0:00

This episode is brought to you by Intel. Deliver all the AI results you

0:02

need, everywhere you need them,

0:05

with Intel and Softchoice, together.

0:08

Contact a Softchoice representative

0:08

today to learn about Intel's

0:11

comprehensive AI solutions, and

0:11

how we can unlock them for you.

0:15

Or visit Softchoice. com to get started.

0:18

Aaron Brooks You're listening

0:18

to The Catalyst by Softchoice,

0:21

a podcast about unleashing the

0:21

potential in people and technology.

0:25

I'm your host, Aaron Brooks. Imagine a world where PCs can

0:30

learn, adapt, and process complex

0:34

data in real time without needing

0:34

constant cloud connectivity.

0:38

This isn't a distant dream. AI PCs are bringing that to life

0:40

with the latest advancements of

0:43

an NPU, or Neural Processing Unit.

0:46

But what exactly are NPUs? And why should you or your business care?

0:51

How can they transform the way you do

0:51

business and how people engage with AI?

0:55

To help answer these questions, I'm

0:55

talking to Cameron Allen, the Partner

0:59

Account Manager at Intel Canada. He'll share how Intel is keeping pace

1:01

with rapid AI advancements and how

1:05

NPUs are making AI solutions more

1:05

accessible and sustainable for everyone.

1:11

So hey Cameron, welcome to The Catalyst. I'm really excited when

1:13

I hear the word Intel.

1:15

I kind of have a bit of a flashback

1:15

to being younger and spending

1:19

countless hours building my own

1:19

gaming rig, trying to figure out

1:22

which CPU I'm going to purchase, what

1:22

RAM type the GPU for my video card.

1:28

I've always loved this

1:28

space of technology.

1:30

And so I'm excited about all the

1:30

advancements Intel has been making

1:33

over the years and changing the tech

1:33

game and devices and processing and how

1:37

it's evolved the impact on the AI boom.

1:39

God, there's just. So much to talk about, but before we

1:40

get into that, I'm really curious how

1:44

you found your love for technology. Sure.

1:47

Absolutely.

1:47

Likewise, I've

1:47

always been around technology.

1:50

I had family members that

1:50

grew up in the tech industry.

1:52

We built PCs. I remember same thing on the, on the

1:53

kitchen counter as a kid, plugging in

1:56

all the fans and, you know, seeing the

1:56

lights and everything were together.

1:59

So I've always been amazed by

1:59

technologies in technology.

2:03

Impact on our lives and it starts

2:03

out as that kind of nerdy techie.

2:07

Let's just see what this can do And you

2:07

know, they say to do something that you

2:10

love for work and it's relayed itself

2:10

into that I started my career in another

2:14

industry and when I was looking to get

2:14

out of that I thought you know What do

2:17

I like and I like technology and so? Absolutely intel is one of those

2:19

companies when you look back at the

2:21

history of it Just the sheer innovation

2:21

of what intel has developed, right?

2:26

You know people take for

2:26

granted things like what?

2:28

Wi Fi, USB, these are Intel

2:28

inventions that are just so synonymous

2:33

with the world around us now. So yeah, I've, I've been in the tech

2:34

industry for about 10 years now.

2:37

For me, it really is about how technology

2:37

enables things and enables people's lives.

2:43

You know, we make people's lives better. And that was what really drew me to

2:44

Intel was the fact that our mission

2:47

statement is using technology

2:47

to make everybody's life better.

2:51

You know, it's not necessarily about business. It's not about profits and losses.

2:55

It's how do we make people's

2:55

lives better with technology?

2:58

I love that. I love that. You bring up a really interesting point

2:59

about Intel that they're in so much

3:03

that we don't see they're like that

3:03

behind the scenes juggernaut that's

3:06

making all these things a reality,

3:06

which is so cool to remind everybody

3:11

just the impact that they're having. Now, I love the fact that you

3:12

brought up your company's purpose.

3:16

We're on this bit of a journey ourselves.

3:18

At soft choice, not just realizing

3:18

our corporate purpose, but

3:21

also our individual purposes. What makes us tick?

3:24

How do we show up to the world,

3:24

to our people, to our customers?

3:28

So I'm just curious if you'd be

3:28

comfortable sharing what your

3:31

personal purposes with the audience.

3:33

Yeah. Again, I think the reason why I joined

3:33

Intel was a very close alignment

3:37

with those, those personal values. I think when you look

3:39

at companies nowadays.

3:42

You can't just have a

3:42

corporate mission statement.

3:44

You need to have a culture. You need to have a mission statement

3:46

that resonates with your customers,

3:50

you know, social responsibility. And, uh, you know, I didn't tell

3:51

sustainability is a big thing for us

3:54

that we're really focused on in terms of

3:54

the manufacturing process that we have.

3:57

It's important to us that we're net

3:57

water positive, that we're going to be

4:01

carbon neutral in the next 10 years. So those types of things

4:03

definitely resonated. I think my personal mission

4:05

is helping to find win win.

4:10

Win solutions for everyone involved.

4:12

I really like being able to take a

4:12

technology, a solution and bring it

4:17

from conception all the way through

4:17

implementation and seeing that moment

4:21

for the customer where they go, Hey, this is what we were looking for. This is what we got.

4:24

And, you know, soft

4:24

choice wins, Intel wins.

4:26

And at the end of the day, the customer

4:26

wins, which is, which is always the best

4:29

part. I love that. I'm totally going to make note

4:30

of the win, win, win strategy.

4:33

I always hear like the win, win

4:33

strategy, third, it's the third win.

4:36

I love that. All right. So I know we geeked out a little

4:37

bit at the beginning around CPUs

4:40

and GPUs, and that's where you're

4:40

getting ready for this conversation.

4:44

I came across something I probably

4:44

should have known, but didn't.

4:47

An NPU, a neural processing unit.

4:50

It sounds so cyberpunk. I love it. What is it?

4:54

Yeah. So the, the NPU, the neural processing

4:54

unit is really what is going to make.

4:58

this AI PC happen. So without getting too technical and

5:00

getting too far into the weeds, uh,

5:04

the neural processing unit is the

5:04

latest technology from Intel that is

5:09

going to enable AI on the PC device.

5:11

So when you think about a chiplet,

5:11

you used to have a CPU, then we

5:15

integrated the GPU, and now we've taken

5:15

this really a quantum leap forward.

5:20

In adding an NPU, so it's going

5:20

to be integrated on the chip and

5:24

it's really going to be driving

5:24

everything in the background.

5:27

So you're going to have things that

5:27

are going to run on the CPU, things

5:30

that are going to run on the GPU, and

5:30

then in the background it's going to

5:32

be this NPU really directing it all. And that's what's going to drive

5:34

these outcomes and is really

5:37

what is going to help enable

5:37

AI on the PC locally for users.

5:43

What would be some of the quote

5:43

unquote things that would run on that?

5:46

Is that the whole natural language model

5:46

and the processing algorithms around that?

5:50

Like maybe a little example

5:50

of what that would look like.

5:53

Sure. Yeah, it really varies. And I think the thing

5:55

to keep in mind is that. AI is going to be a software play, and

5:57

the software is only going to be as good

6:01

as the hardware that it's enabled on. So all the different software

6:02

applications that, you know, your

6:05

customers and your users are making

6:05

use of, those applications will

6:09

be written for a specific chiplet. And again, I don't want to go too

6:11

technical or too nerdy today, but an

6:13

example would be, you know, Microsoft

6:13

Teams is going to run on the CPU, but

6:19

The NPU in the background is going

6:19

to take some of that workload off to

6:22

allow the CPU to run more efficiently. I gotcha.

6:24

So we talked about the gains, you're

6:24

going to see performance gains,

6:26

battery gains, because a lot of that

6:26

workload is being taken off of the CPU.

6:31

And so things like your large language model are actually going to run on the CPU.

6:35

I think a lot of times people associate

6:35

large language models with AI, and that's

6:39

true, but different things are going to

6:39

run on different parts of the chiplet.

6:43

And actually it's about having all

6:43

three of them working well together

6:46

and having that optimization

6:46

to really see those gains.

6:49

We're trying to help customers understand

6:49

that this isn't just the next generation.

6:54

This isn't just a small

6:54

little gain to the chipset.

6:57

This really is the biggest

6:57

architectural change to the

6:59

chipset in the last 40 years. And I think, you know, Over those last 40

7:01

years, we've really gotten customers used

7:05

to the idea that, Hey, every couple of

7:05

years, the new chips that comes out and

7:08

you know, you get your marginal gains. This isn't just a small increase.

7:12

This is a brand new technology

7:12

and a brand new architecture.

7:16

And really in the last 40 years is

7:16

the biggest architectural change

7:19

that we've ever made to the chip set

7:19

by adding this NPU to the chiplet.

7:23

The real question is, would this

7:23

make a difference on my old gaming rig?

7:26

So, yes, I

7:29

mean, ideally, I think the

7:29

rest of the hardware in there might

7:32

be a little bit further behind. So you probably have to update some of

7:34

the other components in there as well.

7:37

Um, but that, that really is the way that

7:37

we're pushing the AIPC and the NPU is.

7:42

Again, there's a little bit of a misnomer. AI has become such a buzzword and

7:44

everyone thinks it's brand new.

7:47

And the reality is it's not AI

7:47

has been around for a long time.

