0:01

This episode is brought to you by Shopify.

0:04

Forget the frustration of picking commerce

0:06

platforms when you switch your business

0:08

to Shopify, the global commerce platform

0:10

that supercharges your selling wherever you

0:13

sell. With Shopify, you'll harness the

0:15

same intuitive features, trusted apps, and

0:17

powerful analytics used by the world's

0:19

leading brands. Sign up today for

0:21

your $1 per month

0:24

trial period at shopify.com/tech,

0:26

all lowercase. That's shopify.com/tech.

0:31

Your brain needs support and New Ali

0:33

brainy Jews are a delightful way to

0:35

take care of your cognitive health. Made

0:38

with scientifically backed ingredients like Thai ginger,

0:40

elephantine and caffeine. briny to support healthy

0:42

brain function and help you find your

0:44

focus, stage show or get energized. Be

0:46

kind to your mind and get these

0:49

new trope of choose at all he.com

0:51

That's olly.com These statements have not been

0:53

evaluated by the Food and Drug Administration.

0:55

This product is not intended to diagnose,

0:57

treat, cure or prevent. Any disease,

6:00

was like an exceptional, like much

6:02

higher than usual, because everybody

6:04

wants to live with AI, like

6:06

what is possible with AI. We

6:09

got 60 great ideas

6:11

submitted through this hackathon.

6:14

It was a 24 hour hackathon. There

6:18

was a great energy, a lot of participation,

6:20

60 ideas came through, and 10 of

6:23

them were actively pursuing as

6:25

a further implementation. You

6:28

know, some of those are like, you know, things

6:31

like, you know,

6:34

we have, you know, obviously, you said this

6:36

like, we have a 100 year old company,

6:38

we have been acquiring a lot of companies,

6:40

like another, as we do this, like, you

6:42

know, there's a lot of technical debt that

6:45

we have built up over time, a lot

6:47

of technology stacks. So with that,

6:49

like, you know, it's very hard for the

6:51

developers to like, you know, look into those,

6:53

you know, tech debt areas,

6:55

and, you know, nobody wants to be spending,

6:58

like not fixing those, you know,

7:00

security issues, that, you know, technical,

7:02

rather they want folks more on

7:05

new feature development, right? So the

7:07

team, one team came up with

7:09

this really cool tool that

7:12

will identify all these technical debt

7:14

areas and automatically fixes and also

7:16

deploy. That's very cool. Like, you

7:18

know, the way I've been, you

7:21

know, the idea

7:23

was good, and you know, that seems very promising.

7:25

It's going to save a lot of time in

7:27

the, you know, development lifecycle.

7:31

We're now taking that into like, you know,

7:33

implementation phase. So I'm really hoping like, you

7:35

know, that will get us to that 2x

7:38

target that we have for, you know, 2024. Along

7:42

those lines, like, you know, the, it's

7:46

mostly the engineering product with the

7:48

hackathon. So like, you know, the not

7:50

a lot of other use cases that we

7:53

have derived through this hackathon, but as I

7:55

mentioned, there are other initiatives that

7:57

are going on across other functional. way

7:59

it is. Super cool.

8:02

Yeah. And I love that you have that

8:04

goal of, you know, two X productivity, um,

8:06

that you're kind of like focusing and driving

8:08

towards, I think it's really important when we

8:10

make, you know, like, for example, an AI,

8:12

um, mandate or like

8:15

any sort of goal that we like really state

8:17

what we're trying to do. So I love that

8:19

you guys are driving towards that. Um, and kind

8:21

of along those lines, another question I would love

8:23

to ask you about is, you know, how have

8:26

you used AI to reduce the time that it

8:28

takes to actually develop new fonts over at Monotype?

8:32

Great question. Um, we actually

8:34

truly see GenA as an

8:36

opportunity to enhance the productivity

8:38

of the type designers, uh,

8:41

with both powering tools like search

8:44

and discovery, uh, driving

8:46

overall efficiency in the business. Um,

8:49

Monotype has built, uh, AI

8:51

enabled tools that have already seen, uh,

8:54

success facilitating exploration in areas

8:56

such as like, you know,

8:58

the suggesting ideal font pairing

9:01

and, um, adding in

9:03

discovery of the new fonts

9:05

through natural language, search, and

9:07

recommendations. So we recently launched

9:09

something like contextual recommendation, meaning

9:11

if you, uh,

9:14

you know, it's almost like a chat deep, if you like, and if you,

9:17

um, type in the context of what

9:19

you want, uh, we will recommend

9:21

like, you know, hey, for this context, this font

9:23

works well. So that has been like, you know,

9:26

the killer feature. A lot of people love

9:28

it. Um, you know, recently at the

9:31

Adobe max, uh, so many

9:33

people came to our booth and appreciated that

9:35

feature as well. I

9:37

love that because I know like for

9:39

myself, you know, running my podcast, whether

9:41

it's me or my designers or my

9:44

editors, we're always like looking for new

9:46

fonts for like thumbnails or for like

9:48

websites or posts and for content and

9:50

running AI box, um, this

9:53

is a huge discussion that I'm always having with

9:55

our designers is, you know, what fonts we're using

9:57

and what like message does a font portray? I.

