0:00

It's time to review the highlights. I'm joined by

0:02

my co-anchor, Snoop. Hey, what up, dog? Snoop number

0:04

one has to be getting the new iPhone 16

0:06

Pro with Apple Intelligence at T-Mobile. Yeah, you should

0:09

hustle down at T-Mobile like a dog chasing a

0:11

squirrel, chasing a nut. That's a nice analogy, Snoop.

0:13

On to highlight number two, at T-Mobile, families can

0:15

save 20% every month versus the other big guys.

0:17

Very impressive. Take it away, Snoop. Head to t-mobile.com

0:19

and get the new iPhone 16 Pro with Apple

0:22

Intelligence on them. Now drop that jingle. See

0:25

how you can save versus the other big guys

0:27

at t-mobile.com/switch. Apple Intelligence coming fall 2024. Hello

0:32

and welcome to VergeCast, the flagship podcast

0:34

of face computers and

0:36

the trademarked VergeCast matrix of

0:39

wearable stuff. I

0:41

don't want to start with a swear word, but if you

0:43

know, it's usually a swear word. Hi, I'm your friend, Neela.

0:45

Alex Kranz is here. I'm also wearing my augmented

0:47

reality glasses right now. Because

0:50

they pass light through them? Yeah, they pass light through them

0:52

and everything is augmented now. I can see. It's

0:55

been augmented into focus. Yeah. I got you.

0:57

David Pierce is here. Hello. And

0:59

Alex Heath is here. Alex, are you actually wearing

1:02

a face computer? I am, but

1:05

you'll never know. It's

1:08

big week in tech news. Quite a lot going on. It

1:11

was MetaConnect. Alex was there along

1:13

with Kylie Robison and Jay Peters.

1:16

Alex, you wore the Orion AR glasses, the

1:19

demo. You interviewed Zuckerberg. We got to talk

1:21

all about that. There's tons of other news

1:23

out of MetaConnect. And then there's Chaos at

1:25

OpenAI, which a company, for

1:28

being as successful as it is, remains

1:30

mired in nothing but pure chaos. We could

1:32

just snip that and play it every single

1:34

week on the first cast. Just that thing

1:36

you just said is like a, it's just

1:38

a universally true statement at this point. Yeah,

1:41

we have to talk about that. And then we

1:44

have Lightning Round. Unsponsored. Still.

1:48

I have walked into rooms at this company

1:50

and demanded why the Lightning Round is unsponsored.

1:53

Was anybody else in the room? No one else was in the rooms.

1:56

This is the, I'm getting, I'm working

1:58

up to the final results. But

2:01

you walk into a room, you know, where eaters

2:03

having a staff meeting, demanding the lightning round be

2:05

sponsored. And they're like, I don't, we're

2:07

doing a roundup on cakes. Like, get out of here. I'm

2:10

working on it every day, closer to the

2:12

goal. Okay. Let's

2:15

start with meta. Alex, you wore the Orion.

2:17

I think the Orion is the thing to talk about.

2:19

They've been working on this for 10 years at reality

2:21

labs, burning billions of dollars a year. They

2:23

did it. They made AR glasses, but you can't

2:25

buy them and they cost $10,000. Kind

2:28

of. You can't buy them. So

2:30

they are glasses in the sense that you can

2:32

put them on and they work to a degree.

2:36

I wrote this in the story and I had

2:39

a hard time writing about this because I've been

2:41

getting a lot of almost products

2:43

put on my face recently, like just

2:45

coming off of snap. We

2:48

do have a series of photos of you that is

2:50

incredible right now. Yeah. I need to

2:52

make like a, like a photo book to just have at

2:54

home and look at it and go like, that's a cry

2:56

for help. Me

2:59

wearing all these face computers. Um,

3:02

yeah, it's, it's a really impressive

3:04

demo. Um, I think

3:07

meta is doing the

3:09

demos because it knows it's impressive,

3:12

but it's not a product. So I mean,

3:14

I wrote this in the story. It's

3:16

not vaporware. Like it's very real and

3:19

it's not a simulation. It's not totally

3:21

on rails. I used it for

3:23

probably a total of two and a half hours between the

3:25

two days we were shooting and

3:28

I had enough opportunity to go off of

3:30

the beaten path a little bit and make

3:32

sure I wasn't just getting piped in a

3:35

complete simulation. And it's a real working piece

3:37

of kit. Um, but

3:39

it's not a product and that

3:41

says a lot about the state of AR

3:43

and these glasses. And

3:47

at the same time, I finally

3:49

feel like I've been running about this for

3:51

so long, banging my head against the wall

3:53

and being like, what am I doing?

3:55

This is never going to happen. I

3:57

finally feel after this week that. AR

4:00

glasses I may actually want to use are

4:02

not a pipe dream. Yeah, I want to

4:04

argue with you about the definition of the

4:06

word vaporware, which is very important to me,

4:08

personally. And holograms. And holograms. So let's start.

4:11

Alex and I have gotten a real fight about the

4:13

word hologram screens this week. We'll get into all of

4:15

this, I promise. But let's start with what it is,

4:17

right? Because AR glasses,

4:20

you're correct. The industry has been talking

4:22

about them for a long time. Magic

4:24

Leap promised AR glasses

4:26

years ago. If you

4:28

will remember their founder claimed

4:31

that he could hack the GPU of your brain. This

4:34

is a real thing. And that was in

4:36

reference to a display technology, because this is the

4:38

challenge. How do we build a display that you

4:40

can look through, that can augment reality, have

4:43

the processing power to see reality and

4:45

augment it, have connectivity, have a battery.

4:48

And no one can solve these problems,

4:50

most of all, the display. The

4:54

thing doesn't exist. The thing that you look

4:56

through to perceive the world and then layer

4:58

information over stuff, you

5:01

could maybe solve battery and processing and all

5:03

this stuff in your way in the backpack.

5:05

But the actual display technology to make it

5:07

good has more or

5:09

less not existed in any realistic way.

5:12

And Magic Leap had to give up on their idea,

5:14

and they tried another thing. And HoloLens was another thing

5:16

with a tiny field of view. And there's a long

5:18

list of things with bad fields of view. And

5:21

it seems like that's the thing that has solved

5:23

most of all here. Yeah. A

5:26

quote that really stuck out to me from the

5:28

product lead when I was getting my demo was

5:31

that the display was a

5:33

scientific breakthrough problem they had to

5:35

solve. And now they're in

5:37

the engineering problem-solving phase of making these

5:39

glasses work at a price point that

5:42

people can actually buy. And

5:44

that really stuck with me. You're right that

5:46

the display is the hardest part. The

5:48

distinction between AR glasses like

5:51

these in the Vision Pro or the Metacquest

5:53

is that yes, you can do mixed reality in the

5:56

Pro and the Quest, in the Vision and the Quest.

5:58

But what they're doing is... fully

6:00

enclosing your face in a computer, piping

6:02

video in and mixing all that in

6:05

the displays. These are literally just

6:07

these glasses, Orion, or letting light in their

6:09

actual glasses. And so the

6:13

challenge there is much different.

6:15

And Meta has

6:17

a lot of good ideas here. And

6:20

they also, I think, realized that

6:23

the specific optical stack that they

6:25

went with for Orion was just,

6:30

inexplicably hard to manufacture. And

6:33

these lenses are made of silicon carbide, which

6:36

is also used in space.

6:40

It's used in EVs. It's used in

6:42

like, Dremels. It's like what you

6:44

use for cutting tools. It's primarily like- It's

6:46

literally like a giant crystal. And so- Yeah,

6:48

it's also used for wedding rings sometimes. So

6:50

basically it's like having two fake diamonds on

6:53

your face. Yeah, and they were talking to

6:55

me about this and this challenge of the

6:57

yields on this. It's like, yeah, you have

6:59

to cut the crystal perfectly. He's like, and

7:01

there's not enough of it in the world.

7:03

So we're growing crystals. We're growing

7:05

like big ass silicon carbide rocks. I'm

7:08

sorry, the image of Mark Zuckerberg walking

7:10

into the basement of his lair, looking at

7:12

his like crystal farm in a haze of

7:14

like purple smoke and being like, I

7:17

will conquer the world. I

7:21

mean, like, go get it, Mark. Very

7:24

distinct image. But

7:26

you have to be that person to do this, I

7:28

think. Yeah, yeah. Yeah, and that's another

7:30

thing of all this that we'll get into. This is

7:32

like, I think this is a device that only Meta

7:35

could have built in a way that they

7:37

can show to the world because for a lot of reasons that

7:40

are very- Because everybody else would expect it to

7:42

ship. Yeah. Like everybody else does

7:45

hardware more often. So we'd be like, okay,

7:47

but when? I think if Meta

7:50

was not a founder mode

7:52

company to the fullest extent, this

7:55

thing would have been killed a long time ago. So can

7:57

I read you? I just want to read you a quote

7:59

from- Ben Thompson who writes a

8:01

newsletter called Shrekery. He tried on Orion

8:03

and he just said this thing that

8:06

it just made me start laughing a

8:08

lot. And I'll explain why it made me laugh so

8:10

hard. Here's the quote, the difference

8:12

from the quest and he is him talking about Orion.

8:15

The obvious limitations of display, particularly

8:18

low resolution felt immaterial. The difference

8:20

from the quest or vision pro

8:22

is that actually looking at reality

8:24

is so dramatically different from even

8:26

best in class pass-through that holographic

8:28

video quality doesn't really matter. Even

8:30

the highest quality presentation layer will

8:32

pale in comparison to reality. Yeah,

8:35

dude. I mean, this is very funny. He's presenting this

8:37

as like a, right? And it's Ben and that's how

8:39

he writes and that's fun. But if

8:41

you'll recall my review of the vision pro was

8:44

looking at display sucks. Like this is the

8:46

end of the road for video pass-through. If

8:49

you want to put a screen in front of your

8:51

head in a VR headset and then do camera based

8:53

pass-through, this is as good as it will

8:55

ever be. And it is nowhere near good enough. At

8:58

the time, many people disagree with me. I believe that

9:00

I have been proven to be correct about the value

9:02

of the vision pro over time. Still

9:06

not pleased with myself that I gave it a seven. Seven out of 10.

9:11

But like, this is the thing, right? Meta

9:13

built the displays, they grew the crystals, Zuckerberg

9:15

is like, I'm spending the money, I can't

9:17

be fired because I own super voting shares

9:19

of this company. Here it is. I

9:22

made the displays. When

9:25

you say it's not a product, did he make the rest of it? Is

9:28

the software good? There is an

9:30

OS. Right. So I guess I'll

9:32

just quickly explain how we got to where we are.

9:34

They started developing this about 10 years ago. This

9:37

is like Zuckerberg's big bet to maybe

9:40

control the next computing platform if you

9:42

buy into the idea that face computers

9:44

are maybe that. In

9:46

2022, it became, I think

9:49

everyone remembers the year of efficiency. Meta

9:52

stock was not doing as well. They

9:54

were cutting back budgets and they decided

9:56

looking at how expensive it was gonna

9:58

be to create a product. Orion, I'm

10:00

told that the cost to build a summer around

10:02

$10,000 per pair. Um,

10:05

and that the display stack, especially the

10:07

lenses, the silicon carbide was just not

10:10

something that would scale. They

10:12

decided to make it a prototype internally. And

10:15

then there was debate after that of, okay, do we show it

10:17

to the world at all? Is it good enough? And

10:19

I mean, a thing that stuck out

10:21

to me that Boz, the CTO told me was like,

10:24

we just, we just didn't think this was

10:26

going to work at all. We, when we set out to build

10:28

this, we thought maybe a 10% chance we actually get to a

10:30

working device. And I think

10:33

he's literally said, I was just like, holy

10:35

shit, it works. And

10:37

I think they were amazed that it works.

10:39

And so they started in earnest on the

10:41

software stack, like three months ago, deciding that

10:43

they were going to show it off at

10:46

connect. So it does have an

10:48

OS, you know, they

10:50

have pretty concrete ideas, especially on the

10:52

interaction elements of it, which we'll get

10:54

into the software is, is bare

10:56

bones, they've got some demo apps. They had

10:58

Instagram, you and I did a call over

11:00

messenger, Eli, um, they've got

11:02

a web browser, but these are like

11:05

the early primitives of how

11:07

you would bring 2d experiences into 3d space.

11:09

It's like a float. It's like vision pro.

11:11

It's like a floating pain. I

11:14

think the work they're doing now that they

11:16

have a working kit is in the

11:18

next couple of years. How do you actually make uniquely

11:21

3d augmented type

11:23

interfaces? Um, which is

11:26

like, you just have to have the thing on your face.

11:28

You have to be able to use it. And for the

11:30

longest time, this hasn't been something that

11:32

even resembles an actual pair of glasses. Like the, the

11:34

thing they showed me that it was even like in

11:36

2022, it had like a backpack,

11:39

I mean, it was just the iteration they'd done,

11:41

even in just the last couple of years is

11:43

pretty remarkable. Yeah. So yeah,

11:47

I, the software had

11:49

enough ideas in it to where I went,

11:52

okay, they have ideas of where this is

11:54

going. It's still super rough, but like the

11:57

AI stuff, especially it's compelling.

12:00

What all did you get to actually do in the

12:02

headset? Yeah, we did web

12:05

browsing, video calling, basic

12:08

kind of Instagram stuff, calling

12:11

in a 2D HD pane, which

12:13

like Neelai beamed in that way and I could see

12:15

him, he couldn't see me because there's

12:17

all these things in Orion that they just turned off

12:20

because they decided to not make it a product. Like

12:22

they have inward facing cameras that could potentially map your

12:24

face to an avatar to show someone and they just

12:26

have them turned off because it's not being used for

12:28

that now. Yeah, but they could also just look like

12:30

garbage. Yeah, so we did that.

12:33

Some meta employees came in as avatars,

12:35

like floating, like think of the horizon

12:37

quest avatar style. Legs or no

12:40

legs? Legs, legs and like full body scale

12:42

across the room. $10,000 a leg is what

12:44

I'm told. Yeah, and then

12:46

we did like a code, they

12:48

have their hyper realistic kind of

12:50

uncanny valley codec avatars, someone called

12:52

in as one of those as

12:54

well. And

12:56

then there were some games and the games were actually, you

12:58

know, I've done a lot of really kind of, I

13:01

would say just gimmicky AR games in

13:03

my career and the games were actually

13:05

decent, like the interactions because they've nailed

13:07

a lot of the interaction elements and

13:10

input elements of the glasses. The

13:12

game experience was actually surprisingly

13:14

good and they have some connected ones where like

13:16

you scan a QR code, you're in this Pong

13:19

game immediately with someone else wearing the glasses who

13:21

just scanned it. And that's how I did it

13:23

with Zuckerberg in the video. You can see on

13:25

all of the verges channels, he

13:27

beat me of course in Pong. Did

13:30

they have a laser tag game? That's

13:32

my AR dream. Yeah, that would be cool.

13:35

They had this kind of space invaders-esque game

13:37

where your head was moved the ship

13:39

and the band, which we'll talk about was

13:41

the lasers for the ship. So

13:44

here's what I'm curious about. They've built

13:46

a lot of stuff in the

13:49

quest, right? Like they've taken Android, they've

13:51

built an entire operating system, they

13:53

have a store, a lot

13:55

of that is complete. And I've always assumed based

13:58

on what Mark and others have said. is

14:00

that they're doing all that work there in that

14:02

form factor, and that lets

14:04

them build the software experiences. And

14:06

then they're building the Ray-Bans and the other

14:08

form factor. And then Orion

14:11

is the goal. Are

14:14

they using any of the stuff from the

14:16

Quest? Any of the

14:18

user interface gestures? Is

14:20

it Android? Yeah, it's Android-based. It's a

14:23

similar kind of app launcher UI. The

14:27

thing with the app launcher, though, that's different

14:29

is it's much more minimal, and your finger

14:31

gesture brings it up and then takes it

14:33

away just as quickly. They're

14:36

very different products. The avatar was

14:38

the same as what you would see on a Quest

14:41

game. But I'm just getting at it. Zuck

14:43

has laid out the idea that the

14:45

meta Ray-Bans are on one side of

14:47

the spectrum, and the headset is

14:49

on the other side, where this is not the

14:51

right form factor, but I can do everything in

14:54

it. And the Ray-Bans are the right

14:56

form factor. I can't do anything in them hardly. And

14:58

in the middle are these glasses. These products

15:00

are going to converge towards the glasses. It

15:03

just seems interesting to me that they haven't used

15:05

all the stuff from the Quest, because that was the plan

15:07

they articulated. Well, like I said, they started

15:10

doing the software for this three months ago. And

15:12

I will say the coolest thing, David, to your

15:15

earlier question that we did was this meta AI

15:17

thing, where they laid out all these ingredients for

15:19

a smoothie on a table. They

15:21

gave me the prompt, which means

15:24

they tuned it a little bit, right? There

15:26

was just a guy in the background. No,

15:28

I messed around with it to know that it

15:30

was actually, it was a model running, but- It

15:32

wasn't Tom Brady being like, all right, smoothie. Yeah,

15:34

but it was like, I asked it to make

15:36

a smoothie out of ingredients, and it popped up

15:38

like a recipe pane above me with, I could

15:40

like click through the different steps. Can I just

15:42

quickly be the turd in the punch bowl on

15:44

this particular demo? Because I've seen a bunch of

15:46

people talking about this particular demo, and I am

15:48

so spectacularly unimpressed with that demo, and I would

15:50

like you to tell me why I'm wrong. So

15:53

first of all, if you go to the challenge, every

15:56

single thing on that table says what it

15:58

is with big ass letters. And you know, it's

16:00

a really easy thing to do. It does. The bag

16:02

of dates says dates in big ass letters. And

16:04

the bag, the box of matcha says matcha in

16:06

big ass letters. Like those are not hard computer

16:09

problems to solve. It recognized the pineapple bit. If

16:11

you're listening to this in your car, I want

16:13

you to know that right now, Alex Heath is

16:15

scrolling through a picture to confirm or deny that

16:17

the bag of dates says dates on it. There

16:19

are two things on the table that don't say

16:21

in words what they are on

16:24

the box. It's the banana and it's the pineapple. And

16:26

it identifies the banana, which is again, a really easy

16:28

thing for a computer to do. And it misses the pineapple

16:30

entirely. Again, I'm just, I'm just putting this

16:32

out there. If you're listening, you

16:34

know, when you're on a Google meet call with

16:36

someone and their face instantly goes to, I'm browsing

16:39

the web face. I've never

16:41

seen anyone move to I'm browsing the

16:43

web face as fast as Alex did

16:45

when David said everything at the label.

16:48

Well, okay, you're right. So David. Big

16:50

box that says dates on it is

16:52

not a hard computer problem to solve.

16:54

David, so the pineapple it missed in

16:57

the video I did was Zuck. It got it

16:59

right the first day I did it. Okay, that's

17:01

good. That's encouraging. You still picked

17:04

the two most distinctive fruits that

17:06

you could have possibly picked. I'm just saying this

17:08

demo is the easiest possible version of this demo

17:10

and still it gives you a recipe that involves

17:12

not everything on the table and a bunch of

17:15

stuff you don't have on the table. So I

17:17

get to the end of that demo and I'm

17:19

like, what problem did we just solve? David, I

17:21

fully agree with your big point. Like, yes, there's

17:23

a lot of smoke and mirrors here. The

17:26

bar is also low. What

17:28

they were trying to communicate is like, oh, what

17:30

happens when you have the visual AI in the

17:32

Ray-Bans with a display? And so

17:35

what I'm talking about is not the fact that

17:37

it recognized everything right or wrong. It's just this

17:39

idea of like that

17:41

visual AI with a display and

17:43

then, oh, I can actually like, I don't have

17:45

to like just listen to the recipe. I can see it

17:47

and I can move it around and I can make it

17:49

big. That was cool. The thing in the demo where it

17:52

popped up above it saying what each thing was, I actually

17:54

thought was like, that's a really cool little bit of UI

17:56

that it was like, this is the cacao. Nevermind that there

17:58

was a box that said cacao right underneath the lip. But

18:00

still like I actually

18:03

thought the UI was very cool for what amounted to a

18:05

not particularly impressive problem. So

18:07

looking at it as like, here is how you

18:09

do this, I think is actually very clever. I

18:11

just was so like, there are

18:13

so many people were like, Oh my God, the smoothie thing. And

18:15

it's like, did we actually do anything here? Also

18:18

to make a smoothie, you just dump everything. I was going to say like,

18:20

yeah, it's blender. It'll

18:25

be fine. There's a technique here that the AI is

18:27

going to walk you through. You

18:29

hit pulse a couple of times and then you hit.

18:31

Shouldn't show this you make bread. So on the smoothie

18:33

thing, just really quick, I kind

18:36

of freaked out, not because of how crazy

18:38

I thought it was hit it correctly, identified

18:40

a pineapple the first time David, they

18:42

had been doing all these little Easter eggs for me throughout

18:44

my, I spent like a full day at meta and I'd

18:46

done a bunch of other demos before this. And

18:49

they were like dropping little things all throughout

18:51

my demos. Like the texts that came

18:53

up in the glasses was like an employee being like,

18:55

Hey, I've got the scoop for you on the next

18:57

AR glasses. I need to go like step

18:59

out of the room to like not get caught. And

19:02

then they, by the time

19:04

I got to the smoothie, I do, I

19:06

make smoothies every morning. And so by the,

19:08

I thought I was like, did they, they,

19:10

do they have cameras in my house? Like

19:13

I got to the point, I just turned around to the room. Cause

19:15

there's like 12 people in the room. And I'm just

19:17

like, are you guys like watching me

19:19

at home? To me on my phone, just

19:21

saying yes or no. Yeah. It was the

19:24

smoothie thing felt special. And so I was actually

19:26

very disappointed to see that everyone got the smoothie

19:29

demo, but that's okay. So

19:31

I want to come back to whether the demo

19:33

is real or fake and the debate about what

19:35

the word vaporware means. But you talked about the

19:38

control band, the wristband several times. Let's

19:40

explain that real quick first. Cause I, there's

19:42

a part of this where the, the

19:46

neural wristband is the most ready to go

19:48

product, which is fascinating in and of itself.

19:50

Kind of the coolest I think. Yeah. It's

19:52

done. It's one part of the dream for

19:54

sure. It's very clear that the band is

19:57

done. I also know this because they're releasing

19:59

a pair. of glasses with a very small

20:01

heads up display. It's not full AR that

20:03

will use the band. I'm pretty sure next

20:05

year. It was very

20:07

clear that the band is done. And that's why I spent more

20:10

time in the story than I originally thought writing about

20:12

the band. Also because it was

20:14

genuinely, I mean, I think I said

20:16

this in the video, it was one of the coolest experiences

20:18

with a new piece of technology I've ever had because

20:22

there's no calibration. You just, you put

20:24

it on, you go through the gestures

20:27

and it has this haptic feedback in the

20:29

band that kind of reinforces when you're doing

20:31

it correctly. I did that

20:33

for like 10 seconds and I was just flying

20:35

through the thing. Um, and

20:37

to the point where I had

20:40

to take them off at one point and put them back

20:42

on and like go back through the setup. And I didn't even,

20:44

I just did it on my own and like it just clicked

20:46

so fast and it's so precise

20:49

and yeah, it uses EMG electro

20:51

myography and it's not reading your

20:53

thoughts, but it's interpreting neural signals

20:55

through the movements of your wrist

20:58

and translating those into input in the

21:00

glasses and met a

21:03

metabata startup, um, that we have covered

21:05

extensively. I believe Addie covered this even

21:07

before they sold to, to Facebook. This

21:09

was like in 2019 control labs and

21:12

it's this tech and it's

21:15

really powerful. I think they've stumbled

21:17

onto something here that I

21:20

haven't seen for headsets, which is how do

21:22

you control them without putting your hands out

21:24

in front of you? You know, like I

21:26

use the spectacles which rely on just hand

21:29

tracking the week before. And

21:31

I was thinking when I was

21:33

like flying to meta the next week,

21:35

like in my airplane seat, it's like,

21:38

you're not going to stick your hands out like

21:40

in an airplane seat to control your glasses. North

21:42

glasses did that too, right? They had, they had

21:44

like a ring that would, they used to control

21:47

it. It was like a joystick on a ring.

