0:01

What is up you guys and welcome

0:03

to the Taiwan show. We're

0:06

here at Computex where I spent

0:08

about maybe two

0:10

hours on the show floor

0:12

and Luke spent even less.

0:15

Probably about the same actually. Really? No. High five. Good

0:18

job. Yeah. We've got a lot

0:20

of great topics for you guys today. Starting

0:23

of course with IKEA

0:27

apparently hiring digital employees to

0:29

work at the Roblox store.

0:33

Dude, is the metaverse actually real?

0:35

People wanted work from home. They

0:38

did, but did they want to work from Roblox? Probably not.

0:41

In other news, Intel wants

0:44

to fight ARM while AMD appeals

0:46

to tradition. But

0:48

the big news of course this

0:50

year at Computex is Qualcomm and

0:52

their Snapdragon X Plus and Snapdragon

0:54

X Elite processors. I

0:57

did get to be hands on with them,

0:59

which is pretty exciting. We're going to be

1:01

talking about that later. What else we got today, Luke? Nvidia

1:04

is number two. How dare you? Look,

1:07

I don't always agree with them, but you can't call them

1:09

that. You

1:13

prepped that. You had that prepared. That was

1:16

preemptive. I don't know how I feel about

1:18

that. What else?

1:20

I don't know. Overpriced maybe,

1:23

but you can't say the quality is... Number two. You

1:26

know, like that, right? You have

1:28

Nvidia graphics card in there. You got some number two in there.

1:32

And we have an update on the car thing stuff. We'll talk about

1:34

that I guess. Yeah, sure. We'll talk about Spotify car

1:36

thing. Oh, how

1:38

do we roll the intro? I guess Dan does it. Sound

1:46

effect Sound effect Sound effect

1:48

Sound effect Sound effect Sound

1:51

effect Sound effect Sound effect

1:54

Sound effect Sound effect Sound

1:57

effect Sound effect Sound

1:59

effect The

2:04

show is brought to you today

2:06

by AG1, Squarespace, and Manscaped. Why

2:08

don't we jump straight into our headline topic this week,

2:11

which is that IKEA

2:15

is launching the

2:17

Co-worker game, a virtual experience

2:19

inside Roblox whose gameplay will

2:22

apparently involve serving

2:24

meatballs and

2:26

decorating showrooms. The

2:29

game is open to anyone, the

2:32

broader Roblox community, gamers, and

2:34

IKEA fans alike. I

2:37

do wonder how much overlap

2:39

there is between those groups. This

2:42

is according to IKEA's press statement. IKEA

2:45

is also hiring 10 people

2:47

for fully remote virtual roles

2:49

in this virtual IKEA universe

2:52

at a rate of 13.15

2:54

pounds per hour, which IKEA

2:56

says is aligned with the

2:58

pay recommendations of the Living

3:00

Wage Foundation for the City

3:02

of London. I pretty

3:04

much promise you that nobody is living in

3:07

London for 13.15 pounds per hour, but that's

3:12

a whole separate conversation. The

3:15

hiring process will involve an

3:17

interview, hiring questionnaire, and

3:19

the submission of a resume, and applicants from

3:21

the UK and Ireland must be at least

3:23

18 years old. Wait,

3:25

so do you have to live in the UK or

3:28

Ireland? I don't know. Probably not. If

3:30

you're working for 13.15 pounds an hour, you

3:32

are probably not living many places

3:34

in the

3:38

UK or Ireland. Yeah, why is there only rules

3:40

for there? I

3:43

don't know. Maybe they're willing to hire children

3:45

as long as they're not in the UK

3:47

and Ireland. I legitimately don't know. Weird,

3:51

okay. Quote,

3:55

we're excited to be the first

3:57

brand to launch paid work on

3:59

Roblox. to showcase how we do

4:01

careers differently, bringing our unique careers

4:04

philosophy to life. At

4:06

IKEA, there is no set route to

4:08

career progression. Our coworkers are

4:11

able to change roles, switch departments,

4:13

and grow in any direction they

4:15

choose both in the game or

4:18

in the real world. To

4:21

clarify, paid players must be 18 plus

4:23

and in the UK or Republic of

4:25

Ireland to apply. Okay, so you

4:27

do have to be in the UK and Republic

4:29

of Ireland. Here's my

4:32

question for you. When you were 18, would

4:35

you apply for this? What

4:38

do you even do? You

4:40

hang out, like you've got the press release

4:42

open. Do

4:44

you hang out in

4:46

the store? Is it a

4:48

store? It seems

4:51

to be like IKEA employee simulator, but some

4:53

of them are paid. I

4:55

don't know how you help

4:57

customers and stuff when

5:00

the game is people

5:03

being what you are, isn't it? I

5:05

don't know. Maybe you manage their shifts, Luke.

5:08

Maybe you're not thinking metaverse enough.

5:10

Maybe you're like the dungeon master

5:12

of this IKEA simulator. You

5:14

create events and stuff. I

5:16

have no idea. Sure. You could spawn customers.

5:19

I might apply for it, but to be

5:21

completely honest, at that time,

5:23

I was sitting in the room, seeking employment

5:25

that paid more than this even back then,

5:28

and there was options for that. Well,

5:30

hold on. Hold on. 13.15 pounds. What

5:34

the devil is that? Hold on. 13.15

5:36

pounds to Canadian dollars. Okay,

5:39

here we go. Okay. That's

5:42

23 Canadian dollars an hour. It's actually not too bad.

5:44

Yeah. When I was that age, I was

5:46

making... Well, hold on a second though, because

5:48

we got to convert. In

5:51

2003, I was making

5:53

16 Canadian dollars an hour,

5:56

so inflation calculator.

6:00

Sure, I'll use a US one. For 2003? But

6:07

that's when I was working. Got it, okay. That's

6:09

when I got my first job out of high school and I

6:11

was making the same the year after. So, 2003,

6:14

$16 an hour is what I was making. Okay,

6:20

27, 27. So,

6:25

I was ahead. All

6:28

right. But in fairness, I

6:30

had to get certifications in

6:32

order to do my job. I was working

6:34

as a lifeguard and swimming lessons teacher. So,

6:37

that was probably a grand

6:39

total of about $1,500 in

6:42

certifications. Not to mention the

6:44

unpaid time. Which is out of reach for some people.

6:46

Right, so I'm

6:49

paying to invest additional time

6:52

in getting those certifications. So, how many hours

6:54

would I have had? At $4 an hour,

6:56

hold on. I

6:59

don't even cover my certification cost. So,

7:02

nah, nah, nah, nah, divided by four. I

7:04

don't even cover my certification cost until I've worked

7:06

hundreds of hours. Right,

7:09

so yeah, realistically, it's maybe

7:11

not that bad. Except that the problem

7:14

is that I've been to London. Yeah.

7:19

And like. I mean, it's not only London though. I

7:22

just mean, my point is just that

7:24

it's not as simple as your cost of living

7:26

is just like currency conversion.

7:29

Totally. Like things just

7:31

cost more over there to

7:34

my crappy Canadian currency. Even compared to

7:36

currency. Like you can't just, yeah, you

7:38

can't just convert your currency. So, the

7:40

fact that they're making 13 pounds or

7:43

whatever it is an hour. If things

7:45

just cost more because

7:48

they're pounds or something then it doesn't really

7:50

help you. So, the math that

7:52

we're doing is not fully representative. Really

7:54

what I wanted to talk about is

7:57

more along the lines of like, would you have been willing

7:59

to work? in a virtual store.

8:01

Would you have applied to work in a virtual store?

8:03

Oh yeah, I mean, I applied for whatever. At

8:06

that age, like, I don't know, who cares? Okay,

8:11

I'm gonna try this one more time without

8:13

just like indiscriminately applying

8:15

to everything. With this, okay, out of

8:18

the 10 jobs you applied to, with

8:21

this rank, which one would you want? I

8:24

would have probably wanted the one that I had at the

8:26

time. At that time, Geek Squad was a separate entity to

8:28

Best Buy and it was actually like super cool. I

8:31

definitely would have preferred the job that I had

8:33

at that time over this, but it is definitely

8:35

also something that I would have considered. I think

8:37

out of the top 10, it

8:41

would probably have been like three or four. It's

8:44

not too bad. You can stay at home at

8:46

that point in time, if you could have spent

8:48

less on gas, it's

8:51

actually like a really big deal. When you're

8:54

making a huge

8:56

amount of income is you're like how much

8:58

money after your expenses you have. And when

9:00

you're making a very low amount of money,

9:02

the amount of money after expenses is really

9:04

small. So if you can reduce those expenses,

9:06

like maintenance on car, gas, stuff like that,

9:08

it's actually super beneficial. Luke, net

9:11

profit versus gross profit tips. Yeah, it's

9:13

a big deal. Yeah, 100%. Honestly,

9:17

not quite this job. Like I don't

9:20

really have any particular love for IKEA. Yeah,

9:22

I don't mean either. Yeah, I don't think

9:24

there's like an IKEA passion that could be

9:26

awakened in the way. And dealing with like

9:28

small children screaming about their digital meatballs not

9:30

being good enough doesn't sound exactly entertaining. Or

9:32

whatever, like I don't even know exactly what

9:34

this is. What the job is, yeah, I

9:36

don't know. But with that said, you know

9:38

what job I would have applied for if

9:40

it existed? If NCIX,

9:43

whose forums I would hang out on

9:45

for hours a day, if

9:48

they had like, even if it

9:50

didn't have an hourly rate, if

9:52

they just had like a commission based...

9:56

You probably would have preferred that over hourly rate. I probably

9:58

would have. rushed it. If

10:00

I could just hang out in the NCIX

10:03

virtual store and people could just ask

10:05

me for help configuring their computer or

10:07

whatever and you know all they

10:09

gave me, all NCIX would have had to give

10:11

me is like a you know like

10:13

a you know little name tag or

10:15

whatever like special recognition that my character

10:18

you know kind of worked at the store.

10:20

Maybe your the name above your head is

10:22

like a slightly different colors. Yeah or whatever

10:24

it is and honestly they

10:26

could have just because they they kind of

10:29

had similar systems for the forum already

10:31

anyway. They were mostly just based on

10:33

like post count in those

10:35

days unfortunately. Not always a great indicator. Not

10:37

always the best but they did

10:39

also have like a system for experts

10:41

I think. The lines of code version

10:44

of full contribution. Yeah yeah but man

10:46

I would have I would have loved

10:48

that. I would have gotten home from

10:50

school and like put

10:52

on my my virtual you know

10:54

tie and button-up shirt and

10:57

I would have gone and sold computers man. I would

10:59

have been so into it. I feel like that that

11:01

might have almost been a problem for you.

11:04

You only spent like too much time doing it. Maybe

11:06

but if I was making money like is

11:09

that really that then I could

11:11

buy more computers. Yeah win-win. I

11:13

feel like that is what the Linus of that time would have

11:15

done. Oh 100 percent yeah 100 percent and I could be like

11:19

only like half tuned into that at the

11:21

time right. I could be like

11:24

half tuned into that and then I could

11:26

be like half tuned into you know my

11:28

other forums like my red flag deals forum

11:30

wheeling and dealing. Only

11:33

you know monitors didn't weigh 50 pounds

11:35

back then then I could have just

11:37

had like I could have got some

11:39

online poker going on up here get

11:41

some like like day trading you know

11:43

forex right here and like yeah

11:46

man. I'm failing to see like

11:48

is this a full-time position or

11:50

not. I

11:52

think this feels intentionally

11:55

vague and nebulous to me. Yeah

11:57

it feels like they just want people like us to talk about

11:59

it to be honest. because it's only

12:01

10 positions. If they were

12:03

doing like actual real work, I suspect

12:05

the amount of people, especially at the

12:07

beginning that flood this game. You just

12:09

call the metaverse, not actual real work.

12:13

I wish I had my swear button. I

12:15

wish I had my bleep button so I

12:17

could tell you what a guy you're being.

12:23

Yeah, what a number two you're behaving

12:25

like. Okay, that is real

12:27

work. They're doing real work just

12:29

because they're sitting with their Cheetos

12:32

and their Mountain Dew in

12:34

their- And their digital meatballs. And their digital

12:36

meatballs. Man,

12:39

my digital meatballs are cooking right now. Yeah.

12:42

I got a computer and a pillow

12:44

on my lap and we are in

12:46

Taiwan heat because we liked the vibe

12:48

of this little outdoor meeting area. We

12:50

could have been inside where the air

12:53

conditioning is, but we love you guys

12:55

so much. We wanted you to share

12:57

in the outdoors of Taiwan with us.

13:00

What do you want to talk about next? Okay,

13:03

let's see. Intel

13:06

wants to fight arm. Let's do it. At

13:08

Computex, Intel gave a deeper look

13:10

into its upcoming lunar lake architecture.

13:13

The company is claiming improvement across

13:15

the board, though with heavy emphasis

13:17

on efficiency and battery life, a

13:20

38% to 68% increase in instructions per clock

13:22

for E cores, compared to a 14% increase

13:24

for P cores. Intel

13:27

is also claiming a 50% performance

13:29

gain for its integrated graphics, which

13:31

will be using a low powered

13:34

variant of Intel's XE2 Battle Major.

13:36

Yeah, Battle Major just kind of snuck

13:38

up on me here. I don't know

13:40

why, but I was kind

13:42

of expecting it to come as a discrete

13:44

GPU first. And then here it is, boom,

13:46

onboard graphics, Battle Major's here, let's go. I

13:50

heard some folks got hands-on with it at

13:52

some point, maybe somewhere. And-

13:54

I wonder,

13:57

because this is pretty abnormal. Usually you'd

13:59

have the- I wonder if this will help drivers

14:01

though. I

14:04

hope so. Help drivers for the discrete GPU

14:07

people. Yeah, no, no, I see

14:09

what you mean where like, if,

14:11

because, because Intel is shipping

14:13

apparently a negligible

14:16

number of discrete arc GPUs,

14:18

according to the latest market share trends,

14:20

which have Nvidia at almost 90% now.

14:22

Oof. Wild, eh?

14:25

But Intel is apparently

14:27

negligible. And

14:29

so when you're shipping a negligible

14:31

number of boards, I imagine that

14:34

the incentive to prioritize development of

14:36

it, even though to Intel's credit, they

14:38

have really chipped away at arc

14:40

alchemists and tried to make it as working

14:44

as it could be. But

14:47

when you're shipping not that many compared

14:49

to when all of a sudden this

14:51

product that is like a do or

14:53

die product for you lunar Lake is

14:55

going up against not just AMD, but

14:57

now Qualcomm Snapdragon X. Full

15:00

disclosure, by the way, guys, not that it actually

15:02

matters for when show will be saying whatever we

15:05

damn well please. But Qualcomm did

15:07

sponsor a video on

15:09

Snapdragon X while we were here, which is part

15:11

of why we were able to poke and prod

15:14

at it a lot. We

15:16

got cool. Yeah, man, I

15:18

love it. Very interesting. I love it. I love it when

15:20

we get a sponsorship that comes with like

15:23

extra access to the product anyways, and

15:25

almost no strings attached. Yeah, when we

15:27

went through the script, they

15:29

were like, can we change that word? And

15:31

I changed it. I did my trick. I

15:33

changed it to something worse. Like,

15:36

oh, nevermind. They were like, nevermind. They

15:41

changed like one word and then like Andy we

15:43

shot we shot like a pickup. Yeah, we

15:46

shot a pickup because we

15:48

actually made too strong of

15:50

a statement about the competition. Oh, and they

15:52

were like, I wanted it to be pulled

15:54

back. They wanted it to be dialed back

15:57

interest. So just yeah, so just Is

16:00

this out already? No, it's not out yet. But

16:02

anyway, Intel, this is a do or die product for

16:04

Intel, where they are gonna be going

16:06

up, not just against

16:09

AMD, who has been kicking their butt a

16:12

little, because AMD had

16:14

been their first generation that

16:16

was super competitive. That

16:18

would have been mobile 3,000 fourth. I

16:22

forget, I can't keep track. Now that the

16:25

Zen generation doesn't line up with the product

16:27

generation, and the mobile and the desktop generations

16:29

don't line up either, it's very confusing. But

16:31

the point is they had one that was

16:33

really competitive, and it was in like a

16:36

handful of devices that functionally weren't

16:38

even really available. Is that 5,000? Doesn't

16:40

matter. The point

16:43

is they were competitive on paper, but

16:45

they didn't exist. And

16:47

then over the last couple of

16:49

generations, they exist and they're

16:51

in some really good designs now. Like I'm

16:54

holding an AMD laptop that

16:57

I freaking love. This is,

16:59

framework disclosure, I guess. This is the FlowX13 from

17:01

Asus. I

17:03

flipping love this thing. It's got an

17:06

eight core Ryzen 7,000 series processor. It's

17:09

got an RTX 4070 in it. Albeit

17:12

a lower TDP one, obviously. It's real

17:14

thin, real small. Yeah, I mean, it

17:16

gets hot enough to, yeah,

17:19

she's cooking, right? Because we're

17:22

using RTX broadcast for noise

17:24

cancellation. It's not like

17:26

the ambient is low here. And I think,

17:28

oh yeah, no, I also have my capture

17:31

card running. So I've got video capture running

17:33

through it. Absolutely

17:36

freaking love this machine, but

17:38

AMD doesn't have the kind

17:40

of fab capacity to

17:43

ship enough laptop chips that Intel

17:45

has to really worry about just

17:48

outright losing. Wallcom

17:52

on the other hand. On the other hand. Wallcom

17:54

on the other hand. Ruchofab capabilities.

17:57

Dude, I can't believe

17:59

how many. design wins these chips are

18:01

in. Like when you

18:03

compare this to AMD's first

18:06

generation of competitive CPUs where

18:08

it was like you could

18:11

tell from the designs that like the

18:15

vendors didn't want to make Intel

18:17

too mad. You know not

18:19

because there's a backroom deal necessarily

18:21

not because Intel is you know

18:23

back to their tricks of you know paying

18:25

people to not use AMD necessarily I mean

18:28

who knows something could emerge you

18:30

know five years from now or whatever I don't I

18:32

don't know either way but what

18:34

I'm saying is that's not the vibe I thought

18:36

the vibe that I got was yeah the

18:39

AMD chip is in the slightly more

18:41

worser one because realistically they know

18:44

they can't get enough allocation to switch

18:46

over a significant volume for this thing

18:48

so they're just kind of putting it

18:50

in some niche devices knowing that that's

18:52

all they can ship anyway and the

18:54

bulk of this generation is going to

18:56

be in like you can you can tell when

18:58

that happens right? It

19:03

ain't happening. Dude these

19:06

Qualcomm designs are

19:08

sick and I don't

19:10

mean that Qualcomm designs them sorry I mean

19:12

like like the Dell XPS with

19:15

Qualcomm's Snapdragon in it. Like the

19:17

laptop itself? Sick yeah it's

19:19

sick the surface the new surface

19:23

it's sick. Dude the

19:25

camera. Dude the camera

19:27

in the surface it's

19:29

sick. Honestly in our video I

19:32

don't think we have the most favorable

19:34

demo of how much better the camera quality

19:36

can be it's like a little better. Are you

19:38

on the floor or something? Yeah but we're in

19:40

like we're in like a demo room and the

19:42

lighting really really sucks to the point where kind

19:45

of anything looks bad but under

19:47

slightly better lighting conditions when we were

19:49

in the meeting room before. That's

19:54

interesting because laptop webcams are

19:56

kind of notoriously terrible. Yeah so we go

19:58

through this in the video because very

20:00

few people are talking about the

20:04

camera image quality benefit of

20:06

Snapdragon, a Windows on Snapdragon.

20:09

Everyone's talking about Qualcomm's talking

20:12

points, right? Which is performance,

20:14

AI and battery life. So is this

20:16

like compute enhanced video? Well, no. What

20:19

it is, is that almost

20:21

all laptop webcams have the ISP

20:24

on the webcam. But

20:28

on Snapdragon, the ISP is on

20:30

die. And it's

20:32

linked using, shoot, what's the, I

20:34

forget what the link is, but instead

20:36

of using USB, it uses

20:38

the same interface as your smartphone. Oh man. So

20:42

they have plug and play compatibility for

20:46

the same camera module. Yeah, that's actually wild. From the

20:48

smartphone industry. Yeah. And they

20:50

have all those years of experience building

20:52

ISPs. So much more work goes into that. For

20:56

smartphones. Disgustingly more work goes into

20:58

that. Dude, dude, the

21:01

new Surface's webcam, you can

21:04

just throw away your add-on webcam. Just huck

21:06

it. That's nice. Yeah, because it's built in.

21:09

Because it's pretty annoying using that. Very

21:11

cool. Like, dude,

21:13

I'm pretty stoked. I've

21:15

heard rumor, genuine

21:18

rumor. I don't know

21:20

if there's articles about this, whatever. This is word of

21:22

mouth rumor. That a decent amount of

21:24

this pressure to make the laptops really sick

21:26

for this generation

21:28

of laptops is from Microsoft. Really? That

21:33

makes sense. I think that they've

21:35

been, it's funny because. Again, total rumor. I

21:39

don't know. I was looking at the

21:41

market share that Apple has gained since

21:43

the launch of Apple Silicon. And it's not as

21:45

much, at least according to

21:47

the source that I was looking at. It's quite

21:49

possible that a company like Microsoft. No,

21:52

I've heard the same thing. Who's going to be

21:55

paying a lot of money per

21:57

year for market research or whatever. it's

22:00

possible they have numbers that I

22:02

don't. But from what I can tell,

22:05

it hasn't really made a

22:07

meaningful difference. And that actually

22:09

kind of jives with Apple's behavior,

22:11

where it's like,

22:13

yeah, I don't know. We'll like

22:15

we'll do a we'll do an update to this one,

22:18

I guess. M4, I don't know. Yeah,

22:20

we'll put it in a MacBook at some point. iPad,

22:23

right? Like, yeah, like it kind of

22:25

jives with Apple's behavior. Whereas like, man,

22:27

when M1 came out, M2 refresh, they

22:30

were like, we're on cadence. Like,

22:32

we're going, we're going, here come all the models.

22:35

We've got the studio, we've got the we've got

22:37

the big one. But they're not converting more users.

22:39

Yeah, but if they're not converting more users, it's

22:41

like Apple just goes through this cycle where they're

22:43

like, oh, yeah, yeah, we're gonna do Mac again,

22:46

we're gonna like kill it. We're not we're focused

22:48

on Mac again. Yeah, sorry, we forgot. Nevermind. Right?

22:50

Like they did it with the trash can, they

22:52

did it with the first generation cheese grater or

22:54

the new generation cheese grater on Intel. They

22:57

did it with the studio like Mac studio is

22:59

that is that still M2? Yeah, probably.

23:03

Yeah, Mac studio is still

23:05

M2 family. That's

23:09

Apple stuff, man. I was I've been trying to

23:11

figure this out. I suspect it's because of whatever

23:14

it's called replay. Oh,

23:17

yeah, yeah, yeah. Our whole pilot

23:19

plus. Yeah, PCs thing. I

23:22

think that's why they're pushing. I mean, yeah, maybe it's

23:24

less to do with Apple's threat. And it's more to

23:26

do with just, good Lord, give

23:28

people a reason to upgrade their damn computers.

23:31

Yeah, I could see that. Anyway,

23:33

coming back to the fight that Intel's in

23:35

with lunar lake, they

23:38

cannot afford for these battle

23:40

mage drivers to suck. And that

23:42

gives me a lot of hope, actually, for

23:44

the for the for the battle mage discrete

23:46

cards. Now, I'm not expecting them to be

23:48

competing with a 5090 or

23:51

whatever. I don't think that's realistic. But

23:53

what I am, what

23:56

I am hoping to expect is

23:58

a real competitive. for the 50 60

24:01

100% because I am I am just sick of Nvidia being like,

24:07

yeah, I don't know. We'll just have the

24:09

most basicest thing that would have been a

24:11

50 class card a generation or two ago.

24:13

We're going to call it a 60 class

24:15

card. It's going to be up to you

24:18

know, what is the 40 60

24:20

max out for the TI like $500 or

24:23

$600 or something like that. What's a 40 60 TI worth? I think the

24:28

price has actually come down a

24:30

little bit. Uh oh my god.

24:33

How why is it so hard to just

24:35

find a thing? It's 400 bucks for an

24:37

8 gig 8 gig 40 60

24:39

TI like come on. Yeah, it's lame.

24:41

So if

24:44

I can get something Arc Battlemage with a

24:46

decent frame buffer with good drivers, I'm excited

24:48

for that. Oh yeah, me too. Um

24:51

I don't know. I don't know if I'm 100%

24:54

confident it's going to be Battlemage or if it

24:56

might be the one after. Celestial. Yeah.

24:59

But I hope it's Battlemage. Anyway,

25:01

moving on uh Intel Lunar Lake.

25:03

So the this is this is

25:05

crazy too. So we

25:08

talked a little bit about how much performance

25:10

gain they're expecting both on CPU and GPU

25:12

to try and take the fight to both

25:14

AMD and especially bulk on this gen.

25:17

Um and you gotta

25:19

kinda wonder how much of that is because this

25:22

is the first generation

25:24

because the first generation of these

25:27

Lunar Lake chips are going to

25:29

be primarily manufactured by TSMC Intel

25:32

says they made this decision because

25:35

TSMC's fabrication processes were simply more

25:37

advanced at the time they were

25:39

designing the chip. What

25:42

a slice of humble pie. That's

25:44

pretty wild. The fact that

25:47

it's not something Intel would normally say no

25:50

or not like II said

25:52

this before Intel's Intel stock has gone

25:54

down quite a lot since I said

25:56

that if I was buying I would

25:58

be buying long Intel. Yeah. This

26:01

actually doesn't change what I said Because

26:04

that kind of add on investment advice

26:06

not investment advice And I haven't bought

26:09

any because I don't allow myself to

26:11

just like trade tech stocks like that

26:14

I do I do own some stake

26:16

in framework laptops as you guys know

26:18

from the series of videos that we've

26:20

done on it and I

26:22

do have a small investment in a

26:25

It's hard to

26:27

even call them a startup. They're very small in

26:29

a startup that is trying to work

26:31

on a NAS operating system that makes

26:33

things simpler But

26:36

that's that's it. I don't let myself just

26:38

kind of play the market But

26:42

I like this humility, yeah,

26:44

I like this willingness to

26:46

make the best product Even

26:49

if it's like even if it's not all

26:52

real men have fabs, you know, like Jerry

26:54

Sanders that that's it. That's a quote Okay,

26:56

that's that's that's like a toxic masculinity quote

26:58

from Jerry Sanders the third Who

27:01

I believe was AMD's founder Second

27:07

or third Jerry Sanders. Yeah.

27:09

Yeah, he's the third Jerry Sanders. There

27:11

you go. I really hope he's related

27:13

to the colonel somehow Thanks

27:18

for that was crazy things have happened good contribution

27:23

Anyway, they went with a rival foundry rather than compromise

27:25

on the design of the chip But

27:27

next year's panther lake will be largely

27:30

fabricated by Intel's own foundry. They've been

27:32

trying to kind of say,

27:34

okay forget it Let's

27:36

just instead of going incrementally. We're

27:38

gonna try to leapfrog and on

27:40

stage. They did show off their

27:43

1.8 a process they showed off. I think it was

27:45

a it was either a wafer or die I can't

27:47

remember but they were showing something on stage saying they

27:49

have power on on it. So Who

27:52

knows maybe panther lakes gonna be good.

27:54

These are starting to move fast. Meanwhile

27:57

at AMD. They have confirmed. They're stupid

28:00

AI 300 series naming scheme for

28:02

the new mobile chips, which are Zen 5,

28:04

which is not stupid. The fact

28:06

that we're getting Zen 5, like right

28:08

away here on both mobile and desktop

28:10

means that AMD is also probably getting

28:13

a lot more fab

28:15

capacity right now. Cause

28:17

yeah, they also announced Zen 5 Epic

28:19

chips. So they must have

28:22

booked some wafers. They

28:24

are claiming a 30% performance advantage over

28:27

Intel's current gen flagship chip, though

28:29

they did not include frame rates

28:31

or percentile frame data. Unlike

28:33

Intel, AMD will be retaining SMT or

28:39

their version of hyperthreading and seem to

28:41

be aiming for pure performance rather than

28:43

efficiency. Now that's something I've seen a

28:45

lot. Like I've

28:47

seen people talk about that a lot, but

28:50

I guess my counter to that is AMD

28:54

already had the efficiency pretty good compared

28:56

to Intel. So if I

28:58

was them talking about my next generation

29:00

chips, maybe less of my messaging would

29:02

be about efficiency, but that doesn't necessarily

29:04

mean that I'm expecting Ryzen AI 300

29:06

series to

29:10

be inefficient. I mean, I

29:13

think anything's gonna look kind of bad compared

29:15

to Snapdragon X. Okay,

29:20

so in the video, spoiler, Alex

29:22

holds up two HP laptops. His

29:25

HP Elite Dragonfly or whatever it's

29:27

called, his Dragonfly daily driver that

29:29

he chose because of its outstanding

29:31

battery life. Like it's one of

29:33

the best Windows laptops for

29:36

battery life and

29:38

HP's Snapdragon X design. According

29:42

to HP's own ratings, which Alex has found

29:44

to be pretty honest in the past, it

29:48

was something like a

29:50

10 hour difference. It

29:53

was like eight to 10 hours difference. This

29:55

is another reason why I feel like Microsoft might be pushing.

29:57

No, no, no, no, no, I'm not done yet. Okay. And

30:00

the Snapdragon one had, I think, a 15%

30:02

smaller physical battery. Oh.

30:08

Oh, yeah. Wow. Yeah.

30:11

Like... It's like this has been the

30:13

fairly notorious thing that gets thrown at Windows

30:15

laptops is that their battery life is

30:17

atrocious compared to Mac's. Yep. So

30:20

it's like this is another one of the solutions to those

30:22

things. Like, yeah, there isn't a huge amount of market share

30:24

for Microsoft to claim back from

30:26

Apple. But they might be

30:28

tired of being like the butt of jokes when

30:30

it comes to laptops. Yep. Like,

30:33

Windows laptops are just considered basically inferior. They

30:35

should fix Windows Modern Standby then. That'd be

30:37

nice. Just an idea, Microsoft. It'd

30:39

be kind of cool. While you're at this

30:41

push, which I totally support,

30:44

I'm super down to have better battery

30:46

life, right? Like... I

30:48

wonder if it'll be better on different chips. It

30:51

could be. I don't know. I

30:53

mean, I hope so because realistically, I'm not going to

30:55

be switching to a Mac anytime soon. So

30:58

I am fully in support of Windows

31:00

laptops getting better. Now, I want to

31:02

make this really clear, guys. The performance

31:04

embargo for independent testing is not lifted.

31:07

So these battery life claims, I mean, they

31:09

could be nonsense. I think that... I

31:13

doubt it. You know? Like...

31:16

I was just saying that the claims

31:19

that they make are decently, consistently accurate-ish.

31:21

Yep. That's on the

31:23

HP side, right? You've got all these partners

31:26

that are staking their reputations. On

31:29

similar claims. On similar claims. Yeah. You've

31:32

got... And what I'll say is that the

31:37

vibe from Qualcomm... There's...

31:44

Very confident. That's an

31:46

interesting... It's always kind of a tell. You

31:48

mentioned that they sponsored. Yep. And

31:51

that they were very not controlling of the

31:53

messaging. Yes. That combination

31:55

is almost always like, oh,

31:57

this is good. Yeah. turn

32:00

they're like we need to control everything you

32:02

say and then we don't work with that

32:04

partner because that doesn't work but when that

32:06

situation happens it's almost always because they like

32:08

don't have anything. Yeah whereas when they're like

32:10

yeah we just want everybody to know. Say

32:12

whatever you want. Just literally

32:15

they're they're they're they're points they

32:17

gave me. It's such a power move. Can

32:20

you talk about performance? Can you

32:22

talk about the NPU? So like the

32:24

on-chip AI, Go Pilot obviously, and

32:26

battery life and I'm like yeah

32:29

I mean how could I possibly

32:31

talk about things that we would talk about?

32:33

How could I possibly talk about this product

32:35

without talking about those things anyway? And dude

32:37

the demo room was crazy because they

32:40

have like DaVinci Resolve running in there. No

32:43

but like they've got Baldur's

32:45

Gate 3 running in there. They've got

32:47

like they had um

32:50

uh what's uh but used to be owned by Sony

32:52

it's now owned by someone else. It's decently common that

32:54

these will softball the applications

32:56

that are running. Yeah. Which is why he's pointing

32:58

this out. Well at that end because

33:00

it's not x86. Yeah. Like this is

33:03

this is Windows on Arm. Things

33:05

that were not made for this. Yeah.

33:07

I don't think Larian like made a

33:09

version of the game for Snapdragon chips.

33:12

Not that I'm aware of. Yeah.

33:14

Yeah so so dude

33:17

oh also they were absolute characters. I have

33:19

no screen share right now so you guys

33:22

are gonna have to wait to experience this.

33:24

But I did a bit where I

33:28

where I I walk into the meeting

33:30

um and here

33:34

I'll try I'll try I'll try and kind of narrate

33:36

along. Um when

33:39

I walked into my briefing on Snapdragon processors

33:41

with Qualcomm I did what I always do

33:43

and plugged in my laptop and

33:46

then we have them like. That's actually

33:48

pretty funny. Mock me for plugging my laptop

33:50

and then they like they coax me back

33:53

into the meeting room with a Snapdragon laptop

33:55

and a sponsorship. That's

33:58

pretty good. So they were. They

34:00

were absolute, like, totally

34:02

fun. Yeah, had a lot

34:05

of fun working with them. We've actually, I think,

34:08

attempted to work with

34:11

Qualcomm once. Ooh, sorry, someone just pointed

34:13

out Ballerskate 3 is already available on

34:15

macOS, so it's ARM. Oh,

34:18

okay. That's good to know.

34:20

But I do believe there's compatibility layers for tons

34:22

of different games and stuff. Yes. Not

34:24

everything, but a lot of things. And

34:28

they had, what's it called? Sony

34:30

used to own it, that video

34:32

editing suite. Thank you, Imperial. Vegas.

34:34

Vegas, yeah. Vegas, they had Vegas

34:37

running. Man, they had so much

34:39

stuff running. Premier doesn't

34:41

work, is that correct? I

34:43

would be very surprised if Premier,

34:46

Premier barely works on Windows x86.

34:50

No, seriously. You know that our

34:52

away teams bring Macs. Ooh.

34:55

They don't bring Windows laptops anymore. Whoa.

34:58

Because Premier is so much more stable on Mac, which

35:01

I don't think is anything to do with

35:03

ARM. I think it's

35:05

just to do with just

35:08

Windows versus macOS. That's wild.

35:13

Yeah, sorry. An

35:17

interesting thing too, you had a narrative

35:19

here of people trying to push for

35:22

more fab capacity. Obviously

35:24

this is somewhat always of a thing, but

35:26

it's been more of a conversation lately. A

35:28

huge amount of conversation both on the floor

35:30

for the limited time that I was there

35:32

and off the floor with a variety of

35:34

people at the show has been data

35:37

centers, building new

35:39

data centers, increasing

35:41

how much you're harnessing current data centers.

35:44

Data centers trying to get out low

35:47

level clients because there's high level clients

35:49

offering three times the rate on things.

35:52

Do you remember? It's crazy. Do

35:54

you remember that data center zone property that I showed

35:56

you like five years ago? Oh dude. And I was

35:59

like, oh, well, you should like. do float plane data

36:01

center, you would have been balling.

36:07

Well, oh, well. And

36:11

it was really small. Like the overhead

36:13

of running a tiny data center, I

36:15

think would be super weird. Because like,

36:17

as far as my understanding goes, scale

36:20

is pretty helpful. Yeah. Yeah. Because you're

36:22

gonna have to have the security. Anyway,

36:24

you're gonna have to have the administrative

36:26

staff. Anyway, you can't

36:28

just have like, you know, few thousand

36:31

square foot data center. But yeah,

36:35

yeah, no dice, no dice. Oh,

36:39

yeah. And back then, like, when even when even

36:42

was that that was during the like 2017 crypto

36:44

boom? Yeah, 100%. Someone's like, it would have

36:47

been crypto

36:49

don't lie. No, 100%. The tenants would

36:51

have been mining 100%. But I was

36:53

also when I

36:56

was talking to Luke about

36:58

it. I mean, we had already

37:00

been kicking around. LMG VPN. Yeah,

37:02

yeah, yeah. We

37:05

already had float plane going. So it was

37:07

one of those things where like, kind of

37:09

like we do, we'd be spinning up a

37:12

thing to be our own biggest customer. And

37:15

I don't know, might have worked out.

37:17

But realistically, I don't think I don't think that

37:20

cash would have been that well invested there compared

37:22

to some of the other things we did just

37:24

like growing LMG and all of that stuff. A

37:26

lot of people are posting a lot of things

37:28

about the data center stuff. Someone someone mentioned data

37:30

centers use a lot of power. Yeah, yeah,

37:33

that's gonna be a huge

37:35

problem. If you've paid attention

37:37

to some of the announcements

37:39

that happened at this show.

37:41

Wattage per like device in

37:43

a data center is erupting,

37:45

dude, the new

37:48

Blackwell GP. Yeah. Okay,

37:50

so I'm just talking

37:52

the one GPU, which

37:54

is to be

37:56

100 dies. Okay, so a B 200

37:58

GPU. is 2B100

38:01

dies and oh God,

38:03

I think it is, is

38:05

it 12 HBM stacks or 8 HBM

38:08

stacks? I forget, 2B100

38:10

dies and some HBM stacks. Nvidia

38:12

is calling that one GPU. So

38:14

they are, they're doing multi-die GPU,

38:18

kind of like we saw on Apple's Ultra SKUs.

38:20

I hope that it works a little better than

38:22

it does on Apple's Ultra SKUs. I don't know

38:24

if you remember this, but we did a video

38:26

on Apple's tools to

38:28

help developers port their games over to

38:31

macOS and performance sucked

38:34

and a lot of the Apple community was

38:36

extremely angry at us for testing on Ultra

38:40

because it has problems and

38:42

I'm like, okay,

38:44

well I didn't know that. Yeah. I

38:47

thought I was putting Apple in the best

38:50

position possible and also if you're mad about

38:52

this, be mad at Apple, not at me.

38:55

I didn't make their interconnect between their

38:57

two, like

39:00

core designs, their two Macs. I didn't

39:02

make that. I didn't make it

39:04

not work properly. Anyway, hopefully Nvidia's interconnect works

39:06

a little better than Apple's, but this B200

39:08

should present as a single GPU, but

39:13

I'm not done yet. A

39:15

B200 is a thousand watts, one B200, but

39:21

I'm not done yet, okay?

39:23

Because a B200 is part

39:25

of a grace Blackwell super

39:27

chip that operates as a

39:30

single chip or whatever, where

39:32

you've got two B200 GPUs

39:35

and then a grace CPU on

39:37

it, okay? And that- Like some

39:39

amount of resentment in the like,

39:41

whatever. And that is 2,700 watts.

39:46

Oh yeah. Because we got two Blackwells and

39:48

a grace, okay? But

39:51

I'm not done yet. Because

39:54

each compute node contains

39:57

two super chips. one

40:00

you think Jake was saying something about

40:03

how some of the partner designs are

40:05

gonna be an ever so slightly thicker

40:07

one you in order

40:09

to get all the cooling in there like they're

40:11

water cool but like I think they're

40:13

creating like a one you plus or something don't quote

40:16

me on that but Jake was talking to me about

40:18

it he's usually yeah he's usually pretty credible about that

40:20

sort of thing but yeah but essentially

40:22

a one you what does that work out

40:24

to math honestly

40:27

I'm so I forgot before

40:29

100 watts what

40:34

wattage density and data centers is

40:36

gonna explode apparently it's an issue

40:38

where like some you

40:41

know fairly established companies that

40:44

run big data are basically

40:47

tapped for how much power they can get into

40:49

their data centers so that's another thing with the

40:51

whole data center game right now is like a

40:53

new data centers are coming up in places where

40:55

they can harness just absurd amounts

40:58

of power and Microsoft's gonna end up

41:00

way ahead of the game with their

41:02

like underwater data centers yeah we're just

41:04

gonna have like a title title data

41:06

center okay title powered like no but

41:08

the whole thing so but like the

41:10

whole data centers under the water for

41:12

cooling yeah and then like basically

41:14

the tide comes in and then like the

41:16

wall the data center comes up and it's

41:18

like a temporary dam and then it just

41:21

runs the seawater back through the thing for

41:23

cooling and power and then rinse and repeat

41:26

yeah we got this yeah like they're

41:28

gonna take up like the entire post

41:31

of freakin like Alaska with data centers or

41:33

something like for real though like

41:35

no but also they're gonna have to

41:37

do something it's gonna be crazy

41:40

yeah so I hear someone did the

41:42

math and flowplane chat this is crazy

41:44

6,000 watts in

41:47

a U of rack space would be over

41:50

a quarter of a million

41:52

watts per rack that

41:55

I mean it would it would melt we're starting to

41:57

talk numbers that are just like absolutely

42:01

stupid numbers. Yeah.

42:03

Like... Multiply that by a data

42:06

center. Dude, my tiny data

42:08

center in White Rock or whatever,

42:10

it could have had one

42:13

rack of raised black walls in it,

42:15

in an empty room. Yeah. Like,

42:18

I don't even think we could have gotten that much power in

42:20

there. Oh

42:23

yeah, I doubt it. Like here. I

42:25

doubt it. How much? Okay,

42:27

how much? Do you know how much power we have left?

42:29

As far as my understanding, are being

42:31

chosen by the proximity to nuclear power

42:33

plants. Specifically

42:35

nuclear power plants. That

42:37

is hilarious. And

42:40

there's conversations going on about like being

42:43

concerned that some nations aren't going to be

42:45

building additional nuclear power plants fast enough to

42:47

power all the data centers that are going

42:49

to be profitable. So that's

42:51

an actual conversation. Our power to the left,

42:54

which is a 20,000 square foot

42:56

industrial building that was formerly used

42:59

by... Steel fabricationers.

43:01

Yeah, some kind of steel fabricator.

43:06

Has about 200,000 watts coming into it. And

43:10

that was like a lot. That's

43:15

industrial, you know? Yeah.

43:17

So I could power not even an

43:19

entire rack of graced black walls.

43:22

So it would just be an entirely empty shell.

43:24

Imagine the lab. With

43:27

one super dystopian is just like in the

43:29

middle of the whole thing. One cabinet. All these

43:31

cables running through. Sitting in the middle of

43:34

it. That's it. That would be wild.

43:38

Hilarious. Just like venting all

43:40

the heat out the roof

43:42

or something like that. You

43:45

could sit over it and cook your food.

43:49

It's crazy, dude. It's freaking crazy. Yeah.

43:51

Data centers and power are going to

43:54

be a huge thing. And if you

43:56

are someone who is sash about the

43:59

general... power usage that we have going on right

44:01

now. It's

44:05

definitely not getting better. With

44:08

that said- AJ just said that the

44:10

Badminton Center has more power than the lab. I

44:15

didn't know that. That actually makes

44:17

a lot of sense. The Badminton

44:19

Center is 40,000 square feet. Right,

44:22

okay. It's two 20,000 square

44:24

foot units. So it probably has considerably

44:27

more power than the lab. Are

44:30

you thinking what I'm thinking? Yeah. Badminton

44:32

data center. Here

44:34

we go. That we use not for

44:36

logical things, but

44:39

for analyzing gameplay. Dude,

44:42

I saw some demos at

44:44

NVIDIA that probably skewed

44:47

the entire attitude of my coverage toward

44:49

them this year. Oh. Okay,

44:52

because like I know a lot of gamers- I didn't

44:54

go there at all. I know a lot of gamers

44:56

are super mad at NVIDIA right now and like rightly

44:58

so. Yeah. The video's

45:00

apathy towards gamers has been apparent for quite some

45:02

time. They didn't mention the keynote at all. I

45:05

don't know. I didn't watch it.

45:07

But dude, they had

45:09

one of their Jetson like

45:12

edge computing devices. It's

45:15

like a new generation Jetson thing. And

45:19

they had it running this like

45:21

stereo camera capture, Majig,

45:24

and on this dataset that was trained

45:26

on Blackwell or whatever. Basically

45:29

the demo was real time

45:31

wire framing. So like

45:33

the stuff that we wanna do. Exactly

45:35

what I wanna do at the Badminton

45:37

Center for like- It's pretty useful. Stroke

45:39

and movement analysis and stuff. So

45:42

one of the dreams that I have is

45:44

every court having basically

45:47

that except trained

45:49

on a dataset of Badminton play

45:51

so that you could like gamify

45:54

your game, your gameplay. So

45:57

it could tell you like, oh, against

45:59

this opponent. you hit a lot of

46:01

smashes, but their smash defense was actually

46:03

really good. And you scored most of

46:05

your winners with your drop shot. You

46:08

did a drop shot 15% of the time. Maybe

46:12

try doing it slightly more often.

46:14

Like, dude, oh, so cool.

46:16

And okay, they had some really cool

46:18

gaming demos running too. And

46:20

so like nothing I said was not true. But

46:23

when I walked in the door, I

46:25

was just, I was a little

46:27

excited. I was a little excited,

46:30

okay? Yeah, yeah, I even think things

46:32

like, this might not be as interesting to you,

46:34

but I think it'd be pretty cool, is like

46:37

endurance evaluation. Like

46:40

having it notice like, oh, okay, once you're

46:42

like 15 minutes into a game, your form

46:45

on like this movement starts to falter. Yeah.

46:47

Like that's really interesting. Dude, you could use it for all

46:50

kinds of stuff. Yeah, yeah, yeah. So

46:52

cool. It's pretty sweet. Yeah. Obviously

46:55

I'm going to need a small

46:58

loan of a million dollars for, you

47:00

know, some grace Blackwell. Yeah, but that's,

47:02

that's like, if you end up

47:05

waiting, you know, a little longer and then it becomes

47:07

the old tech, then it'll be

47:09

a little bit more affordable. Yeah, maybe. And I

47:12

might've put, I

47:16

think about 10,000 Watts of cooling

47:18

and power into the data

47:20

room at the badminton center, just in

47:23

case. We'll get her done. Just in

47:25

case. And

47:27

like, realistically, realistically,

47:33

like you wouldn't, you wouldn't

47:35

have to do the training on site. Like

47:38

you would probably at least time to do that

47:40

anyway. Like, I don't think you would at the

47:42

scale we'd be operating at. I don't think we

47:44

would be doing our own data set training anyhow.

47:47

So as long as we have enough edge computing should

47:49

be fine. Yeah.

47:52

Yeah. And I think at a St, we

47:55

wouldn't be stayed on higher

47:58

some launch lines. 100% there

48:00

was one of them. And I think

48:02

this just very, the text getting pretty

48:05

there. Yeah, pretty fast. But it also

48:07

like, oddly specifically

48:09

validated that thought. It's

48:12

like, maybe you should wait for it. Then

48:14

it's just like shown to you at a

48:16

convention. Like I wouldn't have expected to see

48:18

that. Yeah, media booth, dude, and like, it

48:20

was so good. They had, they had some

48:22

3d models rigged. And they were

48:25

showing, like, just

48:28

live movement to

48:30

rigged 3d model, like, that's

48:33

cool. It's not perfect. Obviously.

48:37

But like, it's

48:39

real time stuff. There's a lot of stuff you're gonna have

48:41

to deal with that something like a public badminton center like

48:46

do what happens if like someone else walks

48:48

into court like weird weird things that you'd have

48:51

to deal with. Oh, yeah. Like actually running

48:53

it in production. But our privacy policy from day

48:55

one is going to be we're filming the

48:57

inside of this facility like it or lump

48:59

it. I'm sorry. Yeah. And you

49:01

know what, we're not gonna we're we take

49:03

privacy extremely seriously, we're not going to misuse

49:05

that data, we're not going to sell it.

49:08

But we're absolutely going

49:11

to use it for

49:13

you. Would you potentially sell it to

49:17

users of the facility? sell

49:19

it to the Oh, oh, like

49:21

the gameplay analysis? Yeah. Okay, then because like

49:23

technically, that's not selling the data. That's selling

49:26

what we derived from it. But

49:28

we would be monetizing. What if what if you

49:30

let people buy like, I don't

49:33

know, replays of their

49:36

games or something? Well,

49:38

that's, that's not the data

49:40

we're collecting. But that would be like,

49:43

yeah, but then you're buying your own. Other

49:46

people could be in the video. Yes, this is

49:48

true. Even someone walking by the car. So we'll

49:50

just have to make sure that our privacy policy

49:52

does account for that. I think I think there's

49:54

got to be a way to make a distinction

49:56

between I don't know, not I don't know legal

49:58

stuff. Yeah, I There's got to be some

50:00

way to make a distinction between like, we

50:03

will be, you know, potentially distributing

50:05

whether we're selling or not, clips

50:07

to users of the facility of their own gameplay

50:09

and you might end up in it versus we're

50:12

going to mass sell everything we have. There's got

50:14

to be a way to... No, I'm sure we

50:16

can handle that. And

50:18

realistically, because of the way the facility is laid

50:21

out, each court will have

50:23

its own camera and yeah, you'll be able

50:25

to see the one behind it, but

50:27

they'll be pretty far away. Oh, I see

50:29

what you mean. Yeah, because it'll be at kind of like a 45 angle. Isn't

50:33

there a wall? No, but well,

50:35

you'll see the one next to it. Oh, I see

50:37

what you mean. Yes, yes, yes, yes, that's sorry. Going

50:39

to be sick. Are you guys ready for some merch

50:41

messages? Oh yeah, Dan, we have no idea what we're

50:44

supposed to be doing, so you feel free to answer. That's

50:51

why I'm buttoning the cards, Dan. There,

50:55

they got lost in transit. Do

50:57

it, do it. Alright,

50:59

let's see here. Luke, how is Final

51:01

Fantasy VI coming along? Yeah,

51:05

Luke. Man, you suck.

51:09

My plan is to play it on the flight back.

51:12

Sure. I brought the controller. Sure. I

51:15

finally tried Cuphead. Yeah? It's

51:17

alright. Yeah. It's alright. Yeah, I played

51:19

the first like third of it, according

51:21

to the percentage. I think I'm

51:23

good. I think they get it. That sounds about right to me, to be honest.

51:26

I think I played probably somewhere about... The art style is super cool. Yeah.

51:30

I think it was worth it. Like,

51:32

it wasn't a full price game. Yeah,

51:34

no. I was happy with the purchase.

51:37

Yes, finally tried it. Are you

51:39

still at like... No,

51:41

I'm a little bit past there. Where are you in the

51:43

game now? You're past Zozo?

51:46

Yeah. Okay, so have you done the Opera

51:48

House yet? I don't think so. Oh, okay.

51:51

I think I like just got through Zozo. Okay.

51:55

Man, you're slow. I'm slow.

51:58

Yeah. Like you're a slow boy. I intended to play

52:00

it. I probably game more than you now. Sad.

52:04

Well, I don't know. If you include Pokemon Go

52:06

walks, probably not. No, I don't think I

52:08

do. Oh, well then. That's

52:10

not a real game, you filthy casual. That's

52:12

fair. No, no, I'm just, I'm just, I'm

52:15

being intentionally. I am

52:17

strongly opinionated. I'm being intentionally

52:19

toxic. I saw

52:22

a post on the Pokemon Go subreddit where someone was

52:24

asking like, should I buy

52:26

this like apartment or something? Buy this

52:28

apartment. And it showed the stops and

52:30

gyms that were around the apartment. And

52:32

there was comments like, if that gym

52:34

is in range of like the couch,

52:37

then like hell yeah, man. And

52:39

like, I do not. Pokemon

52:45

Go is a bad game. I

52:49

think I can say that honestly, like

52:51

barely objectively. It's not, it's not a

52:53

good game. It's great at

52:56

making going out for a walk

52:58

fun. Go outside, please.

53:01

It's so, I feel so separate from the Pokemon

53:03

Go community because there's constantly stuff like that. People

53:06

are mad that they can't remote raid more, which

53:08

is like being able to raid from your couch

53:10

instead of going out and doing it. People are-

53:12

And they understand the accessibility argument. Yes. But

53:15

a lot of those people- That's not the argument for

53:17

the vast majority of people. If that is the argument

53:19

for you, heck yeah. Sorry,

53:21

you can only do it so many times a

53:23

day, but that's also like really expensive. So I

53:26

don't know. Oh boy. Cool.

53:30

Hit me, Dan. Oh wait, we need

53:32

to explain merch messages. Yeah. Right.

53:35

Merch messages are the way to interact with the

53:37

show. They're going to be going to producer Dan,

53:39

who I don't have a button to show him,

53:41

but maybe he does. Wave to the people, Dan,

53:43

maybe, I don't know. Can't monitor the stream. Stop

53:45

it Linus. Yeah, there you go. Stop

53:49

it Linus. Okay. I

53:51

am stopped. Anyway, merch

53:53

messages are the way to interact with the show.

53:56

Don't leave a super chat or a

53:58

Twitch bit or whatever. Leave

54:00

a merch message. All you gotta do is

54:03

go to lttstore.com and

54:05

in the cart, once you've added some

54:07

items, loaded it up with some super

54:09

awesome stuff from LTT store, you will

54:11

see a little box to leave a

54:13

merch message. That will go to

54:15

Dan who will forward it to the appropriate person.

54:17

Pop it up along the bottom of the screen

54:19

for everyone here to enjoy who

54:21

might reply to it himself or he

54:24

might curate it for me and Luke

54:26

to read. We've

54:28

got some cool stuff over on the

54:30

store. Hey, look at that. We

54:33

are speaking of,

54:36

we were just talking about, hold

54:40

on, let me see if I can, where do I

54:42

even find this stuff? Hey, there they

54:44

are. Oh wait, I can't screen share anyway. We're

54:47

relaunching our keyboard pins.

54:50

So they're available in a variety

54:52

of different colors, including,

54:54

oh, we've got

54:57

that one. Our

55:04

RGB color is sick. We've

55:06

also got rainbow and gold

55:09

and purple, yellow and white and pink, blue

55:12

and purple and blue, pink and yellow and

55:14

blue, pink and white and purple,

55:17

gray and white, all these cool

55:19

color schemes. And they are free

55:21

in the bonus bin with your

55:23

purchase in any color and representation

55:26

you like. Let's freaking go. Also,

55:30

if you missed out on the scribe

55:32

driver last week, so that's our fail

55:35

pen made out of failed screwdriver shafts,

55:37

we are working on a restock. So you

55:39

can sign up for a notification on the

55:41

site. Make sure you guys do that. Dan,

55:44

if you wanna show them where the notify button is, that's

55:47

gonna be a good way to ensure that you

55:49

get one. Guys, we don't mess around when we

55:51

say, hey, something's selling really fast, you should get

55:53

it. We're not doing like awful FOMO sales tactics.

55:56

We're informing you that something is selling really

55:59

fast. And if you would like to get

56:01

one, then now is the time to get

56:03

it. Yeah. I have an unfortunate situation where

56:05

I actually wanted to... I can say this

56:07

because he's not getting them, but

56:10

I wanted to get the scribe driver for

56:12

gifts for things like Father's Day. Yeah.

56:14

It's gone. Well, you should have moved fast. I

56:16

should have. I mean, I did warn you. You

56:18

did. Personally. In life. I

56:20

really did warn you that they were running out. And

56:22

I was like, oh. By the way, we

56:25

missed a huge opportunity. I

56:27

saw this on Reddit, I think it was. It

56:30

should have been called the write off. Oh.

56:34

So many, so many layers

56:36

of meaning. Wow.

56:38

Yeah. It

56:40

sucks. You made like a pencil version or something.

56:42

It's not as good. Yeah. It would need to

56:44

be big. Because it was the shafts. Yeah.

56:48

It was the write off. Oh, man. See?

56:51

See how well it works? Oh. Huge

56:54

miss. Yikes. Huge L. It

56:57

sucks when those things happen. I know. It's

57:00

like describe drivers of an interesting name.

57:03

Yep. It's fine. It's

57:05

fine. But it isn't. It's not right

57:07

off. Also

57:10

in other store news, we

57:12

finally have the oh, man. Our

57:17

cable management products under other. Where

57:20

are they? Gear.

57:22

This is a problem. Tools. Where is it?

57:24

I thought it was under tools. Are

57:27

they not under tools? We have too

57:29

many. Oh, we have a whole top level category for cable

57:31

management. Okay. Well, there's a problem. Okay.

57:34

Anyway, the cable tie holders are back in stock.

57:36

We must have air shipped in some cable tie

57:38

holders. So we finally have

57:40

all of our magnetic

57:42

cable management products in stock. The reviews are

57:44

in. They're freaking awesome, which

57:46

we already knew. But hey, now you

57:49

guys have it independently verified. They're all

57:51

four and a half or pure

57:54

five stars. Scribe driver

57:56

reviews are in now as well. Yeah.

57:58

So are in the comments or saying. Flipping

58:00

loving it, the people who are calling

58:02

it the uber-priced merch are

58:05

getting obliterated. Get

58:07

smashed. Obliterated. It's

58:10

actually been very satisfying to see

58:13

how destroyed they're getting. It's

58:15

like, dude, look, I'm sorry

58:18

that you buy all of your t-shirts

58:20

and underwear at Walmart. I'm

58:23

sorry for that. Sometimes

58:27

it can be nice to have

58:30

one that's a little nicer. I'm

58:32

not going to apologize for our stuff being

58:34

more expensive than Walmart. I will even say

58:36

the big crystal pens are

58:39

super based. But

58:41

if you want a nice pen, they cost the money.

58:43

$30 is a really good price. This

58:48

isn't trying to compete with a big crystal.

58:51

And that's okay. Deal with it. Even

58:55

the $20 t-shirts. $20

58:58

p-shirts are pretty cheap. Deal with it. Actually,

59:01

there's pressure on me to increase the price of

59:04

the t-shirts. I'm not surprised. Yeah, we haven't touched

59:06

them since. I know. I've mentioned this a bunch

59:08

of times because it seems crazy to me. Our

59:11

costs have definitely gone up. Oh, yeah. We

59:13

absorbed it for a long time, but at some point,

59:16

t-shirt prices are probably going to go up a little

59:18

bit. It probably makes sense. I

59:20

don't know of any

59:22

other creators that have

59:24

pricing around there. And

59:26

not to throw too much shade, but a lot of them

59:28

are on bad. Yeah. We

59:30

take a lot of flack for the pricing on

59:32

our store that is just stupid,

59:35

honestly. It was notorious

59:38

for a long time. I think it has gotten

59:40

better with creator merch in

59:42

general. I think creator merch in general is

59:44

better now than it was five years ago.

59:46

There's some stuff that's still pretty garbage. Probably.

59:48

We secret shopped someone else's water bottle in

59:50

the tech space a little while ago, and

59:52

it was terrible. I'm not surprised. I couldn't

59:54

believe how bad it was. But it's pretty

59:56

common to get the cheapest. They're

59:59

not necessarily... doing this consciously,

1:00:02

they're working through another place

1:00:05

that is actually like manufacturing the shirts for

1:00:07

them. And that company that they're

1:00:10

working with is providing the

1:00:12

cheapest possible blanks, just like feel terrible,

1:00:14

the printing's really bad, stuff like that.

1:00:17

Sometimes it's not even their fault. Yeah. We

1:00:20

ran into that a number of times where we

1:00:24

would get good samples. My actuals are just bait

1:00:26

and switchee with samples. Yeah. That was why we

1:00:28

started to create

1:00:30

a warehouse because we were so tired of that.

1:00:32

Yeah. We couldn't control the quality of our own

1:00:34

merch. I think that other creator that we secret

1:00:36

shopped is like an investor in that merch company.

1:00:38

I think that's probably the only reason they use

1:00:40

them because... Trig and Floatplane Chat

1:00:43

just said the Scribe Driver pen is

1:00:45

literally one of the lowest priced high

1:00:47

quality bolt-action pens on the market. Yeah.

1:00:51

It's not competing.

1:00:53

And you know what? Some of our stuff is

1:00:55

really expensive. That's true. But

1:00:57

I talked about this in my video on the PlayStation

1:01:02

portal. It's expensive

1:01:05

for what it does. And

1:01:07

if that doesn't have a value to

1:01:09

you, that's totally okay. You can just

1:01:11

totally not buy it. And that is

1:01:13

totally an option. But what it

1:01:15

isn't is overpriced

1:01:18

because I broke it down. And people

1:01:20

were so mad about this, which was bizarre to

1:01:22

me. I broke down

1:01:24

what the retail price for

1:01:27

all of those components would be. And I'm like, yo,

1:01:29

Sony isn't taking any more profit on this

1:01:32

than they are already on their controller. Or

1:01:35

then sellers on eBay are

1:01:37

on batteries and screens

1:01:39

this size and resolution. And

1:01:42

you can be mad that it's locked

1:01:44

down. That it's not hackable. Or you can

1:01:46

be mad that it's an accessory for that

1:01:48

might make it a no-go for you. Yeah,

1:01:50

100%. Yeah. But what you

1:01:52

can't say is that it is

1:01:55

overpriced because that's just what that

1:01:57

costs. It's like if you were to buy a...

1:02:00

It's like if you were to buy a gold iPhone

1:02:02

case and it was $30,000 because it

1:02:04

contains $25,000

1:02:07

worth of gold and like, you know, 50 to 60 hours of craftsman time. Well,

1:02:16

that's not overpriced. You just

1:02:18

bought something a little

1:02:20

crazy. Yeah, it's expensive. It's

1:02:23

not overpriced. And that's a distinction that I would

1:02:25

like to kind of

1:02:28

drive through more in our

1:02:31

future videos. Like, I don't

1:02:33

know, something that I'm

1:02:35

a little sort of disconnected from,

1:02:38

I think, the rest of the tech media

1:02:40

on is the

1:02:43

state of the GPU market.

1:02:46

I'm mad about it that they're really expensive.

1:02:49

But what people I think are

1:02:51

not fully understanding

1:02:57

is the forces at play. I

1:02:59

mean, there are egregious examples. I talked earlier

1:03:01

on this show about Nvidia's 4060 series. There's

1:03:05

not enough RAM on the 4060. At

1:03:08

that price, it should have more

1:03:10

VRAM. Nvidia is absolutely taking that

1:03:13

margin on it. But

1:03:15

with that said, the days of $139.99

1:03:17

like gaming GPU are gone. It

1:03:24

costs so much to take out

1:03:26

at TSMC. So if

1:03:29

you want that $139 GPU, literally your

1:03:31

better option is to buy a second

1:03:33

hand one from a generation or two

1:03:35

ago when that price level might

1:03:37

have been attainable. And

1:03:40

no amount of complaining about it

1:03:43

is going to change that. It's

1:03:45

going to change the market forces that

1:03:47

are at play here. The fact that

1:03:49

TSMC can sell that way for functionally

1:03:53

unlimited monies. Did

1:03:56

you hear TSMC's chairman

1:03:59

or CEO? whatever the guy's title

1:04:01

was, like was

1:04:03

like publicly mulling increasing Nvidia's

1:04:06

pricing. It's just like,

1:04:08

yeah, I don't know. Yeah, we've been looking at how much

1:04:11

money they're making and we

1:04:13

think they could probably absorb a price increase.

1:04:16

And you know what, I don't think

1:04:18

that, I don't think from TSMC's point

1:04:20

of view, who functionally has a monopoly

1:04:22

on cutting edge node technology

1:04:24

until Intel gets their act together. I

1:04:27

don't think from TSMC's point of view, they're going to be like,

1:04:30

oh, but what about the poor gamers?

1:04:32

No. Okay, if you're making

1:04:34

G-Force, we'll give you a better deal. That's

1:04:36

not going to happen. Why would they? Why would

1:04:38

they? It makes no economic sense

1:04:41

to do so. So it sucks.

1:04:43

It's not good. It sucks

1:04:45

a lot. It's particularly very bad for us. We've

1:04:47

talked about this a bunch of times though. I

1:04:50

lament often about how when

1:04:53

GPU crypto mining stopped being as much

1:04:55

of a thing, I

1:04:58

was really hoping for a bit of a market crash

1:05:00

so they would feel it a little bit. But

1:05:02

then the AI rise just, it

1:05:05

happened right at the perfect possible time.

1:05:07

Like the stars aligned and just allowed

1:05:09

them to never feel that hit. They're

1:05:11

either really smart or really lucky or

1:05:14

both. Realistically for

1:05:16

success, you need a combination of the two,

1:05:18

which I guess leads us perfectly into our

1:05:21

next topic. Nvidia is

1:05:24

the number two most valuable

1:05:26

company in the

1:05:28

world with a market

1:05:31

cap of 3.012 trillion

1:05:33

with a T. This

1:05:36

makes Nvidia only the third company to cross the 3

1:05:38

trillion threshold. So it's Microsoft,

1:05:41

Apple, and Nvidia. Nvidia

1:05:43

stock appears particularly attractive to retail

1:05:45

investors. So these are everyday consumers,

1:05:48

rather than just professional

1:05:50

portfolio holders. And

1:05:52

they are likely contributing significantly to Nvidia's

1:05:55

upward momentum. Much of the success is

1:05:57

due to their AI chip portfolio, but Nvidia has also reached a

1:05:59

new level of success. new peak of 88% market

1:06:03

share for discrete graphics. Holy

1:06:05

crap. That is the highest it's

1:06:07

been since the company was founded, which is kind

1:06:10

of wild to me because AMD's products right now

1:06:12

are like pretty good. I

1:06:15

am I am still Radeon challenging.

1:06:17

Are you and I'm

1:06:19

fine. Nice. Well, I committed publicly

1:06:21

to skipping the the 40s. Right.

1:06:23

Yeah. And like, you don't have

1:06:26

any more driver issues or anything? For the number of

1:06:28

times I've been called a liar. I'm

1:06:30

a pretty damn honest person. I

1:06:34

said I'd do it. And I'm doing it. Like

1:06:37

I do. So no problems.

1:06:39

So in in my in the

1:06:41

land PCs in the basement. Yeah.

1:06:44

You know, last hurrah, I picked

1:06:46

up used EVGA 30 series. So

1:06:49

it's not like I don't have any Nvidia

1:06:51

in the entire house or whatever. Yeah. But

1:06:53

when the 40 series came out, and

1:06:55

was so under your

1:06:58

primary computer that you distribute to

1:07:00

multiple screens across the house and

1:07:02

use very often 7900 XCX, maybe.

1:07:05

Yeah. Yep. And I think it's

1:07:07

not a cop out that other computers in

1:07:09

the house have other GPUs. He literally like

1:07:11

when when when

1:07:14

we went over to your place to play hockey tape

1:07:17

to tape. Yep. It was

1:07:19

your upstairs computer. Well, okay. Actually,

1:07:24

no, no, I ended up in the basement, I guess.

1:07:27

Oh, did you? Yeah, because I was

1:07:29

not the land before. Oh, yeah,

1:07:31

yeah, yeah. Yeah, that was my desktop. Yeah, at

1:07:33

the land. It was this for a variety of

1:07:35

reasons. But before I

1:07:37

wanted to be able to use my computer. Yeah. If

1:07:39

other people were in the theater. Yeah. Watch.

1:07:43

But yeah, no, I was using my

1:07:45

upstairs computer. Because the computer's downstairs,

1:07:47

but the screen is up. Well, one of the

1:07:49

one of the more I kind of like it's

1:07:52

like, he like usually uses it upstairs, but the

1:07:54

computer itself is not there. Yeah. And

1:07:56

you know what it was, I had some issues with

1:07:58

it at first. It's

1:08:00

been rock solid. That's really good. That's good to hear. Yeah,

1:08:02

it's been rock solid. I

1:08:06

had CompuTex and Vidia outlined its roadmap

1:08:08

for chip architectures with Blackwell Ultra in

1:08:10

2025, Rubin in

1:08:12

2026 and Rubin Ultra

1:08:14

in 2027. They also

1:08:17

showed off their upcoming RTX AI

1:08:19

PCs with co-pilot plus features, though

1:08:21

they seemed reluctant to use Microsoft

1:08:23

branding of co-pilot plus PCs, or

1:08:26

to acknowledge that these RTX PCs

1:08:28

are powered by AMD Strix Point

1:08:31

CPUs. Fascinating.

1:08:35

Discussion question here. We want the glory and

1:08:38

all of the glory. And Vidia is clearly

1:08:40

killing it, but does that really mean it's

1:08:42

worth one and a half Googles slash alphabets?

1:08:44

I feel like this is gonna pop. I

1:08:47

just don't know when, but not investment advice. I don't

1:08:49

know. It just seems like

1:08:51

too much. It feels like another

1:08:56

version of Nortel. Yeah,

1:08:59

I was looking at this pop up

1:09:01

in my Twitter feed, a

1:09:04

comparison of Cisco's meteoric

1:09:07

rise and then leveling

1:09:10

off. I mean, what goes up

1:09:13

exponentially must come

1:09:15

down at some point. But would

1:09:17

I bet against it today? Dude,

1:09:19

if you shorted it when that first

1:09:21

massive spike happened and then

1:09:23

this event happened and it went up

1:09:26

again, you're hurt. Yeah. Like

1:09:29

that would be a really quick way for me to

1:09:31

go completely bankrupt in

1:09:33

like days. Yeah. So, yeah,

1:09:36

I don't know. I don't know, man. If you got

1:09:38

the tea leaves on you, I suspect there's some big

1:09:41

money to happen there, but I don't know. Not

1:09:44

me. Not

1:09:46

me at all. And like

1:09:48

I was number one. Number

1:09:51

one is Microsoft. Not

1:09:54

by much. So

1:09:57

let's see. Let's see who ends up with the

1:09:59

higher valuation. The shoveler. or the shovel maker? I

1:10:01

was just going to say the shovel seller has

1:10:03

not yet surpassed the shoveler. Yeah.

1:10:05

We'll see. But

1:10:07

I mean, in fairness, Microsoft has a lot

1:10:09

of other stuff going on. What I want

1:10:11

to know is like what's next for NVIDIA

1:10:13

because NVIDIA

1:10:17

is one of those companies that

1:10:19

values partnerships until

1:10:21

they don't really feel like it anymore. Until

1:10:24

they don't really feel like it anymore. And I

1:10:26

wouldn't say that they've necessarily Sherlock'd too

1:10:28

many partners. But they

1:10:31

definitely go to loners slash squeezers.

1:10:33

There's a reason NVIDIA

1:10:37

and Apple don't get along. Too

1:10:42

similar. You can only

1:10:44

fit so much ego at one

1:10:46

negotiating table. And

1:10:49

so yeah, I want to see what's

1:10:51

next. Like, okay, look at the way

1:10:53

Bitname behaves about their mining

1:10:56

hardware, right? Yeah. Or has

1:10:58

where they'll like literally first.

1:11:00

Yeah, they'll create a new

1:11:02

generation thing. And like,

1:11:04

at what point with their billions

1:11:07

and billions and billions of dollars, does

1:11:09

NVIDIA kind of just go? Why

1:11:12

are we selling this hardware? I

1:11:15

mean, it's not like the thought never occurred to

1:11:17

them data centers. And it

1:11:19

what about what about, you

1:11:21

know, NVIDIA Enterprise Now instead

1:11:24

of GeForce Now? Why?

1:11:27

That's not even that's actually in

1:11:29

my opinion, that's stronger branding than

1:11:32

GeForce Now. Enterprise Now is actually

1:11:34

wicked brand. Holy

1:11:37

crap. It's not like they couldn't afford to

1:11:39

build a startup. And you're just like, that's

1:11:42

the product that one I don't know the

1:11:44

naming of that for a startup company as

1:11:46

a product for a startup company is very

1:11:48

strong. And so I just like I'm just

1:11:50

looking at it going, why

1:11:53

wouldn't they just take their first

1:11:55

100,000 Blackwell chips and

1:11:57

just give themselves even

1:12:00

like a three month lead. Yeah,

1:12:02

just lease it. I mean,

1:12:04

that's the same thing that Ari does

1:12:07

with their top level cameras. You

1:12:09

can't buy their top level camera. They

1:12:12

can't bin enough perfect sensors at that

1:12:14

size or whatever, whatever the reasons for

1:12:16

it are. The cost would be so

1:12:18

high that they just like, no, you rent this.

1:12:21

We literally will not sell it to you.

1:12:24

So that way we can ensure anything, shot

1:12:27

on Alexa, whatever that, hey, hey, Andy,

1:12:29

Andy, what's the one you can only rent?

1:12:33

The Ari 65 or something like that. Yeah, you

1:12:35

can't buy it. Oh, are you checking enterprise now

1:12:38

on GoDaddy? Yeah, it's already bought. Of course it

1:12:40

is. It's being camped by someone. Yeah,

1:12:42

yeah, yeah, yeah. Well, Nvidia can afford it. Yeah.

1:12:47

Dude, I- I'm kind of

1:12:49

wondering if someone just sniped that. Like, well, we- No,

1:12:51

no, no, no, no. There's no way. That

1:12:53

would have been pretty fast. Two dictionary words,

1:12:56

like 12 characters. It was octave. Yeah,

1:12:58

it's probably already taken. Man,

1:13:03

I don't know, man. Does it

1:13:05

mean it's worth one and a half alphabets slash

1:13:07

Googles? Like that's, here's the

1:13:09

thing though. Like AI is going to

1:13:11

be this tool that underpins everything. It's

1:13:13

going to underpin alphabet and Googles efforts

1:13:16

to better target their advertising. It's

1:13:18

going to underpin Facebook's efforts to,

1:13:20

you know, better understand everything about

1:13:22

you in your life and

1:13:24

target you. Like it's going to- And to be fair, a lot

1:13:26

of it already does. I

1:13:29

think it's just like development in this space is accelerating

1:13:32

extremely rapidly. Enterprise

1:13:35

trust in proposals

1:13:37

and willingness to accept proposals of

1:13:40

work in this space are

1:13:43

almost blank check right now. Yeah, I mean,

1:13:45

look at Humane. The

1:13:47

fact that they raised that kind of money

1:13:49

for what was just obviously a stupid product.

1:13:54

Do you see their, oh, that's one of our other topics

1:13:56

today. Do you see they're trying to sell? Yeah.

1:14:00

They have apparently been

1:14:02

in talks with HP. I

1:14:05

love our headline here. I'm gonna credit Jessica

1:14:07

with this. Might've been Riley

1:14:10

though. They both worked on the Doctor of

1:14:12

Creek, I think. Humane and HP, a match

1:14:14

made in Hades. Humane,

1:14:16

the company that made a

1:14:18

reportedly bad AI pin, is

1:14:21

trying to sell its business to HP

1:14:24

for $1 billion, or

1:14:26

roughly what Humane was valued at before

1:14:28

their product launched and subsequently bombed spectacularly.

1:14:31

According to an article from the New

1:14:33

York Times, the company has sold maybe

1:14:36

10,000 pins, adding

1:14:38

up to around $7 million in revenue. Several

1:14:44

current and former employees told the Times

1:14:47

that Humane's founders essentially

1:14:49

banned internal criticism, disregarding

1:14:52

warnings about product battery life, and

1:14:54

even dismissing a senior software engineer

1:14:56

who raised concerns about the pin.

1:14:59

Add that to the fact that Humane has

1:15:01

warned its dozens of users to not use

1:15:03

the charging case anymore because of a fire

1:15:05

safety risk, and I have

1:15:08

no idea where they are getting

1:15:10

this valuation from. Like,

1:15:12

honestly, I think I would try to buy

1:15:14

ICQ before

1:15:17

I would try to buy it Humane. I'm

1:15:20

actually not sure if I'm unserious

1:15:23

about that anymore. ICQ just

1:15:25

shut down. Is it for sale? Could

1:15:27

I buy ICQ? Owned it.

1:15:30

I don't know. Wait,

1:15:32

what's ICQ new? Hold

1:15:35

on a second. ICQ

1:15:37

new. Oh, hold on. Mail.ru

1:15:41

group in 2010. Oh,

1:15:43

I feel like the politics around buying

1:15:45

it from a Russian entity right now. It's VK,

1:15:47

dude. Wait, VK? VK

1:15:51

is like Russian Facebook. Oh

1:15:53

yeah, yeah, okay. You're not buying this. Yep,

1:15:55

sorry. I

1:15:57

don't think they'll sell it to you. That's a bummer.

1:16:00

I don't know if you'd want to even if I even

1:16:02

if they would Yeah,

1:16:06

I I don't think we need to give money to

1:16:09

For the Russian government. What is I

1:16:12

think you now though. I'll tell

1:16:14

you what mr. Putin Pull

1:16:17

out of Ukraine and I

1:16:19

will consider buying your ICU final

1:16:23

offer I Like

1:16:26

that yeah Yeah,

1:16:31

see chat likes the deal chat

1:16:34

likes the deal Mr.

1:16:36

Putin, it's a good deal your move. It's a

1:16:39

solid deal Line

1:16:46

us out here solving geopolitical problems. Yeah, a

1:16:48

hundred percent All

1:16:50

right, it sounds like VK has been

1:16:52

using ICU is like teens You

1:16:56

can chat with friends in VK

1:16:58

messenger and colleagues in VK workspace.

1:17:00

Oh Sorry, no,

1:17:02

no, no, that's not branding. That's them

1:17:05

saying well They might have just they

1:17:07

might have just used ICU tech and

1:17:09

then just like deployed it We don't

1:17:11

even a different product. Yeah, like Skype

1:17:13

or whatever. Yeah. Yeah Dan

1:17:19

we never finished doing merge messages No,

1:17:22

that's right, we've got one more to do and then we're

1:17:24

about 15 minutes to sponsors as well Let's

1:17:30

see. Hey Linus. Can we get some of the details of how

1:17:32

the GPU factory tour video was made? What was planning and communication

1:17:34

like considering you only had three hours to work with keep up

1:17:36

the great work? Was

1:17:40

a lot of fun very chaotic We

1:17:43

had we had a three-person team Oh

1:17:46

shoot Andy. Do you remember the name of that power color

1:17:48

rep? That was super helpful and high-energy and

1:17:50

basically just like was a total Chad helping

1:17:52

us get through there Yeah,

1:17:56

great guy guy.

1:18:00

He does PR there. I'm really

1:18:03

bad with things, but yeah, he was

1:18:05

awesome. Apparently he

1:18:11

started working there when he was like 18. He was

1:18:13

just like, I'm going to work at Power Color

1:18:15

because I'm a GPU enthusiast and he's just still

1:18:18

there now, which is cool. So

1:18:21

he was a big part of helping us

1:18:23

get that tour arranged framework was

1:18:25

a big part of helping us get that tour

1:18:27

arranged because Power Color ended

1:18:29

up being one of the, well, let

1:18:33

me put it this way at frameworks

1:18:35

volumes and with how high touch

1:18:37

they are, right? Like you, you'll, you

1:18:39

saw it in the factory tour. They had to

1:18:41

create a custom rig for testing

1:18:43

frameworks, stupid, rando, you know, GPUs

1:18:46

with their interface they invented, right?

1:18:48

Like this is not, I'm going

1:18:50

to say stupid. I don't mean

1:18:52

stupid. I mean, inconvenient, right? When

1:18:55

you're set up for PCIE thingies,

1:18:57

go into motherboard, weird

1:19:01

pad interface with pogo pins and

1:19:03

needing a rig that can, you

1:19:05

know, like walk onto it and stuff that

1:19:07

that's like extra work. And I'm sorry,

1:19:10

how many of these are you making?

1:19:12

You know, at the kinds of volumes that

1:19:14

the semiconductor industry works in someone

1:19:16

like a framework is a very, very, very

1:19:18

small fish. They're doing great. They're

1:19:20

growing Rome wasn't built in a day,

1:19:23

but they're a very small fish. And so

1:19:27

the fact that Power Color was willing to

1:19:29

take them on was, was a big deal

1:19:31

and was a big part of bringing the

1:19:33

framework 60 into market. So anyway,

1:19:35

framework has that relationship and they

1:19:37

helped to make that connection as

1:19:39

well. And then once we got in there, man,

1:19:41

it was like, normally, we would want

1:19:43

to go around first, make

1:19:45

a bunch of notes frantically. And then I would

1:19:48

lock myself in a boardroom or something, put my

1:19:50

headphones in and kind of rewalk for like a

1:19:52

couple of hours. And then I would like write

1:19:54

everything and I would ask to have someone kind

1:19:56

of nearby that I can sort of holler at

1:19:59

and ask questions. And

1:20:01

then we would go back through and we

1:20:03

would shoot like a scripted, like,

1:20:05

like a roll read. And

1:20:08

then we would also capture all the B roll. This

1:20:10

time, basically, I just

1:20:13

got a verbal briefing on what we were

1:20:15

about to see. And

1:20:17

then I got a like,

1:20:22

like a more detailed, like I kind of put in

1:20:24

my head sort of how much time I wanted to

1:20:26

spend on each one. And then I

1:20:28

got any little details that pertained to each thing

1:20:30

as we went. And then I

1:20:32

didn't have a script. So

1:20:34

people often ask like, what, which videos

1:20:37

are scripted? Which ones are unscripted? Tell

1:20:39

me, does it matter? Could you

1:20:41

tell that that video was unscripted? Maybe

1:20:44

I should maybe I should script less. I

1:20:47

should find one that was scripted. Okay. Could

1:20:50

you tell the Noctua booth was scripted? Yes.

1:20:54

Could you tell that the factory tour was not scripted? I

1:20:56

didn't notice. Useless.

1:21:00

Useless. Okay.

1:21:06

I'm trying to hear from the other

1:21:08

people, from the people in the chat. I don't know

1:21:10

which one they're responding to. They're just saying yes. Useless.

1:21:15

Might as well talk to talk to Twitch chat at

1:21:17

this point. Anyway, the

1:21:20

point is that they would give me any little

1:21:22

details that I needed for that particular chunk and

1:21:24

then I would do it. And

1:21:26

then Andy would go and he would

1:21:28

shoot like any B roll bits that

1:21:30

go along with it. And then Alex

1:21:32

was there to be a backup note

1:21:35

taker during the tour or

1:21:38

during the briefing. And then he

1:21:40

was a note taker. He was our only note

1:21:42

taker during the actual actual tour. And

1:21:45

then he was also noting down the clip numbers

1:21:48

for every section. Some

1:21:51

eagle eyed people noticed that that

1:21:53

video was actually uploaded within 24

1:21:55

hours of us completing

1:21:58

the tour. It

1:22:01

was kind of crazy to get it done because we wanted

1:22:03

it to be part of our CompuText coverage this year. So

1:22:06

it was a little crazy to get

1:22:08

that done, but Alex noting all the

1:22:10

clip numbers was how Dennis, who was

1:22:13

like, I have not edited an LTT

1:22:15

video in about two

1:22:17

years. Let's see how this goes.

1:22:20

That was how Dennis managed to turn around that

1:22:22

edit in just a few hours. I

1:22:24

did a review, we had to change some stuff. I had

1:22:26

to shoot or record a small

1:22:29

audio pickup on my lab

1:22:31

microphone. He exported again and boom, it

1:22:34

was up. I've got three comments. One

1:22:36

is a challenge from ScrappyDP saying,

1:22:39

you scripted the walkthrough in the beginning of the factory

1:22:41

tour. I scripted the walkthrough?

1:22:44

This is a claim. It's lovely

1:22:47

chat. I scripted the walkthrough. Of the

1:22:49

factory tour. Well, when

1:22:51

I say scripted, I mean word for word

1:22:53

scripting. Yeah, not notes. Yeah, I

1:22:55

can show Luke the notes that we had at the

1:22:57

beginning of the tour. This

1:23:02

is the notes that Alex gave me before

1:23:05

we started. These

1:23:07

are the notes that I took. Hold

1:23:09

on. Oh, wait, no, Alex has added

1:23:11

a bunch of stuff to this. None

1:23:15

of this is a script anyways. Yeah, so

1:23:17

here's the SMT line. Yeah, I

1:23:19

will read you word for word. This is what the video

1:23:21

would have sounded like if I had read off a prompter.

1:23:25

This is for the SMT line portion.

1:23:27

Yes. Spray anti-static.

1:23:31

Now that all of the components are validated,

1:23:33

time to put them mounted to board. Anti-static

1:23:35

shower, mount to boards, 9994. Each

1:23:39

PCB gets pushed into the machine, 9999. Intake

1:23:43

of PCBs, soldering paste

1:23:45

dash 0003. Scanning

1:23:48

for paste inspection. Pick

1:23:50

and place, 77,000 components per hour. Cutting

1:23:53

edge machine nearly doubled the speed of their older machines

1:23:55

on a typical line. This investment was

1:23:57

made because small pilot batches are often done.

1:24:00

here 15 minutes from their R&D headquarters

1:24:02

and speed and efficiency is a major factor

1:24:04

in product development dash 0021. VRMs.

1:24:10

GPU die placement dash 0030.

1:24:13

Oven that bakes the solder to attach everything.

1:24:16

Inspection CT scanning dash 0038.

1:24:19

Final of SMT 0044. 0048. Some get stopped for

1:24:22

inspection. 10

1:24:27

to 15% get sent to X ray manual

1:24:29

component placement put on tray add

1:24:32

capacitors. Flux spray 0068 solder

1:24:35

waterfall not quite a slide 0074 final

1:24:38

inspection. Probably

1:24:41

three quarters of that was added by

1:24:43

Alex while we

1:24:46

were going through on the actual. So like

1:24:48

help. So a fraction of that was

1:24:51

was actually done

1:24:53

before we started the tour. Google Doc can

1:24:55

go back. OK, here's the

1:24:57

version. Hold

1:25:01

on. Here's the version before

1:25:03

Alex added notes. Uh,

1:25:06

I also want to say that you

1:25:08

you did the reveal before you asked

1:25:10

if people could tell. So

1:25:15

there was there was a considerable amount of people saying

1:25:17

that they could tell on both and

1:25:19

I'm not saying that you're lying. But

1:25:22

I do think that in a lot

1:25:24

of cases, this one might notice

1:25:26

that something feels different, but

1:25:28

they don't necessarily notice why. OK, and then

1:25:30

you're like, oh, this one wasn't scripted. And they're like, aha,

1:25:32

I knew that that was exactly what I was. I'll

1:25:35

tell you what, why don't we play the game

1:25:37

with some more videos? Did you watch any of our other

1:25:39

coverage? No. What if I can just jump in

1:25:41

here? What

1:25:44

are the numbers? What are the numbers mean? Those

1:25:47

are clip numbers that Alex was adding as we were

1:25:49

shooting. So almost anything with

1:25:52

a clip number was added as

1:25:54

we were going through. So there

1:25:56

was very, very, very little in the way of

1:25:58

of a script. Okay,

1:26:02

here, I'm just going to add a filter

1:26:04

for, let's say views more than 500,000. That

1:26:09

should filter out anything that's just

1:26:11

like unlisted or whatever. Okay,

1:26:15

guys, NVIDIA, NVIDIA tour,

1:26:17

scripted or not scripted? Go. Why

1:26:22

don't we do our sponsor spots while we wait for

1:26:24

them to think? Do you want to pull it, Dan?

1:26:26

Okay. Yeah, or do sponsor spots?

1:26:28

Let him do the poll before you start going. Okay,

1:26:31

give me one second. You can set up the poll and I'll do sponsor

1:26:33

spots. The show

1:26:35

is brought to you today by AG1. When

1:26:38

cooking up your morning breakfast, chances are you're not

1:26:40

including certain ingredients that supply your body with the

1:26:42

energy it needs to get through the day. Our

1:26:44

sponsor AG1 can help with a simple and convenient

1:26:47

way to get the vitamins your breakfast may have

1:26:49

otherwise been lacking. As a Nesquik

1:26:51

cereal lover, yes. Their

1:26:54

team is always trying to find better ways to source,

1:26:56

test, and find the best quality And

1:26:59

unlike other supplements that only reference

1:27:01

research for specific ingredients, AG1's whole

1:27:03

formula is backed by multiple studies

1:27:05

done by third parties in double-blind

1:27:07

placebo-controlled environments. They're transparent with their

1:27:09

ingredients too, so if you're not convinced, you can do

1:27:12

your own research. A few of our team members even

1:27:14

drink AG1 every day and have done so even before

1:27:16

they were a sponsor. It's easy and

1:27:18

satisfying to start your journey with AG1, and

1:27:20

you can even get a free one-year supply

1:27:23

of vitamin D3K2 and five free AG1 travel

1:27:25

packs with your first purchase at

1:27:28

drinkag1.com/wanshow. The

1:27:30

show is also brought to you by Squarespace. There

1:27:33

are a lot of different shapes that

1:27:35

exist, like triangles, octagons, decagons, bestagons, but

1:27:37

we're really into squares these days thanks

1:27:39

to our sponsor Squarespace. Did they really

1:27:41

script that in here? Okay, sure. Squarespace

1:27:44

is an all-in-one platform that makes getting your

1:27:46

website up and running easily

1:27:49

and quickly. Just start with

1:27:51

one of their award-winning templates and start

1:27:53

building from there with intuitive drag-and-drop elements,

1:27:55

a streamlined yet comprehensive asset library, and

1:27:57

other features that are anything but square.

1:28:00

Their templates are the foundation of their website builder

1:28:02

and they're flexible enough to fit your unique needs.

1:28:05

Once your website goes live, you can use

1:28:07

their advanced analytics tools to gain insight into

1:28:09

where you can optimize and keep growing. And

1:28:12

if you require assistance, Squarespace has useful guides

1:28:14

and a 24-7 support team. So

1:28:17

visit squarespace.com/when to receive 10%

1:28:20

off your first purchase. And

1:28:23

someone's mowing the sidewalk. Sorry

1:28:25

for that unpleasant noise. The show is also brought

1:28:27

to you by Manscaped. A bunch of

1:28:30

our team is at Computex this week and

1:28:32

a big question we had was, how would

1:28:34

we take care of their bits and bobs?

1:28:36

I can tell you right now with some antiperspirant,

1:28:38

but anywho. Luckily,

1:28:40

there's Manscaped. Their performance package

1:28:42

5.0 Ultra has a whole slew

1:28:44

of tools to help keep your landscape groomed.

1:28:46

It comes with the Lawnmower 5.0 Ultra

1:28:49

meant for keeping your sack whacked with

1:28:52

dual blade heads. Um,

1:28:55

okay. By whacked, they mean like

1:28:57

weed whacked, not like struck. The

1:29:00

Weed Whacker 2.0 for trimming down

1:29:02

those nasal nuisances. And also their

1:29:04

Crop Soother and Crop Preserver to keep

1:29:06

your grass smelling fresh and fragrant.

1:29:09

They even include some bonus goodies like

1:29:11

a pair of their boxers and a

1:29:13

toiletry bag. So go check it out

1:29:16

for yourself at manscaped.com/wanshow and use code

1:29:18

wanshow for 20% off and

1:29:20

free shipping. While we're

1:29:22

at doing promotions, let's just promote the

1:29:24

end of Dan Week. Dan Week

1:29:26

is almost over. We had some Dan-centric

1:29:28

releases this week, including... Dan, are you

1:29:31

able to show your screen to

1:29:33

show things? Is that a thing we can

1:29:35

do? Yeah, I can do

1:29:37

web pages if you give me a minute or so. Yeah,

1:29:41

sure. In that case, why don't we do

1:29:43

the poll results while we wait for that? I

1:29:45

don't actually see the poll. I can't do a poll.

1:29:47

I'm going to read through some responses, though. Integrations at

1:29:49

the top. Elijah said... This

1:29:52

is why I said you need to give him time. Because we

1:29:54

started doing the sponsorships. Oh, because he had

1:29:57

to live do the clips. Oh, well...

1:30:00

I don't know, maybe he should grow a third arm. Yeah.

1:30:02

Did he ever think of that? Dan, do you ever think of that growing a third arm?

1:30:06

It's a reasonable response. Do you think you'd be

1:30:08

willing to like help the company, Dan? All

1:30:11

right. Sorry. Inject

1:30:15

some of that forced evolution virus. Dan's

1:30:18

doing great. Let's be encouraging for

1:30:20

Dan. Let's go Dan. Good

1:30:22

job, Dan. I

1:30:25

don't need you, Penny. Do you want to do that topic while we wait

1:30:28

for Dan? Sure, yeah. Sure. Copilot

1:30:31

Plus isn't for everyone. Yeah.

1:30:33

Source, the Verge, Source

1:30:35

2, Microsoft. Everyone. Yeah.

1:30:38

Strangely, Intel, AMD, and Nvidia

1:30:41

seem to be hinting that the

1:30:43

much-hyped laptops powered by their new

1:30:45

chips won't launch with Copilot Plus

1:30:47

features enabled, despite all meeting Microsoft's

1:30:50

required specs, seemingly. Microsoft

1:30:52

has likewise made statements seeming to

1:30:54

indicate that Intel and AMD laptops

1:30:56

will get access to Copilot Plus eventually.

1:31:01

Some, the Verge's Sean Hollister

1:31:03

in particular, have speculated that

1:31:05

Qualcomm might have some kind

1:31:08

of yet unknown timed exclusivity.

1:31:11

In other Copilot news, Microsoft

1:31:13

appears to be making its controversial

1:31:15

recall feature opt-in instead of

1:31:17

opt-out, which is good,

1:31:21

but I still don't like their attitude about it here.

1:31:24

The fact that they tried to make it opt-out, the

1:31:26

fact that it was ever going to be opt-out, did

1:31:29

not indicate a good

1:31:32

faith effort for Microsoft to maintain

1:31:34

the privacy of its Windows customers.

1:31:36

Look, if Windows was

1:31:38

free, like formally

1:31:40

free, I might

1:31:43

feel slightly differently about this. I

1:31:45

still wouldn't like it, but I'd be like, that

1:31:48

makes sense. Windows is

1:31:51

not free. Windows costs

1:31:53

$100. I

1:31:55

know because I just bought a copy of Windows.

1:31:58

Seriously. I bought a copy of

1:32:00

Windows this week. For what? For a cool

1:32:02

video that we did here. We went to the

1:32:04

tech mall. Oh! Yeah, buddy. I'm very

1:32:07

excited about this. We went to the tech mall,

1:32:09

and it was actually Jake's idea. He was just,

1:32:12

he was cruising in the tech mall, and he

1:32:14

found this shop, little hole in the wall shop.

1:32:16

Like, I mean, literally hole in the wall. I

1:32:18

could almost put my arms from one side to

1:32:20

the next side that had this

1:32:24

sick hardline water-cooled system

1:32:26

in the front of it, with

1:32:28

like this 3D printed League of Legends mask in it.

1:32:30

And it just like, it looked flippin'

1:32:32

awesome. And normally you walk

1:32:34

up to those shops, and you're like, I'll take

1:32:37

that one. And they're like,

1:32:39

no, man, that's like for demo. I

1:32:41

can build you like a basic gaming

1:32:43

computer. Yeah. I was like, I want

1:32:46

one like that. And for my

1:32:48

budget, I picked the starting price of

1:32:50

a main gear, what is

1:32:52

it? Whatever their hardline one is.

1:32:55

Main gear, hold on. Oh,

1:32:59

hold on. Yeah, this is my auto complete here. Yeah, I got this.

1:33:02

Hold on. My internet's a little, a little,

1:33:05

it's a rush. A main gear, a main

1:33:08

gear rush, Apex rush, starting

1:33:10

at 51.19. So

1:33:13

I gave them a budget of 170,000 new Taiwan dollars, and

1:33:17

basically was like, bro, I'm

1:33:19

trusting you. I

1:33:21

want it to look really cool like this one. And

1:33:24

I want a game at 4K. I want to be able to play any

1:33:26

game. So I gave him

1:33:28

functionally the same, the same budget. Yeah.

1:33:31

And I just looked it up. It's super similar. I wanted

1:33:33

to know what's, what are you better off with? Are you

1:33:35

better off with an Apex rush?

1:33:38

Or are you better off with some hole

1:33:40

in the wall shop in Taiwan? And

1:33:44

I guarantee you, no matter what you thought,

1:33:46

the answer will surprise you. Really

1:33:49

cool video. Really? Lots of

1:33:51

fun. Yeah, I had an absolute blast. And you

1:33:53

know what? I was having so much fun doing it.

1:33:55

I was like, man, I wish

1:33:58

I could do this kind of stuff more often. Okay,

1:48:03

sure. We're gonna have some merch

1:48:05

messages. Forget Dan time. Okay. I

1:48:08

can ask the merch messages if you want to set up the poll. Yeah,

1:48:11

I mean, I don't have the dashboard up. I

1:48:13

can kind of do both at the same time. I'm

1:48:16

asking a question. I'm doing it. I'm doing the

1:48:18

thing. Let's go Taiwan show. What's the most pointless

1:48:20

thing you saw at Computex this year? Man,

1:48:22

I didn't see that much. Me neither. I

1:48:24

didn't I didn't really make it to the show. Yeah,

1:48:27

I got a I'm gonna I'm gonna completely take

1:48:29

over this because I didn't neither of us really

1:48:31

saw too much of the show. I got my

1:48:33

first swag item I've gotten in

1:48:38

probably like seven years. Oh,

1:48:40

yeah, it felt like I just snorted

1:48:42

meth and cocaine at the same time. Oh,

1:48:44

just the hit. Oh, it was amazing. The

1:48:47

Russia free stuff. Just garbage. You know that

1:48:49

you get so much free stuff from work.

1:48:51

Oh, no, but not just trash, you know,

1:48:54

like it was a branded

1:48:56

bag. That's like this big like there's

1:48:58

no chance I want that. It's like

1:49:00

basically one use the little clasp on

1:49:02

it is already breaking inside was some

1:49:04

coconut flavored snacks that I don't want.

1:49:06

And then there was a voucher and

1:49:08

the voucher looks like a bill for

1:49:10

300 NT. Taiwanese dollars. And I was

1:49:15

like, that's weird. That's not a denomination that I

1:49:17

think exists here. And then I realized that it's

1:49:19

like, oh, if you spend I think it's like

1:49:22

16,000 NT, then

1:49:24

you can use this voucher for 300 at

1:49:27

like Taipei 101 or something. I'm like,

1:49:29

well, that's not happening. So it's all just useless.

1:49:32

And I was like, Yes, we're back. We're

1:49:34

back at shows where they just give you things for no

1:49:36

reason that no one wants. It was amazing.

1:49:40

There's it was actually like, I don't think I

1:49:42

can support any of this. Oh, of course not.

1:49:44

Okay. Yeah, cool. Yeah. Dan, the

1:49:46

polls wrong. Also,

1:49:49

this is hilarious. The poll

1:49:51

is was the factory tour scripted, which I

1:49:53

already told you guys it wasn't scripted. So

1:49:55

the fact that so

1:49:58

many of you are wrong either. Oh

2:08:00

goodness. Sheesh. I'm going to

2:08:02

get you guys out of here in about 45 minutes. So

2:08:05

why don't you do another topic and then we'll move into

2:08:07

After Dark. Okay,

2:08:09

topic time. Google leaks leak.

2:08:12

An internal Google database of thousands

2:08:14

of security incidents flagged by Google

2:08:16

employees between 2013 and 2018 has

2:08:18

given rare insight into how Google

2:08:20

manages these issues. The document, acquired

2:08:22

by 404 Media but confirmed

2:08:24

to be authentic by Google, includes reports

2:08:27

of accidental recordings of children's voices, a

2:08:29

major leak of the itineraries and home

2:08:32

addresses of Waze carpool users, and

2:08:34

an exclusion software failure for Google Maps

2:08:36

resulting in the accidental creation of a

2:08:39

database of thousands of geolocated license plate

2:08:41

numbers. These incidents were typically

2:08:43

reported, investigated, and resolved at the time that

2:08:45

they happened, but only a few were previously known

2:08:47

to the public. There were also

2:08:50

more long-term breaches, such as an incident

2:08:52

following Google's acquisition of socratic.org, where around

2:08:54

a million user emails from the company

2:08:56

were publicly exposed for over a year.

2:08:59

One relatively harmless incident involved the 2017 leak

2:09:02

of the announcement of

2:09:04

Nintendo's Woolly World 2. A

2:09:06

Google contractor downloaded the private announcement

2:09:09

from Nintendo's YouTube channel using admin

2:09:11

privileges and sent a photo of

2:09:13

the video to a friend who then posted the

2:09:15

screenshot to Reddit. Not only

2:09:18

did this friend admit to

2:09:20

getting the photo from a

2:09:22

friend who works at Google,

2:09:25

the URL and the image

2:09:27

started with www.admin.youtube.com. In other news,

2:09:30

replace your own job with AI. In

2:09:33

an interview with The Verge, Zoom

2:09:35

CEO Eric Yuan recently suggested that

2:09:37

Zoom should eventually have a feature

2:09:39

that allows you to, instead of

2:09:42

replace the background, replace

2:09:44

the foreground, delegating your

2:09:46

meeting to an AI-powered digital

2:09:48

twin who could create a summary

2:09:51

for you. This AI double would

2:09:53

likewise respond to emails and answer calls.

2:09:56

This seems to be more than an

2:09:58

anthropomorphized note-taker, however. custom

2:12:00

software. Hendrick theorizes that

2:12:02

Spotify wasn't interested in making this

2:12:05

easy open sourcing more widely known

2:12:07

because the device is essentially a

2:12:10

potato with a weak

2:12:12

Amlogic processor, 4

2:12:14

gigs of eMMC storage, and only

2:12:16

512 megabytes of RAM.

2:12:18

Its hardware is too limited to

2:12:21

do much above its current

2:12:24

intended purpose. I mean, part

2:12:27

that bothers me isn't that it can't do much

2:12:29

beyond it. It's that Spotify tried

2:12:32

to brick it. People have already shown that

2:12:34

they can do what they want with it.

2:12:36

Yeah. Yeah. So the intended purpose was fine.

2:12:38

Yeah. It doesn't really matter that it's a

2:12:40

potato if all it needs to be is

2:12:42

a potato. Potatoes are

2:12:45

cool. I like potatoes. I'm dead. Apple's

2:12:48

holding WWDC next Monday and the company

2:12:51

reportedly plans to announce how it will

2:12:53

integrate AI, Apple

2:12:55

intelligence into

2:12:57

their upcoming products. Yes. Their

2:13:00

AI system will be called Apple

2:13:02

intelligence. It's cool. Because that won't

2:13:04

be confusing at all. I

2:13:07

mean, it's just as useful the name as

2:13:09

artificial intelligence. Okay. That's true. We've got a

2:13:11

video coming on how AI is a lie.

2:13:13

Yeah. And like really is. It's

2:13:16

really cool what we're doing, but

2:13:18

it isn't. There's a lot of cool

2:13:21

AI. Yeah. It was written by Emily.

2:13:23

So she basically like went through and was

2:13:26

like, here's what you think AI is and here's

2:13:28

what it is. And here's what it's really good

2:13:30

at. And here's what it's like not good at.

2:13:32

And that sounds like a good video. I really

2:13:35

liked the de-Google-fy your

2:13:37

life video. Yeah. Seems

2:13:39

like other people did too. I'm excited

2:13:42

for it too. Yeah. I'm genuinely pretty

2:13:44

stoked for part two. Non-AI announcements around

2:13:46

half of the presentation will apparently include

2:13:48

improved customizability for iPhone, as

2:13:51

well as new software for the Vision Pro,

2:13:53

Apple watch and Apple TV. And

2:13:56

finally, oh my goodness. Oh

2:13:59

man. We got a few more things to

2:14:01

get through pretty quick here. Twitter says, okay,

2:14:05

to unclad chests.

2:14:08

The website formerly known as Twitter has added

2:14:10

language to its rules that formerly allows

2:14:14

consensually created adult content on the

2:14:16

platform, so long as it is correctly

2:14:18

and prominently labeled. Twitter has

2:14:20

typically been tolerant of sexual content, though the

2:14:22

site didn't officially condone it. And

2:14:24

to be clear, that is true all

2:14:27

the way back. Therefore, the primary change

2:14:29

will be the creation of an official

2:14:31

tagging system for graphic and explicit adult

2:14:33

content. Some critics have called the new

2:14:36

policy a way for Twitter to reject

2:14:38

responsibility for dealing with the sharp increase

2:14:40

in pornographic spam that Twitter has seen

2:14:42

during Elon Musk's tenure as CEO. But

2:14:45

I actually have an alternate

2:14:47

take. What I'm hoping

2:14:49

is that this just allows things to be tagged

2:14:51

more easily. And so maybe I will

2:14:53

actually see less of it. Yeah,

2:14:56

blurred out or hidden

2:14:59

behind reveal buttons. Our

2:15:01

last topic is first move to Europe, second,

2:15:04

delete Facebook. European users

2:15:06

of Facebook and Instagram were recently notified

2:15:08

by meta that the company will be using their

2:15:10

data for AI training. The policy is

2:15:13

global. It's just that

2:15:15

only European users were informed and only

2:15:18

European users had the ability to opt

2:15:20

out of this use for their data due to

2:15:22

EU legislation. And another unpopular

2:15:24

policy change, Instagram is currently

2:15:27

testing unskippable ads. They nice.

2:15:29

Very good. Nice. Oh, actually, there is one more

2:15:32

thing. The hashtag fixed TF two movement. Um,

2:15:35

yep, it's pretty bad. So

2:15:37

it doesn't seem like they're gonna do anything.

2:15:39

Yeah, quarter million people have signed the

2:15:41

petition and valve is going to do whatever they want

2:15:43

to do whenever they want to do it because they're valve and

2:15:46

they print infinite money and we

2:15:48

all love them for it. All right, time for when show

2:15:51

after dark. Could you please

2:15:53

ask Andy to give me roughly two

2:15:55

stops of neutral density filter? I

2:15:58

could ask Andy for that. But is

2:16:00

not here. Okay, I guess.

2:16:04

Oh, wait, no, no, no, no, hold on. Hold on. Dennis is here. Dennis.

2:16:06

Oh, uh,

2:16:10

can I tell you? Oh, wait.

2:16:13

Do we have an ND filter on there? I don't see an

2:16:15

ND filter. Should be digital. I

2:16:17

can do it. It's fine. Is there a

2:16:20

digital ND on that camera? I don't think

2:16:22

so. Yeah, no, this FX 30. Oh, now

2:16:24

it's dark. Uh, no, no,

2:16:26

it's at Wancho after dark. I

2:16:28

just made it dark. He made it dark.

2:16:30

It's okay. It's okay. Thanks,

2:16:33

Dennis. Oh, yeah, he could have just closed the aperture.

2:16:35

That would have been an option. Oh,

2:16:37

well. Alright. What's

2:16:40

up, Mr. Daniel Besser? Who even are

2:16:42

you? Looks like we got some merch messages. My

2:16:45

wife loves your clothes so much it upsets

2:16:47

her that it isn't readily available and places

2:16:49

women frequently shop. How do you fix this

2:16:51

gap between product superiority and access to the

2:16:54

target market? That's

2:16:56

a really great, great question. If you can answer it.

2:16:59

I think I have a position for you at

2:17:01

Creator Warehouse Inc. Yeah. Oh, I see where you're going.

2:17:03

We're having we're having trouble with marketing. We kind of

2:17:05

yeah, I'm just gonna be honest. We kind of suck

2:17:07

at it. And I don't mean that as

2:17:09

a knock against the team. It's not like I have some

2:17:11

silver bullet and I'm some kind of genius

2:17:13

and I know how to direct a consumer market.

2:17:16

If I was running the marketing. Yeah, exactly.

2:17:18

Like we're really good at being a

2:17:21

marketing vehicle for our sponsors who were

2:17:23

really good at making really good stuff.

2:17:25

Yeah, we're great at making content, but

2:17:27

we actually we don't

2:17:29

know much about the dark

2:17:31

arts of of effectively

2:17:34

using a, you know, Instagram

2:17:36

advertising budget. Yeah. It's

2:17:38

a whole other it's a whole other field

2:17:41

essentially. Genuinely. Yeah. So sorry.

2:17:43

I wish I had more for you. Hi

2:17:46

DLL. Given AMD and

2:17:48

Intel's focus on delivering powerful and

2:17:50

efficient mobile APUs. Do you see

2:17:52

discrete mobile GPUs sticking around

2:17:54

much longer? I

2:17:57

do. They're

2:17:59

great. but a discrete chip is

2:18:01

still going to be the way for a long

2:18:03

time. I think that the split is

2:18:05

going to continue to

2:18:07

widen for integrated in low-end

2:18:10

gaming. Like low-end gaming discrete

2:18:12

GPUs are already functionally not

2:18:14

a thing. But for people

2:18:16

who want real performance on the go, you're going

2:18:18

to need a dedicated GPU for a very long

2:18:21

time. Realistically,

2:18:25

does running your components hot? Home office

2:18:27

gets hot during the summer for work

2:18:29

hours, nine or ten hours a day,

2:18:31

will harm them. Or what's the best

2:18:33

solution for a hot summer days in

2:18:35

a small office? Being

2:18:39

on at all harms your

2:18:42

components. They wear out over

2:18:44

time. We did a good video on TechWiki

2:18:46

on why do computers die?

2:18:49

And basically, even though we think of

2:18:51

them as having no moving parts, like

2:18:54

electrons and stuff, like charges

2:18:59

are being moved around and

2:19:01

stuff. So

2:19:03

they wear out in time. And the hotter they

2:19:06

run, the faster they

2:19:08

wear out.

2:19:10

And so less is better and

2:19:13

more is worse. But

2:19:15