0:00

Good morning! Very daily show. I'm

0:02

your prime and and I'm tell

0:04

me how today we have an

0:07

insane bonus episode for you. We

0:09

are interviewing Mark Zuckerberg. Let's ride.

0:15

Mark and welcome to the show! A

0:18

good to see You guys can see

0:20

as well or at let us jump

0:22

right in. On Tuesday you posted a

0:24

video to Instagram where you reviewed the

0:26

Apple Vision Pro a competing headset and

0:28

compared it to your own. The Quest

0:31

Three. What we're all wondering is why

0:33

did you decide to make that video?

0:35

You know, having a C O publicly

0:37

review with a rivals product isn't exactly

0:39

typical. You know the seal? Papa John's

0:41

isn't going over to dominoes and eating

0:44

their pizza. Man,

0:47

Well I mean look in in my

0:49

household, it's basically like my wife Priscilla

0:51

screaming at me to like It's like

0:53

markham come to sleep and I'm like

0:55

no, there's someone wrong on the internet.

0:57

Gotta gotta do something about this. So.

1:02

The i mean that's that's kind of the

1:04

the daily but know for for real though

1:06

I'm in a you know basically. going

1:09

into it but I hadn't tried the vision

1:11

prefer while right i'm an apple wasn't like

1:13

in a rush to to give me one

1:15

but going into that. I

1:17

sort of expected. That.

1:19

For a device that airing their device three thousand

1:21

five hundred dollars ours is Five hundred dollars I

1:24

I I kind of thought because there's would be

1:26

better but a bunch of things than ours but

1:28

ours would still be the better value for most

1:30

people given a seven times less expensive. but when

1:32

I actually got a chance to try there's out.

1:35

As wow. It's. Actually,

1:37

ours isn't just the better value, it's the better product

1:39

are you know? And and I think that you know

1:42

I just saw the media coverage around this. Was.

1:44

Just sort of breathlessly assuming that you know

1:46

I think because it's Apple and because it

1:49

was such an expensive product like it must

1:51

be you know the better one in quality.

1:54

Even. Though a lot of people are saying hey no

1:56

huge prego by quest three because it's a it's a

1:58

better price and I'll I just wanted to. But

2:00

my flag down and say hey no you don't discuss

2:02

this thing costs seven times more. Doesn't mean it's better

2:04

if you actually go, use case by use case on

2:07

this thing Quest Three stacks up really well and I'm

2:09

just really proud of the work that we did. so

2:11

I just wanted to put that out there and say

2:13

that. It was definitely you. You

2:15

fly saddam a little bed. brought the

2:17

receipts on some level though you do

2:20

have to be a little please that

2:22

Apple is getting into the space because

2:24

what everyone is talking about headsets Now

2:26

they weren't as much this time last

2:28

year. Do you think you're benefitting from

2:31

this Apple halo effect? What?

2:34

I mean I think competition is doesn't have been

2:36

good for people graduated. So I mean I'm a

2:38

very competitive person. Seven having a a competitor who's

2:40

doing a good job. Was to

2:43

motivate our seventh and I'm it's good for

2:45

the teams recruit more of a sense of

2:47

urgency. So I do think that people really

2:49

when when there's competition night and I I

2:51

don't think that that's that's true. And yeah,

2:53

but I do think that because people sort

2:55

of assume that if Apple comes into space

2:57

their gives you a good job that it's

2:59

a case your maybe I hadn't considered wearing

3:01

one of these headsets and now I will

3:03

on and. Your. Hey, the quest there is

3:05

actually better and it's five hundred dollars so I'll go

3:07

for that one. I didn't. A lot of people are

3:09

going to buy quest because of that so that's the

3:11

to be going to be a pretty exciting thing but

3:13

all I mean we're in this for the long term.

3:16

I think it's gonna be. Into. Their can

3:18

be a lot of ups and downs in ah,

3:20

just ah, so I'm not. I don't worry too

3:22

much about any given year. I'd try to do

3:24

the best stuff that we can. I mean, speaking

3:27

of a given year I'd do. I'd take us

3:29

back to two thousand and seven for a second.

3:31

There is this infamous Steve Ballmer interview where he

3:33

literally last at the I phone after it's introduction

3:35

because he thought it's way too expensive it didn't

3:38

have a keyboard das and he said it would

3:40

never really gain market share. Are you at all

3:42

nervous that when you start comparing a pride from

3:44

Apple that is less than a month old to

3:47

your product? Which has ten plus years of

3:49

development behind it that you might be

3:51

making the same laughing at the i

3:53

phone my Steve Ballmer type mistake. I

3:57

don't think so. I mean look for so I'm not laughing

3:59

at them. I take apple seriously. I think it's

4:01

like a good company. They they do good work.

4:03

I mean, I'm just laying out the facts as

4:05

I see them today of I'm not saying that

4:07

they're not going to improve, but hey, we're going

4:10

to improve to India were company that moves pretty

4:12

quickly. sigh I actually think we'll probably grow faster

4:14

arms, but. I. Do think that

4:16

there's this whole vibe with Apple

4:18

fanboys, which I find kind of

4:20

a little. Funny and it where

4:22

it's like a lot of people. Skip said

4:24

if you dare to question if Apple is

4:27

going to lead in a new space and.

4:29

I. Don't know. I'm and I just think if you

4:32

look at the history of computing. You every

4:34

major generation has sort of an

4:36

open model where it's more partnership

4:39

oriented, more favorable to developers, and

4:41

the may be closed model where

4:43

it's just like it's usually just

4:46

Apple, basically just integrating older stuff

4:48

closely together and. Yeah.