21:49

It was cool. It wasn't cool. Yeah, but this

21:51

is like moving your hands

21:53

out in front of you. This is

21:55

like hyper precise click like input

21:58

that you can do with. very,

22:00

very small hand gestures in your pocket

22:02

or behind your back. And

22:04

I just, I

22:06

mean, I like audibly gasped when I started

22:09

using it. This is like cool

22:11

technology we've seen for ages, right? Well,

22:13

Control Labs has been around. Yeah, yeah,

22:15

Control Labs, we've seen like horrible

22:18

products from small startups using this stuff

22:20

before, but it's also like just regularly

22:22

used in prosthetics now, particularly

22:25

like hand prosthetics and arm prosthetics to

22:27

give that movement. So it's like, it's

22:29

really cool to see it come to

22:31

consumer products. Yeah, I've never seen EMG

22:33

in a consumer products. So

22:36

it was really cool. It was like one in 2014,

22:40

called like the frog or something. Does nobody remember

22:42

this? No. I swear

22:44

that there was a real product. You know that face when

22:46

I immediately started scootling? Yeah, we were like, that's my face

22:48

right now. No,

22:50

but Alex, the thing you said in your story

22:52

that most got me was the

22:54

moment where you're like, you had your hands in

22:56

your jacket pockets and you're controlling the thing. And

22:59

that's the moment where you're like, oh, this is

23:01

obviously the correct answer for

23:03

this. And I think I'm fascinated by

23:05

the whole kind of kit that

23:08

they put together where it's the glasses, there's

23:10

the wireless compute puck that feels like you

23:12

put it in like your jacket pocket or

23:14

your backpacker or your backpacker or whatever, and

23:17

then it's the band. And I think at

23:20

least for now, that strikes me as

23:22

like the

23:24

exact right three parts in the right

23:26

way. I think the Vision

23:28

Pro's cable thing is wrong. The

23:32

only hand motions that the cameras can see in

23:34

front of your face is wrong. I

23:37

honestly believe Meta got that part of this

23:39

thing exactly correct. And it just feels like

23:41

you instantly are using it in a way

23:43

that feels way more natural than even

23:46

like Nila's experience at the Vision Pro where you

23:49

still have to sort of keep your hands on

23:51

the front plane of your body and in view.

23:53

And that is just an amount of thinking that

23:56

you don't have to do with this and that

23:58

feels much better. Yeah, I totally. I

24:00

want to talk about the puck just I guess on

24:02

the band though too. What

24:04

meta hopes is that over time the band

24:06

becomes its own platform and that it

24:09

interfaces with appliances, it interfaces with

24:12

your car, other gadgets. And

24:14

the idea is like this is probably a 10 to 15

24:16

year out thing honestly for it to really, you know. It

24:19

interfaces with your car? Yeah,

24:22

so the idea is like you're walking around with

24:24

these glasses, you look at your car and

24:27

you want to start the engine as you're walking

24:29

up to it. You do a little tap because

24:31

it sees that you're looking at your car, the

24:33

band is connected, there's an API, you're logged in

24:36

and you turn your AC on or you turn

24:38

your... This is the idea of a man who

24:40

can grow crystals in the basement of his mansion.

24:42

Well, it's honestly, there's several leaps

24:44

to get there. It's not like the ecosystem

24:46

of it all will be tough to develop

24:48

until there's scale with these glasses, but

24:50

it's pretty obvious to think about, okay, if

24:53

I'm wearing contextually, spatially wear glasses

24:55

at all times and I'm walking

24:57

around my house and I look at my thermostat and

24:59

I want to change the setting, why

25:01

can my band not, because it's all connected,

25:04

why can't my band not adjust the thermostat?

25:06

And that's the idea they have. I think

25:08

it's really cool idea. Why can't I turn

25:10

reality into a bitmap display with your windows

25:13

and a mouse? Yeah,

25:15

with a camera and a controller

25:17

and good enough processing, you

25:19

can get a pretty long way towards

25:21

that. Yeah. So people will see

25:24

the band soon, like next year. And

25:27

I'm excited about that. I got to do an

25:29

update real fast. I found the band.

25:31

It was called the Myo. It cost

25:33

$200. They

25:35

first started reporting on it in 2013, came out in 2016, and

25:39

you wore it on the middle of your form to control your

25:41

laptop. And also

25:43

to keep this sweat away from you while you play

25:45

tennis. It's so ugly. Oh,

25:47

I remember this thing. I tried this thing. That thing

25:49

looks like a ball of angle. It

25:52

really does. And it uses EMG

25:54

currents? Yeah, yeah, it used EMG. That

25:57

particular technology, the EMG technology... has

26:00

been around a while, we've seen it in other

26:02

stuff, but it's always been kind

26:04

of, nobody's gotten the software part of it

26:06

right, and those interactions. And the form factor,

26:08

it looks like a fit. Yeah, like this

26:11

thing is not bulky, and it

26:13

just sits on your wrist. Did you have

26:15

to wear it tightly around your wrist? It's

26:17

tight, but it's not like, there are sensors

26:19

that indent a little bit into your skin,

26:21

but it wasn't uncomfortable. I wore

26:23

it, I left the demo room with it

26:26

still on to go interview Zuckerberg, and

26:28

they came running after me, and we're like, you still

26:30

got the band on. Keep running. Like,

26:32

please let us keep that. So if we're

26:34

talking about the three parts of this,

26:36

I guess on the puck, I just like,

26:40

David, I agree with you that this feels like

26:42

the right combination for now. They obviously don't want

26:45

the puck to be in the picture long term.

26:47

Sure, but I think we're a long way away

26:49

from that. Yeah, they made the decision, which I

26:51

think is the correct call, is that they want

26:54

the glasses to be as light as possible and

26:56

to not burn your face, which is what

26:58

I've experienced a lot, not literally burning, but just they

27:00

get really hot. And so

27:03

they'd made the decision to offload most of

27:05

the app logic onto the puck. They invented

27:08

their own wifi protocol, and it

27:10

has to be 12 to 13 feet max

27:12

away at all times where the glasses just don't work.

27:15

Why do we call it a puck and not a computer? I

27:18

mean, why do they call them glasses and not a

27:20

computer? Yeah, but like the- Is

27:22

it a circle? Yeah, it looks like

27:24

one of the big, iPhone

27:27

brick chargers you could like put on the

27:30

back of the iPhone with like MagSafe, like

27:32

a big ass anchor. Yeah, it looks like

27:34

you could put it in your back pocket, but you want

27:36

to wear a belt. I could

27:39

put it in my back pocket because I have huge back

27:41

pockets, but yeah, I mean, the idea is like- Are you

27:43

wearing Shinkos? Did we just learn something about flex? All

27:46

right. Yeah, you know. I know you live in LA,

27:48

but like, how bad has it gotten over there? It's

27:50

pretty bad. Wow. Yeah. Yeah.

27:53

So the puck is the right trade off

27:55

to make because the- These

28:00

glasses only weigh 98 grams, whereas

28:02

the Vision Pro is like five, six

28:05

X that. And like I

28:07

said, I wore them for two hours

28:09

and I never felt like uncomfortable. And

28:12

the funny thing about the two hours, I was

28:14

informed as we were wrapping, like,

28:16

you know, there's a bunch of men of employees in the

28:18

room. I think I was the

28:20

second journalist ever to try them. I've

28:23

also written about these extensively. So they were

28:25

really like on edge. And they,

28:28

towards the end, like I took them off and I could

28:30

just hear everyone like breathe a sigh of relief. And

28:33

they were like, you just broke the record for

28:35

longest demo. Actually longest

28:37

continuous wear of them at two hours. We were

28:39

like, yeah, we thought they would just like crap

28:41

out by now. So

28:44

that was- Was there a battery indicator or anything?

28:47

No, no, there wasn't. And they didn't swap them

28:49

out or anything. That's cool. Yeah.

28:52

Okay, you wear glasses. How heavy

28:55

versus the ones you're wearing now, which

28:57

seem fairly lightweight. Like you noticed, but like

28:59

I have a pair of chunky LA

29:01

glasses that honestly felt in the same

29:04

ballpark. Like it wasn't too

29:06

heavy. Cause like it was on the ears. It was just

29:08

those ear pieces are so good. No pressure on the ears. I mean,

29:10

if anything, it's a little in the front of your face. But, and

29:13

the thing that struck me was, you know, they

29:15

told me definitively, like the frames will be half

29:17

as thick in the consumer version, which means- Okay,

29:19

wait, that's it. That's my cue to our you.

29:21

There we go. There we go. Okay,

29:24

so just describe the product that you saw,

29:26

which is a very impressive demo, right? Lots

29:28

of people saw the demo. Everyone

29:31

is suitably impressed. Great.

29:33

The thing that they accomplished that

29:36

no one else has been able to accomplish

29:38

is getting that display to work in that

29:40

form factor. 98 grams on your face, two

29:42

hours of battery life. You can look at

29:44

reality through the lenses and it will put

29:47

from what I understand, labels above boxes of cocoa.

29:50

That's wonderful. Mark Zuckerberg can play along with you. They

29:53

had, right? They've iterated

29:55

it and developed the existing

29:57

EMG control band. We

30:00

put a computer in a giant battery and

30:02

are sending signals from it wirelessly to something

30:04

else. Challenging, because I got to do it

30:06

in real time, but

30:08

a thing that existed. We

30:11

put it all together in a product, and we sell it

30:14

far, far away. I'm

30:16

giving them credit for a bunch of stuff, some

30:18

stuff they iterated, some stuff they bought, some stuff

30:20

they had to build from scratch and spend billions

30:22

of dollars that basically no one else has managed

30:25

to crack yet. All great. It's

30:28

a long way from here to there. Three years. So

30:31

that's their promise. And so the

30:33

thing I would say is it's a vapor till it

30:35

ships. And Meta has the extraordinary

30:37

opportunity to show a thing

30:39

that isn't shipping. They're

30:42

all proud of it. You can see they're bursting

30:44

with pride. In

30:47

particular, built a display technology they can

30:49

demo for two hours with a variety

30:51

of people doing stuff that no one

30:53

else has been able to do. Magic

30:55

Leap hacking GPU of your brain,

30:58

I think, is the closest to this. And they

31:00

had people sitting down next to a box the

31:02

size of a refrigerator, and they couldn't ship it.

31:06

What is the, and

31:09

Meta doesn't have to do that, right? They're selling the Quest.

31:11

They make a bunch of money selling ads. They got Kristen

31:14

Bell being an AI voice. They have a

31:16

whole business that's running that can subsidize this

31:18

thing. It's

31:21

still vapor, in my opinion. They have to

31:23

ship it. They're just able to be confident

31:26

that they've gotten this far. But

31:28

it's unclear to me what this actually looks

31:30

like as a product, because I sincerely

31:32

doubt it will have a wireless compute puck. And

31:37

once they start shipping the neural wristband,

31:39

it won't take on a life of its own in some other

31:42

way. And it will be completely divorced

31:44

from the Quest ecosystem. There's

31:47

a set of unanswered questions here

31:49

that I think it's very tempting to pre-answer

31:51

or imagine. And that honestly is the fun part,

31:54

and I'm excited to do some of that with

31:56

all of you. But it's also like three

31:58

years a long time. a lot of things can change

32:01

in three years and a lot of

32:03

other things have to go right. Like the lenses

32:05

in your story, they're planning

32:07

to ship regular lenses, right?

32:10

They're gonna go from silicon carbide to glass. Well,

32:12

they wouldn't tell me the material. And the truth

32:14

is that they're parallel passing like

32:16

four or five different options. So

32:19

there- Right, but if your big innovation

32:21

is the display and you're

32:23

like, we have four options to bring this

32:25

display to market, you actually haven't picked. I

32:27

think you're being simplistic when you say display

32:29

because the display is the projectors, the wave

32:31

guides and the lenses. The

32:34

lenses specifically are what they

32:36

can't manufacture in a cheap enough

32:38

way at scale. So the lenses will be

32:40

different. The projectors are on the

32:42

path of what I saw going

32:44

to be the same kind of technology. They invented them

32:47

from scratch. The wave guides the

32:50

same. They wouldn't tell me what the

32:52

lenses are, not because I don't think they

32:54

don't know necessarily, but because they haven't

32:56

fully decided because they're locking these things

32:58

in on like a

33:01

timeline that is aggressive, but also

33:03

they don't wanna speak before they,

33:06

like if they would have said, let

33:10

me back up, like they decided to

33:12

not ship Orion two years ago. So

33:14

the timeline for this hardware stuff is

33:16

farther out. When they say, this is

33:19

coming in a few years, it's

33:21

not like we have not set anything up for this

33:23

to happen in three years. They have, they're on a,

33:26

believe me, heads will roll if there is not

33:28

a consumer version of these era classes in a

33:30

few years. Oh sure, but

33:32

those are the stakes. The stakes are either we're

33:34

all double clicking on cars, do

33:37

you know, see what happens? Or

33:39

heads will roll. Like I don't wanna shy away

33:41

from the stakes. The part

33:43

of me that says it's great to imagine how these

33:45

products might work, especially products that

33:48

are meant to deliver AI as a

33:50

new platform that replace smartphones, on

33:53

balance, I don't believe you know,

33:55

probably doesn't work is like the right answer. And

33:58

the fact that they got the display to work. Everything

38:00

that he ever described was Orion.

