0:00

podcast is brought to you by the

0:02

OBSBOT Tiny SE High Refresh Rate 1080p

0:04

webcam. Get 5 % off at Amazon or

0:06

their official website through the links in

0:08

the description. And then it is also

0:10

brought to you by cdkeyoffer .com. Use

0:12

offer code brokensilicon to save 25 % on

0:14

all Microsoft software. We'll talk about these

0:17

sponsors more later. But for now, let's

0:19

get on with the show. Welcome

0:47

to Broken Silicon, a gaming

0:49

hardware podcast. I am

0:51

your host, Holy Moly Tom.

0:54

I'm joined today by co -host. Easter

0:57

Dan, we'll just keep it simple. See,

1:00

that says the two things going on today. We're

1:02

probably going to have to make this one a little

1:04

shorter than usual, although it's not abnormal for us

1:06

to only go for like an hour, 20 an hour,

1:08

30, which I assume is what we'll about do.

1:11

And it should be easier because usually we record

1:13

at six or eight PM and I have

1:15

to admit, I've heard this on other podcasts too.

1:18

The earlier in the day we record,

1:20

we're recording very early afternoon instead of

1:22

in the evening, the quicker we

1:24

are to get through the information. Sure.

1:26

And I don't think it's just

1:28

like a, uh,

1:31

you're sharper. So you

1:34

speak quicker or have more

1:36

energy. I do think there's just

1:38

less talking off subject that I

1:40

already are. Oh yeah. Sure. And,

1:42

uh, I don't know. We, we've

1:44

got a a bit of a, I

1:46

don't know if you have a time crunch this

1:48

today, a bit of one, I guess, but

1:50

it depends who you ask. My girlfriend would say

1:52

we definitely do have a time crunch because

1:54

we have to go to her family's Easter thing.

1:56

Of course, Dan, you can see is in

1:59

the, in room he was in for a while,

2:01

a couple of years ago, a year ago,

2:03

a year ago. Yeah, something like

2:05

that. When he was staying with me in

2:07

Nashville. So Dan and my family is in

2:09

Nashville. Of course, our parents

2:11

are not to be seen, but they are

2:13

on the premises. their presence is

2:15

felt. Their presence is felt much

2:17

like the Lord Jesus on Easter. All

2:20

right, everybody. Let

2:22

us just jump into it. I don't

2:24

have any opening reader mails here. Perfect

2:26

again. Perfect. And

2:29

no corrections, because we're never wrong,

2:31

man. But of course, we

2:33

are often wrong, and you can submit corrections. And

2:35

the Moore's Law said, Discord, we

2:37

just did not get any this new cycle. But

2:40

yeah, let's get to the first story

2:42

here, which I mean, you could say

2:44

is a corrections of sorts. Just tons

2:46

of reviewers trying to correct Nvidia, the

2:48

story. Number one, RTX 50 60 Ti

2:50

reviewed and launched. So I'm doing a

2:53

quote, a few reviews here or a

2:55

couple of reviews and then get into

2:57

the editorialization. First from Linus tech tips,

2:59

which I highly recommend. I thought they're

3:01

basically refusal to do a review is an excellent

3:03

video. Linus said, Nvidia, you're

3:05

acting like that chick that used to be

3:07

hot. And she thinks she's still hot and

3:09

thinks I'm going to drop everything when she

3:11

comes calling, but I'm married with three kids

3:13

now. And not only do I love my

3:15

wife, but if we're being honest, you peaked

3:18

in a high school anyway. I'm not playing

3:20

your games anymore. And honestly, that's a hard

3:22

thing for me to say. I want to

3:24

have positive things to say about graphics cards.

3:26

I just don't for this review. Also, I

3:28

don't believe he could get through that bit

3:30

they did without laughing because how silly it

3:32

sounded. Now from Kit Guru. Personally,

3:35

I think there is a mixed bag

3:37

that really sums up this. card,

3:39

the 50, 60 TI. Let's start with

3:41

the negative points. It is just

3:43

a 31 % gain over not the

3:45

40, 60 TI, but the 30, 60

3:47

TI for rasterization was really seems

3:49

like quite underwhelming of a result considering

3:52

that that GPU launched at the

3:54

end of 2020. I think we'd all

3:56

hope for further progress in five

3:58

years in the context of today's market

4:00

though, it is delivering 15 % gains

4:02

over its predecessor, the 40, 60

4:04

TI 16 gigabyte at a lower

4:06

price. And that actually sounds not

4:08

that bad compared to just how

4:10

horribly bad other Blackwell GPUs are.

4:12

I mean, consider that the 50

4:14

70 was just one to 5

4:16

% of the 40 70 super. And

4:19

so there you go. At the end

4:21

of the day, the 50, 60

4:24

TI 16 gigabyte seems to offer decent

4:26

performance and a decent price at

4:28

MSRP. But the way Nvidia has handled

4:30

this launch is a cause for

4:32

tons of legit criticism. Namely, they are

4:34

refusing to provide eight gigabyte GPUs

4:37

for testing, almost certainly in order to

4:39

hide the horrible performance issues. In

4:41

fact, the biggest performance surprise in a

4:43

pause of light is how close

4:45

the 50, 60 TI 16 gigabyte got

4:47

to the 50 70 in tracing,

4:49

especially in hardware and box testing, focusing

4:52

on newer and more demanding titles.

4:54

But really this just illustrates Nvidia's

4:56

continued folly of a 70 class

4:58

card with 12 gigabytes with a

5:00

60 Ti's get 16 for some

5:02

of their models. It also implies

5:04

what we can expect for ray

5:06

tracing performance, namely. horrible ray

5:08

tracing performance for the 8 gigabyte

5:11

model. And meanwhile, American and

5:13

American centric reviewers, we must note, decried

5:15

the lack of supply for the 5060TI's launch.

5:17

However, it must be noted that the

5:19

situation is quite a bit different in

5:21

the European Union, where we see plentiful stock

5:23

even in second and third tier markets,

5:25

like some Eastern European countries. This fits What

5:27

we have reported on more is lies

5:30

that weeks ago by our sources in

5:32

the supply chain, we were told that new

5:34

launches would start to focus on Europe.

5:36

Will the U S gets cheaper to make

5:38

yet still overpriced for their quality basic

5:40

models to try to cope with tariff

5:42

surcharges. We also discussed how tariff uncertainty

5:44

results in hesitance or the real inability

5:47

for companies to ship goods into the

5:49

U S at all until the many

5:51

confusions and uncertainties clear up a bit.

5:53

Although finally the real big supplies for

5:55

this launch is that it seems gamers

5:57

are also even in Europe where there's

5:59

plenty of full supply, rejecting Nvidia's pricing, the

6:01

wholesale, the European union might have had

6:04

a decent launch, but you know what? They're

6:06

still sitting there and no one's buying

6:08

them up despite shortages. People do not want

6:10

Nvidia's crap anymore. And so the way

6:12

I would summarize that quite long right

6:14

up is, and we did a whole

6:16

dice ring talking about both the 50, 60

6:18

TI and what's going on with Nvidia's

6:21

mind share, but wreck, you know, just $2

6:23

a month. just $2 actually just gets

6:25

you access to that in hundreds of

6:27

dye shrinks. Um, so just, you know, consider

6:29

supporting some patron for that. But we

6:31

talked about that in great detail, like

6:33

what's happening to the mind share. And there's

6:35

a section where we say, well, the

6:37

50, 60 TI in a vacuum, isn't

6:39

really the worst card. Maybe it shouldn't be

6:42

the 50, 60, maybe it should be

6:44

the 50, 60, maybe it should just

6:46

have 12 gigabytes of lower price, but really

6:48

a near $400 card converted in European

6:50

union, maybe not in the U S

6:52

right now. 16 gigabytes for that level of

6:54

performance. Isn't that bad? But the problem

6:56

is, well, there's no, no founders addition.

6:58

I mean, that's a pretty big, but the

7:00

problem is the way Nvidia is hiding

7:03

the eight gigabyte card for this launch. I

7:05

think earns it a bad review anyways.

7:07

And so reviewers hated how Nvidia is

7:09

handling this. People are tired of Nvidia's

7:11

nonsense. There is supply in Europe, but people

7:13

don't want it. Um,

7:15

yeah, Dan, let me stop rambling.

7:17

What do you think? I

7:20

mean, like you said, at

7:22

what the purported MSRP of $430,

7:24

it's fine. But the problem

7:26

is there's not really an attempt

7:28

by Nvidia, even like day

7:30

one, to supply anything at MSRP.

7:32

Not in the United States,

7:35

in Europe they are, though. Yeah.

7:38

But Americans, but we do have to think

7:41

globally. Yeah, I know, but there

7:43

is. not much anchoring it to

7:45

that MSRP without a founders model. There's

7:47

no eight gigabyte models out there.

7:49

So nothing's really anchoring it to that

7:51

$430 MSRP. So a lot of

7:53

these cards, they're reported MSRP is closer

7:55

to like, like the day one

7:57

reviews that we saw were like $500

7:59

MSRP models. So in a

8:01

lot of ways, from what I'm,

8:04

my perspective, we're getting a 7800

8:06

XT with a few extra bells

8:08

and whistles a year and a

8:10

half later for about the same

8:12

price. And at 400 or $430,

8:14

I think the 50, 60 TI

8:16

is fine. That's a

8:18

decent progression in price performance

8:20

versus what we had last

8:22

generation. But I don't know,

8:24

it's just not there. The

8:26

models that they say exist

8:28

don't really exist in any

8:30

market. And at least in

8:32

the US, the supply is

8:34

terrible, which I don't think

8:36

you can live. completely at

8:38

the hands of NVIDIA. There's

8:40

some extenuating circumstances, I would

8:43

say, but, uh, yeah. Uh, okay.

8:45

I mean, fine GPU, bad

8:47

launch, I would say. Yeah.

8:51

Again, I just, if it had

8:53

12 gigabytes was called the

8:55

50, 60 for like 380 even

8:57

350. And it's

8:59

like, okay, you know, it's not, I

9:01

wouldn't give NVIDIA a glowing review, but

9:03

I'd be like, It's more, but boys,

9:05

it's stronger than the 40 60 and

9:07

it actually has 12 gigabytes, whatever. But

9:09

instead again, we have really a card

9:11

that just points out how much Nvidia

9:13

sucks right now. It has 16 gigabytes

9:15

in the EU. Maybe you can get

9:17

it for a good price in the

9:19

US. You can't. Um, and it just

9:21

kind of makes the 50 70 look

9:23

worse because it already is in some

9:26

scenarios performing close to a 50 70

9:28

because it actually has enough RAM unlike

9:30

the 50 70, but then you look

9:32

at that and you're like, well, maybe

9:34

I should just skip black wall altogether.

9:36

And I really think that is the

9:38

major story for the 50, 60 TI

9:40

is even in markets where it is

9:42

available at MSRP, people aren't flocking to

9:44

Nvidia. The 50 70 is coming

9:46

in stock at MSRP in the US every

9:48

now and then below MSRP in the EU.

9:51

For me, it's just the death of Nvidia's

9:53

mind share. Again, we did a whole

9:55

dice rink about this. And I don't want

9:57

to Relitigate all that, you know, for

9:59

those who've already listened to it on the

10:01

Patreon, but I really do have to

10:04

say that again, I'm a contrarian. So I

10:06

love having fun little opinions and seeing

10:08

if they play out. But I just, there's

10:10

a part of me that goes, I

10:12

think people are these people in the comments

10:14

that say Nvidia, people just flock back

10:16

to Nvidia when stuff is in stock. They're

10:19

not in Europe. And

10:21

Nvidia is more popular in Europe,

10:23

relatively speaking to the US versus

10:25

Radeon. And I

10:27

don't know. I think people are

10:30

severely, severely underestimating how much damage

10:32

to NVIDIA's mind share they have

10:34

done over the past three generations.

10:36

And it just seems like Blackwell,

10:38

where everything's gone wrong at once

10:40

is the culmination of that. Yeah,

10:43

it really is just Blackwell

10:45

overall has deficiencies that aren't

10:47

typical of NVIDIA. And I

10:49

wonder if that's just catalyzing

10:51

the collapse of their mind

10:53

share with at least

10:55

this generation, I'm curious if it is

10:57

a one -off and if they recover with

10:59

the next generation of people care or

11:01

people are just turned off from Nvidia

11:03

right now. Because Blackwell, I

11:06

mean, you have the issues

11:08

with drivers, you have the issues

11:10

with ROPs, you have, like

11:12

you have these amateurish issues and

11:15

they're just not really delivering

11:17

that much of a generational uplift.

11:19

Ironically, except with the 50

11:21

60 ti, which is probably

11:24

relatively speaking one of the

11:26

better cards this generation from

11:28

Nvidia, but it really does

11:30

just seem like Yeah, people

11:32

don't want them Yeah, um, I

11:35

don't know Let me see here. I

11:37

have a variety rights and it says

11:39

did Nvidia hire the former radion PR

11:41

team for black? Um,

11:43

I think they may have hired

11:45

the rdn a1 driver team Yeah.

11:48

I don't think they have the

11:50

same issues that a radion had where

11:52

they just had dubious claims everywhere

11:54

that people are like, I don't know

11:56

if I trust these. Yeah. Yeah.

11:58

It's not like they have the same

12:00

issues of radion. Although you could

12:02

point to RDNA three, of course, having

12:04

those ridiculous claims for the performance

12:06

of Navi 31, but even that wasn't

12:08

really Nvidia scale. They were off

12:10

20. percent NVIDIA is usually off by

12:12

a factor. Didn't they claim

12:14

the 50, 60 TI is 50 times faster

12:16

than the 40, 60 TI in some

12:18

insane scenario. I can't remember

12:21

anymore. Was the 50, 60 TI

12:23

supposed to be the one that

12:25

gave us 40, 90 performance? No,

12:27

no, no. was the 50, 70. That

12:29

was the 50, 70. Okay. So believe that

12:31

there's a new claim out there right

12:33

now. that they're saying

12:35

it's 50 times faster than the

12:37

30, 60 Ti or 40, 60

12:40

Ti. So again, well that's good.

12:42

Nvidia for RDNA three claiming 70

12:44

% instead of 40 % is still

12:46

not Nvidia's tier, but no, I

12:48

actually don't think, you know, they

12:50

hired the radion PR team and

12:52

maybe the RDNA one driver team.

12:54

I just think that Nvidia's PR

12:56

team has cried wolf one too

12:59

many times. And you see this,

13:01

it's like, I've asked, people

13:03

right at AMD, like,

13:06

you know, when I leak the IPC for

13:08

Zen five being like at least 15 %

13:10

or higher, or I think I said 16 %

13:12

or higher was the final thing. And then

13:14

it was like, whatever they said, it was

13:16

16 or 17%. I don't even remember at

13:18

the top of my head, but I remember

13:20

AMD engineers going, Yeah, I think

13:23

that you could fairly claim 15 or something,

13:25

but maybe they'll try 20. I hope they

13:27

don't because I don't think most people will

13:29

see it as a 20 % IPC increase.

13:31

This happens all the time when I talk

13:33

to people at Intel or AMD or Nvidia

13:35

where they're like, well, this is where the

13:37

performance is. I don't know what marketing will

13:40

claim. And I wonder if there's going to

13:42

be massive pushback next generation from engineering

13:44

that will actually be listened to at Nvidia.

13:46

Like we've been telling you for three generations,

13:48

not to say this two to four X

13:50

crap. Now no one even

13:52

believes us when we're not lying

13:54

like and so we can't even

13:56

market you guys have realized you've

13:58

created a situation where you can't

14:00

even market Nvidia graphics cards anymore. Oh

14:03

Yeah, I mean, I made jokes

14:05

about this. I think Starting with

14:07

ampere like their claims are so

14:09

outlandish that they don't even feel

14:11

dishonest at a certain point You

14:13

just kind of see them and

14:15

laugh and you're like, okay, Nvidia

14:17

cool. We're getting what

14:19

from a over the course of

14:21

two to three generations, what a

14:23

10x uplift in performance according to

14:25

their charts, if you do the

14:27

multiplication. And everybody knows it's dishonest,

14:30

but at a certain point that does

14:32

bite you in the ass where, yeah,

14:34

I don't even know why you would pay

14:36

attention to their marketing anymore. It's

14:39

completely decoupled from reality at

14:41

this point. So why not

14:44

just wait for reviews? Well,

14:46

it has at least created a whole

14:48

cottage industry for people like me and Hardware

14:50

and Box and Gamer's Nexus to read

14:52

through the data and find the two things

14:54

where we can go, oh, I found

14:56

it. There's the real performance. So it does

14:58

allow us to make videos selfishly. I

15:01

don't mind it that much. You've become a

15:03

Jensen Wang oracle. Yes, like it allows

15:05

us to profit, I suppose, off of their

15:07

lives by finding a way to tell

15:09

you all the truth. But outside of that,

15:11

I think it's pretty bad. This

15:14

piece of content is brought to

15:16

you by the OBSBOT TinySE. The

15:18

TinySE supports an unprecedented 1080p at

15:20

100 frames per second or 720p

15:23

at 150 frames per second, outperforming

15:25

the overwhelming majority of affordable small

15:27

webcams on the market. You can

15:29

enjoy crystal clear and ultra smooth

15:31

video that captures every nuance in

15:33

motion effortlessly. And it is not

15:35

just about the frame rate, by

15:37

the way, it's a breakthrough in

15:39

cost effective low light performance with

15:41

a 1 and 2 .8 inch

15:44

stacked CMOS, dual native ISO staggered

15:46

HDR and 12 -bit dynamic color range

15:48

in order to ensure excellent video

15:50

quality in any lighting conditions. Additionally,

15:52

the tiny SE goes beyond basic

15:54

AI by including more advanced functions

15:56

like zone tracking and body part

15:58

tracking as well. And you know

16:00

what? I'm using one right now.

16:02

Genuinely, we at Moore's Laws Dead,

16:05

me and Dan, the co -host

16:07

of Broken Silicon, use OSPOT cameras

16:09

because We have concluded they are

16:11

just the best use for what

16:13

we want out of a camera

16:15

for our purposes. And so I

16:17

really do recommend them. And if

16:19

you want to check them out

16:21

for yourself and also support the

16:23

channel, follow the links in the

16:26

directions below. Again, support Moore's Laws

16:28

Dead by checking out the Osbot

16:30

Tiny SE through the links below

16:32

today. The

16:35

Forbidden Juice writes in, beginning of Gamers

16:37

Next, it says, just like Gamers Next's video

16:39

on Nvidia driver woes, it's a quite

16:42

damning for Steve to say that they are

16:44

trying their best to be like Intel

16:46

alchemists at launch. I have to say you

16:48

must feel vindicated now. Having said for

16:50

the past years, Hey guys, Nvidia has made

16:52

bad drivers before. I always did believe

16:54

that to be the case. And this launch

16:56

just proves it looks like Nvidia not

16:58

only phoned in their hardware, but their software

17:00

as well. Yeah. And I've seen carbon

17:03

cry, you know, who's part of the Moore's

17:05

law said team pointing that out to

17:07

me as well, that we should highlight this,

17:09

that since the start of Moore's law

17:11

is dead, When people are like AMD's drivers

17:13

are bad. I've been like, they're not,

17:15

they're fine. I mean, I even switched to

17:17

AMD for like five years or so

17:19

because our Nvidia cards during the 500 series

17:21

were just unusable with their drivers. And

17:24

then AMD was great. You know, every now

17:26

and then there's a miss. And

17:28

then Nvidia, I'd say after Kepler had

17:30

pretty solid drivers and then something just

17:32

started to go wrong here with Blackwell

17:34

really badly. But there's always blips here

17:36

and there. And I've never felt like

17:38

you can just say Nvidia's drivers are

17:40

more stable, except for like, again, maybe

17:42

RDNA one through three to a degree,

17:44

but it's a slight degree. And I

17:47

don't know if I use the term

17:49

vindicating, but I would remind people Nvidia

17:51

has always had periods where they've worse

17:53

drivers than AMD, sometimes far worse drivers.

17:55

And this narrative being popped, I love

17:57

it. Cause now we can just focus

17:59

on if the card is good and

18:01

I don't have to deal with people

18:03

saying, Oh yeah, but you know, it

18:05

really does feel sometimes like, yeah, if

18:07

it gets repeated in forums enough, does

18:09

it become the truth? Because that does

18:12

seem to be the origin of the.

18:15

Nvidia just has blanket more stable drivers

18:17

than AMD. This was like, if,

18:19

if I tell myself this enough on

18:21

Tom's hardware and Reddit, maybe it

18:23

becomes true. I mean, to a degree,

18:25

I feel like there's people who

18:27

have always bought Nvidia and they have

18:29

to have something they can tell

18:31

themselves for why Nvidia is more popular

18:33

than AMD outside of they just

18:35

happened to be like they played the

18:37

market better, you know, with marketing

18:40

and how they position their cards for,

18:42

from like, I don't know, I

18:44

don't know if the 200 series, but

18:46

certainly whether you like it or

18:48

not, they are me actually out started

18:50

out sell the five 400 and

18:52

the 500 series started out sell AMD

18:54

or be close and like 500

18:56

series through max, which interestingly, I

18:58

would say pop probably the worst. Probably

19:00

the worst architecture NVIDIA has come out with

19:02

since I've been in PC gate, a

19:04

fan of PC games to like pre nine,

19:07

nine thousand GT GT X to like

19:09

find a worse one. Like maybe I think

19:11

the five thousand was pretty bad supposedly,

19:13

but you know, but that's like what to

19:15

that's like Oh five or something. We're

19:17

going, we're going far back. We're going for

19:19

like five, like somewhere between Oh three

19:21

and Oh six or something. Okay. All

19:23

right. Let us now then move on

19:25

to story number two, our ticks, 5

19:27

,000 supply rush leak. Now this was

19:30

something that wasn't initially necessarily going to

19:32

be a story, but late Friday night,

19:34

I started getting these quotes from sources

19:36

in retail. And so I just want

19:38

to read two of them. The first

19:40

one here. on Friday, which is April

19:42

18th. We, the speed and a major

19:44

online retailer received about 10 times as

19:46

many black wall cards as usual to

19:48

our warehouse. So it seems that NVIDIA

19:50

is realizing they have a problem here.

19:52

It was mostly RTX 50 70 and

19:54

50 70 TI's with some 16 gigabyte

19:56

50, 60 TI's. Nope, no eight gigabyte

19:58

models a handful of 50, 80s. And

20:00

so yeah, they're mostly targeting RNA for

20:03

performance here. No RTX 50 90s. And

20:05

I did follow up on this, like

20:07

10 times as many, like maybe to

20:09

this warehouse, a shipment of like I

20:11

know somewhere around like 300 graphics cards

20:13

when usually they'll receive 20 or 30.

20:15

So this isn't like they sent twice

20:17

as many and it won't make an

20:19

impact. This is one warehouse. Something's

20:21

going on here. Also

20:23

another source. We, this the

20:25

person that works at a major U S

20:27

retailer says, just received a couple hundred black

20:29

hole GPS or a location today. Again, this

20:31

was the 18th. And I also asked a

20:33

friend of mine that works at a different

20:35

location. This place is brick

20:37

and mortar stores. They received over a hundred

20:39

there for their location as well. Usually

20:42

received 10 to 30. So I don't know

20:44

what's going on here, but for sure,

20:46

but at least short term, it seems like

20:48

all of a sudden in video, I

20:50

don't know if there's applying as many cards

20:52

as AMD, but yeah, this would be

20:54

about as many as four has been receiving.

20:56

a couple of times a week, at

20:58

least in one shipment. And they're all cards

21:00

that mostly seem to focus on RDNA

21:02

for performance. So I don't know how short

21:04

lived this will be. I don't know

21:06

if Nvidia just realized they screwed up a

21:08

few months a month ago. And this

21:10

is the first cards who arrive off of

21:12

boats in the United States or whatever.

21:14

But I think this is worth pointing out.

21:17

If you're looking for graphics cards, they're

21:19

appearing in stock. And I have noticed that

21:21

going on Best Buy's website, for example,

21:23

I have started to see them. So just

21:25

an FYI to everybody. At least right

21:27

now, this second Nvidia is all of a

21:29

sudden ratcheting up supply on an extra

21:31

couple of cards a week, hundreds more to

21:33

multiple locations. I mean, if my mental

21:35

math is right, that sounds like it's at

21:37

least getting into the same ballpark as

21:39

what AMD is supplying to brick and mortar

21:41

stores, which in every shipment. Yeah. Yeah.

21:43

Normally you would think they would ship a

21:45

lot more, but, uh, I feel like

21:47

that's something like this should have a material

21:49

impact on, uh, supply

21:52

and hopefully maybe for the next

21:54

couple weeks, you see prices

21:56

go down a little bit or

21:58

maybe not balloon to like

22:00

$1 ,000. So hopefully for like

22:02

$1 ,000 plus for like a 5070.

22:04

So hopefully in a bad year, we

22:06

have a couple more weeks of decent

22:08

supply if you need to buy something.

22:10

Like I could see this being the

22:12

last time you actually want to buy

22:14

for a long time. Yeah. So I

22:16

guess that's kind of the conclusion we

22:18

would make. Keep your eyes open. this

22:21

month, this month being late, like

22:23

late April. Keep your eyes open. Late

22:25

April, mid, mid May, maybe. And

22:27

we're hearing similar things in Europe as

22:29

well. Of course, as I've already

22:31

harped about on this episode. So like

22:33

just Europe, US, keep your eyes

22:35

open. Something's going on here where things

22:37

are being supplied in higher numbers

22:39

than have been for months. If you

22:42

miss the first wave this wave

22:44

will probably cost a little more on

22:46

average But then again, I don't

22:48

know prices are seemingly going down in

22:50

Europe like this might be your

22:52

chance right here before who knows what

22:54

happens with tariffs again All right,

22:56

let us now move on from that

22:58

brief update to story number three RX

23:01

90, 60 XT and RX 90, 70

23:03

GRE specs leaked and release date has

23:05

been pushed back. So first quoting from

23:07

my leak from the, about the 90,

23:09

60 XT and 90, 70 GRE a

23:11

couple of weeks ago. So the RX

23:13

90, 60 XT will be five to

23:15

12 % weaker, most likely than the

23:17

7 ,700 XT, but probably also a solid

23:19

at least 10 % faster than the

23:21

40, 60 TI. So yes, this leak

23:23

came out before the 50, 60 TI

23:25

launch, you would say. It's probably going

23:27

to come in slightly weaker than the

23:29

five 50, 60 TI, but slightly stronger

23:31

than the 40, 60 TI. Um,

23:33

and so yeah, it's stronger than the

23:35

old $400 cards. Now I say 400

23:37

and it's the pricing and I highlight

23:39

that price point because this is the

23:41

part that has surprised me from what

23:43

I was told. I was told that

23:46

the 90, 60 XT might cost below

23:48

400, possibly well below between 329 to

23:50

379 for the 16 gigabyte model with

23:52

me. Suggest, I would suggest

23:54

probably 350 will be the price or,

23:56

you know, And I know that a lot

23:58

of us were hoping that the 90,

24:00

60 XT was going to like beat

24:02

the 7 ,700 XT, but I got to

24:04

say if it did, it'd probably be

24:06

400. If it's three 30 or $350

24:08

for almost 7 ,700 XT, which means almost

24:10

6 ,800 performance, except that actually at 16

24:12

gigabytes of RAM, that's pretty good. That's

24:15

about the same price the 7 ,600 XT

24:17

was with the same amount of RAM

24:19

for 30 % more performance, uh, and, or,

24:21

and, or maybe even more than that.

24:23

Um, and the eight gigabyte model sounds

24:25

like it will probably be two 99, which

24:27

again, I don't know much to say

24:29

about that. The most I can say about

24:31

a two 99, eight gigabyte, 90, 60

24:33

XT is that they are giving you

24:35

better than 40, 60 TI performance for

24:37

the old price of the 40, 60. So

24:40

I guess it's fair from the perspective

24:42

of you're getting like a 6 ,800 level

24:44

of performance for 1080p gaming, but Generally

24:46

speaking, I won't recommend this if the

24:48

price is even within a hundred dollars

24:50

of the 16 gigabyte model And so let

24:52

me see here. Well, okay anyways moving

24:54

on to the editorialization With gamers no

24:56

longer holding space for Nvidia can gamers

24:58

now hold some for radion? Well, yes and

25:00

no the 90 60 XT seems like

25:02

a winner with decent performance and a good

25:05

MSRP and our sources say that significant

25:07

volume will be available more than they

25:09

had for the 90 70 and the

25:11

90 70 XT launch but It will still

25:13

not be enough, most likely as Nvidia

25:15

vacates the market. We'll see how that

25:17

develops though. And AMD simply cannot keep

25:19

up with market demand in less Nvidia starts

25:21

quadrupling supply, even from 10 X in

25:23

supply, by the way, expect to see

25:25

MSRP cards sold out in minutes, queues

25:27

at micro centers and continued shortages. But the

25:30

MSRP and launch supply seem to show

25:32

that AMD is still trying to exploit

25:34

Nvidia's weakness, which is good. Although admittedly alongside

25:36

the good news, we also have some

25:38

bad. We can confirm there is something

25:40

weird going on with the 90s, 60s encoders,

25:42

most likely with sources telling us that

25:44

AMD won't confirm to them if it

25:46

has hardware encoding. Now most people don't

25:48

really use encoders or need them. Most CPUs

25:50

now have good encoders and they're integrated,

25:52

uh, graphics. And let's, I guess you're

25:55

on a M four, but people who

25:57

do encode decode a lot should. Keep

25:59

mind with this when they look at reviews

26:01

and watch for if they include them. Uh, worse

26:03

also though, the 90, 60, 60 is seemingly

26:05

being delayed. I had heard they're trying to

26:07

launch by April 24th. I think months ago, me

26:09

and other people leaked April was going to

26:11

be the launch tape. And now I'm hearing that

26:14

it has been pushed back into later in

26:16

quarter two. And most sources tell me this is

26:18

due to tariffs, like everything we can't avoid

26:20

bringing up tariffs. Unfortunately. Um, we're

26:22

on video responded by launching the 50,

26:24

60 is planned. Uh, I

26:27

don't know. So I don't know. Hopefully that

26:29

means there'll still be supply as well, which is the

26:31

final thing I would note about this, which is like

26:33

a week and a half ago or something, right? I

26:35

was told there should be more supply than the 90

26:37

70 and 90 70 XT. Um,

26:39

but now that I'm hearing it's being delayed

26:41

due to tariffs, I don't know, at least in

26:43

the U S if the supply will be

26:45

as good as we thought, hopefully it will be

26:47

though. Oh, and I forgot to mention this,

26:50

90 70 GRE, 48 compute units, one 92 bit.

26:52

The great radion edition continues. I don't

26:55

have much to say expect that

26:57

to be I

26:59

guess probably 20 to 30 %

27:01

faster than the 90, 60 XT.

27:03

Well, having less RAM, I haven't

27:05

been given a price. It's not,

27:07

it's mostly for the Chinese market.

27:09

They can even sell it there.

27:11

I don't know. Um, for

27:14

this day and age, what would I

27:16

even price that? If it's three 50 for

27:18

a 16 gigabyte card for 29, I

27:20

guess the same as the 16 gigabyte 50,

27:22

60 TI, it was 20 % faster. I

27:25

don't know. I'd probably do three 99. Anyways,

27:27

Dan, what do you think of all this? I

27:31

mean, if

27:33

the 9060 XT is

27:35

close, if the

27:37

16 gigabyte model is like

27:39

$350 and at the

27:41

upper end of that performance

27:43

expectation, it seems pretty

27:45

great. Although, once again, in

27:48

this day and age, it's really hard to talk

27:50

about anything in reference to MSRP because I

27:52

think everything is just going, at least in the

27:54

US, is going to go way

27:56

above MSRP because of

27:58

the insane volatility we've decided

28:00

to throw into the

28:02

market for no reason. So

28:05

yeah, if it's relatively speaking

28:07

about that same golf in

28:09

price between that and the

28:11

50, 60 TI, I think

28:13

the 90, 60 XT is

28:15

probably the better choice, but

28:18

we'll see how that shakes

28:20

out. Yeah. Well,

28:22

on this note, Part DB and

28:24

me, right? Senate says the arcs

28:26

90 cent of GRE sounds dumb

28:28

from a naming perspective alone. So

28:30

now we have a supposedly great

28:33

radion edition that is weaker than

28:35

the car that has no suffix,

28:37

the 90 70. And it counters

28:39

to the previous example where the

28:41

GRE models were typically more powerful

28:43

than the vanilla ones with no

28:46

suffix on models in the US.

28:48

Yes. The 90, no, the 7 ,900

28:50

GRE was, I guess, Weaker

28:52

than the XT, but the XT

28:54

did have the XT moniker and

28:56

there was no 7900 non XT

28:59

But I believe the 6700 GRE

29:01

could mean anything like there was

29:03

a 10 gigabyte model with a

29:05

You know the GRE is the

29:07

I don't know they don't make

29:09

any sense to me I do

29:11

agree the great Radeon edition suffix

29:14

makes I mean just really this

29:16

is becoming the SE like

29:18

50, 60 from Jeremy where SE

29:20

is just like they had the 50,

29:22

60 TI, the 50, 60. And

29:24

then the worst, worst, worst yields 50,

29:27

60 SE, but it's sold to

29:29

China. So they, they want to pander

29:31

to GRE, which used to meant

29:33

mean great Radeon edition. See, this is

29:35

where I go. I still don't

29:37

think NVIDIA has hired Radeon marketing because

29:39

this is all types of levels

29:42

of blunders that I can't even get

29:44

into. Well, they need the GRE

29:46

because I mean, it needs

29:48

to be a 70 because it's just that. So

29:51

they can say the Chinese

29:53

market has a 70 for them.

29:56

Because I mean, they used to do

29:58

like have like five if

30:00

I'm remembering correctly, like this might

30:03

make more sense as like a

30:05

90, 65 or whatever. I

30:07

just put on it like Nvidia

30:09

does with 50 90 D the call

30:11

call it these 90 70 dragon,

30:14

which sounds cool. The dragon great radion

30:16

sounds like this should be the

30:18

best model. Dragon just means a special

30:20

model because in China, they don't

30:22

care as much about VRAM. And so

30:24

you can remove, you know, 20 %

30:26

of the price and 10 % of

30:28

the performance and sacrifice VRAM inches

30:30

popular in China. I would just call

30:32

it the dragon addition to fully

30:35

pander to that market and not try

30:37

to cling to a rabbit moniker

30:39

that doesn't exist anymore and is really

30:41

counterintuitive. Well, now we're on to

30:43

the, we're just calling it great radio

30:45

on edition, which I don't know,

30:47

sounds like 1984 ish to me for

30:49

some reason where it's like there's

30:51

some weird double speaking there, but I

30:53

guess that's what we're calling it.

30:56

Yeah. All right. Smash actually

30:58

writes in and says, well, AMD is

31:00

released any more variants of the 9 ,000

31:02

series probably, but he says, such as

31:04

new models, white versions in the light.

31:06

Cause it seems like all AIBs are

31:08

recent releasing like 20 different variants for

31:10

50 series GPUs. Or is this a

31:12

case of Nvidia playing them off to

31:14

not allow things for AMD? Look,

31:17

I think there is something. I mean,

31:19

look at MSI said they're not even

31:21

making a RDNA for graphics cards. This

31:23

Jen. like trying to pander

31:25

to getting more NVIDIA supply at the

31:27

same time. Yeah. I mean, I think

31:29

NVIDIA is like multiple times trying to

31:31

make aces, like have more models for

31:33

them than Radeon. So there's some of

31:36

that, but I will say this. If

31:38

we agree NVIDIA has like 80 % of

31:40

the market, wouldn't you expect there to

31:42

be like four times as many models

31:44

to pander to that portion of the

31:46

market? Just trying to

31:48

point out the obvious here.

31:50

Like, Like maybe it isn't

31:53

a conspiracy. Maybe Nvidia just gets more models because

31:55

they have more market share. Yeah.

31:57

I mean, you, why

31:59

would you just, why would you have

32:01

the same number of production lines

32:03

for four times fewer cards? Yeah. I

32:05

don't think you would. a quick

32:07

jumper rights. And it says, Tom, I

32:09

think there's a silver lining to

32:12

the ongoing GPU shortages. First it's giving

32:14

AMD an opportunity to build a

32:16

stronger position and gain market share. We

32:18

saw with CPUs that it took

32:20

time for a hardware manufacturers to properly

32:22

embrace AMD, but now they're implementing

32:24

numerous optimizations for AMD platforms. Another positive

32:26

development is that reviewers have started

32:29

conducting value for money analysis based on

32:31

real world availability data. This seems

32:33

like a healthy shift in the review

32:35

landscape. Yeah, sure. After all, it

32:37

doesn't make sense for reviewers to strongly

32:39

recommend GPUs that are perpetually understocked

32:41

or where most available inventory consists of

32:44

overbuilt overpriced AIB models. You think

32:46

these developments might lead to a more

32:48

balanced GPU ecosystem in the long

32:50

run. I think the continued

32:52

shenanigans have forced reviewers to become

32:54

smarter and how they analyze things.

32:56

And I think that's a good

32:58

thing. We're to be a fluke

33:00

like how. the

33:03

30 series panned out and then the

33:05

40 series is normal. The 50 is

33:07

normal. And then there was a fluke

33:09

again. Maybe no one really learned the

33:11

lesson, but certainly I guess I would

33:13

agree that I like that radion has

33:15

an opening right now. And I think

33:17

it's probably good that every reviewer has

33:19

been forced to finally get through their

33:21

heads. You can't just go 20 %

33:23

less, 20 % more. I say good. And

33:26

these reviews are really turning into like

33:28

an analysis of the market, which they

33:30

always should have been. Well,

33:32

they have to in this day

33:34

and age where in the past

33:36

they really didn't have to as

33:38

much because I don't know when

33:40

I was younger You could typically

33:42

find a couple models that were

33:44

close to our at MSRP and

33:46

GPUs would typically fall in price

33:48

within a year of being released

33:50

And that's just not the market

33:52

we exist in anymore. So MSRP

33:55

is continue to be less and

33:57

less relevant. So you have to look

33:59

at them from the holistic view,

34:01

like what are these actually selling for?

34:03

Where in the past, you can,

34:05

even if you're doing a somewhat dumb

34:07

analysis, that dumb analysis pretty much

34:09

works because if you're hunting hard enough,

34:11

you could find it a $350

34:13

MSRP model for any GPU. If the

34:15

MSRP was a was $350 for

34:17

now, hell, the cheapest thing you

34:19

might find for a 350 MSRP is like 600.

34:25

J H H 50 series available on

34:27

eBay writes in Justin Wang says

34:29

50 series available on eBay. Okay. With

34:31

AMD having promised that the future

34:33

is fusion, why is no one holding

34:35

them accountable for their lack of

34:37

success and doing already R and D

34:39

into the field of fusion? You

34:42

know, guys, this is a joke. We

34:44

will include the AMD future is fusion logo

34:46

for this one, but I have to

34:48

say I don't know why I let this

34:50

question get through. Cause I don't think

34:52

there's much for us to say. Well, do

34:54

you think they're going to achieve cold

34:56

fusion at AMD Tom? The only thing I'll

34:59

say is that is one of those

35:01

like science magazine articles we see every five

35:03

seconds, you know, of like in Paris,

35:05

they've almost achieved cold fusion. You know what

35:07

I mean? Yeah. Yeah. The

35:09

battery was annoying one for

35:11

me. Yeah. That big story about

35:13

fusion with the, the What

35:15

was it we've got we've achieved

35:17

ignition with fusion and that

35:19

was going everywhere for two weeks

35:21

All right, let me know

35:23

when I'm charging a car and

35:25

because of fusion powering some

35:27

power plant All right, let us

35:29

now move on to

35:31

story number four. Are you

35:33

too depressed to go outside because

35:36

of how expensive Microsoft software

35:38

is? Well, there's absolutely no need

35:40

for that. Just go to

35:42

cdkeyoffer .com. That's right. This piece

35:44

of content is once again sponsored

35:46

by cdkeyoffer .com. And I say

35:49

once again, because they've been

35:51

a fantastic sponsor of Moore's Laws

35:53

Dead and its community for

35:55

many years. And that's because they

35:57

always deliver the best pricing

35:59

reliably for Microsoft products like

36:02

Windows operating systems, Office software, and

36:04

they also sell games and other

36:06

things as well. So make sure

36:08

you check them out, especially during

36:10

their Easter sale going on right

36:12

now. And if you do, use

36:14

offer code broken silicon to save

36:16

25 % on all Microsoft software and

36:18

then dye shrink to save 3

36:20

% off on everything else on

36:23

the website. The community uses them.

36:25

I use them for my new Zen

36:27

4 X3D desktop, and Jesse here needs to

36:29

stop moping around, and I think use

36:31

them as well. So that's once again, support

36:33

Moore's Laws Dead by going to cdkeyoffer .com

36:35

through the links below today. Okay,

36:38

Jesse, it's okay. Moore's

36:43

Laws Dead Zen 6 leaks

36:45

seemingly confirmed. All right,

36:47

first here, we're just going to read

36:49

from an AMD press release. AMD today

36:51

announced its next generation AMD epic processor

36:53

codename venus is the first HPC

36:55

product in the industry to be taped

36:58

out and brought up on the TSMC

37:00

advanced two nanometer process technology. This

37:02

highlights the strength of AMD and

37:04

TSMC semiconductor manufacturing partnerships to co

37:06

-optimize new design architecture with leading edge

37:08

process nodes. It also marks a

37:10

major step forward in the execution

37:12

of the AMD data center CPU

37:14

roadmap with Venice on track to

37:16

launch next year. AMD has also

37:18

announced the successful bring up in

37:20

validation of its fifth generation Epic

37:22

CPU products and TSMC's new fab

37:24

facility in Arizona, underscoring his commitment

37:26

to U S manufacturing. saying,

37:29

we are proud to have AMD be

37:31

a lead HPC customer for our advanced

37:33

two nanometer process technology in the TSMC

37:35

Arizona fab. It's good. They're making it

37:37

there though. I suppose it wouldn't be

37:39

a 2025 news episode without more on

37:41

our several Zen six leaks. It seems

37:43

this time we have something perhaps more

37:45

exciting than the leak itself. It's confirmation

37:47

breaking the tradition of Apple being the

37:49

first to TSMC's newest node. That's a

37:51

big deal. AMD and TSMC announced that

37:53

it is AMD's Venice to be the

37:55

first HPC product out in N two.

37:58

This confirms information about chiplets of Zen

38:00

six using N two for the core

38:02

chiplets. By the way, let me just

38:04

jump in here. How many of you

38:06

losers in the comments called me stupid

38:08

for saying AMD is going from four

38:10

nanometer skipping three to two. I was

38:12

right again. You were wrong. I'm sorry.

38:14

I have to dunk sometimes on you

38:16

losers because I'm so sick of this,

38:18

like making stuff up. There's no way

38:20

this would happen. It is happening. Well,

38:23

this channel is first and we also

38:25

were called stupid for saying that some Zen

38:27

six products were launched in 26. They're

38:29

going to nerds. They're going to subscribe to

38:31

us on YouTube. All right, go on,

38:34

Dan. Well, I guess that was confirmed within

38:36

a couple months or not even a

38:38

couple of months, a couple of weeks, right?

38:40

Yeah. were talking about this story. I

38:42

think last news episode to be fair, we've

38:44

been talking about this in a series

38:46

of Zen six leaks and there's been Zen

38:48

six leaks. Many a long time

38:50

ago years ago where we were suggesting it could

38:52

be two nanometer as well But yes, we

38:54

I would say we doubled and tripled down over

38:56

the past two months. Yeah, so

38:58

Fist bump fist bump remember supports on

39:01

patreon Supports on patreon people that are

39:03

this is real journalism We're doing to

39:05

bring in this stuff and I don't

39:07

know how much I really have to

39:09

say about it though You know, we've

39:11

already confirmed all of this. It's nice

39:14

to see it publicly confirmed Um,

39:16

I guess the only thing I would

39:18

especially note, it's not really new, but exciting

39:20

as them saying 2026, you know, who

39:22

knows maybe that's just tape out, but I've

39:25

long heard that some products should launch

39:27

for Zen six, uh, next year. I even

39:29

heard for a while that some could

39:31

launch this year, maybe, but I don't think

39:33

so. But I think, yeah, I think

39:35

it's pretty clear AMD is going to have

39:37

Zen six out before Nova Lake, most

39:39

likely, which we're about to get to. And

39:42

so there's just, AMD is not

39:44

sitting still, as I keep saying, there's

39:46

always something new around the corner. Buy

39:49

now if you need something now. If

39:52

you don't, there will be Zen 6. If you

39:54

say, going to wait till Zen 6, but I

39:56

could use something now, but I'm to wait. Just

39:58

remember, then there will be Zen 7 within two

40:00

years after that. So. Well, but Zen

40:02

7 is when they'll stop, right? Yeah,

40:04

that's when they will just stop making new things

40:06

because of course. And then that's like, we have

40:08

a wins the next Sandy Bridge where I can

40:11

just hold on to something for 10 years. never

40:13

again. Cause there's competition. Although

40:15

I don't know if you bought a

40:17

5 ,800 or 5 ,700 or yeah, no,

40:19

like 5 ,800 years ago, it's still probably

40:21

fine for most things. Yeah. I think

40:23

the only thing that's changed is I

40:25

think people forget that when Sandy bridge

40:27

came out, most people gamed at 60

40:29

frames per second. And of course that

40:31

might still work even in a lot

40:33

of games now with an Sandy bridge,

40:35

I seven, maybe not an I five

40:37

and one. 44 Hertz was pretty much

40:39

a thing. And so that's remained the

40:41

thing until now. I think a thing

40:43

that's changed is it went from most

40:46

people gaming at 60 some at 144

40:48

to most gaming at 1 a 44

40:50

mini wanting to game at 240 360.

40:52

There's like 500 Hertz monitors now. And

40:54

so I do think every now and

40:56

then people need to remember like, Oh,

40:58

my CPU is, you know, whatever. Like

41:00

I, I, I'm right now using a

41:02

Zen for X 3d CPU upgraded my

41:04

desktop and most of my games can

41:06

run to 40 Hertz, but some I

41:08

might cap at 200 or lower. And

41:11

it's like, well, you see, you could use

41:13

a faster CPU. And it's like, well, to

41:15

be fair, we used to be okay with

41:17

120 and now we're like at 240 Hertz

41:19

and way more complicated games. And so I

41:21

guess there's more of a push to upgrade

41:23

now though, because the monitors just keep getting

41:25

so much better. There's a real monitor stagnation

41:27

for like a decade of while ago. Yeah.

41:30

Um, all right. AI Hollow Knight writes

41:32

and it says, can you 3D stack

41:34

TSMC and three family nodes with TSMC

41:36

and six, there's a way too much

41:38

demand. I think for N four and

41:40

N three wafers, but N six is

41:42

a ton of unused capacity. Uh, example,

41:44

could a three nanometer Zen six CCD

41:46

still use N six for the V

41:48

cash? Or does it have to be

41:50

N four? Um, I guess

41:53

I can't confirm the V cash thing, but

41:55

like which node they're going to use

41:57

for this generation of V cash for So

42:00

that wouldn't surprise me if they stuck with six

42:02

nanometer or something. Although, yeah, I don't know. We'll

42:04

have to see on that one. I guess, I

42:06

don't know. I don't know what note it is.

42:08

Don't write that down. But I will say this

42:10

AI Hollow Knight. I feel like I keep seeing

42:12

in the Moore's Oz at discord, like saying, what

42:15

are they going to do about capacity? What are

42:17

they going to do about capacity? And I would

42:19

remind you, I have leaks that they are probably

42:21

going to use six nanometer for some of the

42:23

IO guys for desktop. And they're

42:25

using three nanometer for the others. And they're

42:27

using three nanometer for a lot of the

42:29

APUs. So I know you're worried about capacity,

42:31

but for two nanometer, those chiplets are like

42:33

not that big. I think I, yeah, least

42:36

there's 75 millimeter squared. And

42:38

everything else is pretty much using

42:40

other nodes. And they're still

42:42

leaving six nanometer capacity for like the

42:44

PlayStation six and some other products. Give

42:47

it to give it a year and a

42:49

half. Like two nanometer will be in demand, but

42:51

they're still going to use a lot of

42:53

three nanometer. And I, I'm not

42:55

sure how, and they're still using four

42:57

nanometer for a much of graphics

42:59

cards for the next two years. I

43:01

don't, and the PS4 pro. So

43:04

I, I don't know. Let's

43:06

not assume there will be

43:08

the same supply constraints. Will it

43:10

be some, yes, but there's

43:12

always supply constraints when new products

43:14

use bleeding edge nodes. And

43:17

I just think that's something you have

43:19

to accept. There's always going to be

43:21

a supply constraint. when Zen set two

43:23

uses seven nanometer and then they move

43:25

to, you know, five and then four

43:27

nanometer and that's be, they accept it

43:30

because that's the beginning of the generation

43:32

and people are still buying Zen three

43:34

AI Hollow Knight. And so that's why

43:36

it's new. It's going to have supply

43:38

constraints and then it won't for most

43:40

of its product life. Yeah.

43:43

And we still live in

43:45

a world where AMD is

43:47

trying to penetrate the market

43:49

more. So, going

43:51

for unlimited supply or might be a

43:53

stupid idea for them. Yeah. And I

43:55

think this is just, let's, we don't

43:57

know how much supply they're buying up.

43:59

The fact that you're leapfrogging apple on

44:01

this one is actually pretty crazy to

44:03

me. I mean, who knows, Zen six

44:05

could be the generation where they really

44:07

go for it with OEMs. And you

44:09

know, they, they want to win. They

44:11

want to keep the performance crown over

44:14

Intel and I, you know, so we'll

44:16

see, you know, maybe they'll be expensive,

44:18

but they just buy up a ton

44:20

of the capacity. Like they're actually willing

44:22

to outfit Apple for some of this.

44:24

Uh, and let's wait and see how what

44:26

happens here. It's already crazy. They're going

44:29

with two nanometer. I don't think they go

44:31

with two nanometer to like make 10

44:33

wafers. Like this would probably be some massive

44:35

strategy here and they're using the plant

44:37

in Arizona. So that's more capacity. too. So

44:39

I don't know. I wouldn't assume they're

44:41

just doing this to win a benchmark. They're

44:43

also probably doing this because they have

44:45

some bigger plan to actually have enough supply.

44:49

AI generated fish writes in and says

44:51

piercing together what we know so far

44:53

about Strix Halos launch. What's the bottleneck

44:55

in production here? It's not using standard

44:57

desktop chiplets. Is that it? Or the

44:59

low OEM demand due to high pricing

45:01

and unproven demand for a new type

45:03

of APU. I want to get excited

45:05

about Medusa Halo, but AMD does not

45:07

have a history of height, but does

45:09

have a history of hyping things. And

45:11

then their GPUs are in one to

45:13

three laptops, which are in low supply

45:15

and weirdly specced out. So I'd say

45:17

two things. AI generated fish. Number one.

45:19

It's new. So same thing happened with

45:21

Renoir, a new APU platform, AMD competing

45:24

in a market or a segment they

45:26

hadn't before, before it was just standard

45:28

laptop. Now it's, I mean, Apple mega

45:30

APUs, Apple like mega APUs. And anytime

45:32

there's a new platform, it's going to

45:34

take time to develop. And over time,

45:36

Renoir had hard to get, Cezanne, not

45:38

as hard to get Rembrandt, very hard

45:40

to get it first, then easy to

45:42

get Phoenix, a little easier to get

45:44

and hot point, easy to get Strix,

45:46

easy to get. Like it's probably going

45:48

to be a similar cadence, maybe faster

45:50

or more accelerated. But I think with

45:52

the Halo series, we're just like, this

45:54

is brand new stuff. So just let

45:56

them cook. You know, yeah. Yeah.

45:59

Yeah. I mean, hopefully Medusa Halo,

46:01

uh, there's, they already have their

46:03

relationships built or to introduce more

46:06

models that they designed with strict,

46:08

with Strix Halo in mind. Like

46:10

a D -Day metaphor, Strix Halo

46:12

is the first guys on the

46:14

beach. Medusa may land easier.

46:17

Yeah. What do you want me to say guys?

46:19

I'm getting my, I'm a white guy. My

46:22

thirties. I'm going to start making world war two

46:24

references. It's going to happen pretty soon. It's

46:26

just to be, I don't think it's going to

46:28

be the history channel that that jokes too

46:30

old because the history channel isn't the history channel

46:32

anymore. It's pond stars now, but the history

46:34

channel is now pond stars and aliens and aliens.

46:36

I don't, I don't think I'll be watching

46:38

that. I'll be watching just tons of YouTube documentaries

46:40

probably, but, but here's the second point. It

46:43

does have unique CCDs. And I

46:45

don't think that helped. And

46:48

I will be curious. And I

46:50

think we can kind of see a

46:52

lineage to Zen 6 here. The

46:54

CCDs and the way the fabric connects

46:56

to the IO dye is different from

46:58

desktop Zen 5 with Strix Halo. And

47:01

it's lower latency. There are some benchmarks

47:03

where a wildly outperforms desktop Zen 5.

47:05

And so I think there's something there

47:07

where they learn things making Strix Halo

47:09

that didn't get into Granite Ridge desktop

47:11

Zen 5. that will probably be in

47:13

Zen six and hopefully all of Zen

47:15

six. So it is all interchangeable. And

47:17

so even though I know there's like

47:19

an end to X and then an

47:21

end to P variant for Zen six,

47:23

I do hope they're more interchangeable so

47:25

that like, at least some models are

47:27

easier to supply because I will admit

47:29

the unique CCD thing probably is not

47:31

helping supply. Yeah.

47:33

I mean, if they're more modular

47:36

next generation, that would be great. So

47:38

they could hopefully shift supply more

47:40

easily back and forth to whatever is

47:42

in demand. Um, bull

47:44

dead writes in and says, is

47:46

the reason that we're not seeing

47:48

all that much NPU utilization is

47:50

because software developers are waiting on

47:52

first party, possibly even open source

47:54

software to implement NPU and GPU

47:57

load balancing because AMD just launched

47:59

a new LLM client whose selling

48:01

point seems to just be NPU

48:03

IGPU load balancing. Uh, and I'm

48:05

going to skip some of this here. Once

48:07

installed, you launch guy and begin exploring its

48:09

various agents and capabilities or two versions. Um,

48:12

Yeah, so I Don't know

48:14

Dan. Why do you think nothing

48:16

uses these NPUs that take

48:18

up die space yet? I

48:22

Don't know I mean

48:24

I don't I just

48:26

don't think that I

48:28

think There was over

48:30

-enthusiasm about the hardware

48:32

before anybody actually thought

48:34

of a software solution

48:36

solution it was solving

48:38

like I think there's a

48:40

ton of enthusiasm because it's useful

48:42

in research and then people didn't

48:44

realize that it's not that useful

48:46

outside of a few applications. I

48:49

don't know. Yeah, I mean, because

48:51

look at like the whole

48:53

Microsoft recall thing with Windows 11

48:56

that they like just backtracked

48:58

on. I don't and they still

49:00

I don't know someone's probably gonna write

49:02

in the comments that they actually are

49:04

implementing it tomorrow And I was an

49:06

idiot for saying this I but clearly

49:08

this just all the AI features I

49:10

will say aren't being rolled out as

49:12

smoothly as initially people thought they would

49:14

and I don't really have a solid

49:16

answer, right? I

49:18

I think I think

49:20

at the end of the day, no one's finding

49:22

a super useful thing to use these for,

49:24

or the MPUs are still aren't strong enough to

49:26

use them well. And I can

49:28

say that like using the MPU and

49:31

not even using it, using the co

49:33

-pilot button on this mini's for mini

49:35

PC I've had over here when I

49:37

mess around with it that doesn't use

49:39

the MPU. I mean,

49:41

I would tell it what its name

49:43

is. It would forget it in

49:45

30 minutes and it There were just

49:47

some, it was more accurate for

49:49

sure than Google's Gemini AI, but

49:52

it hedged so hard to not

49:54

be inaccurate that it wasn't useful half

49:56

the time when I used a

49:58

co -pilot to search things on the

50:00

web for me. So I just think

50:02

that there's a thing here where

50:04

it's just, it's just not really that

50:06

ready and either it needs more

50:08

processing power than the NPUs can provide

50:10

or At the end of

50:13

the day, no one, it's a chicken and egg

50:15

problem where like developers don't know, is the MPU

50:17

good enough? Should we put everything into making a

50:19

40 tops requirement? Do the thing we want it

50:21

to, and we know new ones may be there,

50:23

but we don't know if anyone will use this. I

50:26

don't know. I think there's a chicken and egg of

50:28

like trying to find a useful use of this and

50:30

deciding if this is where we stop. Does that make

50:32

sense? Yeah. All

50:35

right. Let us now

50:37

then move on. to

50:41

story number five Intel's next generation

50:43

Nova Lake CPUs will seemingly use a

50:45

new LGA 1954 socket. So I'm

50:47

going to quote here from Tom's hardware.

50:50

And then I actually have a

50:52

couple of leaks shipping document source from

50:54

NBD LTD purport that Intel might

50:56

switch to the LGA 1954 platform for

50:58

its next generation, Novel a processors

51:00

on desktop via all rack. This is

51:02

Accompanied by PCH tooling, likely intended

51:05

for the 900 series chipsets. Importantly, these

51:07

listings do not indicate an imminent

51:09

launch, especially since Noble Lake has officially

51:11

been confirmed as a 2026 product.

51:13

What's really official that Intel says, but

51:15

Noble Lake is officially part of

51:18

Intel's product family. So they have acknowledged

51:20

it exists. At least I agree

51:22

with that. Such a supersede aero lake

51:24

next year for a liminary silicon

51:26

confamurations, alleged two clusters of eight core

51:28

coyote co B cores. These are

51:30

an Arctic Wolf E cores. This is

51:33

of course, exclusively leaked by this

51:35

channel. Accompanimented by four low power LPE

51:37

cores in the SOC title, adding

51:39

up to 52 hybrid cores. Intel's engineers

51:41

explore, yeah, let's skip ahead

51:43

here. The information within the manifest implies

51:45

that Intel's actively distributing LGA 1954 testing

51:47

hardware to its global facility specifically. These

51:49

are not fully fledged motherboards, but appear

51:51

to be some form of specialized interposer

51:53

to test voltage regulation for the upcoming

51:55

platform. Either way, these kids are designated

51:58

as Novel Lake S. the short as

52:00

N V L S the shorthand for

52:02

no link S. So this heavily does

52:04

suggest that. No, I want to say

52:06

throw two last minute leaks here. Cause

52:08

I, this is a last minute story

52:10

added to the notes. And then I

52:12

reached out to a couple of sources. One

52:15

source tells me that it's not

52:17

necessarily finalized, but they have heard about

52:19

an LGA 1954. I'm actually looking

52:21

at what they said to me on

52:23

the side monitor right now. Um,

52:25

and that it was originally possibly some

52:27

embedded thing, but they might use it also

52:29

as they socket and eventually platform for

52:31

Novel Lake and possibly even the arrow Lake

52:33

refresh this person claims. So though I'm

52:36

wondering if they mixed that up. And the

52:38

only reason they would do that is

52:40

if they made the NPU so much stronger

52:42

that they needed, or they claimed to

52:44

need new pants. I don't know. Amy seems

52:46

to be able to do a lot

52:48

without needy new pinouts, but whatever, um, or

52:50

not new pinouts, new pin configurations. Um,

52:52

but I have another source here. Um,

52:54

an Intel source that tells me

52:56

they were debating whether Nova Lake should

52:58

continue on LGA 1851 or 1954.

53:00

Last I heard, and I have specifically

53:03

heard of a 1954, meaning this

53:05

person can't confirm if they're going to

53:07

use it, but they've heard of

53:09

this configuration and seeing this being shipped

53:11

with Nova Lake. That's why we

53:13

made this main story too. Cause sources last minute confirmed

53:15

this to me. So like, yeah, it sounds like at a

53:17

minimum, Nova Lake's going to require a new socket. So

53:19

what do you think of that, Dan? Not

53:22

that much to say other

53:25

than, good job Intel, just

53:27

keep that platform longevity not

53:29

being a thing. When

53:32

you're on the back foot,

53:34

really losing market share and trust

53:36

with consumers, having platforms

53:39

that last a quarter or

53:41

half as long as your competitor

53:43

really doesn't help. Yeah,

53:45

I mean... One thing I would

53:47

say is I continue to

53:49

believe that this isn't necessarily like

53:51

a nefarious thing. It's an

53:53

incompetent problem. It's like, you

53:55

can't convince me that it was all

53:57

due to milking. I think that was

53:59

explained some of it, but I don't

54:01

know. When you go back and look

54:03

at that, they had Kaby Lake and

54:05

then they went to coffee like one

54:07

and then coffee like two and then

54:09

comment like, I believe these all require

54:12

different sockets. To me, there's just

54:14

something there where there's a lack of uniform

54:16

leadership and there's no one at the top going,

54:18

Hey guys, we're going to be competing with

54:20

AMD really hard. Maybe we should just go to

54:22

the Comet Lake socket now so that all

54:24

these are on the same socket because surely that's

54:26

just something we should get ahead of. That's

54:28

something I believe AMD is done where they go.

54:30

Does Zen six really need a new platform

54:32

for consumers? Are they really going

54:34

to buy expensive DDR six right when it

54:36

launches? Maybe we should just, if it can support

54:38

DDR five, do that, which is what I

54:40

believe they're doing with Zen six on AM five

54:42

to support it. And then you look at

54:45

Zen seven and they're like, you know, okay, well

54:47

Zen seven is definitely going to require DDR

54:49

six. Well, what was then eight probably required? Let's

54:51

plan ahead for a socket when you need,

54:53

you know, and Intel doesn't do that. Yeah.

54:55

It could, or from the outside, it

54:57

does seem to speak to a lack

54:59

of cohesiveness or communication between the various

55:01

teams working on different architectures that Intel

55:04

where I don't know, it seems like

55:06

they're all just doing their own thing

55:08

at a certain point. Yeah.

55:10

Or that's a possible explanation. I don't

55:12

know if there's anything to that, but

55:14

yeah. And it's like, we've seen pictures.

55:16

Actually, I may know a source and

55:18

maybe get, be able to physically get

55:20

a hold of one of those meteor

55:22

like desktop chips. And it's just like,

55:24

so you thought meteor like might be

55:26

on LGA 1700, but then it was

55:29

1851. And then you would go

55:31

to Aero Lake, but then it wasn't. So

55:33

that was kind of a waste to have Aero

55:35

Lake use a new platform. If you're just

55:37

going to go to another one, to me, you

55:39

know, all of this is just that, them

55:41

not planning ahead and going, huh, hmm,

55:43

if Meteor Lake's not going to use 1851,

55:45

maybe we should just skip 1851 and whatever

55:47

we're going to use after that should be

55:49

what Aero Lake's on, especially if there's a

55:51

small chance Aero Lake refresh will even be

55:53

on it. That's just ridiculous. Um,

55:56

PRTD. B N M E writes in and

55:58

says, in light of the LGA 1954

56:00

rumors, they remind me of an aspect of

56:02

Intel's market dominance. It's stagnation 10 years

56:04

ago. Intel had LGA 1156, then 1155, then

56:06

50, then 51, where the difference between

56:08

those sockets were just a few pins, sometimes

56:11

even going down and then back up.

56:13

It is clear to me now that the

56:15

reason they did that is to force

56:17

motherboard upgrades once every other generation. See back

56:19

then, I think that is. pretty much

56:21

true. Yeah. Well, the physical sockets

56:23

themselves are similar to each other. My question

56:25

is that with the LGA 1954 rumors, what

56:27

is preventing Intel from designing a single socket

56:29

that can last for many generations? Why is

56:31

it that they always limit their socks to

56:33

two generations of a similar architecture? Well, they

56:35

do not have the market of mantra diamonds

56:37

that once had 10 years ago to justify

56:39

it anymore. Again, what we

56:41

just said, I probably should have

56:43

read this question before he kept

56:45

talking about it. You know, like,

56:47

I think it's like no one

56:49

told one team that, Hey, we

56:51

should make 1954 be what you

56:53

have, uh, aerolake use. And we

56:56

should skip 15 1851. If meteor like

56:58

isn't actually going to be launched on it.

57:00

They didn't. And so now there's just,

57:02

again, it's a huge company, a bloated company

57:04

still that just has people working on

57:06

different products. None of them talk to each

57:08

other. All of a sudden we just

57:10

get new platforms every year or so because

57:12

they aren't talking to each other. That,

57:14

that, that's my answer. Yeah,

57:16

I mean it really speaks

57:18

to the fact that

57:20

What I mean Intel is

57:23

bigger than AMD and

57:25

Nvidia combined And they put

57:27

out worse products than

57:29

both of them. It's it's

57:32

kind of crazy. Yeah

57:34

All right. Well, let us

57:36

now move on to

57:38

story number six. All right

57:41

PlayStation 6 handheld leaked

57:43

again uh, quoting from tech

57:45

power up, uh, with carbon cry noting that

57:47

it was heavily edited for readability. The

57:49

mysterious handheld is said to be powered

57:51

by a 15 watt SSC manufactured on a

57:53

non -specific three man meter node. According

57:55

to Kepler L two on Twitter that posted

57:57

that play stations are in hand world

58:00

is capable of running PS five generation

58:02

games is something we've already confirmed. However, at

58:04

reduced resolution and affirming due to bandwidth

58:06

and power restrictions, they also see a PS

58:08

six portable gaming performance being somewhere between

58:10

the Xbox S and the PlayStation five,

58:12

depending on what Sony decides to do. But

58:14

Yeah. So there's that. There's also PlayStation

58:16

portable leaks that keep percolating on NeoGaff

58:18

as well. Half of these are just people

58:21

talking that seems semi -informed by friends,

58:23

by the way. Um, so, but long story

58:25

short, multiple leaks are continuing to come

58:27

out that echoed the exclusive reporting of Moore's

58:29

law instead. I believe about a year

58:31

ago now on an upcoming PlayStation handheld,

58:33

specifically Kepler thinks that the handheld will play

58:35

PlayStation five titles, lowers those solutions. And

58:37

they say the SSC is designed for

58:40

15 Watts, a similar power level to steam

58:42

decks, Van Gogh, APU with a better

58:44

note in underlying architecture, PlayStation six portable.

58:46

Therefore could truly push handheld gaming forward, assuming

58:48

it launches within a few years, as

58:50

one would hope given the length of

58:52

the current generation consoles. And so

58:54

what I mostly take this is,

58:56

is again, the leaks we've put

58:59

out are being further confirmed. And

59:01

I would just add that. Yeah,

59:03

the PlayStation 6 one is also heavily

59:05

rumored in circles that seem to know things

59:08

ahead of time as well And I

59:10

continue to believe that is the move Sony

59:12

should make is they should skip a

59:14

PlayStation 5 portable Especially if it would be

59:16

different hardware and just make a ps6

59:18

model so that it can keep going forward

59:20

Keep so that you have that one

59:22

design that can go on for the next

59:25

10 years instead of a ps5. That's

59:27

about to be replaced in two years Yeah,

59:29

I mean with a ps6 portable And

59:32

if it could run the same

59:34

titles as the PS six, just at

59:36

a lower resolution, that would be

59:38

pretty cool. Yeah. Um,

59:41

all right. Let me see here.

59:43

Let me skip this one and

59:45

just go to the final story,

59:47

which is quite a long one.

59:49

Let me say story number seven

59:51

tariff update and video misleads customers.

59:53

Andy projects losses and Apple air

59:55

lifts iPhone. So I'm just going

59:57

to quote little things here. from

59:59

various sources and get to the

1:00:01

editorialization. First sources from Reuters. NVIDIA

1:00:03

did not warn at least some

1:00:05

major customers in advance about new

1:00:07

U .S. export rules. It was told

1:00:09

about a week ago requiring it

1:00:11

to obtain licenses to sell its

1:00:13

China -focused artificial intelligence chips, like

1:00:15

the 5090D, I believe. According

1:00:17

to two sources familiar with the matter, Reuters

1:00:19

quoted it again from a different

1:00:21

article. Meanwhile, Nvidia said it would take

1:00:24

$5 .5 billion in charges after the

1:00:26

U .S. government limited exports of its

1:00:28

H20 artificial intelligence chips to China,

1:00:30

a key market for one of its

1:00:32

most popular chips. And at the

1:00:34

same time, AMD expects charges of up

1:00:36

to $800 million due to the

1:00:38

latest curves by the Trump administration exports

1:00:40

of advanced processors from China that

1:00:42

companies sell on Wednesday. Now moving on

1:00:44

to the verge who quotes is

1:00:46

being quoted that razors up plumbing blade

1:00:48

16 and other laptops are no

1:00:50

longer available for pre -order or purchase

1:00:53

in the United States. When asked recently

1:00:55

if tariffs might affect razors prices

1:00:57

or availability as public relations manager Andy

1:00:59

Johnson told the verge, we do

1:01:01

not have a comment on this stage

1:01:03

regarding tariffs. Yeah. Seems like you're

1:01:05

just scared to admit what you're already

1:01:07

doing. Now reporting from heist .de as

1:01:09

the Taiwanese financial newspaper commercial times

1:01:11

reports, Acer, ASUS Lenovo, Dell and HP

1:01:13

are suspending deliveries to the US.

1:01:15

This means that the availability of new

1:01:17

products there is likely to be

1:01:20

limited for the foreseeable future. And finally,

1:01:22

quoting Reuters again, tech giant Apple.

1:01:24

Chartered cargo flights to very 600 tons

1:01:26

of iPhones or as many as

1:01:28

1 .5 million of them to the

1:01:30

United States from India after it stepped

1:01:32

up production there in an effort

1:01:34

to beat President Donald Trump's tariffs sources

1:01:36

told writers. So yep, tariffs are

1:01:38

here and everybody's winning, right? Sarcasm aside,

1:01:40

we see exactly what we expected

1:01:42

and reported on previously. Companies are tallying

1:01:44

millions and billions in near immediate

1:01:46

losses. Others are rushing to ship in

1:01:49

their products. If their margins could

1:01:51

sustain things like airlifting and others are

1:01:53

taking the more obvious response. to

1:01:55

just wait. No one is sure what will

1:01:57

happen with tariffs. No one can even say precisely

1:01:59

how they work and how they will be

1:02:01

enforced, at least of all the U S government

1:02:03

can't even seem to say it seems what

1:02:05

is clear is the continued tech blockade of China.

1:02:07

And it is a blockade at this point,

1:02:09

by the way. Now even products

1:02:11

designed to get through sanctions are barred. AMD and

1:02:14

Nvidia are expecting significant losses. And it seems

1:02:16

Nvidia might have hoped, and I think this is

1:02:18

an interesting thing read in between the lines

1:02:20

here. NVIDIA might have hoped to

1:02:22

prevent the tightening of sanctions. And

1:02:24

maybe that's why they didn't tell their partners

1:02:26

yet. Cause he, I believe he had like

1:02:28

a, let me continue. See the way sanctions

1:02:31

work is that companies are told ahead of

1:02:33

time to prepare or to raise concerns, but

1:02:35

for Reuters sources, NVIDIA didn't give advanced warning

1:02:37

to at least some of their customers. And

1:02:39

that likely means that Jensen Wang hoped to

1:02:41

convince Trump to allow age 20 sales into

1:02:43

China, but failed at least for now. Cause

1:02:45

I believe they had some dinner or something.

1:02:49

Then after the dinner, he had to

1:02:51

tell them too late. It'll be interesting

1:02:53

to see if Singapore becomes Nvidia's largest

1:02:55

market for future earnings. Like it already

1:02:57

seems to be to smuggle and stuff

1:03:00

to China. Also DHL. Yes. The massive

1:03:02

shipping company DHL just suspended deliveries to

1:03:04

the U S for anything that costs

1:03:06

over $800, which is a lot of

1:03:08

things you use DHL for. So that

1:03:10

was the last minute edition. By the

1:03:12

way, that is just Guys, not just

1:03:14

like companies, DHL doesn't want to bother

1:03:16

with shipments to America anymore. Unless it's

1:03:19

cheap ones that I think get through

1:03:21

that de minimis thing, get through that

1:03:23

tear. Yeah. Uh,

1:03:25

well, I, I guess those are

1:03:27

the, uh, the versions of

1:03:29

getting around tariffs that, uh, everybody

1:03:31

is discussing. There's some companies

1:03:33

that are like, screw it. Uh,

1:03:36

we're just not dealing with

1:03:38

them right now. Uh, you have

1:03:40

I apple that's hoping to

1:03:42

get around tariffs by uh,

1:03:44

changing shipping routes. And then Nvidia,

1:03:46

that's just hoping they could, uh,

1:03:48

reason with Trump, uh, last minute

1:03:50

and even lying to customers. Cause

1:03:52

they felt like they had to

1:03:54

be able to, but now they

1:03:56

didn't. And it's going to backfire

1:03:58

on them even harder. Or

1:04:01

who knows, maybe it won't, maybe

1:04:03

in a week and a half though,

1:04:05

those sanctions will just be reversed

1:04:07

because everything is stupid and erratic. Yeah.

1:04:09

And so. Like one spiel I

1:04:11

want to say too is like I've

1:04:13

actually our tariff content has had

1:04:15

higher upfoot ratios than average So anyone

1:04:17

who thinks that this is a

1:04:19

popular subject is wrong It is a

1:04:21

popular subject because this is affecting

1:04:23

everybody and people want I think there's

1:04:25

a hunger out there for people

1:04:27

who want people who cover the tech

1:04:29

space to not be Babies and

1:04:31

scared to talk about the number one

1:04:33

thing affecting gaming hardware this year

1:04:35

because it is tariffs and so more

1:04:37

so I said we're not afraid

1:04:39

to do that And it's reflected

1:04:41

mostly positively. Now, of course, there

1:04:43

are some people that are attack us.

1:04:45

And, you know, there's the usual

1:04:48

canned things like graphics cards are exempt

1:04:50

or not. Or like all

1:04:52

these other things. And, you know, half of those

1:04:54

annoying comments are probably just confused because we're

1:04:56

all confused and the news changes every day. And

1:04:58

they're clinging to the thing that they think

1:05:00

will make them right for what they are pretending

1:05:02

as a sports team. It's not a sports

1:05:04

team. This is stuff where things are being hurt

1:05:06

and people are being hurt. So. you

1:05:08

know, actually think about what you're rooting for.

1:05:10

This isn't sports anymore, people. This is real

1:05:12

policy that is changing things. And I guess

1:05:14

that's my main point I want to make

1:05:16

right now is when people say things like,

1:05:18

that's not true. This was just said, for

1:05:21

the most part, when me and Dan talk

1:05:23

about tariffs on this channel, we're talking about

1:05:25

things that have already happened. You can't tell

1:05:27

us we're wrong. HP did suspend

1:05:29

sales. And even if they started them again,

1:05:31

there will be a gap in sales

1:05:33

that will raise prices. DHL

1:05:35

did stop shipping for things over $800.

1:05:38

You can't say we're wrong. We're

1:05:40

reporting on things that have already

1:05:42

happened, like companies focusing new supply

1:05:44

of Nvidia cards, mostly to Europe. That

1:05:47

has happened. That is not a

1:05:49

theory, right? And so that's what

1:05:51

I would say is the chaos is

1:05:54

the problem. You can, I'm

1:05:56

sure on Monday before this episode comes

1:05:58

out, then on Tuesday after and Wednesday,

1:06:00

these will just change and yo yo

1:06:02

back and forth. But try to notice

1:06:04

like we're reporting on Apple. They have

1:06:06

airlifted iPhones until probably until December, but

1:06:09

NVIDIA and AMD are taking immense losses.

1:06:11

Will they recover them in the future?

1:06:13

Sure. Say they will. If you want

1:06:15

to believe they will, but we're recording,

1:06:17

we're reporting on the problems that have

1:06:19

already happened in telling you how that's

1:06:22

going to affect you now. So those

1:06:24

comments that say we're wrong, we're not

1:06:26

wrong. We reported on something that happened

1:06:28

a day or a week ago and

1:06:30

how that's going to affect you in

1:06:32

the future because we're not mostly bothering

1:06:34

to try to guess how this will

1:06:37

pan out. We just know

1:06:39

it will probably be bad even if things go

1:06:41

in a certain way. