are you going to do and they'll

be like i'm going to study these things

i didn't you know to be fair i

didn't even know about a levels really i

never even thought about it and then i

there was an opportunity to do them

obviously and then i just basically didn't

really know and so then just kind of

copied and then i

Then everyone was going to university and

I was like, oh,

maybe I'll go to university then.

Um,

and then didn't really know what to do.

Um,

but I was already quite hands-on

technically music played around with

things.

So actually I ended up studying,

but it was probably a waste of time.

I don't think I should have done that.

And then I was working all throughout

university and I just tend to have a

good, a different work ethic.

So I think I just go straight into

work.

Yeah.

Um,

Nowadays, if I could advise the old me,

I'd probably think about practicing a

vocation rather than theorizing over one.

And I think practicing something is better

than the theory a lot of the time.

And then the theory you can learn because

it supports the practicing of something.

Yeah,

apprenticeships or would be probably

something I would probably choose to do

nowadays,

like kind of get in somewhere to do

that.

But also I'd lower my expectations.

I think everyone's expectations are quite

high nowadays.

They want to leave and like,

I don't know,

they want to be an influencer or they

want to be whatever, right?

I don't know what they want to be,

but they expect a lot rather than just

being like content to have a job,

I think is the first thing.

Treat it as a vocation where you learn

a lot about yourself,

how to interact with people,

how to be disciplined,

how to be prepared and organized,

how to communicate well.

So there's a lot of other skills outside

of just the applicable skills that you

kind of learn.

And so I think getting into work is

probably the prime.

Just get into a job.

As soon as you can get exposed to

working environments and then just keep

pushing yourself really.

And I think you'll kind of just do

well anyway,

if you follow the market and opportunity,

you'll tend to do well without needing a

vocation specialism.

I think that's my, anyway,

that's a very specific thing to me.

So there you have it, folks.

If you want to be a successful CEO,

don't worry about your studies.

Just copy what John did and you'll do

well.

You don't need to worry about anything.

Don't read about anything.

Don't get educated on anything.

Just listen to our podcast.

Just listen to podcasts,

turn up to places, demand things.

But you need to employ me,

be really demanding and then follow them

around, stalk them, send them letters,

take pictures of them.

John,

I'm going to have to stop you there.

Do not do that.

But definitely do open doors for

yourselves.

by making yourself visible and going to

see, like, you know, talking to people.

But, you know, John,

I think the point that you're making is

a very valid one.

Because, you know,

depending on some jobs do require for you

to have that knowledge.

If you want to become a doctor,

you can't just turn up in a hospital

and be like, oh,

I'm going to become a doctor.

So some jobs do.

Well, you say that.

You say that,

but I feel like I have met a

few doctors where a part of me did

feel like, have you just turned up?

That's fair enough.

I don't know what kind of places you're

going to, John, but that's fair enough.

But I'm going to go back to the

point that you said,

because a lot of companies now are saying

the enterprise leaders,

these reports are coming out.

They're like, oh,

eighty percent of the companies do provide

some sort of air training now what that

might be is up for grabs but then

about sixty percent of people are still

saying they have an active skills gap um

so i think you but you joining the

workforce earlier on will expose you to

these things because as we just said the

education system is a bit slower in

catching up and what's actually happening

in reality

And even in school and also in

universities,

because most of the times what you do

in a university come out for work.

I've seen this myself and say, oh,

forget about everything you studied in

university.

It's not really going to help you.

The stuff that you do, teamwork,

how to solve a problem will still help.

But the technical bits that you learn has

gone out of the window.

So yes,

apprenticeships are a good thing because

people say, oh, yeah,

I do know how to use chat GPT.

But prompt engineering is like a

standalone skill which has a very short

shelf life because it exists right now and

the interfaces are getting better with

this stuff and they're becoming more

forgiving by the models themselves.

The skill will be for people to,

of course,

you definitely need to learn how to use

these tools,

is to identify what the output should be

of it.

What's the right thing to do?

That's probably going to become more

valuable.

And, you know,

we talked about how to implement it

properly.

Governance and compliance will always be

there.

Yeah, so there's a bunch of this stuff.

So, John,

let me ask you to maybe perhaps we

can start wrapping up,

bringing it back as if you're a listener,

and you're an early career,

because I know, John,

you've got many fans.

I'm not your only fan,

but you have many fans in an AI

exposure.

Or let's say you're a manager,

and you need to decide what to invest

your training budget in.

What advice can we give them from the

data and the discussions that we've had?

um well i guess it depends a little

bit but i i would say personally i

always wait more on um soft skills and

capabilities over just technical ones so

because they're a little bit more

universal so i think be really good at

facilitating being really good at critical

thinking be able to solve problems and

work a problem through like those skills

are universal doesn't matter

what it is you're trying to apply the

ability to, to scope and think properly.

Actually, I will share something.

It's like a live sharing thing, random.

It's a little bit random.

Let's see if you can see it.

Can you see it?

Add to the stage.

There we go.

So basically this is where we are today

is in like a lot of the execution

is by people and obviously the definition

and the verify, you know,

mostly historically has been quite

lightweight.