10:00

love that you guys have that tool. This

10:02

is definitely something I could use. I think

10:04

anyone that has designed anything in their life

10:06

knows the feeling of, you know, you're on

10:08

like, Canva or Photoshop or Illustrator or whatever,

10:10

and you, you type something

10:12

out and you're like, what font

10:14

should I use for this? So you highlight it,

10:16

then you go over to like the font tab,

10:19

and you just hit the down arrow over and

10:21

over again, is it like switches what the font

10:23

looks like, you're just trying to find like the

10:25

perfect font, and it takes forever. So I think

10:28

that AI tools genius. I think that's amazing. Yeah,

10:30

we're trying to eliminate that font when you in

10:32

every single application possible. I'm

10:34

so grateful that you're you are doing

10:37

the world a massive service. Honestly, that's

10:39

awesome. So I think that's gonna be

10:41

super cool. You know, and

10:43

you guys have 1000s and 1000s of fonts. So

10:45

this makes a ton of sense. You can find the perfect

10:47

font using AI. That's amazing. Something

10:49

else I'd be curious talking about kind

10:52

of like AI and fonts. Recently, chat

10:54

GPT has unveiled GPT four Oh, and

10:56

I, I did a whole

10:58

podcast episode about this covering a bunch of

11:00

the features coming in some of the use

11:02

cases, a really interesting one that I saw

11:04

and actually mentioned monotype in the episode because

11:08

it made me think of you guys. And I'm wondering, like, you

11:10

know, where your thoughts are at on it. Someone

11:12

showed a use case where they went to

11:14

GPT four Oh, and I think it's like

11:16

Dolly on there or whatever. But they they

11:18

were like, Hey, like I need a futuristic

11:20

space kind of font for this specific thing.

11:22

And it generated like an image, like obviously,

11:24

it's not like an actual real font or

11:26

anything. But it generated an image with like,

11:29

letters for like an actual,

11:31

like futuristic space font. I'm wondering, are you

11:33

guys able to use AI to actually help

11:36

you generate fonts? And if so, like, what

11:38

are some of the challenges you're facing around

11:40

that? Yeah, so this is

11:42

very interesting problem for us and many others

11:44

as well. Right. So the

11:46

when you think about AI generated

11:49

fonts, though, you know, whatever

11:52

is available today, even GPT four

11:54

Oh, it could only

11:56

generate images. And even those images, I

11:59

don't think you know, with my own

12:01

experience, I've been experimenting a lot from

12:03

the beginning, even with the recent four-wall.

12:08

Those are not usable at all, in my

12:10

opinion, right? Even the images, right? You

12:13

know, there are a lot

12:15

of initiatives like that. You

12:17

know, Adobe has, you

12:20

know, been working on that, like

12:22

many others, trying to put

12:24

the text onto the images,

12:26

and which is not even

12:28

close to being perfect. Yeah.

12:31

And of course, you know, being

12:33

the world's largest font company, we

12:36

have been investing a lot into generating

12:39

AF fonts, the real fonts. And

12:43

that's really quite

12:45

a challenging task for sure. You

12:49

know, we have a very high

12:51

hope that, you know, I just got update

12:55

this morning from our team, like, you know, it

12:57

looks like they are making a great progress, and

12:59

we're going to achieve our Q2 and Q3 goals,

13:02

which means we're

13:04

getting closer to

13:07

that goal to create like another

13:09

high quality. So if something comes

13:11

out of Maratype Studio, that

13:14

means it has to be really high

13:16

quality, like, you know, the fonts that, you

13:18

know, everybody appreciates, like,

13:20

you know, we can't release any

13:22

low quality fonts, you know,

13:25

because Maratype meant for like, you know, very high

13:27

quality fonts, right? So

13:31

with that, you

13:33

know, there are a few things we

13:35

need to consider as like, you know, there

13:38

are a lot of concerns

13:41

around how

13:44

are we going to train the debt, like, you

13:46

know, these models, like, you know, we have largest

13:48

library available, like, you know,

13:50

we are assuring our partners, whoever is like,

13:53

you know, working with us, like, we're not

13:55

going to leverage any other

13:57

fonts to train our models. You

14:00

know, we have a

14:02

very strict guidelines, we published even guidelines to

14:04

like, you know, everybody to be like, you

14:06

know, what Marathat is gonna follow

14:09

when we are, you know, how ethically

14:11

we're gonna be when we're training our

14:13

models and, you know,

14:15

unbiased and things like that. You

14:19

know, bottom

14:22

line here is like, no, we're gonna get

14:24

there soon, but not there yet. Yeah,

14:27

yeah. No, for sure. I

14:30

think when we get there, I

14:32

think, you know, the first step

14:34

is obviously towards that contextual recommendation

14:36

that I have mentioned, like another

14:38

if someone needs a text in

14:41

the image, that is evident, right? Everybody

14:43

needs a text in the image. We

14:47

probably as an expert will be

14:49

able to help with anybody

14:51

who is in that image creation business.