7:50

It just used to take massive

7:50

supercomputers and data centers

7:53

to be able to process all of that. And so while you could run.

7:57

These AI applications on an older

7:57

PC, it's just going to get throttled

8:01

because it wasn't made for that. And so this is really about getting

8:03

the right hardware for the new

8:06

technology and for the

8:06

new applications that are coming.

8:09

I gotcha. So it'll help with a lot of the

8:09

frustration normal users have on

8:13

their phones being slow because there

8:13

are a couple revisions behind and

8:16

the applications coming out can't

8:16

take advantage of that hardware.

8:19

And so it's conceptually like

8:19

that, but on a grander scale.

8:23

Exactly. Like that PC just was never made

8:23

for what we're asking it to do now.

8:26

And speaking of PCs and the buzzword AI,

8:26

I understand that Intel's got AI PCs now.

8:33

Maybe help me understand a little bit

8:33

about what an AI PC is and how it's

8:36

different than traditional laptops. For sure.

8:38

Yeah. So at its root,

8:40

an AI PC has that NPU in it.

8:42

The broader definition though,

8:42

is really bringing AI to The

8:46

end user to the PC level. Like I mentioned, AI is not anything new.

8:50

AI has been around. It's just taken, you know, massive

8:52

amounts of hardware to really analyze

8:55

and to really process these things. The AI PC is going to do those things at

8:57

the client level for the individual users.

9:03

The other cool thing is because it's going

9:03

to be local, it's going to be your AI.

9:08

So an A IPC is gonna learn

9:08

the way you respond to emails.

9:12

It's going to look at the way

9:12

you do things, and it's gonna

9:14

start building that knowledge. And you're gonna have these

9:15

almost SLM small language

9:19

models able to run on the pc. And so an A IPC will be really

9:20

individual to the user, and you'll

9:23

be able to train it on yourself.

9:26

And so it'll become very much a

9:26

reflection of you and become your own ai.

9:31

This is like a bit of a digital

9:31

personal assistant it is Yeah So I

9:34

think the one that obviously people

9:34

recognize or associate with most is

9:37

co pilot and it really is a co pilot I

9:37

think that's an important distinction.

9:42

It's not an autopilot. It's a co pilot.

9:45

So based on what you were doing with

9:45

it It's gonna work Alongside you and

9:48

that that's one application and then

9:48

it'll also expand into all that other

9:52

software that we talked about depending

9:52

on the application It's going to

9:54

go into your security applications.

9:56

It's going to go into your adobe It's

9:56

going to go through all the different

9:59

applications that you're using and

9:59

it's going to be able to do it all

10:01

locally So the benefit there again

10:01

of what makes an aipc an aipc Is you

10:07

don't need to do those applications

10:07

or you don't have to run those

10:09

workloads on the edge or in the cloud.

10:11

So you're going to get latency improvements. You're going to get efficiencies.

10:15

It's going to save on power and it's

10:15

going to be more secure because it's

10:18

running on the device and not in a

10:18

data center on a cloud somewhere.

10:21

Very interesting. One of the things that every time I hear

10:22

somebody say it goes into your apps,

10:27

we have customers that instantly go

10:27

to, Whoa, I don't want you in my apps.

10:31

And so when you think about

10:31

the ethical side of A.

10:34

I. And the privacy related

10:34

issues that somebody doesn't

10:37

want the copilot of the A.

10:40

I. P. C. Poking around and like, how do

10:41

you have that conversation with

10:44

somebody that's got those concerns?

10:46

Yeah, it's a valid concern. I think it's maybe not why we're seeing

10:47

a hesitation in adoption, but certainly

10:51

why larger enterprise customers are

10:51

taking a long look at this, right?

10:56

This is something new, and we're

10:56

still really waiting for that

10:59

first big customer to say, yeah,

10:59

we're going to adopt this fully.

11:02

So it's a valid concern. I think when you look at it, though,

11:03

like any application, there's going

11:07

to be security in the background. And I would argue that it's actually going

11:08

to be More secure because AI is going to

11:12

be in all of your security applications. So we look at something like CrowdStrike

11:14

is a great use case that we're coming

11:17

across so far where any security

11:17

threats are actually more easily

11:22

detected and they're more efficient

11:22

by utilizing AI threat detection

11:26

efficiencies go up significantly.

11:29

By using AI. So absolutely privacy is always going to

11:30

be a concern, especially in healthcare,

11:34

in legal settings, you know, user

11:34

privacy is absolutely going to remain

11:38

an issue, but that's not anything new. And it's something that we've always

11:40

been able to address and work through.

11:42

I think the benefit of the

11:42

applications will far outweigh

11:46

any of the security risks.

11:47

Yeah, I think, you know, when

11:47

I think about security, it's always

11:50

this teeter totter, and there's

11:50

the feeling of security, and then

11:53

there's the reality of the security,

11:53

and feeling always trumps reality.

11:57

It's why people are more afraid to

11:57

fly than they are to drive, even

12:00

though it's statistically safer. Um, and you reflect back to any

12:01

new technology, people are a

12:05

little bit leery because they

12:05

want to see that it's safe to use.

12:08

So I, I totally get what you're

12:08

saying, and it makes a lot of sense.

12:11

You did say something interesting

12:11

around the use case of security.

12:14

Now, my experience with AI over the last

12:14

couple of years and talking to customers

12:18

is this is such a use case driven

12:18

dialogue and the importance of that.

12:22

So you're pointing it at an outcome,

12:22

not just using it because it's fun.

12:26

Maybe you can share with us a couple

12:26

of other use cases for an AI PC that

12:30

you're seeing in your customer base.

12:31

For sure. Um, I think when we look at, at

12:32

Intel as well, and in the kind of.

12:35

Let's call it commercial space.

12:37

It's not necessarily rocket science or

12:37

anything exciting, but it's cool little

12:40

things that we're starting to see. So we've kind of classified when we

12:41

talk about personas and use cases,

12:44

we're really looking at your knowledge

12:44

workers, your everyday PC users, some

12:49

of your more high end than mobile two

12:49

and one on the go type of professional,

12:53

and then what we're calling our extreme

12:53

users, which would fall into the kind

12:56

of, uh, Heavy data users, your creative

12:56

types, ones that would traditionally

13:00

use maybe more of a workstation. And then in terms of outcomes,

13:02

we're trying to classify it down to

13:05

productivity gains, time saving gains,

13:05

and then what we're calling, wow,

13:09

you know, the brand new stuff, and so

13:09

some of the use cases we're seeing in

13:12

commercial that we've started using at

13:12

Intel, there's a couple applications.

13:15

One is called iterate. ai. It's a software that helps

13:17

basically draft emails.

13:20

And again, going back to that,

13:20

your type of AI, and it's going

13:23

to be personalized to you. It would know how.

13:26

Camera would respond to this email. Whereas if you're using it, it's here's

13:27

how Aaron would respond to the email.

13:30

And it does a fairly decent job

13:30

of drafting long emails, complex

13:36

emails in your tone of voice. The other one that I think is already

13:37

fairly prevalent that a lot of people

13:40

are using things like meeting recaps. So every meeting that we have at Intel.

13:44

The video is uploaded, the transcription

13:44

is uploaded, and it spits out here's a

13:48

one page summary of what this meeting

13:48

was about, here are the action items,

13:52

who's responsible for them, when they

13:52

said that they're going to get them done.

13:55

So as opposed to me being, you know, note

13:55

taker in the meeting and having to send

13:58

it out and spend an hour after the meeting

13:58

summarizing, AI does it for me, it sends

14:02

it to all the pertinent people, and I'm

14:02

able to get that hour back in my day.

14:06

And I think the next step here is.

14:09

AI is not going to solve all the world's problems. We still need to figure out what

14:10

we're going to do with that.

14:13

And so when we say, Hey, I get an

14:13

hour back in my day, that's great.

14:16

Is that an hour that I use to

14:16

browse YouTube and talk by the

14:18

water cooler or do I actually do

14:18

something productive with that?

14:21

And I think where we

14:21

extrapolate this out is okay.

14:24

An hour back in your day. That's great.

14:26

Do that five days a week,

14:26

do that 52 weeks a year.

14:30

How much more can you get done in that time? And I'll take it one step further.

14:33

That's great in a corporate setting. Hey, I answered more emails.

14:36

I finished that PowerPoint faster. How about a doctor?

14:39

How about a doctor who's taking patient

14:39

notes and has to summarize patient visits?

14:43

If that doctor gets three hours back

14:43

a day that he doesn't need to spend

14:46

transcribing from his dictaphone and

14:46

he can see 20 more patients a day.

14:52

Extrapolate that out over 52 weeks. You're talking about hundreds of patients

14:54

a year That we are able to interesting

14:58

help enable those outcomes and again

14:58

going back to what we said at the

15:01

beginning technology enabling people's

15:01

lives Think about wait times in hospitals

15:05

if we can shrink that because doctors

15:05

are more efficient That's an outcome.

15:09

That's something that's

15:09

really cool that I think AI in

15:12

real life is going to cause. I love that.

15:14

We've come across so many customers

15:14

that have the need for high compute

15:18

power, but also the need for that

15:18

personalized AI, whether it's architecture

15:21

firms, the legal industry, there's

15:21

so many different types of use cases.