4:50

In In Mobile, Apple One, And

4:52

I mean that there's technically more entered

4:54

signs out there, but you know, basically

4:56

Apple if a business perspective one and

4:59

and and dollar developed countries where people

5:01

can afford their products. Seven people basically

5:03

prefer the Be I phone on. But

5:06

that's not the way it always works. written

5:08

Pcs before that I you know it is in

5:10

the nineties. Your Microsoft and and Windows were

5:12

really the leader in their the Open model so

5:14

I just don't think that the future is

5:16

written on this yet. and. I.

5:18

Think that. For. The next generation Apple's

5:20

clearly gonna be the close model your Mehta

5:23

and in our approach is gonna be more

5:25

open and partnership oriented are more developer focused

5:27

and yeah I think we have a really

5:30

good chance at winning and and I I.

5:32

I. Am really committed to doing that. I've and

5:34

I'm like I am a competitive person. I'm like

5:36

passionate about what we're doing and liked of were

5:38

playing to win. So no I mean you know

5:41

that it's I can't like sit here and guarantee

5:43

that the future will go one way or another.

5:45

but I think people who will who try to

5:47

say that there's no chance that it goes the

5:49

other way it's are are just wrong. Much as

5:51

I think that this is this is an An

5:53

open game an arm and I, I feel pretty

5:55

good about the team in the effort and know

5:58

the products that were putting out. Apple fanboys. The

6:00

put on notice absolutely and we know you're

6:02

a competitive person to and we're going to

6:04

get to that later in the sell by.

6:06

Do want to talk about meal and eyes

6:08

Personal experience with the class three We actually

6:10

try it on. Yesterday I had a great

6:12

time while I was in it but as

6:14

soon as I took it off and looked

6:16

at Neil across the room kind. Of waving

6:18

his arms around at he was playing bead saver

6:20

at one point and I thought to myself wow.

6:23

This is pretty socially isolating actually.

6:25

if you're not physically wearing it,

6:27

Is this the future that you

6:30

are that you think consumers and

6:32

people really want. Well.

6:35

I think that there's a mix of it. Would

6:37

you spend a lot of time in front of

6:39

our computers today? You know it's I prefer space

6:41

for you having a big Tv, You're an entertainment

6:43

station in your living room where you kind of

6:45

do immersive things for hours at a time like

6:48

play games I think about. That is a thing

6:50

that I think people once, but you don't. I

6:52

see these videos of people like. Wearing.

6:54

These devices Walking down the street. That is not

6:56

what we're designing. the sports. I do think that

6:58

there's going to be a mobile version in this

7:00

next and the sex and generation of computing and

7:03

that's more going to be glasses And we're getting

7:05

started on that with the X Ray Ban. Met

7:07

a. Smart glasses that we put out that

7:09

are doing really well. You know it's own. The second

7:11

generation that we just put out, it's in. We have

7:13

met a I in there. Was

7:16

a lot of really exciting useless I'm really

7:18

excited to roll out.the. multi modal part of

7:20

that which which basically and and a I

7:22

speak as like you can ask questions using

7:24

both the camera is a suits look at,

7:26

look at what's around and you can speak

7:28

to at south in we have people doing

7:30

stuff like this are you going for a

7:32

walking tour around Stanford campus and it's like

7:34

look at a building and they they just

7:36

asked the class it's a case of books

7:38

and tell me what is the significance of

7:40

this building in the history behind it? Okay

7:42

what about the sculpture that's here to tell

7:44

me about that and I'm. Just that's a

7:46

pretty wild and and can be more social

7:48

stuff and and I think is is more

7:50

the things you're in to see more out

7:53

in the world but I mean just look

7:55

at people do spend. Many hours

7:57

in front of arms, you know, in front

7:59

of her. Big story today either either during

8:01

entertainment or productivity type things and. Energy

8:04

increasingly even fitness oriented devices with things

8:07

like Peloton Hundred the a big screen

8:09

component of that soaps and will the

8:11

future involved those activities. But and I'm

8:13

a more immersive version of that. Yeah,

8:16

I think so and I think part of what

8:18

we need to do is make that social rights

8:20

a part of it as the mixed reality. You

8:22

can see this the physical world around you and

8:24

blend the digital objects into that. That's a big

8:27

part of what we did that with. I was

8:29

that there's the big jump from quest to to

8:31

three is now we have this high resolution color

8:33

mixed reality pass throughs you can see the world

8:35

around you and put digital objects facts. but. Your

8:37

big focus for us and room when I

8:40

talk about the Metaverse that is about that

8:42

kind of social sense of presence and being

8:44

there with other people would just knowing that

8:46

we can designed into our products to make

8:48

them all sort of multiplayer and social and

8:50

I edged think I'm in are not like

8:52

the dna of our company so I you

8:54

know we're going to really focus on that.

8:56

I'm not. Scenario where I think Will Pride

8:58

do the best work in industry so in

9:00

your ideal world someone would be using the

9:02

quest headset in their home and then were

9:04

out the glasses is is that kind of

9:07

the relationship between. Those two. Yeah.

9:09

I mean I think that you can think of

9:11

the glasses. As. The Next Computing

9:14

Generations phone right? It's the

9:16

it's a Yam mobile truly

9:18

you're using it on the

9:20

Go computing platform where as

9:22

as the headset is can

9:24

be more the Next Generations

9:26

computer or Tv screen on.

9:29

So the bigger screens that you have an

9:31

yeah I mean I think for most people

9:33

the phone as pie the more important device

9:35

in their life and I think that that's

9:37

probably can be true in this next generation

9:39

that the glasses are probably going to be

9:41

the more important in more ubiquitous thing but

9:43

I think that they're both meaningful and I

9:45

also think technologically there's a pass where you

9:47

want to build kind of this. Immersive

9:49

and three d content and

9:51

you can build it first

9:54

in mixed reality. Because

9:56

the devices, by virtue of the fact that they're a

9:58

bit bigger, have more compute. So it's

10:00

easier to build stuff there. And

10:03

then by the time we get to glasses that can put holograms

10:05

in the world, we'll have built

10:07

up the whole developer ecosystem and platform to have

10:09

a bunch of the content there. But

10:11

the thing that I think has been really cool is

10:14

I thought that smart glasses weren't going

10:16

to be a ubiquitous

10:19

consumer product until

10:21

we got to that hologram in

10:23

putting digital objects, just overlaying

10:25

them physically on the world. But

10:28

then this crazy thing happened in the

10:30

last year, which is this emergence of

10:32

these AI models where you can talk

10:34

to them and ask questions. And

10:38

they can do a lot of things for you. And now

10:40

I actually think that something like the Ray Ban

10:42

Meta smart glasses is going to have much broader

10:44

appeal than I would have guessed even a few

10:46

years ago. And when we started that, we

10:48

thought of it as the intro

10:51

project working our way up towards holograms,

10:54

whereas now I actually think a lot of people are

10:56

just going to want that. Right. Because

10:59

it's an affordable product. It's a

11:01

stylish product. You don't need

11:03

to cram a ton of technology into it. You can just

11:05

like, yeah, it's like it has cameras, it has speakers, it

11:07

has a microphone. You can talk to AI. You

11:10

can take phone calls on it. You can listen to audiobooks while you're

11:12

out there. It's

11:15

just pretty capable already. Right. Many

11:18

people in the tech industry have been questioning whether

11:20

AI will render the end of

11:23

the smartphone and they're building these different hardware

11:25

gadgets from a pin to Peppers

11:28

little thing. I don't even know how you would describe

11:30

it. But but you think

11:32

that the smartphone because of AI

11:34

will replace or I'm sorry, the

11:36

glasses will replace a smartphone. Well,

11:40

I guess maybe I should clarify that. I

11:42

think what tends to happen is that the

11:44

last generation of computing doesn't go away. It's

11:46

not like when we got phones, like people

11:49

stopped using computers, but you do shift

11:51

in terms of use. Right. So like

11:54

before, we probably used computers more and then we got phones. Now

11:56

I do a lot of stuff on my phone today that I

11:58

would have done on my computer. computer before, and that's

12:01

clearly the primary thing. So what I think

12:03

is going to happen is that glasses are

12:05

going to become the primary mobile device. I'm

12:07

not predicting that in 10 years, we're not

12:09

also going to have phones. I just think

12:11

that a lot of the time, they're going

12:13

to sit in our pocket when we're just

12:15

doing stuff that we can do more naturally

12:17

and in a more socially good

12:20

way with glasses without having to

12:22

have something in your hand or

12:24

take your eyes away from the

12:26

interaction we're having and look down

12:28

at a phone. But

12:31

I'm certainly not predicting the complete

12:33

disappearance of phones. I think that that's

12:36

not in the cards, I think, anytime soon. And

12:39

to some degree, I think the glasses will interact with

12:41

the phone because if you're asking the

12:43

glasses a question, the AI basically

12:45

connects to the phone and then uses the phone's

12:47

connection to the internet to then talk to our

12:49

infrastructure to ask the question and get it resolved.

12:52

So we're kind of betting that you will have

12:54

a phone. You will probably just take it out

12:56

a lot less. I've been so surprised to see

12:59

the creator kind of economy embrace this

13:01

product as well. I don't know if you saw

13:03

the trend that went pretty viral on TikTok of

13:05

people kind of dancing around while wearing the glasses.

13:08

That's completely organic, by the way.

13:10

And I was just surprised to see them

13:12

kind of immediately glom onto this new piece

13:14

of technology and immediately integrate it into their content.

13:16

Were you surprised about that or was that the goal

13:18

all along? Well,

13:22

it's always awesome to see. My

13:25

history and where I come from,

13:27

I'm more of like

13:29

an engineer, write a build software.

13:31

So building stylish hardware is

13:33

relatively new for me. Mark,

13:36

you got a quiet luxury vibe to you. Give

13:38

yourself some credit. OK, I appreciate that. But

13:45

that's wearing other people's stuff. Don't

13:50

get me wrong, I have a lot of confidence. I do believe

13:52

that given enough iterations,

13:54

I can Get

13:57

there on building the stuff that we want to. But for

13:59

this one, we did some. given what should, we worked with

14:01

Ray Ban. Or it and you're there

14:03

like. More. A year The company

14:05

that builds the Ravens as the meeting

14:07

in a fashion and in kind of

14:10

glasses company in the world. So they

14:12

do great work on style and luxury

14:14

which it's really been a very fascinating

14:16

partnership I'm in. The technology behind even

14:18

producing can normal glasses before you put

14:21

on electronic senate is actually quite sophisticated

14:23

and green. A We've learned a lot

14:25

about working with them about how do

14:27

you integrate. And of this

14:29

new consumer electronics technology with all the technology

14:32

that goes into creating like. Really?

14:35

Sophisticated lenses and

14:37

com. And and making it's

14:39

that earth. the materials in that and the

14:41

glasses feel premium but or light self is

14:44

all the stuff that I think like neither

14:46

one company could have done on their own

14:48

arm. But. Through.

14:50

This partnership I think we produce something that

14:52

like wings. State. Vr technology

14:54

Around day I am. With.

14:57

Something that has been a fashion icon for

14:59

decades. And I I think it's pretty neat

15:01

that yeah, with someone I I don't know,

15:03

it's It's tough for me to predict exactly

15:05

what the uptake would beats, but I'm. But

15:08

it's. I mean, not that out

15:10

of the question. I think a lot

15:12

of that is because you know that

15:14

to design is really nice. Okay, I

15:16

do. I transition to a pretty under

15:18

the radar topic right now. We've met

15:20

a few times already. I'm Michelle and

15:22

that is artificial Intelligence. On your most

15:24

recent earnings, call the phrase a I

15:26

with Medicine seventy times for says a

15:28

seven for the Metaverse details or company

15:30

name the matter, and to keep saying

15:32

you're not creating a Ice and the

15:34

Metaverse, but it still feels like that's

15:36

where the industry focus is right now.

15:38

The Metaverse. Feels pretty twenty twenty two

15:41

right now. How much did your strategy

15:43

sift after? A i kind of blew

15:45

up when Sacha Be T was released.

15:49

ah well you know we've been working on a

15:52

i and the metaverse for for a long time

15:54

avenues have been the two major kind of big

15:56

tax best for for the company for like a

15:58

decade and you the reality is But we're spending,

16:00

if you just look at how much we're investing, it's like

16:02

$15 billion a year, at least

16:05

into each of them. So yeah, I mean,

16:07

what we talk about will vary at different

16:09

points. And yeah,

16:11

like what we see happening in

16:14

the industry will help sharpen our product

16:16

strategy and our sense of

16:18

what we can be building. Right

16:21

now, I definitely think where we are in the

16:23

hype cycle is much more, you know,

16:25

people want to talk about AI. Whereas,

16:28

you know, yeah, in like in 2122, I

16:30

think people are pretty excited about talking about

16:32

the metaverse. Now, I think

16:34

we're in this mode where people are waiting for

16:36

the products to get like more fully adopted. But

16:39

like, we're just kind of quietly churning and building this

16:41

stuff in the background. Like I don't care that much

16:43

whether something is cool one year or not. But

16:46

yeah, I mean, at a moment when people have more questions

16:48

about the AI work that we're doing, I think it's natural

16:50

that you'll hear me be talking more about AI. When

16:54

the metaverse space heats up a bit, you

16:56

know, people try to call it different things because they don't want

16:58

to embrace our branding around this. But

17:02

when that kind of heats up, then you know, you

17:04

go through these moments where, okay, now we're talking about

17:06

our headsets or that stuff more. So I just think

17:08

this will kind of go back and forth a little

17:10

bit over the next several years. But

17:12

the reality is that these have been huge focuses for

17:15

a while and I think

17:17

are, you know, going to continue to

17:19

be for like the next decade or so until these

17:21

things both reach their full potential. Let's

17:23

talk about Sam Altman a bit.

17:25

The open AICO, he's been reportedly

17:28

trying to raise $7 trillion to

17:30

build a semiconductor infrastructure needed to

17:32

power artificial intelligence systems. NVIDIA CEO

17:34

said this week that Altman's plan

17:37

was overkill. But what do

17:39

you think? Is Sam insane or is this

17:41

scale of investment needed to support the development

17:43

of AI? I

17:46

don't know. I mean, Sam is, I think, actually

17:48

like a very earnest and thoughtful guy. And

17:51

I think his level of ambition for the

17:53

space, I think, has been very helpful for

17:55

building it out overall. And

17:59

I mean, the number is... astronomical. So I mean, it

18:01

is hard to kind of wrap my head around

18:03

that. But honestly, I haven't gotten a chance

18:05

to talk to Sam about it. So I don't, I

18:08

don't want to kind of weigh in, weigh in

18:10

too much on that. But I think the place

18:12

that he's right is that like, you know, if

18:14

you look year over year, these companies

18:16

keep on building these larger and larger

18:18

training clusters of computers to

18:22

basically train the state of the art models. So it's like

18:24

a couple years ago, maybe it was like 1000 chips

18:27

that you kind of wired together, then it's 10,000 then

18:31

it's 100,000. You know, at some point, it's gonna

18:33

be a million, right? And then it's like, at some point, okay, yeah,

18:35

you really need to produce a ton of chips. So at

18:37

that point, I don't

18:40

think he's crazy. I

18:43

don't know what it will

18:45

take to kind of accelerate all of

18:47

this to the point where kind of

18:50

the technological curve can

18:52

keep up. But I think that this

18:55

is one of these things where like, building software

18:57

sometimes hits roadblocks, when you get into physical stuff,

18:59

right, and all the stuff that we do, it's

19:01

like when we're building the apps, right, you're building

19:03

like Instagram, or WhatsApp, or Facebook, you know, we

19:06

can just ship updates, like every week. But you

19:08

know, when we're building like a new headset, it's

19:10

like, okay, we work on that for two years,

19:12

and we ship one headset, right? And then we

19:14

work on another one for two years, ship another

19:17

headset, it's just that the cycle time is slower.

19:19

So I think the software part of AI will

19:21

probably want to move fast. But I

19:23

think at some point, Sam is right that we're

19:25

going to run into some physical constraints on like, how

19:28

many chips can you churn out? And, you know,

19:30

how much power you need from like, power

19:33

plants to be able to train the next bigger

19:35

model. And I don't know, I think it is

19:37

worth thinking about how to kind of alleviate some

19:39

of those constraints, because I think it's just going

19:41

to enable, and this technology is going

19:43

to enable all these awesome use cases that I

19:45

think are going to make our lives better in

19:48

so many ways. I do just want to get

19:50

into kind of the philosophical argument around AI a

19:52

little bit. On one side of the spectrum, you

19:54

have people who think that it's got the potential

19:56

to kind of wipe out humanity, and we should

19:58

hit pause on the most advanced system. And

20:00

on the other hand, you have the market

20:02

reasons of the world who said stopping AI

20:04

investment is literally akin to murder because it

20:07

would prevent valuable breakthroughs in

20:09

the healthcare space. Where do

20:11

you kind of fall on that continuum? Well,

20:15

I'm really focused on open

20:17

source, which is, I'm not

20:20

sure exactly where it would fit on the continuum. My

20:23

theory of this is

20:25

that what you want to prevent is

20:28

one organization from

20:30

getting way more

20:32

advanced and powerful than everyone else. Here's

20:35

one thought experiment. Every

20:38

year, security folks are

20:41

figuring out what are all these bugs

20:43

in our software that can get exploited

20:45

if you don't do these security updates.

20:47

Everyone who's using any modern technology is constantly

20:50

doing security updates and updates for stuff. So

20:53

if you could go back 10 years in time

20:55

and kind of know all the bugs that would

20:57

exist, then any given organization

20:59

would basically be able to exploit everyone

21:02

else. And that would be

21:04

bad. It would be bad if someone was

21:06

way more advanced than everyone else in the

21:08

world because it could lead to some really

21:10

uneven outcomes. And the

21:13

way that the industry has tended to deal with this is

21:15

by making a lot of infrastructure open source. So

21:18

that way, it can just get rolled out

21:20

and every piece of software can kind of

21:22

get incrementally a little bit stronger and safer

21:25

together. So that's the case that I

21:27

worry about for the future. It's not like, I mean,

21:30

I don't want to write off the potential

21:32

that there's some runaway thing, but right now I

21:34

don't see it. I don't see it anytime soon.

21:37

The thing that I worry about more

21:39

sociologically is just one organization basically having

21:41

some really super intelligent capability that isn't

21:43

broadly shared. And I think the way

21:45

you get around that is by

21:48

open sourcing it, which is what we do. And

21:50

the reason why we can do that is because

21:52

we don't have a business model to sell it.

21:54

So if you're Google or you're open AI, this

21:56

stuff is expensive to build. The business model that

21:58

they have is they kind of. build a model,

22:00

they fund it, they sell access to it. They

22:02

kind of need to keep it closed. And

22:06

it's not their fault. I just think that that's

22:08

where the business model has led them. But we're

22:10

kind of in a different zone. We're not selling

22:12

access to this stuff. We're building models and using

22:14

it as an ingredient to build our products, whether

22:17

it's the Ray-Ban glasses, or an AI

22:19

assistant across all our software, or eventually AI

22:21

tools for creators that everyone's going to be

22:23

able to use to let your

22:26

community engage with you when you

22:28

can engage with them and things like that. So

22:31

open sourcing, it actually fits really well with our model.

22:35

But that's kind of my theory of the

22:37

case, is that, yeah, this is

22:39

going to do a lot more good than harm. And

22:42

the bigger harms are basically from having the

22:44

system either not be widely or evenly deployed

22:46

or not hardened enough, which is the other

22:48

thing, is open source software tends to be

22:50

more secure historically, because you make it open

22:52

source. It's more widely available. So more people

22:55

can kind of poke holes at it, and

22:57

then you have to fix the holes. So

23:00

I think that this is the best bet for

23:02

keeping it safe over time and part of the reason why we're

23:04

pushing in this direction. I'd love your

23:06

take on AI in the job market.

23:08

There's a lot of anxiety among people

23:10

that AI could replace them, especially with

23:13

all the tech layoffs of the last few

23:15

years and companies signaling huge

23:17

investments in AI. From your

23:20

point of view as the person who

23:22

is literally building these systems, say on

23:24

the average white collar employee working in

23:26

professional services as an accountant, say, should

23:28

I be worried? Should I pivot my

23:30

career to become a chatbot prompt engineer

23:32

or whatever? I'm

23:37

not sure I'd pivot my career yet. I

23:40

mean, look, there's a lot in what you're saying. Over

23:42

the long term, I'm actually quite bullish

23:45

that all these tools will give more

23:47

people the potential to kind of do what

23:49

they care about. When I see

23:51

my kids, it's like they want

23:53

to do all this creative stuff, and then we're

23:55

like go eat broccoli. And I

23:58

do think that they. A

24:02

lot of people, I think, maybe are not in

24:04

the jobs that they would ideally want to be

24:06

in. I think in the future,

24:10

we'll just have the capacity to do more of that. That

24:13

is the long term. Then there's always the question of how

24:15

do you get there. Yeah,

24:17

I think that there will be a

24:20

bunch of transformation that will mean that

24:22

certain jobs, we don't do them

24:24

in the way that we're doing them now, certain things might

24:26

get automated. I don't know,

24:28

I just think that the limit to

24:30

human ingenuity, it's either

24:33

limitless or extremely

24:35

high. I just don't think we're anywhere

24:37

near fulfilling that potential yet. I'm

24:40

much more optimistic about that. In

24:42

terms of the layoffs and stuff like that, I

24:45

actually think that that was more due to companies

24:48

trying to navigate COVID and the pandemic.

24:52

It was this weird wave where at

24:56

first, we saw e-commerce go through

24:59

the roof. It's like, okay, yeah, because a lot

25:01

of people aren't going out and physically shopping. It

25:03

was really hard to predict, okay,

25:05

is that going to continue?

25:07

Is all this e-commerce just better than

25:09

physical shopping? What's the balance after? I

25:12

think across the economy, a lot

25:15

of companies just overbuilt. Then

25:18

when things went back to pretty close

25:20

to exactly the way they were before,

25:22

I think a

25:25

lot of companies realized, hey, we're not

25:28

in a good financial place because we

25:30

overbuilt. Then you have

25:33

this wave of layoffs that we're

25:35

basically responding to that. Then

25:37

you also have this generation of companies that knew

25:40

nothing except growth. The

25:42

idea of doing layoffs was just

25:44

like, oh, no, this is crazy. We

25:46

couldn't even comprehend that, but then they had to do it.

25:49

Then I think a lot of companies went

25:51

through this zone of thinking, hey, it didn't

25:53

end the company. It

25:56

was obviously really tough. We parted with a lot of

25:58

talented people we cared about. But in

26:01

some ways actually becoming leaner kind

26:03

of makes the companies more effective. So

26:05

now I think that there's another kind of wave

26:07

that a bunch of companies are going through. And

26:09

we did this last year, but I think some

26:11

companies are kind of still in this mode of

26:14

thinking, OK, well, we're thinking about

26:16

how efficient and lean we should be. Maybe

26:19

we should be a somewhat different shape than we are now to

26:21

do the best work that we can. But

26:24

I don't think that that, at least

26:27

for us, the AI stuff was

26:29

not a major driver of that. It

26:31

was like first just kind of this overbuilding and then this

26:33

sense of like, let's do the best work

26:35

that we can by making a lean company. As

26:38

lean as you can be for our scale. Yeah,

26:40

it does feel like the layoffs of early 2024

26:42

do have some this AI tinge to it. It's

26:47

not the back to normal COVID stuff

26:49

because when you listen to these CEOs

26:51

on their earnings calls and when they're

26:54

doing the layoff announcement, it's more strategic

26:56

reorganizing of their focus where they're

26:58

investing more in AI. And they're not

27:00

saying I'm laying you off because of

27:02

that, but it's just a trimming of

27:04

certain departments and they're hiring in others.

27:06

So to me, I feel like this

27:09

wave of layoffs is a little bit different than

27:11

the ones coming out of like early 2023 like

27:13

yours. Yeah,

27:17

that's possible. I mean, maybe

27:19

I'm just mostly commenting on

27:21

what I see in our company and sort

27:23

of our kind of peers and the companies

27:25

around me in the tech industry, but there's

27:27

certainly broader stuff. So I just maybe I'm

27:29

not seeing that one as clearly. Yeah.

27:32

Last question on AI here. You have

27:35

a lot of listeners who have may

27:37

have only interacted with AI on a

27:39

very limited basis. If you wanted to

27:41

below their minds, delight their imaginations, what

27:43

would you show them right now? Would

27:46

you exemplify the power of AI to

27:48

someone who hasn't really experienced it so

27:50

far? I

27:52

Mean, I Think the Ray-Ban glasses are

27:54

going to enable, like already do and

27:56

will enable some pretty mind blowing things.

28:00

And I think people always tend to. Think. That

28:02

these physical kind of manifestations of things are

28:04

more impressive than just the software Only ones

28:06

I think having glasses that can see what

28:08

you see here we you hear you can

28:11

ask questions or what's going on around you

28:13

and I do the ability to At night

28:15

I sang with walk around through some tough

28:17

historic place and just ask until I looked

28:20

suffocate. Explain to me what what what matters

28:22

about best you go into a store and

28:24

rak. Okay because this a good price for

28:26

this thing or years ago. What is this?

28:28

Where else can I get this? I'm. I

28:31

not that stuff is just gonna be super valuable. Yeah.

28:36

And it's only going to get crazier.

28:39

So. I mean some of the stuff

28:41

that we're working on. I'm There's always this question

28:43

of like how do you talk to the I. Were.

28:46

It's. A Today okay you type on a

28:48

computer or phone and on the glasses,

28:50

speak to it. And one of the

28:52

wilder things that were working on his

28:54

arm. Is as neural interface.

28:57

And I don't mean like a chip that you

28:59

jack into your brain. I think that the i

29:01

mean maybe in the future seminal do that but

29:03

i i i don't have i would one of

29:06

us version one of that were at every level

29:08

of into such but I figured you might when

29:10

wait until that one's pretty mature literati our to

29:12

have the ah yes that's i guess I don't

29:14

wanna sell so it's wait until attorney take up

29:16

rate that thing every year. arms that like but.

29:19

I. Think he did. The one that we're working on is

29:22

basically a wristband. And. The Weather works

29:24

as it's a neural interface in the sense

29:26

that it arm in the way that you're

29:28

kind of brain communicates with your body and

29:31

Mike controls the where you move is it

29:33

sends these. These like

29:35

nervous system signals are to to your

29:37

your to fear did muscles and you

29:39

can actually pick those up with armed

29:41

with it with the Dmz wristband and

29:43

it turns out that there's like all

29:45

this extra bandwidth and and like all

29:47

the sex for signal that your brain

29:49

concerned that. Is not currently being used

29:51

to move your fingers and whether you want and you can.

29:54

Just said. You'll. be able to

29:56

in the future essentially just like type

29:58

and control something like by kind of

30:00

thinking about how you want to move your hand, but

30:02

it won't even be like big motions. So I can

30:04

just sit here and I'm basically

30:07

typing something to an AI. And

30:10

then right now with the glasses, it basically can

30:12

speak back to you. But in the future, I

30:14

think you'll have something where it can speak to

30:16

you or there will be a little display so

30:18

it can show you a notification. I

30:22

think you'll have this completely private

30:24

and discrete interface where you

30:27

can walk around throughout

30:29

your day, just text different things and you

30:31

can text your friends or

30:33

you can text your AI and then basically

30:36

just get the response back in real time.

30:38

So I think that

30:40

that is just going to be insane. And the

30:43

research is like, we've been working on this for a while.

30:45

So it's actually, it's not a

30:47

one year project, but we're quite a

30:49

few years into it. We're actually kind

30:51

of close to having something here that

30:53

we're going to have in a product

30:55

in the next few years. So I'm

30:57

pretty excited about that. I mean, you're

31:00

making teachers' lives harder. Imagine someone with

31:02

their hand in their pocket researching AI

31:04

stuff while taking a test. turnitin.com is

31:06

going to have to really boister their

31:08

defenses for sure. All

31:10

right, Mark, let's talk about you. People

31:13

want to know more about you. So you've

31:15

been leading Facebook, NowMetta, for 20 years,

31:17

literally more than half of your life.

31:20

I'm sorry to remind you. How has

31:22

your management style evolved over the past

31:24

two decades? Oh,

31:27

man. I mean, that question

31:30

assumes that I had a management style when

31:32

I got started. I'm not assuming anything. The

31:34

reality is like, when I started Facebook, I

31:36

wasn't trying to start a company. I just

31:38

thought it was this cool thing to

31:40

build. And I

31:43

was pretty convinced when I was in

31:46

college, my aspiration was I like building

31:48

stuff and I thought I was going to just go

31:50

through college and then go be an engineer at Microsoft. And

31:53

my mom was like, nah, you're going to make a company. I was

31:55

like, no, I'm not going to make a company. My mom actually bet

31:57

me that I was going to drop out of college. I was like,

31:59

mom. Like it's like what is it like in a

32:01

good way or a bad way and she's like, I don't know. We'll

32:03

see My way

32:06

before you went to college. She said you're gonna

32:08

drop out. Yeah, shout out mom

32:10

Wow Yeah, no mom mom

32:14

And my younger sister bet that she was gonna finish

32:16

college before me which also happened

32:18

obviously So yeah, so I

32:21

started this company But I

32:23

didn't started as a company right? I just This

32:26

partially why it was so messy in the beginning

32:28

I we basically ended up turning it into a

32:30

company because it's a guy in order to scale

32:33

it We needed money, which means we needed a

32:35

business model and revenue and all

32:37

that stuff. So basically You

32:39

know pull the group people together to try to do

32:41

that, but I didn't know anything about building a company

32:44

I mean that like wasn't my goal And

32:47

it just you know, it just turns out if you want to

32:49

do big things It like helps

32:51

to have the resources and in order to be able

32:53

to get the resources you want to have a business

32:55

model that supports That so I guess

32:57

over time I've sort of had to just learn a lot

32:59

of stuff and I've done that mostly by Working

33:02

with really good people. I mean one of my mentors

33:04

and you know my business partner for a long time

33:07

Cheryl Sandberg, you know, we joke that she

33:09

like raised me as a manager and

33:12

like as a parent almost and I

33:16

Think that that's true in a lot of ways. It's I

33:19

mean a lot of my kind of leadership

33:21

philosophy and management Just

33:23

came from working with her and other people who I

33:25

really respected who had more experience But now we're sort

33:28

of in this funny place where like and if you

33:30

look at early Facebook in like 2004 2005 We

33:35

were basically just like a bunch of kids running

33:37

around just like building stuff that we thought was

33:39

cool and then like had some more professional Executives

33:42

helping to run the pulse, but

33:44

now all the kids have grown up, right? So,

33:46

you know, you have people like like Chris Cox

33:48

our chief product officer has been there Yeah, I

33:50

think he joined in like 2005 or 2006 Andrew

33:55

Bosworth who runs reality labs similarly,

33:58

you know his has been in

34:00

there for, it's 18 or 19 years at this point.

34:05

So, and along the way, I think we all just

34:07

kind of learned how to run a company together. And

34:10

we're all pretty close friends. And

34:14

it's just interesting, because you can't hire that. If

34:17

you're just thinking about today, it's like, okay, if I was trying to

34:19

build a team, or what advice would you give people? It's like, I

34:21

don't know, you can't, nothing

34:25

that kind of replaces all those years of

34:28

just experience, getting to know each other and going

34:30

through all these times where everyone thinks you suck.

34:32

And then other times when people are

34:34

just giving you way too much credit for stuff that

34:36

clearly doesn't get actually pan out for

34:39

a few years, but you believe in it. And

34:41

we've just been through all those experiences. So,

34:44

I don't know. I

34:46

think it's, the secret

34:48

to all this stuff is it's not one person. And

34:51

as much as I think the world

34:53

wants to focus on who's

34:55

the singular figure, it's

34:57

always more than one person. And

35:00

it's the team dynamic. And

35:02

I think that that matters a lot. So, I

35:05

mean, if you can just kind of stay focused,

35:07

you find stuff that you like, you stay engaged. I

35:11

don't mean by the way that it's like, okay, you set up

35:13

the team and then you're kind of passive. Right, I think like

35:15

probably one of my most

35:18

controversial leadership or

35:20

management things is like, I

35:22

don't actually believe in delegating that much. I kind of

35:24

think like the way a founder should

35:26

work is you should basically make as many decisions and

35:28

get involved in as many things as you can. And

35:31

like, I mean, you need to know where your limits

35:33

are and where like you're just thrashing people because you're

35:35

involved in something in a half ass way. And like,

35:37

you can't, you're not actually, you

35:40

don't have all the context, but I don't know. I

35:42

mean, you need all these other awesome

35:44

people because there's still gonna be, no matter how much time

35:46

I put into all these things, there's still gonna be so

35:48

much stuff that I can't get to. And we need awesome

35:50

people who can like do all the really important stuff that

35:52

like I'm not doing. But

35:56

I don't know, that's something that I guess I've just

35:58

gotten more confident in over time is just so. of

36:01

feeling like, hey, yeah, I can go deep on all this

36:03

stuff and push it in a

36:05

direction that I think. And yeah, not everything is going to

36:08

go well in the near term, but you just learn.

36:10

Rinse and repeat. Do good work over a

36:12

long period of time. It sounds like

36:15

you're having a great time still leading

36:17

this company. You've been at it for

36:19

a very long time. But we got

36:21

to talk about some of these hobbies,

36:23

Mark. Probably the one that people most

36:25

know you for is your love of

36:27

mixed martial arts, but you also fly

36:29

helicopters. You hydrophile, you love smoking meats,

36:31

and you're actually raising your own cow

36:33

right now. What's the current

36:35

obsession for Mark Zuckerberg? Oh, gosh. Right

36:40

now, I'm focused a lot on knee rehab. So

36:42

I can get back to fighting. It's

36:45

like, man, I mean, I spent

36:49

a bunch of time fighting, but rehab

36:52

takes a lot of time because you want

36:54

to stay in shape and you

36:56

can't do everything that you were doing before.

36:58

So it's actually not super time efficient, but

37:01

it's really important. So that I think is

37:03

my main thing. I

37:07

really like fighting in combat sports. That's

37:09

been awesome. But no, I mean, look, I mean, I've

37:11

loved... The

37:14

whole meat thing has sort of been a

37:16

fascination for a long time. I

37:18

mean, look, ever since that, yes,

37:21

Sweet Baby Ray is just sitting

37:23

out here with friends and smoking meats. I

37:28

love it. And I always

37:31

joke with Priscilla

37:33

and my daughters now that

37:36

if I'm ever done with meta, I'm going

37:38

to run Mark's Meats. And

37:41

if you're like a kid, like my daughters,

37:43

it's kind of hard to wrap your head

37:45

around like what meta is, because it's just

37:48

this very abstract app. It's like,

37:50

okay, you're like a young kid. So my

37:53

daughter is for a

37:55

while, she just thought that I was a cattle

37:57

rancher. She's like, all right, like, dad, that's... That's

38:00

clearly what you do. It's like

38:02

Mark's Meats isn't ready for prime time yet, but you're really

38:04

focused on it. I get it. It's like

38:06

you've got a few of these cows. I

38:09

just had this project where

38:11

I want to see if we can produce some

38:13

of the highest quality meat in the world. I

38:18

think most commercial

38:20

cattle operations are

38:22

basically constrained where you can't put

38:25

more into the cow in

38:27

terms of quality of ingredients than

38:29

you can then turn around and sell the cow for

38:31

when you're done. I

38:34

obviously don't have that issue. I'm not trying to

38:36

do this commercially. I'm just trying to

38:38

create the highest quality stuff that we can. I

38:41

think it's pretty fun to vertically integrate. I'm like, all

38:43

right, let's see if we can... What

38:46

is the densest, most

38:48

nutritious thing that

38:50

we can feed these cows to basically get

38:53

them to be the healthiest and

38:55

gain weight as quickly as possible and just

38:57

be the most delicious cows? It's like, all

38:59

right, well, as a human, what do you

39:01

think is the thing that basically you just

39:04

sit and eat a lot? It's like beer and

39:06

nuts, basically. Nuts,

39:09

super dense. Beer

39:14

induces appetite, which I think people

39:16

are familiar with. Yeah,

39:18

induces appetite, I think, is

39:20

the formal jargon in the

39:23

cattle industry for get them to eat a

39:25

bunch, which is

39:27

why for a long time, a lot

39:29

of the highest quality Wagyu beef,

39:32

they basically fed it beer. There's

39:34

this whole thing about, okay, does it make it

39:36

more relaxed? The scientific literature on that

39:38

is a little mixed. I've read a

39:40

bunch of this. I'm

39:43

not convinced on the relaxation point, but

39:45

I do think it's clear that it makes them eat more

39:47

and then they gain more weight. I

39:50

just think it's super fun. It's like, all right, let's

39:52

grow our own beer. Let's

39:55

grow our own macadamia nuts, process that,

39:58

have our kids be a part of it. part of

40:00

the process of figuring out what it's

40:02

like to run this kind of a

40:04

process. Easier for them

40:06

to do that than be involved in the software business. So

40:08

I think that that kind of makes sense. And

40:11

I don't know, it's just fun. Maybe

40:13

one day it'll be a thing. But for now, I'm

40:16

just producing some meat that I like eating. Mark, the

40:18

women in your life are telling you something. Your mom

40:20

said you're going to drop out of college and your

40:22

daughter said that you are going

40:25

to be a cattle rancher. So I would

40:27

take that advice to heart. Yeah,

40:30

I mean, they're usually right. We will workshop

40:32

the name, though. Mark's meat. There's still some

40:35

room to grow there as well. Oh, no,

40:37

it's intentionally really lame. All right, all right.

40:39

And you absolutely named it then. It also

40:41

sounds like you could be

40:44

the Andrew Huberman of like cattle ranchers.

40:46

You could launch a podcast about just

40:48

cattle health. Because I was learning

40:50

a lot right there. You had me and wrapped it. So

40:52

maybe you and Andrew team up for a little cattle

40:55

pod. The

40:57

cattleverse. The cattleverse. The cattleverse. You heard

41:00

your first. Exactly. Mark, this was an

41:02

awesome conversation. I really appreciate you taking

41:04

the time to stop by and nerd

41:06

out with us. I'm going to actually

41:08

plug your social accounts right now. You

41:11

can follow Mark on the social networks

41:13

he invented at Zuck. That was a

41:15

very cool sentence to say. And

41:18

then also check out all the cool work

41:20

happening at the Chan Zuckerberg Initiative to see

41:22

all the work that you're doing with your

41:24

wife Priscilla. Mark once again, thank you

41:26

so much for joining us. Awesome.

41:28

Thanks for having me on guys.