14:55

So we can partner with them

14:57

to provide this text generation capabilities

15:01

alongside their models. So that's our

15:03

goal. So

15:06

as well as everybody else who is

15:08

in this business. Yeah, I

15:10

love that. I think that's such a fascinating future.

15:12

I mean, I could see a future I know this

15:14

is tons of attack and this

15:16

is not necessarily right around the corner, but where,

15:19

you know, I want like a font for a

15:21

specific thing and I describe a font and then

15:23

monotype uses like an AI model to help generate

15:25

the perfect font, probably trained from your own data

15:27

set of your fonts and it's kind of using

15:30

variations of stuff that you've created. But like you

15:32

mentioned, the key here is like, monotype

15:34

is known for like Times New

15:36

Romans, these iconic fonts, like really,

15:39

really high quality stuff. So making sure

15:41

that whatever you create is up to

15:43

that standard, obviously is the big challenge.

15:45

But I think it could be a

15:47

really fascinating future. I know that,

15:49

you know, talking about some of these interesting concepts and

15:52

this R&D and working in this

15:54

space, I know you've talked about doing some

15:56

collaborations with universities. I wonder if you can

15:58

discuss some of the nature. and

16:00

the goals of your collaborations that you currently have

16:02

going on? Yeah, because

16:04

the generating a funds is

16:07

such a difficult task.

16:11

We thought reaching out to some

16:13

of the top large universities across

16:15

the globe would

16:17

be beneficial. We are actually working with

16:20

multiple universities across

16:23

different regions. So it's not only like Latin

16:25

funds that we need to create, we have

16:27

to create the word funds like another. Creating

16:30

a Latin versus creating a Japanese and

16:32

Chinese is quite different. It

16:35

requires different skill set. So

16:38

for that reason, we have been

16:40

working with

16:42

multiple universities like another, as I

16:44

stated earlier, we do

16:47

follow this framework called technology

16:49

readiness framework, which

16:51

NASA uses, ERL, and

16:55

it has eight levels. If

16:57

you think about it, one, two, three is like, those

17:00

are the experiments that you do

17:02

within the labs. And three to

17:04

five is where you do like build MVPs

17:06

and five to eight is basically where you

17:08

scale up your part. With

17:11

that framework in mind, I would

17:13

say probably we're at between one

17:15

to three, some universities are like,

17:17

still discovering, like coming up with

17:19

their own methodologies, approaches

17:22

in doing this, some are a little

17:24

bit further out, like where we can

17:26

take and implement other MVP and

17:29

put it out in front of some of

17:31

the designers. We're

17:33

very excited, like we're fortunate to

17:35

find some of these universities to

17:38

work on this project alongside our

17:40

internships. That's

17:42

super, super exciting. You

17:44

guys are doing some really innovative stuff. I'm excited to see

17:46

what comes out of this. One

17:49

thing amazing about yourself is you are

17:51

an investor in AI. This is an

17:53

area you focus on. You're an investor

17:55

in AI box, my software company. I'm

17:57

curious from your perspective and kind of

17:59

your advantage. that

30:00

Dragon. So simple. Anybody, you know,

30:02

without even like, you know, much

30:05

of the technical knowledge, they could

30:07

easily use the platform. I

30:09

would like to see more of those. Right?

30:13

So I think that's where, you know,

30:16

there's entrepreneurs will shine. I

30:19

love it. Yeah. Amazing advice. Venkat, if people

30:21

are interested in keeping up with you and

30:23

your journey, what you guys are working on, Monotype,

30:26

what's the best place for them to do that?

30:30

Obviously monotype.com. You

30:32

know, we update monotype.com all

30:34

the time. Like another best place

30:36

to visit or reach out

30:39

to me on LinkedIn. Amazing.

30:42

Okay. We'll do, and I will leave

30:44

links in the show notes to the

30:46

listener to monotype.com, to Venkat's LinkedIn if

30:49

you're interested in connecting with him. Thank you so

30:51

much for coming on to the listeners. Thank you

30:53

so much for tuning in to the podcast today.

30:55

If you enjoyed the episode, make sure to review

30:57

us wherever you get your podcasts. And I hope

30:59

that you all have an amazing rest of your

31:01

day.