15:25

One of the things I found interesting, you mentioned. About the personal assistant, if you

15:27

will, that sits on the device you're

15:31

calling an AI PC is the idea of

15:31

putting it in tone and understanding

15:36

reactions because some of the feedback

15:36

we've had so far to date is, oh,

15:40

it's fine, but it feels very plastic.

15:42

The way that I get a feedback

15:42

from the large language modeled

15:45

isn't sounding authentic. And so if you can get past

15:47

that, I think that just.

15:50

Super drives up adoption. Are you finding your team using it?

15:53

Like, how are you personally using it? What's your favorite use

15:55

case outside of the meetings?

15:57

Like what's, what's got you excited?

15:59

For me, Copilot and being

15:59

able to access large language models,

16:03

but also just vast amounts of data.

16:05

So, uh, another example, you

16:05

know, Intel's a huge corporation,

16:08

Softchoice is a big corporation. We have all kinds of documents,

16:09

SharePoint sites, HR documents,

16:14

different policies using AI.

16:17

to quickly find answers to fairly simple

16:17

questions, you know, whether it's parsing

16:23

through an HR document of, hey, how many

16:23

vacation days do I get instead of the

16:26

individual having to find the document,

16:26

look through it, you know, control F,

16:31

they hit a button on their PC, and again,

16:31

this will be another thing that makes

16:34

an AI PC and AI PCs, it's going to come

16:34

with a copilot button going forward.

16:38

So you'll hit it. The co pilot button and say, Hey,

16:39

how many, uh, vacation days am

16:42

I entitled to at soft choice? And it will know, Hey, Aaron's

16:44

been at software for 20 years.

16:48

So he's entitled to X number of weeks.

16:50

It'll look in the system and say,

16:50

Hey, he's taken four already.

16:54

He's got X number of days

16:54

left in his vacation time.

16:56

And so again, it's simple uses, but

16:56

these are things that will just run in

17:01

the background very smoothly that you

17:01

won't even have to start thinking about.

17:04

It's going to become. Second nature.

17:06

It's very cool and new right now. But I think about my daughter

17:08

who's seven years old.

17:11

This is just going to be innate for her. She's going to grow up in a

17:13

world where this is just the way things have always been.

17:17

And again, you think about

17:17

Intel technology creations.

17:20

USB is just everywhere now. At one point, USB was cool and new.

17:24

Eventually AI will get to this. Oh yeah, it's just, it's

17:26

just AI. It's just, it's just there.

17:29

I'm still waiting for the day. Like as these more of these personal

17:30

assistant type of technologies

17:33

come out, we start getting into

17:33

naming them and having different.

17:36

personality types, kind of like

17:36

the series and the Alexa's and like

17:40

all those different elements of it.

17:41

Yeah. Going back to the tone that you

17:42

mentioned, we like to use the example,

17:45

everyone remembers Clippy, the

17:45

little paper clip in Microsoft Word.

17:48

That was AI, right? It was a rudimentary version, but that

17:50

was, that was the first AI assistant.

17:54

And what you're going to be able to see

17:54

now on the AIPC, again, without getting

17:58

super technical is something called a

17:58

reg model, which will basically bridge.

18:03

An online large language model

18:03

with your personal device.

18:08

And so you'll be able to access large

18:08

language models, pull answers from,

18:13

from a cloud, from, from the internet,

18:13

from a public domain and cross

18:17

reference it with your personal AI and

18:17

get tone response and things that are

18:23

personalized and applicable to you.

18:26

Hey there, IT leaders. Ready to accelerate your AI journey?

18:30

In today's fast paced world, enterprises

18:30

that embrace AI are pulling ahead.

18:34

With Intel's AI expertise, unmatched

18:34

partner ecosystem, and best in class

18:39

hardware and software, you can deliver

18:39

AI results everywhere you need them.

18:43

Where might you need AI? Whether it's in the cloud,

18:44

data center, edge, or on the

18:48

PC, AI needs to be everywhere.

18:50

Intel makes it all happen with

18:50

comprehensive solutions like data science

18:54

tools, development frameworks, edge, data

18:54

center, and cloud solutions, plus AI PCs

19:00

powered by Intel Core Ultra processors.

19:03

Why partner with Softchoice? Moving forward with AI is best

19:05

done with an IT solutions provider.

19:09

Softchoice offers seamless

19:09

integration, expert assistance,

19:12

and extensive industry knowledge. Softchoice. We're committed to customer

19:14

satisfaction and helping you leverage

19:17

Intel's cutting edge technologies

19:17

to drive your business forward.

19:21

Deliver all the AI results you

19:21

need, everywhere you need them,

19:24

with Intel and Softchoice together.

19:26

Contact a Softchoice representative

19:26

today or visit Softchoice.

19:30

com to get started. Now, back to the show.

19:34

So I know, uh, last

19:34

year we actually had a couple

19:36

of podcast discussions around

19:36

the democratization of AI.

19:40

And I think this is

19:40

obviously in support of that.

19:43

However, I'd be curious how Intel is

19:43

ensuring that these capabilities are

19:47

accessible for all users, not just

19:47

those that have the high end hardware.

19:51

Um, so maybe talk a little bit about that. For sure.

19:54

I think firstly, it's

19:54

important to recognize the awareness

19:56

of what AI is and where it runs. There's been a lot of talk that, oh, in

19:58

order to do AI, you need to have a GPU.

20:04

And that's not necessarily true. At Intel, our tagline is bringing

20:05

AI everywhere and for everyone.

20:09

And so it's about making sure that

20:09

you have the right hardware and

20:12

running the right application on the

20:12

right workload on the right chiplet.

20:15

So. Our focus is making sure that

20:16

there's awareness around the fact

20:19

that AI does not equal a GPU.

20:22

Depending on the application, it's going to run either on the CPU, it's going to run on the GPU,

20:24

or it's going to run on the MPU.

20:27

And Intel really is one of the only

20:27

company that can offer a solution

20:31

on what we call the AI Continuum. And so that's in the cloud.

20:35

That's in the data center. That's at the edge and now it's

20:36

on the PC and on the device.

20:39

So regardless of what workload you're

20:39

looking to run, we have a solution

20:43

for you and it's not about shoehorning

20:43

or pigeonholing a customer into, Oh,

20:47

you need to buy this solution again,

20:47

going back to right customer outcomes.

20:50

It's what do you want? What does success look like for you?

20:53

Okay, cool. Here's what we can do and here's how

20:54

we can enable it through software.

20:57

I think that's the other part

20:57

that gets lost sometimes is

20:59

it's not just a hardware play. It's the two of them together.

21:01

It's hardware and software. So, hey, here's the hardware that

21:02

Intel has that will enable this.

21:06

And here's the software

21:06

optimizations that we can help

21:09

you with to get to that solution. And so yet we use the

21:10

term democratizing AI.

21:13

I think Intel has always been a proponent

21:13

of open ecosystems, again, without getting

21:18

super nerdy, but, uh, you know, the X

21:18

86 architecture is an Open architecture.

21:23

It is literally the defining

21:23

architecture of our industry.

21:26

It is open in the long run.

21:29

AI is going to be no different. Rising tides are going to lift all boats.

21:32

And if we're going to see these

21:32

futuristic, great outcomes that AI is

21:36

going to change the world, it's not

21:36

going to happen in a silo and it's not

21:39

going to happen in a closed ecosystem. It's going to happen in an open

21:40

democratized ecosystem where everybody

21:44

gets access regardless of the workload,

21:44

regardless of the hardware that they're

21:47

running. It's totally true. And it's one of the things I find special

21:49

about Intel is just how closely they do

21:53

work with other technology providers.

21:55

It's such a cooperative

21:55

model that Intel has built.

21:59

And you just mentioned everything

21:59

from end user devices to

22:02

cloud and Intel is everywhere. And so I love that concept of

22:04

being able to put the right

22:07

process into the right environment. processor in order to

22:10

get the right outcome.

22:12

So you did three wins. I did three rights. There you go.

22:15

There you go. Right.

22:16

Yeah. We play well with everybody in the sandbox. Right. And that's again, what I love

22:18

about Intel from both a historical

22:22

standpoint, as well as where we're

22:22

going in the future, it really is

22:26

about bringing everybody along with us,

22:26

making sure that it's win, win, win.

22:30

And, you know, we see those

22:30

outcomes benefiting everybody.

22:34

Love that. So on the conversation of

22:35

cost, I know life is expensive.

22:38

You mentioned you have a young one. I have two that are not so young.

22:41

They're 19 and 22 and they cannot leave

22:41

the house because life is too expensive.

22:46

If we look at technology,

22:46

it's no different.

22:48

There's a real price increase in scarcity

22:48

of GPUs and they're high costs and

22:53

there's always mining and logistic issues.

22:56

So I'm just curious how Intel is

22:56

planning to address that market challenge

23:00

while still supporting the high demand

23:00

that's clearly coming around AI.

23:04

So, I mean, the good news is with our

23:05

new Meteor Lake chipsets,

23:05

which were the 14th gen chipsets

23:09

that launched last December, we're

23:09

not seeing massive price increases.

23:12

It's inflationary, small price increases.

23:15

So from that level, there's not too

23:15

much to be concerned about right now.

23:18

Absolutely. You're right around the GPU scarcity. There's going to be GPU

23:20

haves and GPU have nots.