You could argue that the due diligence on

the stories,

what we're actually really trying to

build, you know,

is more reliant on the executors in the

end.

And a lot of things like everyone biases

for execution because it feels like you're

being busy.

We've got a bunch of people.

We need to make them active.

Let's just build things.

And a lot of the wrong things get

built.

That's just kind of a common pattern.

And how people think about what it is

they're building and challenging.

Like, why do we need to build this?

What's really the requirement around it?

What users get impacted?

Is it the most valuable thing for us

to build today?

Should we be building something else?

know those things that tend to be

traditionally product managers and

delivery management roles i think though

when the execution drops and you now have

ai to do a lot of the execution

um you've got to change you've got to

think about how do i verify the quality

of what it's doing and how to define

very clearly about what it is that it

really needs to go and do

and that's why you're seeing a lot of

slop you know the whole ai slop is

because people don't know how to define

very well um and because they don't have

the skill because they've been executors

and they also don't really think about the

verification because again they've

probably not really had to worry about

that what's the definition of done what's

the quality assessment of the work that's

been produced maybe there were other

people a qa team or whoever else right

so all of a sudden the accountability is

starting to shift around we're like i need

to get very good at defining

because it's basically the definition is

so loose it's just going off and you

know doesn't have enough and builds all

kind of random things that maybe aren't

aligned to what we need so i think

this is where the shift is going to

go and these are for me i think

the soft skills if you can master those

That's great.

I think then also the verification,

that's like quality, security,

the non-functional.

How does all that get done in the

new world?

What does that look like now?

You've kind of got AI.

Is that different solutions to that

problem that aren't the same company?

Are there new vendors out there?

Is it like a different set of capabilities

and data and metrics?

And how do you start to weave that?

those types of things into the definition

at the beginning.

And so that you can kind of measure

succinctly the quality in the end of what

you try to do versus what got produced.

i think there'll be all these new skills

and new jobs and so i think if

i was to say invest time i'd probably

think about good product management ba

type skills first if you wired that way

or thinking about probably the data side

and the verification side and the quality

side and the security side if you've wired

more probably on the technical aspect and

you want to kind of lean more into

that that would be my advice i don't

know what you think

I know it's a bit random,

and I just went through this as a

slide.

No, this is absolutely good, right?

Basically,

what we're getting is that you need to

move towards things that are hard to do,

that are not defined that well,

like the ambiguous stuff that, you know,

things like you already mentioned learn

how to audit the output of an AI

you know ask how do I know that

the answer that I got given is actually

correct or not but the only way you

can tell it's correct is if you have

that domain knowledge so that it's not

going away you know that you still need

the database administrators for example to

tell you that the thing that's giving you

is incorrect imagine you don't even have

any knowledge you run a command that drops

the whole database lovely

that's a big bigger problem right and then

also being able to like we it will

still hallucinate because that's the way

it's been developed the AI model so you

need to figure out uh can I check

the citations can I source the documents

can I look at the API documentation so

this stuff I think it's uh it's absolutely

valid that what you're saying is you you

need to

Early careers,

pick roles that are exposed to AI that

will give you those skills to learn and

to figure out.

And also in your free time to use

AI tooling to do what you want to

do.

It'll probably help you in the long run.

So don't specialize in AI.

uh interface i guess right specialize in

the judgment that's behind it yeah that's

what we're saying that is what we're

saying so that is fair enough john um

i love your words that are going to

encourage the the the younger people

graduating i wish i heard your speech when

i was graduating john and i'll be a

different man today

I mean,

we still don't really know what you would

have done.

I mean, we know you've got good grades,

but it seems like you said you wouldn't...

I told you I'd be a carpenter or

a glassblower.

You said you wouldn't ignore the question,

and then you managed to somehow ignore the

question at the end.

No, no,

I said I'll still be in engineering,

because, you know, engineering is still...

You'll still be in engineering.

I like problem-solving, all right,

building things.

So you'd be doing this...

It doesn't have to be software.

You'd be doing...

the more on the how do you prove

out the quality and the security or the

model building,

you'll probably be more on that side with

you.

no no yeah not even that build trains

why not john let's build some machines you

know but you oh right so even build

these yeah no yeah i was still probably

more kind of engineering even yeah yeah

probably more in the i studied aerospace

so i still have that but you know

i turned up in i ended up in

software which is which is which i love

and uh probably like yeah model building

Great.

Like it.

I think using those models to build other

things is where the benefit is.

How many people are going to take these

models and build?

I can't build a better model than the

people that already the models out there

using these models to see what we can

design, see what we can build.

is where we will see all these benefits

and that's what's happening all across in

all industries and you know just use these

models to figure out what type of material

should i use when i'm designing a cog

for an engine oh perhaps you know you

can come up with a different material that

we've never seen before i don't know but

i'll be something useful nice and if you

can't do any of that just become an

influencer uh start a podcast yes i'll be

a supermodel

You know, you can't do that.

So anyway, on that note,

let's head off and we'll be back as

always on the next episode.

See you later.

Thank you.

Cheers.