38:03

So I think Zuck loves the fact that he showed

38:05

off the thing that Apple has been describing for years,

38:07

well before Apple was ready to do it. I

38:10

do wonder, like, when

38:13

Apple's like, we need chips, they go to

38:15

TSMC and they're like, we're gonna buy all

38:17

the chips. We're gonna build

38:19

you a factory Corning, but

38:21

it's your, we don't wanna own it. We just wanna

38:23

build it for you so we have the capacity for it. You

38:25

run it. We're not putting that on our books. That's

38:28

the level that Apple makes investments at. And

38:32

I don't think they were able to pull off this form

38:34

factor. And even with the

38:36

Vision Pro, the form factor compromises are like

38:38

now in retrospect, like ludicrous, right?

38:41

Like the battery pack, all the stuff, right?

38:43

The hand tracking. I'm

38:45

actually, even the Orion, right? That

38:47

uses your eyes to the pointers still feels like a

38:49

miss to me. That's still overloading

38:51

an input within output. I disagree because of

38:54

the band. The eyes, when they're your drag

38:56

and not your click, it's,

38:58

and the click is not sticking hands up.

39:01

It works pretty well. Like I

39:04

thought the same thing with the Vision Pro. If

39:07

you try Orion at some point, I'd be

39:10

curious to hear how you think when you

39:12

try it because the eye tracking was surprisingly

39:14

good with the band. When you

39:16

were having like the conversation with Neelai, were you able

39:18

to also do other stuff at the same time? Like,

39:21

yeah. Cause I'm just curious like how that works if

39:23

you're like looking at him and then you need

39:25

to move something over here. And that means

39:27

you have to drag it. So you have to kind

39:29

of like look away from him in a potentially

39:31

awkward way. You could stick your hand out

39:33

and pinch it and drag it. Or you

39:36

could glance at it really quick, tap your

39:38

fingers together to. Oh, cause

39:40

it does hand tracking too. It does hand tracking. It

39:42

will merge hand tracking with all of it anytime

39:44

you put your hands up. That's smart. Well, the reason

39:46

I brought all this up is like Apple

39:49

solved, Apple punted on the big problem

39:51

and the display problem and then tried to solve a bunch

39:53

of the other problems. Or at least show you

39:55

what a bunch of the other problems could look like if they were solved.

39:58

Like hand tracking like windows. that

42:00

kind of look like the size and

42:02

shape of Orion. Like this stuff is just

42:04

out there. I just, as the verges foremost

42:06

phase computer expert, again, this is a cry

42:08

for help. This is not a brag. Um,

42:11

I actually will push back on you

42:13

guys. I do not think anyone has

42:15

anything working in this form factor as

42:17

a standalone kit that is not completely

42:20

on rails or something on a desk

42:22

that you look through. Yeah.

42:24

I don't think that actually exists. Can I just for

42:26

the sake of our own comments, can I just say

42:28

a bunch of nouns so that people, what?

42:32

How a lens X

42:34

real to pro air or whatever it's called the

42:37

HTC other companies.

42:39

Yes. I know other companies have made glasses before.

42:41

I know that the hall, Alex

42:43

and I have both changed spark plugs

42:45

wearing a hollow lens, like

42:47

the, the form factor innovation here is real,

42:49

like getting it down to glasses, I'm not

42:51

trying to take away from that. I'm just

42:53

pointing at like it's

42:56

vapor till it ships. Like it's one of our

42:58

rules. Right. Um, and the, the

43:00

distance from what we see today

43:02

to a product, all

43:04

of the questions I have to answer in the middle,

43:07

this is what I was asking at the quest is

43:09

like, they've answered a bunch of these questions with the

43:11

quest. I suspect they think that some of those answers

43:13

are the wrong answers actually. And they

43:15

need to go in a different way to enable some

43:17

of the things they want with true air glasses. I

43:19

think that's right. I also just think that before this

43:21

week, I honestly didn't

43:24

know where they were at

43:26

with this in terms of how

43:29

they were going to do it. And I feel

43:31

like after this week, they actually have the right

43:33

approach of how they're going about it, which is

43:35

just relentlessly focusing

43:37

on the form factor, which

43:39

meta has made a lot of hardware

43:41

mistakes over the years and has I

43:44

think just not had the right priorities in terms of

43:47

what they're building and what it's for, or even

43:49

understanding what it's for. It took them 10 years

43:51

really to understand that the quest is really just

43:53

a gaming device, even still. How

43:56

much do you think that's the Ray Ban lesson?

43:58

That's part of it. But I just, I think.

44:00

I think there's clarity here about

44:03

why people would want these. And

44:05

like something Zuckerberg told me was, I

44:07

thought pretty, um, illustrative of this

44:09

as like, we want them to be just

44:11

as good when they're off, because we know

44:14

that you're not going to have the AR

44:16

on maybe most of the day. And

44:18

so focusing on that

44:20

first and letting the technology then kind

44:22

of fill in, I think

44:25

is the right approach versus just how they've approached

44:27

hardware so far, which is like, let's just put

44:29

as much technology in something as we can and

44:32

see how people will use it. And here they're like, no,

44:34

we know how people will use this. It's

44:36

utility. It's like heads up,

44:38

lightweight interactions, video calling, um,

44:41

and AI. And the

44:44

form factor is, is primary.

44:46

And that's what like, that's the intro

44:48

of the piece. Like I walk in and I'm

44:50

looking at them, you know, I'm probably four feet

44:53

away and I, you can't tell, you can't tell

44:55

that they're AR glasses. And it's like, even

44:58

that in as a prototype, and

45:00

I struggle with this cause yeah, Neelai this is vapor until

45:02

it ships, but it's, it's also not

45:04

like, it's somewhere between a mirage

45:06

and a product. Like it's, it is real. Like

45:09

I touched it, I used it, but

45:11

it's not, it's not productized. So

45:13

seeing that and going, wow, like the form, like, and it's

45:16

going to be half as thick in a few years. Like,

45:19

okay. I finally see where this is going, especially with the

45:21

band. And I feel like before this week, even

45:23

I as someone who's been reporting on this a lot, couldn't

45:25

really see where they were going. And I

45:27

think they feel so confident in where they're going and

45:29

the clarity of what they're going to do that that's

45:31

why they shut it off this week. And also I

45:33

think just thinks it's really cool. Yeah.

45:36

And he's right. It is really cool. No one

45:39

else has done. I, again, I'm just not taking

45:41

anything away from, no one else has pulled off

45:43

the display. Like hundreds of millions of dollars, if

45:46

not billions of dollars have been thrown at

45:48

this problem of, can I put lenses

45:50

in front of your face that can convincingly

45:53

in 3d augment reality, not just show

45:55

you a TV like the X real glasses

45:57

do. Um, and not require

45:59

you to wear a. a whole lens and still have no field

46:01

of view, like billions upon

46:03

billions of dollars in the scenario, even thrown out

46:05

as problem. And only meta has allowed people to

46:07

wear it for two hours. And it's not an

46:10

overestimation to say that they have

46:12

probably spent snaps market cap on

46:14

Orion in the last 10 years. And

46:17

it's like no other company would do this. Like,

46:19

like Apple killed the car project because it's like,

46:22

we don't see a realistic

46:24

timeline to shipping. We're not

46:26

even sure what it is. Like, this is a

46:28

uniquely meta thing. And it's, it's like a

46:30

personal thing for Zuckerberg, which is just like,

46:33

screw Apple and Google. I'm never going through this

46:35

shit again on mobile,

46:37

you know, that he's gone through. Actually, hold right there,

46:39

Alex. I want to talk about Zuck versus

46:41

Apple and Google. And I want to talk to the Ray Bands a bit, but

46:43

we got to take a break. So let's take a break and then come back

46:45

and then we can talk about Zuck's war with

46:47

Tim Cook. We'll be right back. It's

46:52

time to review the top three highlights of the

46:54

day. I'm joined as always by my co-anchor, Snoop.

46:57

Hey, what up, though? Snoop number one has to

46:59

be getting the new iPhone 16 Pro with Apple

47:01

Intelligence at T-Mobile. Yeah, you should hustle

47:03

down at T-Mobile like a dog chasing a

47:05

squirrel chasing a nut. Nice analogy, Snoop. Dogs

47:07

do love to chase squirrels. And squirrels love

47:09

to chase those nuts. On the highlight number

47:11

two, at T-Mobile, families can save 20% every

47:13

month versus the other big

47:16

guys. That is very impressive. You know, y'all can

47:18

take some of those savings and buy some Snoop

47:20

merchandise. That's exactly what I'm planning on doing with

47:22

my saving, Snoop. Now take it away, Snoop. Get

47:25

to t-mobile.com and get the new iPhone

47:27

16 Pro with Apple Intelligence on them.

47:30

Now drop that jingle. The

47:36

24-month bill credit is eligible traded on Go 5G Next.

47:38

Well-qualified customers plus tax and $30 connection charge. Canceling entire

47:40

account before receiving 24 bill credits and credit stop and

47:42

bounce on required finance agreement to bill credits. And if

47:44

you pay off the vicerily, see how you can save

47:46

versus the other big guys at t-mobile.com/switch. Apple Intelligence coming

47:48

fall 2024. All

47:52

right, we're back. Keith, right

47:54

before the break, we're talking

47:56

about Zuck and you said he

47:58

doesn't want to go through this shit again. people

50:01

start using it and they ship a cheaper

50:03

one and these become, you know, they

50:06

start selling. Um, maybe they don't let met

50:08

his apps on there at all. And there's

50:11

this just real tension there of

50:13

who has the distribution control. It's

50:16

why Google pays billions of dollars to Apple

50:18

to have search and Safari. Whoever

50:21

controls how you access your thing ultimately

50:23

has leverage over you. And

50:25

Zuckerberg's been feeling that pretty acutely for 10

50:28

plus years. And so, yeah, this

50:30

is like a, this is a, as much

50:32

about like inventing the future as it is

50:34

correcting the past. And, um,

50:37

I don't think any company without that unique

50:39

context would be doing this. And

50:41

I think that, so that's a uniquely meta thing. Can

50:43

I just read a quote from the interview that I

50:45

thought was, was very telling. Um, you,

50:48

you, you guys spent a long time talking about basically

50:50

what AR and AI are going to do

50:52

to the next generation of gadgets. And then he, what

50:54

he says is, uh, for what it's worth. I also think that

50:56

all the AI work is going to make phones a lot more

50:58

exciting. Uh, you know, blah, blah, blah. AI is cool. If

51:02

I were at any of the other companies trying

51:04

to design what the next few versions of iPhone

51:06

or Google's phones should be, I think that there's

51:08

a long and interesting roadmap of things that they

51:10

can do with AI that as an app developer,

51:12

we can't like to me, that's the whole thing.

51:15

Right. He's like, if I am convinced that if

51:17

meta had its way, the puck

51:19

wouldn't exist on Orion, the puck would be your

51:21

phone. That's what those companies

51:23

want. And they just can't do it because the only

51:26

companies that are ever going to be allowed to do it

51:29

are Google and Apple. And, and like, and so if I'm

51:31

Mark, I'm like, oh, a gigantic part of what I am

51:33

doing should be tethered to this thing that I'm just not

51:35

allowed to touch. And that would piss me off. They

51:40

are so mad that they can't automatically sink photos

51:42

off your Ray-Bans into your iPhone camera

51:44

roll. They can do it on Android. It's super

51:46

annoying. It's so dumb. They can do it on

51:48

Android because the APIs are there. Android is a

51:51

little looser. IOS, no, like the reason the first

51:53

ones, the first Ray-Bans also had a lot of

51:55

issues. They had a

51:57

lot of Bluetooth pairing issues and like limitations with

51:59

what they have. that could do with Apple. And

52:01

Apple is because of the EU gradually being forced

52:04

to open up its APIs. But

52:06

yeah, they're pissed they can't sync photos

52:08

to a camera roll. But why doesn't

52:11

Meta complain more, I guess?

52:13

We saw this with Epic. They complain all the time.

52:15

No, no, but they complain all the time. We saw

52:17

Epic put their money where their mouth is. Does

52:20

Meta do that? It doesn't seem like

52:22

they do. Well, Meta's also monopolistic. Yeah,

52:24

I was like, is it because they

52:26

know that Lina will be like, yeah,

52:28

yeah, that's true. And

52:30

this is what I mean by the politics

52:32

of it all. And Alex, in your talk,

52:35

Mark, there's just

52:37

hard shots to the EU throughout your interview,

52:39

where he's like, they should figure out what

52:41

they want. And it's like, dude, what they

52:43

want is slightly better Bluetooth and headphones. They

52:46

want to open, interop between messaging

52:48

platforms. And so that is

52:51

bad for WhatsApp. Straight

52:53

up, I think they're trapped between, we would

52:55

love for you to kick open the doors

52:57

on iOS, but also don't do that to

52:59

any of our platforms that have the same

53:02

kind of restrictions. I mean, they're being forced to

53:04

do interop with WhatsApp. We just wrote about, you're

53:06

going to be able to message people outside of

53:08

WhatsApp. Yeah,

53:11

yes, there's truth to what you're saying. I would

53:13

argue Apple is still super locked down, and

53:16

it's opening up over time. But I don't

53:18

think they think Apple will open up soon

53:22

enough. And so it's like, we need to just invent

53:24

the next thing. Well, and I think my read of

53:26

this is that if you're meta,

53:28

the easiest way to get where you're

53:31

going would be to start all

53:33

of this cool AR stuff as an accessory to

53:35

your phone. And the reason that the Ray-Bans have

53:37

worked is that they are an accessory to your

53:40

phone. They're a vastly underpowered

53:43

accessory to your phone just because of what they're allowed

53:45

to do, but also the restrictions of the technology and

53:47

stuff. But the 1.5 step

53:49

of this transition you would want to be, your

53:53

phone becomes the computer, and then you

53:55

have a bunch of new wearable accessories

53:57

that you use. You're just not allowed

53:59

to use. to build that. And so

54:01

what all these companies have to do

54:03

is basically like engineer a giant societal

54:05

shift out of nowhere because the like

54:07

gradual change that actually would make this

54:09

make sense is just walled off to

54:12

all but two companies on the planet. And if I were

54:14

anybody but one of those two companies, I would be really

54:16

pissed about that. So what has Zuckerberg

54:18

done? He has gone and linked up

54:20

with the Italians and the French because

54:22

it's technically Elisaur, Luxottica,

54:24

the parent company of Ray Ban is

54:27

French and I'm glad, but what did you

54:29

think, Eli, of his comment? So they've just

54:31

done a 10 year deal with Elisaur, Luxottica.

54:34

The idea is that metas tech stack is

54:36

something Elisaur will be able to put into

54:38

any of its lines. It owns a ton

54:40

of brands, Oakley, Ray Ban. It's a monopoly.

54:42

It's a monopoly of eyewear. You and I

54:45

are both wearing glasses made by them right

54:47

now, presumably. Yeah. Yeah. Pull

54:50

over in your car. Look at your glasses.

54:52

Probably made by Xlora. They just bought Supreme.

54:54

So Supreme glasses. He

54:57

confirmed in our review that they're buying a steak in

54:59

Elisaur, Luxottica, which was news. But

55:01

his comment about he

55:03

thinks that Elisaur will be for

55:05

Europe and smart glasses what Samsung

55:07

was for Korea and phones. Incredible.

55:10

If you think about that, it's a good line.

55:13

It's a great line. It is

55:15

extraordinarily presumptuous in the best possible

55:18

way. Like if

55:20

you're at Samsung, you're like, yo, we

55:22

were Samsung before. We make ships.

55:28

Our nuclear reactor division is doing

55:30

just fine. So

55:33

that is confusing. But

55:35

what he means is Android is

55:37

the operating system enabled Samsung

55:40

to enter the mobile phone market. Samsung

55:42

is now even more Samsung than it was before. I would

55:44

point out, by the way, a Saturday

55:46

Samsung alert, if you are a virtualist, you

55:49

know that Samsung required all of its executives

55:51

to work six days in the office to

55:53

insert a sense of urgency and crisis into

55:55

the company. They don't make anything. So all

55:58

they can do is come up with deals.

56:01

They cut the price of one of their gaming monitors by 50% and

56:03

gave you not one, but two TVs for free if you

56:05

buy one. Incredible. It's on the website. Saturday

56:08

Samsung, everybody. You had to go on that tangent. You

56:10

had to. It's just incredible. Every

56:12

part of Saturday Samsung is better than last.

56:15

Makes me happy. But

56:17

what he's saying is there was a

56:19

hardware vendor in Samsung, and

56:21

Android came along, provided at the operating system and

56:23

the opportunity to go address the world market, and

56:26

now they are Samsung. Again, extraordinarily

56:28

presumptuous in the best way. Are

56:30

you saying that Est is

56:32

a hardware maker and something

56:34

is the operating system that

56:36

will let you go become

56:38

Samsung? Well, you're missing some

56:40

key components, right? Est doesn't

56:43

make chips. Samsung makes fucking

56:45

DRAM. Like the

56:47

hardware maker. They make OLED displays

56:49

and chips. The first

56:51

processor in iPhone was a

56:53

Samsung ARM processor. So

56:56

the industry was built on top

56:58

of Samsung from the jump. I'm

57:00

pretty sure I'm looking at these new transparent

57:03

Ray Bands they announced. I'm sorry, Tom Warren,

57:05

translucent Ray Bands. She

57:08

really ran into like, I work with a

57:10

bunch of nerds this week. Yeah. We're going

57:12

to go on speeds. And I think it's

57:14

a Qualcomm chip inside, actually. I'm just saying,

57:16

like all of that, like Samsung's latent capabilities

57:19

were brought together because

57:21

Google provided them an operating system that let them

57:23

become dominant in phones. Maybe Microsoft

57:26

could have provided an operating system. You remember

57:28

Samsung made a bunch of weird Windows phones

57:31

in the early days. But

57:34

does Luxottica have chips

57:36

and RAM? Or does it just have design

57:39

chops and retail distribution? I think that whole

57:41

comment was basically another

57:43

shot at the EU.

57:46

Because the big criticism of all the EU

57:48

regulatory stuff is the

57:50

United States, deeply

57:53

weird libertarian unregulated United States makes

57:55

the tech companies. And you make

57:57

the taxes. And all you have

57:59

is spot. which only ever complains about

58:01

not having access to iOS. And

58:04

that basically is the shape

58:06

of the complaint. We have European listeners are

58:08

gonna argue with me about something or the other, it's fine.

58:10

But that is the shape of the complaint from Silicon Valley.

58:12

And I think Mark is saying, look, I'll take one of

58:14

your companies and I'll turn them into Samsung, just get the

58:16

hell out of my way. And I

58:19

don't know that that comparison holds up because

58:22

again, Luxottica does own everything, but instead of

58:25

DRAM fabs, they own Supreme.

58:28

Yeah. It's just a

58:30

very different thing. But

58:32

potentially like as the proprietor of

58:34

the Neel eyes insane

58:37

theory of wearable bullshit. I'm

58:39

ready. I'm ready for this. If you're opening a

58:41

store, I'm walking right in. You have to acknowledge

58:44

that making the DRAM and owning Supreme might be

58:46

equally important in this. Like. Oh,

58:48

they are. They think that. You

58:50

can't sit here and tell me that

58:52

fiddliness and value are opposing things unless

58:54

you agree that looking good on your

58:57

face is important. And there are a

58:59

small number of companies that are very,

59:01

very good at making things that people

59:03

like wearing on their face. And

59:06

I think you could argue that a lot of companies

59:08

in tech have tried to like acquire

59:10

or develop that capability and can't do it.

59:13

There's no way Meta would have made these.

59:15

There's no way they would look anything

59:17

like these. It honestly, it might be easier to

59:19

learn how to make DRAM as a company than

59:21

to learn how to be cool. Yeah. Okay.

59:24

So you just clip that line and

59:27

just put it on TikTok and just let

59:29

it live its own life. I sincerely believe

59:31

that. The history of technology said that is

59:33

probably true. David Pierce, easier to make DRAM

59:35

than glasses that look good. I

59:37

mean, the ones that they wear, they got just a

59:40

little, right? The Orions, because they look like buddy Holly

59:42

glasses from the 50s. Which are actually

59:44

kind of coming back. Like that was another thing. That

59:46

was another thing Zuckerberg said is like, it's kind of

59:48

nice that chunky glasses are back in style. Like,

59:51

because. This is like a Zuckerberg long game,

59:53

right? Like. Yeah. He bought off

59:55

a bunch of fashion houses in the background. I was like,

59:57

make the glasses bigger. And like, you know, it's like that.

59:59

that scene from Devil Horse Prada. He turned all

1:00:01

the knobs on Instagram to make vintage cool again

1:00:04

and now it's back. I mean, he's right, he's

1:00:06

right. I was at the snap event the week

1:00:08

before and I saw someone wearing what I thought

1:00:10

were the new spectacles and they were Prada glasses.

1:00:13

I'm just saying, Cerulean was picked

1:00:15

in this room for you, Alex. That's what's

1:00:17

happening here. No,

1:00:20

but like, look, I'm saying, I

1:00:23

agree. I'm very excited to get my

1:00:25

clear Ray-Bans on Monday. They look so good.

1:00:27

They're pretty cool. I'm in the small size

1:00:30

and I have a huge head and I'm furious about this. You're 49 or

1:00:32

50 or 52. Whatever,

1:00:36

what's the bigger? Go bigger, figure number. 52 is

1:00:38

like, you know. 60, just as

1:00:40

big as they can be. No, no. 52 is

1:00:42

like, most brands are like, sorry. Yeah,

1:00:45

I know, which is why I can't. So you're probably a 50. I'm

1:00:47

like a 49, 50. These are

1:00:49

comfortable. Stop measuring my head, first of all.

1:00:52

I've just been enough with you in person to kind

1:00:54

of, you know, map it in my mind. I

1:00:57

have a side hobby of head measurement. It's

1:01:00

not problematic at all. I'm

1:01:04

just saying as big as they can get, and I know Bos

1:01:06

has an equally sized noggin. And I

1:01:08

know that he should have built the bigger ones. Because

1:01:10

the regular ones come in two sizes and these only

1:01:13

come in medium. But I'm getting them on Monday. I'm

1:01:15

very excited. It's the first time I've got them. I

1:01:17

think the clear ones look sick. I'm excited to do

1:01:19

all the stuff with them. The jump

1:01:22

that I'm curious about, right? Because now meta AI

1:01:24

is more conversational. It can

1:01:26

remember things. They've added some capability

1:01:28

to it. That's all running

1:01:31

in their cloud still, right? You talk

1:01:33

to the glasses, it talks to your

1:01:36

phone. It uses your phone's connection. They've got a

1:01:38

decent amount that's on device now. On the glasses,

1:01:40

on device, yeah. Because they're

1:01:42

doing with llama these super fine-tuned

1:01:45

distilled models. So the

1:01:47

latency on these is like pretty incredible in

1:01:49

terms of how quick the AI is. It's not humane

1:01:52

pen, wait a minute to set a timer. It's

1:01:54

like answer something in like

1:01:56

a second, right? Or like, tell me what's

1:01:59

in there. in this photo and it does

1:02:01

it in a couple seconds. And

1:02:04

I got the demo of the new AI stuff

1:02:06

that they're adding to the Ray-Bans. So did Kylie,

1:02:08

so did Jay. And it's

1:02:10

like, wow, okay, they're understanding

1:02:12

this form factor and what's unique about it. So

1:02:14

being able to like look at a phone number

1:02:16

and say call that phone number and it just

1:02:19

calls it and the call comes in on your

1:02:21

glasses because it's paired to your phone or scan

1:02:24

that QR code like you're looking at a

1:02:26

menu, the webpage for the menu just gets

1:02:28

sent to your phone. Breaking

1:02:31

down some of the interactions

1:02:33

you would do with a phone that require

1:02:35

extra steps and just like taking a step

1:02:37

away is like kind of an

1:02:40

aha moment when you do it. You're like, oh,

1:02:42

like this is where it's going. Yeah,

1:02:44

and it's the first time that I

1:02:46

feel like AI makes sense in a wearable.

1:02:50

And did you guys watch the

1:02:52

keynote in the live translation demo that Zuk did?

1:02:54

Yeah, that's when he brought out the MMA fighter.

1:02:56

Yeah, he said, yell some words at me in

1:02:59

Spanish. That is wild, that is

1:03:01

wild. That is like Putin's earpiece

1:03:03

on stage, like at scale for

1:03:05

like anyone and

1:03:08

it works. I mean, it was a

1:03:10

thing like Google promised a couple of years ago, right?

1:03:12

Yeah, I think the really interesting part of this is

1:03:16

these ideas are all just the ideas.

1:03:18

Like these are the demos we have

1:03:20

been hearing about or promised for a

1:03:22

decade, if not more, right?

1:03:25

Live translation, like Apple has live translation on your iPhone,

1:03:27

you just open the app and it'll do it. Google

1:03:30

can do it. It's the form factor, like

1:03:32

you're saying, Alex, that like, okay, we've put

1:03:34

this in glasses, your phone is away, there's

1:03:36

no screen, there's just an ambient computer, another

1:03:39

thing that has been promised for a billion

1:03:41

years and it's just paying attention and helping you

1:03:43

out as you go through your day. Does

1:03:46

that ambient computer require sending an awful

1:03:48

lot of data to Meta? It sure

1:03:50

does, right? It just really, really

1:03:52

does. But that's the trade off. All the

1:03:54

way down to one day, I'm

1:03:56

guessing Orion will have my dream feature, which is

1:03:58

I will look at some. and it will tell me

1:04:00

their name. And that will absolutely

1:04:02

require Meta searching its gigantic worldwide surveillance

1:04:05

database of names and places. But like

1:04:07

they're gonna build that. What's the technical

1:04:09

name? I was waiting for you to

1:04:11

bring this up in my Orion demo.

1:04:15

The VP of wearables at Meta

1:04:17

said, we're excited about name

1:04:19

tags. Yeah, it

1:04:21

is the feature. If you can do name tags, it

1:04:24

doesn't matter. I will wear the backpack. Yeah, right?

1:04:26

Like all day long. Everybody at CES. I

1:04:30

wanna double click on what you said, that the

1:04:32

phones have live translation because I think this is

1:04:34

important when we're talking about like these form factors.

1:04:36

Yes, they do. Like especially iOS 18, it's really

1:04:38

easy to just do like live

1:04:40

language translation from the action button. There's

1:04:44

something about not having your phone out though

1:04:46

that feels like, oh, I would actually use

1:04:48

this. Think about like talking to

1:04:50

someone in another language and holding

1:04:52

your phone between you. It's like an

1:04:54

interview. Like it's just awkward. But it's like, oh wait,

1:04:57

when this isn't a form factor where like they can't

1:04:59

even hear that what they're saying is being translated

1:05:01

to me in my language and

1:05:03

I'm not even pulling a device out. Like

1:05:06

all of a sudden these features make sense as like something

1:05:08

you would use in the world. And I

1:05:10

think that's what these mean. Like these, it's like

1:05:12

taking all these concepts and like putting in a

1:05:14

device that like, oh, like you'd actually do this.

1:05:17

Yeah, cause they are cuter than the Pixel Buds. No

1:05:20

disrespect. Well, so here, these are,

1:05:22

they're just, again, the ideas, the

1:05:25

end state, the goals, very

1:05:27

familiar. Everybody shares them. The

1:05:29

demos as described, very

1:05:32

familiar. Everybody has the same ones. It's

1:05:35

really down to where is the form factor? How fast

1:05:37

can we get there? Can we ship it at scale?

1:05:39

Are the phone makers, the operating system vendors gonna

1:05:41

get in our way, which is a big deal.

1:05:44

And you're kind of like, okay, well, the closest

1:05:46

Apple is fundamentally is AirPods. Yeah.

1:05:50

Right? Like fundamentally the closest they are

1:05:52

to this kind of idea is AirPods. They haven't put

1:05:54

glasses in their face yet. They have an Apple Watch

1:05:56

and they don't really do any AR stuff. Yeah.

1:06:00

You can tap away with an Apple watch. I

1:06:03

wear my metal ray bands. The number one feeling

1:06:05

I have is boy, I wish this actually tied

1:06:07

in with my phone. So I could use Siri

1:06:09

because then I would have access to all my

1:06:11

stuff. Yeah, Siri could just like,

1:06:14

you know, like text the wrong person at

1:06:16

any point. I could just play my music.

1:06:19

Okay, so that's AirPods for Apple.

1:06:22

Alex, you were just bringing up the Pixel Buds. Yeah. That

1:06:25

is Google's, although does Google remember that it

1:06:27

made the Pixel Buds? It made the Pixel

1:06:29

Buds too, which... Sure, Neelai. Yeah,

1:06:31

the review came up the same time,

1:06:33

the metal event. So Neelai,

1:06:35

AirPods, great example. With my AirPods, I can

1:06:37

say, hey, remind me about whatever. I can

1:06:39

say whatever I want it to remind me

1:06:41

about at a certain time, right? And I'll

1:06:43

get the notification and the reminders app. With

1:06:45

the ray bands, you can say, hey, look

1:06:47

at this and remind me about it at

1:06:49

whatever time. Right, because it can see. Yeah,

1:06:52

that's like, that's wildly useful to me. I cannot

1:06:54

tell you how many times throughout the day, I

1:06:57

just wanna have a snapshot of whatever I'm looking

1:06:59

at and go back to it later. And

1:07:02

those ideas, because you literally have a camera

1:07:04

on your face, you can't

1:07:06

do that with AirPods. And so you're right, the

1:07:08

closest we have are these earbuds that are internet

1:07:11

connected and can hear our

1:07:13

voice. There's something about adding the cameras though,

1:07:15

that's big. So the prevailing theory

1:07:17

of why the camera control button is not

1:07:19

just on the pro phone, but also on

1:07:21

the regular 16, is that Apple

1:07:23

at visual intelligence, and then you'll be able to

1:07:25

quickly access the camera and look at stuff. When

1:07:28

you say theory, you mean Apple just said that

1:07:30

out loud in that many words. Well,

1:07:32

they haven't shipped yet. It's vapor till it ships. No, but

1:07:34

I mean, whether it's good

1:07:36

or not is a question, but that is clearly, that

1:07:38

is half the purpose of the thing. They have just

1:07:40

said as much. At this point,

1:07:43

my theory, whether Apple has its own

1:07:45

unified theory of what it's doing, it's

1:07:47

like, sure. I've used Control Center and

1:07:49

iOS 18 enough to be like, did

1:07:52

anybody pay attention to that slide? But

1:07:54

also there are increasingly convincing and well-sourced

1:07:56

rumors about things like AirPods with a

1:07:58

camera. which will be that look like

1:08:00

and an Apple watch with a camera,

1:08:03

because I agree with you Heath, like the camera is

1:08:06

the thing, both

1:08:08

as the input and as like the activity,

1:08:10

right? Like I hear from people all

1:08:12

the time for whom the reason to have the smart glasses

1:08:14

is to take pictures of your kids, like or your pets

1:08:16

or whoever, like I'm convinced my dog knows when I take

1:08:18

out my phone to take a picture of her. I'm a

1:08:21

hundred percent sure and she runs away and now she doesn't

1:08:23

anymore, I can take a picture of her, it's awesome. I

1:08:25

can't wait to get these glasses. Did you buy the glasses

1:08:27

too David? I already have two pairs. I

1:08:30

don't have a great answer for why I have two pairs. Do you

1:08:32

have like the large size or the small size? Because if you don't

1:08:34

need the large size, I'm in the market. I

1:08:37

have, I don't know, I also have a big

1:08:39

head. So we'll have to figure this out together.

1:08:42

Alex, how big do you think David's head is? We're getting

1:08:44

49? I'm gonna guess he's like a 48, 47. Neil

1:08:49

has like a 65. I'm saying, the

1:08:51

better. It's not insulting to me,

1:08:53

80, let's go. The bigger, the better. All

1:08:56

right, so we talked a lot

1:08:58

about the meta stuff. I

1:09:01

wanna make sure we talk about two other pieces of

1:09:03

Meta Connect that are important and

1:09:05

then spend a little time in OpenAI before

1:09:08

we go to the later end. The other piece

1:09:10

of Meta Connect that I think is deeply fascinating is,

1:09:14

you guys did talk about AI a lot and

1:09:16

how it will be expressed. It seems like they're

1:09:18

chasing a big distribution advantage with glasses that maybe

1:09:21

other people will catch up to. But then

1:09:23

there's like training data and open source

1:09:25

models. Zuckerberg said,

1:09:27

and I think this is one of the shots

1:09:29

at Google. Like I think all of the, we're

1:09:31

doing live demos, we're shots at Apple for doing

1:09:33

infomercials. I think when he says things

1:09:36

like, we are the Linux of open source, because

1:09:38

llamas, like that's a shot at Google, right? We're

1:09:40

just gonna have more distribution, we're gonna have more

1:09:42

development, we're gonna be better at everybody else because

1:09:44

open always beats close, which is usually Google's model.

1:09:47

And that's just not how Google's doing it. They

1:09:51

had to suck up a lot of training data

1:09:53

and you asked about people wanting to opt out.

1:09:56

There's all the celebrities on Instagram right now saying

1:09:58

opt me out of this. His

1:10:00

answer, explain his answer, then

1:10:02

we should talk about it a little bit, because it was really interesting. Yeah,

1:10:05

he said the quiet part out loud,

1:10:07

which is that everyone

1:10:09

thinks their data is more valuable than it

1:10:11

actually is for these models. In aggregate, it's

1:10:13

valuable, but when

1:10:18

you're a company like Meta, and it's

1:10:20

like the entire world is your corpus

1:10:22

of information and the internet at large,

1:10:26

what exactly do you have to have? And

1:10:29

I think we were talking a lot about

1:10:31

publishers and about how Meta's basically thrown its

1:10:33

hands up on news. And

1:10:35

Rupert Murdoch is in Australia saying, pay

1:10:37

me or I'm gonna take your news

1:10:39

off. And Zuckerberg's like, fine, whatever. People

1:10:41

don't even like this. So

1:10:45

we were talking a lot about publishers, but I

1:10:48

think he's kind of just saying that,

1:10:50

look, yes, people have feelings about their

1:10:52

data being

1:10:56

used to train these models. And

1:11:00

yes, we're all doing it, but we

1:11:02

also, if you don't like it, go

1:11:04

away, whatever. So

1:11:08

it's really interesting to me about this is

1:11:11

that's the lines to the publishers. We

1:11:13

will derank your news, we won't have news on

1:11:15

threads however you want to interpret all

1:11:17

of the many things Adam Osteria said about that. There's

1:11:21

no opt-out for any user of these services,

1:11:23

unless you're in the EU from having yourself

1:11:25

trained on. And guess where Meta AI is

1:11:27

not available in the EU. Right.

1:11:30

And so, and he did say, I remain what, eternally

1:11:32

confident that we will launch this in the EU. Something

1:11:35

like that. On stage. But there's just,

1:11:37

I would just draw a connection. I think it

1:11:39

is very funny to dunk on

1:11:41

people posting what amounts to chain letters on

1:11:43

their Instagram stories, being like, dear

1:11:46

Zuckerberg by the statute of Rome,

1:11:48

I command you from not. We were doing

1:11:50

this on our Facebook walls like 15 years

1:11:52

ago. Yeah, this keeps happening. This is a

1:11:54

chain letter. Like that's what it is. And

1:11:56

it's like, people think the law is magic

1:11:58

words. and he just

1:12:00

like issued an incantation to the internet. Like someone

1:12:02

has to do with his sentence. Usually

1:12:05

chain letters have a lot more like

1:12:07

promising death than these. Yeah. And

1:12:09

I really miss that. I wish if it was like send

1:12:11

this to 12 of your friends or Zuckerberg will appear in

1:12:13

your house in his Latin shirt and beat the shit out

1:12:15

of you. That would be great. Have

1:12:18

you ever fallen down the whole of courtroom

1:12:20

videos of people invoking the

1:12:23

sovereign citizen defense in

1:12:25

like local courtrooms? And then they're

1:12:27

like, shut up. And the judges are like, that

1:12:29

doesn't mean anything. You're going to

1:12:31

jail. They're like, by what authority?

1:12:33

And the judge is always like, because I'm the

1:12:36

judge. And it's just like that's

1:12:38

happening on. And it is very funny. It's Michael

1:12:40

Scott on the office yelling, I declare bankruptcy. It is

1:12:42

exactly that. And then somebody pulls him aside and says,

1:12:44

that's nothing. And he goes, I didn't say it. I

1:12:47

declared it. That's what these people are doing.

1:12:49

So it is very funny. And Tom

1:12:52

Brady, who has a meta deal to use

1:12:54

his face for a chat bot, has done

1:12:56

it. He did. They

1:12:58

canceled the deal because that idea was bad. And

1:13:00

they came up with a better idea of what

1:13:02

if we just used the celebrities' voices and said

1:13:04

they were the celebrities instead of pretending

1:13:07

they were other characters? Yeah, it was Tom

1:13:09

Brady as like Barry, your workout partner or

1:13:11

something. Not great. But

1:13:14

it was his face. And then Kristen Bell, you

1:13:16

can just chat with Kristen Bell. She put up one

1:13:18

of these that Kylie noticed because she's a Kristen Bell

1:13:20

fan. And so she's like, you have a deal with

1:13:22

meta AI and you have an Instagram post saying, don't

1:13:24

use my stuff for meta AI. That

1:13:28

was before she saw the zeros on the check. Yeah,

1:13:31

money. Money is cool. Yeah, they're paying for

1:13:33

these celebrity voices. They're

1:13:35

not paying for the rest of us. So

1:13:37

here's what I will just say about that thing.

1:13:41

The idea that you can

1:13:43

negotiate with meta feels

1:13:45

intuitive to everyone, right?

1:13:48

Or any platform. I'm

1:13:50

posting my stuff here. I should be able to tell

1:13:52

you how I want you to use it. But

1:13:55

you don't. And the answer is, well, you already signed the Terms of Service

1:13:57

and you go away. But everyone knows

1:13:59

nobody reads. that shit. When

1:14:02

half of your users are like collectively

1:14:05

trying to renegotiate the terms of their

1:14:07

agreement with you, that means there's a

1:14:09

problem. Like kind of they're doing it

1:14:12

on your platform though. That's the thing.

1:14:15

Like, well, that's fine. But like

1:14:17

you can't, there's, there's so other

1:14:19

things in the world where like

1:14:22

you don't get to negotiate on this

1:14:24

level except for weird internet terms of

1:14:26

service that everyone agrees no one reads.

1:14:29

Right? Like, and I would just say like, if

1:14:32

you are one of the many, many, many young

1:14:34

staffers in DC who listens to the show, which

1:14:36

is a weird audience that we have and I

1:14:38

love you. I'm glad that

1:14:40

we're making the commute to your basement office in the

1:14:42

capital better. Um, you

1:14:44

should look at all the people posting.

1:14:46

I don't want you to train on

1:14:48

my data on metas own platform and

1:14:51

the basic response of, yeah, go

1:14:53

fuck yourself and be like,

1:14:55

Oh, this is actually what regulations are for. Right?

1:14:58

Like the people are expressing that

1:15:00

they do not like that the

1:15:02

terms of this arrangement are not

1:15:04

in their favor and no

1:15:07

one as an individual has enough power to

1:15:11

change it. And Mark Zuckerberg is

1:15:13

like, everyone overvalues the vote. They're working this

1:15:15

context. So I don't have to pay attention

1:15:17

to it. Like that is actually

1:15:19

the problem that like, it doesn't

1:15:22

matter if you're a Republican or a Democrat or a

1:15:24

libertarian or whatever, that thing is

1:15:26

the problem that governments are meant to

1:15:28

equalize. How is this any different from

1:15:30

using data for ad targeting? Think about

1:15:32

when we copy and pasted protest messages

1:15:34

on Facebook 15 years ago, what was

1:15:36

it about? It was there. I'm putting

1:15:38

this all in the same line. Right.

1:15:41

This is all the same line. If my

1:15:43

point is when everyone is like, I signed

1:15:45

this agreement, I didn't read it. Right. And

1:15:47

now I've handed over whatever leverage I might've

1:15:49

had because I signed this agreement that I

1:15:51

didn't read. And the second I made aware

1:15:53

that I, that by saying,

1:15:55

abracadabra, this is like, go back in

1:15:57

my favor. I will just hapless. start

1:16:00

yelling African Dabra as loudly as I

1:16:02

can. Like that is, that's the problem.

1:16:05

Yeah, but that's also just like no one

1:16:07

reads Yula's and that's, that's. But like, so

1:16:09

who should read the Yula for you? Should

1:16:11

we hire, and you can't do AI. And

1:16:13

then what? Do you have

1:16:15

bargaining power if you disagree with the terms?

1:16:18

No. If the user, I do use

1:16:20

the Instagram. Using Instagram's terms of service and ships

1:16:22

it to you. Are you like, you know, Tim,

1:16:25

paragraph five, couple notes. Are

1:16:27

you saying that social media needs to unionize?

1:16:30

I am saying terms of service agreement should

1:16:32

be illegal. Like flatly. Because

1:16:34

they're a contract. Christen Bell for president. They

1:16:36

are contracts no one reads. And it's these

1:16:38

moments, I just have a lot of empathy.

1:16:40

I have written the don't

1:16:43

believe the Instagram meme story five

1:16:45

times in my career. And

1:16:48

one time I got a call from Instagram, and I'm like, thank

1:16:50

you so much. And I was like, I just, I don't like

1:16:52

this. Because

1:16:54

it happens over and over again for a

1:16:56

reason. Which is whenever you

1:16:58

offer somebody back the mechanism

1:17:00

of control over their own information. They're like,

1:17:02

yeah, I'd like that back, please. Yeah.

1:17:05

Universally. Yeah. And

1:17:07

no one reads the agreements. The agreements change all the time.

1:17:10

The agreements always change in favor of the platform using your

1:17:12

data for more and more and more things. And

1:17:14

then you get Zuckerberg saying, everyone

1:17:16

overestimates the value of their individual data. And it's

1:17:19

like, who? What is

1:17:21

the mechanism that you would rebalance

1:17:23

this equation with? There isn't

1:17:25

one. There isn't one. And I think Zuckerberg's also

1:17:27

thinking, yeah, guess what? You're using Instagram and you're

1:17:29

gonna keep using Instagram. He's not incentivized to say,

1:17:31

actually. Yeah, we should deliver more value to you.

1:17:34

Yeah, he doesn't care. He's like, yeah, give me

1:17:36

all of your data for free so I can

1:17:38

make lots of money and afford my cool shirts.

1:17:40

Like, of course he is. Look, I'm not saying

1:17:42

we all shouldn't dunk on a bunch of people

1:17:45

posting, advocate, abrac to Instagram. Like, we absolutely should.

1:17:47

We should. It

1:17:49

is one of the funniest repeat memes

1:17:51

that can possibly exist, is

1:17:54

people basically doing sovereign citizen in Instagram.

1:17:56

I'm like, very funny. But if

1:17:58

you just, I'm just trying. But one step back

1:18:00

from a place of empathy, it's

1:18:02

a bunch of people communicating that they don't

1:18:04

like the terms of the agreement. And there's

1:18:07

no mechanism to channel that into any change.

1:18:09

That agreement is not changing. I fully agree

1:18:11

with you. And that's how I frame it

1:18:13

to him. I'm like, do you sympathize with

1:18:15

this icky feeling people have about this trade-off

1:18:17

and the sense that value is not flowing

1:18:19

back to them? And he said

1:18:21

the thing about how everyone overvalues their data. You

1:18:24

can run. Well, there's also a sense of,

1:18:26

yeah, if you feel this way, you would

1:18:29

stop using the product. We were doing this

1:18:31

about ad targeting and our data being used

1:18:33

for ad targeting 15 years ago and

1:18:36

copy paste messages on Facebook. Everyone's

1:18:39

still using, well, not everyone. 3

1:18:41

billion people apparently are still on Facebook. So

1:18:45

for them, it's like the platforms, they

1:18:48

don't see any data to

1:18:50

suggest that this value exchange is actually

1:18:52

unfair because in their mind, we would

1:18:54

stop using the services. Yeah. Yeah,

1:18:56

and that just presumes that people have that power and

1:19:00

can afford to do that. What do you mean? You

1:19:02

can delete Instagram? Yeah, I would say these things are

1:19:04

businesses for a lot of people, right? Like this is

1:19:06

a way they advertise their businesses. This is a way

1:19:08

they advertise themselves if they're influencers.

1:19:10

But most people in the

1:19:12

Verge's newsroom are no longer posting on X. People

1:19:15

can make decisions about the apps that they post

1:19:17

on. People can. Even when they have career implications.

1:19:19

I think there's also a lot of journalists who

1:19:21

are still on X. I mean, I think there

1:19:23

is a power imbalance there that Neelai is speaking

1:19:25

to of, these people have no power

1:19:27

in this relationship. It

1:19:30

is if you wanna use this, suck

1:19:32

it up. And I think people slowly get tired of that.

1:19:35

Right, I think it's just that there

1:19:37

are very few other relationships you have.

1:19:39

If you are a creator,

1:19:41

and creators are mostly just business people at

1:19:44

this point. They're all running content businesses on

1:19:46

the terms of the platforms. If

1:19:48

you're a celebrity, you're just a content business on

1:19:50

the platform. Which is why I think the

1:19:52

celebrities are so fast to be like, don't use my stuff. They

1:19:55

know, they do not want their voice and likeness to

1:19:57

use without payment. All

1:20:00

I'm saying is from a place of empathy, one step back, there's

1:20:03

no mechanism to channel this frustration

1:20:05

into anything. And that is

1:20:07

actually, again, I don't think

1:20:09

you have to be very political on either side of

1:20:12

the spectrum. It's like, oh, this is actually the thing

1:20:14

we should, this is what it's for, right?

1:20:17

A lot of people, like most people in the

1:20:19

society would like to renegotiate the terms of their

1:20:21

agreement with the large platforms. That

1:20:23

is called a privacy law, right?

1:20:26

Where we're just gonna reset the floor of the agreement.

1:20:28

A lot of people wanna reset the terms of how

1:20:31

their content can be used for training. That's

1:20:33

just an AI training law. Like

1:20:36

that, it's just very simple. And

1:20:38

like, I don't know what those laws

1:20:40

should do or how they should read or whether

1:20:42

you are a conservative and

1:20:44

you think the answer is like,

1:20:46

I don't know, some weird public-private

1:20:49

partnership that the Heritage Foundation

1:20:51

runs. Like those are those ideas. Or you're a

1:20:53

hardcore liberal and you're like, I will start a government

1:20:55

commission and we will, the government

1:20:58

will set the rates, which is what

1:21:00

we do in copyright law. Like there's a million solutions to this. And

1:21:02

I think that's a very liberal solution in copyright law. I'm just

1:21:05

saying there are government entities that set rates in other parts of

1:21:07

this world. But it's

1:21:09

just, I would just point like, we're

1:21:11

talking about AI and distributing it and who

1:21:13

has the power, where the models run and

1:21:15

how powerful the experiences can

1:21:18

be that you had, Alex. And

1:21:20

underneath it all is like, hey,

1:21:23

do all of these people feel ripped off? Is

1:21:26

that okay? And Zuckerberg's answer is Zuckerberg's answer.

1:21:28

I asked Sundar Pichai this about YouTube in

1:21:30

the same kind of framework of a question.

1:21:33

How do you feel about OpenAI training on

1:21:35

YouTube? And he was like, I

1:21:37

think that would be inappropriate. And it's like, do you understand

1:21:40

why all of your creators think that's inappropriate? And he just like,

1:21:42

sighed at me. And I

1:21:44

suspect something else is gonna happen there.

1:21:47

Like some set of lawsuits or some

1:21:50

Scarlett Johansson OpenAI situation

1:21:53

where it just becomes less and less tenable to

1:21:56

be this blithe about it. We

1:21:59

should get off the line. point because we could do this for hours

1:22:01

and we'll end up talking about the feta verse and we'll be a whole thing.

1:22:03

Oh my God. Let's do it. But the

1:22:05

thing is no one has

1:22:07

any reason to believe that any of what you just

1:22:10

said is true. And what's actually happening is everybody is

1:22:12

now saying the quiet part out loud. Eric Schmidt is

1:22:14

out here being like, Oh, you want data from the

1:22:16

internet? Just steal it. Mustafa Suleiman was like, Oh, it's

1:22:18

all free. It's on the internet. You can just have

1:22:20

it. Like this is what everybody thinks. And there has

1:22:22

been absolutely nothing to convince them

1:22:24

otherwise so far. So if I'm Mark Zuckerberg, why

1:22:26

I'm going to look around and be like, Oh,

1:22:28

I'm going to be the good guy here. And

1:22:31

it's going to cost me in this race

1:22:33

to AI. That is suddenly the only thing anyone cares about.

1:22:35

Like, no, of course he's not going to do that. He's

1:22:38

going to say, Oh, you like Instagram? Keep using Instagram. We

1:22:40

can do this for another hour, but also his new attitude,

1:22:42

which is like, I no longer apologize for everything. Apologizing was

1:22:44

a 20 year mistake that I'm not going to make anymore.

1:22:46

No, dude,

1:22:51

you grow the hair out. You get a little older, you know,

1:22:53

like I'm thinking about growing my hair out.

1:22:55

I watched that interview. I was like, Oh, I could get

1:22:57

the, is it the cauliflower hair? Is that what they call

1:23:00

it? Look, when every, every 12 year old in your boy

1:23:02

is gifted a gold chain and you have to make a

1:23:04

decision and I made one decision

1:23:06

and I'm saying I could remake that decision. You could.

1:23:08

The hair is getting long. You could, yeah. I

1:23:11

feel like you're like, you're a few months away

1:23:13

from curls. Oh, this hair is very

1:23:15

curly. I could tomorrow. We're not talking

1:23:17

about this right now. We are changing the

1:23:19

subject. So French and we need to go

1:23:21

to a break. Can we just before we get a break, I

1:23:23

do want to talk about open AI. Oh

1:23:25

yeah. It, it God seems

1:23:28

like chaos over there. What is going on over there? Ox. Did

1:23:30

you guys know that open AI is a nonprofit? Do

1:23:34

they know that? I'm not sure

1:23:36

they do. I think they're finding out and

1:23:39

that's what's going on. Honestly, like at a

1:23:41

very high level, but everyone's leaving like Mira

1:23:43

Marati, the CTO is leaving Greg

1:23:45

Brockman, the president, apparently

1:23:48

he just vanished. Yeah. There's this

1:23:50

great photo of all of the exact, Greg,

1:23:54

Mira and Sam together for this big

1:23:56

New York times profile. This was like

1:23:58

right before the boardroom boardroom coup. which

1:24:00

Neelai was almost a year ago when

1:24:02

you were at Disney. No way. Yeah.

1:24:05

Isn't that crazy? You were at Disney not

1:24:07

getting to ride the rides almost a year ago,

1:24:09

cause we were reporting on this, which is, yeah,

1:24:11

that's to me. But there's this photo of him

1:24:13

and all these execs. And now they just Photoshopped

1:24:15

out everyone except Sam Altman, who's the only one

1:24:17

left. And what's

1:24:20

happening is that OpenAI just raised the largest

1:24:22

round of funding of all time. I think

1:24:24

they just went above Elon's mega round for

1:24:26

XAI, which was I think six

1:24:28

billion on purpose. Of

1:24:31

course, Optics, they're valued at $150 billion, which

1:24:34

is more than the market cap of Goldman Sachs. Guess

1:24:37

what? OpenAI is legally still a

1:24:39

nonprofit. They get tax

1:24:41

breaks in

1:24:43

the state of California as a

1:24:45

nonprofit. They make billions of dollars

1:24:47

a year. And so what's

1:24:49

happening is this very chaotic fast transition

1:24:52

from what was a research

1:24:55

lab, we're going to invent AGI, all

1:24:58

the Ilia stuff to in the last

1:25:00

year when Sam won the board coup.

1:25:03

No, we are a large

1:25:05

commercial for-profit next

1:25:08

big mega tech company. And

1:25:10

they hired a CFO, they hired

1:25:12

a CPO who used to run product at Instagram,

1:25:15

not the other one who used to run product at Instagram who's

1:25:17

an anthropic. And they are

1:25:20

very chaotically turning into a

1:25:22

real big company. And

1:25:24

I think what you're seeing is a lot

1:25:26

of the previous era being

1:25:29

abruptly managed out, quitting,

1:25:32

what have you. And it's the Altman

1:25:34

show. And he famously does not

1:25:36

have equity in OpenAI, has been doing

1:25:38

a ton of interesting

1:25:41

dealings with other startups and

1:25:43

investments around OpenAI. It

1:25:45

is now reportedly going to be getting a

1:25:47

large stake in this new for-profit company that

1:25:50

they're creating. I will

1:25:52

love to see how what the tax

1:25:54

implications are of this transition, more to

1:25:56

come on that. And the state of

1:25:58

California really... doesn't like

1:26:00

when nonprofits become for profits. It's

1:26:02

a very, very contested thing. Um,

1:26:05

and yeah, I think that's what's happening at

1:26:08

high level. So Mira, the CTO is out,

1:26:10

um, who was memed with her reaction to,

1:26:12

did you use YouTube to train Sora? They

1:26:15

were very unhappy about that from what I understand.

1:26:17

And then they're ahead of research who literally just

1:26:19

did an interview about the new reasoning model with

1:26:21

Kylie a couple of weeks ago, um, who

1:26:24

won out against Ilya in the research kind

1:26:27

of group battle a year ago. So yeah,

1:26:29

it's just like stuff's happening very fast and

1:26:32

yeah, big picture. This is

1:26:34

open AI entering a decidedly

1:26:36

commercial big tech company

1:26:38

phase. They are speed running. Should that be

1:26:40

terrifying for us? Given that they were supposed

1:26:42

to be the safeguards of AI. Well, yeah,

1:26:45

now he was doing, Billy is doing super safe, whatever

1:26:47

AI. So it's going to be super safe. But he

1:26:49

doesn't have the money. He

1:26:52

raised a bunch of money. Yeah, he raised a bunch of money. Um,

1:26:54

but not as much. No, not as

1:26:56

much. But I mean,

1:26:58

they are speed running the Facebook, Google playbook, which

1:27:01

is move incredibly

1:27:03

fast. Let regulation

1:27:05

catch up. Let, you know, just

1:27:07

outrun your competitors. And it's

1:27:09

intense. I mean, the vibes that we hear is

1:27:11

that working inside open AI is just, it's way

1:27:14

different than it was a couple of years ago,

1:27:16

the pressure to ship the pressure to make things

1:27:18

commercial. And so that research

1:27:20

culture that built the company originally is

1:27:22

being just broken down. And it's

1:27:25

very messy right now. Yeah. My

1:27:27

sense of it, uh, David mentioned Mustafa

1:27:29

Saliman earlier, who is now

1:27:31

the CEO of Microsoft AI is

1:27:34

Microsoft is reacting to

1:27:36

their former research investment becoming a product

1:27:38

company by being like, now we have

1:27:40

a CEO of products, uh, which

1:27:43

I'm very curious to see how that relationship manages

1:27:45

over time. All right. We should take a break.

1:27:48

We're going to go through the lightning round. It's

1:27:50

lightning speed because we are so over at this

1:27:52

point. We'll be right back with the lightning round.

1:27:58

It's time to review the top three highlights. the

1:28:00

day, I'm joined as always by my co-anchor Snoop.

1:28:03

Snoop number one has to be getting the

1:28:05

new iPhone 16 Pro with Apple Intelligence at

1:28:07

T-Mobile. Yeah, you should hustle down to T-Mobile

1:28:10

like a dog chasing a squirrel chasing a

1:28:12

nut. Nice analogy Snoop, dogs do love to

1:28:14

chase squirrels. And squirrels love to chase those

1:28:16

nuts. On to highlight number two, at T-Mobile

1:28:18

families can save 20% every month

1:28:20

versus the other big guys. That is very impressive.

1:28:22

You know, y'all can take some of those savings

1:28:25

and buy some Snoop merchandise. That's exactly what I'm

1:28:27

planning on doing with my saving Snoop. Now take

1:28:29

it away Snoop. Head to

1:28:31

t-mobile.com and get the new iPhone 16

1:28:33

Pro with Apple Intelligence on them. Now

1:28:36

drop that jingle. Alright,

1:28:59

we're back. We are so over.

1:29:03

We are often asked how long the show is supposed to be

1:29:05

and the answer is not this long. Six

1:29:07

and a half minutes. But we're never going to tell you how

1:29:10

long it's supposed to be. Actually a 90 second show. We've

1:29:12

gotten some incoming on Lightning Round sponsorships. Again, I

1:29:15

don't make the deals. I just walk

1:29:17

around Voxmute and demanding why the deals aren't made, which

1:29:20

has proven to be ineffective. So

1:29:22

I'm going to try some new strategies,

1:29:24

but today we remain unsponsored. Which

1:29:27

is a personal failure. For

1:29:30

somebody who complains about Creator, like influence and

1:29:32

brand deals as much as I do, I'm

1:29:34

horrible at this. Which makes it

1:29:36

that money. Anyway, the Lightning Round is available to sponsor. I

1:29:38

might say your company's name

1:29:41

if you give us a bunch of money or I might not

1:29:43

because of ethics policy. You won't know

1:29:45

until you pay me money. That's Lightning Round. Alright,

1:29:48

Alex, let's start with you because I think

1:29:50

you've got the most relevant Lightning Round item

1:29:52

after that previous conversation. Yeah, Tripp, Nickelton New

1:29:54

York Times has this great story about what

1:29:56

Johnny Ive has been up to post Apple

1:29:59

and this design. firm he founded called love

1:30:01

from lots of nuggets in here.

1:30:03

I thought it was very interesting that he

1:30:06

says in the story he interviewed Ivan June and

1:30:08

this came out the week of iPhone 16. Um,

1:30:12

the timing was interesting. Um,

1:30:15

but yeah, Johnny's he's working. I

1:30:17

think the time's kind of buried the lead

1:30:20

here. Um, he's literally, it's like the last

1:30:22

paragraph story. He

1:30:24

confirms that he's working on some kind of

1:30:26

AI device with open AI. It sounds like

1:30:28

they haven't quite nailed down what it will

1:30:30

be. Um, it was the first time he'd

1:30:32

confirmed it on the record. Um, my understanding

1:30:34

is that love from a super small, it's like 40

1:30:37

ish 50 people. And there's

1:30:39

only 10 people working on this project. Yeah.

1:30:41

And a wild detail in this story

1:30:43

is, is as Mr. I've

1:30:46

climbed a wooden staircase to the studio second floor

1:30:48

that morning. And by the way, he bought like

1:30:50

a whole block, a city block of downtown San

1:30:53

Francisco for their office. Um, he

1:30:56

said he spoke about love firms, clients, love

1:30:58

firms, clients, which pay the firm as much

1:31:00

as $200 million annually. And

1:31:04

that includes like Airbnb Ferrari. He just designed

1:31:06

a $3,000 jacket with Montclair. Love

1:31:10

from is like 50 people. So

1:31:13

how much money are they making? Johnny. I,

1:31:15

well, this is crazy block in San Francisco.

1:31:17

Like they doing good. Brian Chesky is

1:31:19

paying him $200 million a year to

1:31:21

like riff on what the Airbnb logo

1:31:23

should be. Can I, can I just tell

1:31:26

you a story about the Airbnb Johnny

1:31:28

I've connection, which I was reading

1:31:30

this and it says what he worked on. Like

1:31:32

they did the experiences where you can like live

1:31:34

in a pineapple house or something, whatever.

1:31:37

I mean, $200 million nonsense is happening. Right.

1:31:40

Um, Airbnb last year,

1:31:42

the year before, after

1:31:44

the Johnny I've stuff was out, um,

1:31:47

like announced a bunch of redesign stuff and like an

1:31:49

app redesign and all this stuff and we emailed them

1:31:52

and we're like, was this the stuff Johnny I've worked

1:31:54

on because obviously that is very relevant and interesting detail.

1:31:56

And they're like, yeah, yeah. And then we like, And

1:44:02

that's it for the VergeCast this week. Hey, we'd

1:44:04

love to hear from you. Give us a call

1:44:06

at 866-Verge-11. The VergeCast

1:44:09

is a production of The Verge and Vox

1:44:11

Media Podcast Network. Our show is produced by

1:44:13

Liam James, Will Poor, and Eric Gomez. And

1:44:16

that's it. We'll see you next week.

1:44:22

It's time to review the highlights. I'm joined by

1:44:24

my co-anchor, Snoop. Hey, what up, Dope? Number one

1:44:27

has to be getting the new iPhone 16 Pro

1:44:29

with Apple Intelligence at T-Mobile. Yeah, you should hustle

1:44:31

down at T-Mobile like a dog chasing a squirrel,

1:44:33

chasing a nut. That's a nice analogy, Snoop. On

1:44:35

to highlight number two with T-Mobile, families can save

1:44:37

20% every month versus the other big guys. Very

1:44:40

impressive. Take it away, Snoop. Head

1:44:42

to t-mobile.com and get the new iPhone 16 Pro

1:44:44

with Apple Intelligence on them. Now drop that jingle. See

1:44:48

how you can save versus the other big guys

1:44:50

at t-mobile.com/Switch. Apple Intelligence coming Fall 2024.