23:24

Geographically, there are going to be

23:24

countries that just eat up the supply

23:28

of GPUs, which is another reason why

23:28

we believe in an open supply chain

23:32

in a sustainable supply chain and

23:32

why we want to drive the message that

23:36

AI is CPU, GPU and MPU altogether.

23:40

The reality is that a lot

23:40

of AI is inference based.

23:44

So we use the analogy of

23:44

take a weather forecast.

23:46

If you're a meteorologist,

23:46

you need one person.

23:50

That is a trained meteorologist. You need one person to

23:52

train that AI model.

23:54

Maybe it's on a GPU. But then the rest of us, we just

23:56

need to look at the weather forecast

23:58

on our phones and consume that. That would be the CPU, right?

24:01

Most of the users aren't training models.

24:04

They're using the models. Gotcha. And so that's where we want to,

24:06

again, democratize AI around.

24:10

It's happens at an

24:10

inference level on the CPU.

24:13

Um, everybody can have access to

24:13

it depending on their devices.

24:17

And so you're absolutely right. GPUs are going to be scarce.

24:21

They're going to be centralized

24:21

in countries that have in

24:24

countries that have not. And unfortunately, Canada is a

24:25

have not when it comes to GPUs.

24:28

And so why we're really focused

24:28

on that awareness message of.

24:33

AI is more than just a GPU. Now, I'm now you got me curious.

24:36

Why is Canada have not, uh, comes

24:36

to population comes to, you know,

24:40

economically we can get into all the

24:40

geopolitical reasons behind that.

24:45

Sure. Um, I think where Canada is going to

24:45

see the benefit and where Intel is

24:49

helping to drive this is from a North

24:49

American standpoint, you know, we're

24:52

investing heavily in our supply chain

24:52

in the United States, building foundries

24:58

so that we can create these chips. locally in North America without

24:59

going into all the geopolitical,

25:03

you know, reasonings behind it. There's a need to have

25:05

a diverse supply chain.

25:08

I think when you look at 2020, if you

25:08

remember trying to buy a fridge or any

25:13

type of electronic device at the outset of

25:13

the pandemic, you couldn't get it because

25:17

the reality is now chips are everywhere.

25:19

Chips aren't just a PC play. You've got chips in your dishwasher.

25:22

Chips in your fridge. Think about a car.

25:25

There's something like 70

25:25

chips in a car nowadays.

25:28

And so this isn't just

25:28

about making PCs accessible.

25:31

This is about making technology accessible. And that's why we're investing in, you

25:33

know, fabs and, and production facilities

25:37

in Phoenix and Ohio so that we can create.

25:41

A more sustained supply chain for,

25:41

for users in Canada will benefit

25:44

from that, but you know, the GPUs

25:44

will be eaten up by the regular

25:48

suspects on the geopolitical stage.

25:51

Understood. Well, well handled question, my friend.

25:54

I like that. So as we wrap up today, I did want to get

25:54

a little bit personal insight from you.

26:00

I'm gonna do like a hypothetical with you. So imagine you're on stage a year from

26:01

now, maybe two years from now, and you're

26:06

talking to a group of peers and customers.

26:09

What do you foresee the one or two

26:09

things you're going to be talking

26:11

about that you're super excited

26:11

about a couple of years from now?

26:14

Yeah, that's,

26:15

that's a great question. I'll reframe it a little bit first around

26:17

the timeframe, because I think it's not

26:20

necessarily going to be a year from now. The one example that I like to

26:22

give is when we talk about how

26:26

long it's going to take AI. So I'll give you a couple of dates.

26:29

January 1st, 1983 was the day that

26:29

the internet first went online.

26:35

September 4th, 1998.

26:38

was the day that Google

26:38

launched, effectively bringing

26:41

the internet to the masses. So if you do the quick math

26:43

on that, that's a 16 year gap.

26:46

And so we don't think it's going

26:46

to take AI that long to get to

26:49

that point, but just to reframe,

26:49

Hey, AI is not going to happen.

26:52

overnight. I know it feels like it. We all kind of roll our eyes

26:54

at another AI presentation.

26:58

The reality is, is that this isn't

26:58

a flash in the pan and it's going

27:01

to take years to figure out what

27:01

those use cases really become.

27:05

Going back to the security question,

27:05

seeing where this goes, how this develops.

27:09

And so, you know, I gave a talk

27:09

recently where I laid that out.

27:13

If you take the same timeframe, that

27:13

16 year timeframe from when we launched

27:16

our AI PCs on December 14th, last year.

27:19

That gets us into August, 2039. So you and I can sit down in

27:22

2039 and we'll regroup and

27:26

debrief on how AI turned out. But getting back to your question, I think

27:27

when I look at where I want to see this

27:32

in a couple of years, I think it's really

27:32

going to change critical thinking skills.

27:36

And again, I take it back to my

27:36

daughter and, you know, the next

27:40

generation of people coming up. If we go back, you know, five,

27:41

10, 15 years, everyone said,

27:44

Oh, you got to get into coding. Right? The way the curriculum changed and

27:46

the way that we as a society approach

27:49

things, that was a big advent. I think the next one is going

27:52

to be prompt engineering.

27:56

Because the way that you use AI

27:56

is going to be dependent on what

27:59

you feed it and how you prompt it. You know, AI is only as good as

28:01

the applications you use and only

28:04

as good as the user using it.

28:06

And so when you take a philosophical

28:06

stance back on this, people are going

28:11

to have to change the way they think. And that's a big shift right

28:13

at a very granular level.

28:18

We need to be more focused

28:18

on prompt engineering.

28:21

We need to be more critical in our

28:21

thinking to get those outcomes.

28:26

Another thing I'll mention here is

28:26

AI is not going to replace people.

28:30

People who can use AI will

28:30

replace people who can't use AI.

28:34

And that really gets to that,

28:34

that prompt engineering, right?

28:38

And so again, from a philosophical

28:38

stance, I think it's really cool.

28:42

when you think about how is society

28:42

going to improve because we have

28:48

an entire generation of people who

28:48

think more critically, who look at

28:54

things from a different lens and

28:54

approach a problem differently.

28:58

These are fundamental skill sets that I

28:58

think, again, are beneficial to society.

29:02

And I take it right back to the

29:02

reason why I love Intel is it's

29:05

about improving people's lives,

29:05

not about improving the technology,

29:08

not about improving the business. If we improve people's lives, how do

29:10

people then extrapolate that out and

29:14

improve their community, their city, their

29:14

country, you know, the overall society.

29:20

And so I think that's where, again,

29:20

we believe AI is, is for good.

29:24

There's always going to be bad actors.

29:27

There's always going to be the threat

29:27

of, of negative things happening with

29:30

it, but technology and AI at the end

29:30

of the day, I believe are always going

29:35

to be positive beneficial forces.

29:38

And so yeah. The cool part for me is, like I said,

29:39

how is society going to benefit from

29:44

an entire generation of people that

29:44

think differently about a problem?

29:49

Right? Like if you take a, if you take

29:49

a positive approach to that,

29:52

that's a cool thought, right? It's easy to be pessimistic.

29:55

It's easy to be, oh, everything's

29:55

going to be terrible.

29:58

But you know what? What if it's not? What if we all decide that,

29:59

hey, this is going to be really

30:01

cool and really awesome and we

30:02

choose to believe in that instead? That's a win for you, a win for

30:05

me, and a win for everyone else.

30:07

That's triple win, right? That's what we're shooting for.

30:09

The triple win. I appreciate you taking the time today,

30:10

Cameron, and by my math, by Season 21

30:15

of The Catalyst, we will have you back

30:15

to talk about how AI has progressed.

30:19

Absolutely. We'll do the before and after. Thanks again for having me

30:21

here, and this is great. Maybe we'll get a couple

30:23

touchpoints in between now and 2039.

30:27

Absolutely. Look

30:28

forward to chatting then

30:28

as well. Alright, take care of yourself my friend. Likewise.

30:33

NPUs are set to revolutionize

30:33

personal computing, making AI

30:37

more accessible and efficient. They could empower business and

30:39

people to achieve unprecedented

30:42

levels of innovation and productivity. But as we embrace these exciting

30:45

advancements, it's critical to

30:48

ensure that these capabilities

30:48

are available to everyone.

30:51

Not just those with the latest technology. And I love the conversation

30:54

with Cameron for two reasons.

30:57

One, he introduced me to a

30:57

wonderful new saying, win, win, win.

31:00

We win, you win, and the industry wins.

31:03

And secondly, AI is not

31:03

going to replace people.

31:06

People with AI will replace people.

31:09

Well, that's it for now. If you enjoyed this episode, please

31:11

leave us a review on Apple podcasts.

31:15

Join us again in two weeks for

31:15

another episode of The Catalyst.

31:19

The Catalyst is brought to you by

31:19

Softchoice, a leading North American

31:22

technology solution provider. Written and produced by Angela

31:24

Cope, Felipe Dimas, and Braden

31:27

Banks, in partnership with

31:27

Pilgrim Content Marketing.

31:32

This episode is brought to you by Intel. Deliver all the AI results you

31:35

need, everywhere you need them,

31:38

with Intel and Softchoice together.

31:40

Contact a Softchoice representative

31:40

today to learn about Intel's

31:44

comprehensive AI solutions and

31:44

how we can unlock them for you.

31:47

Or visit Softchoice. com to get started.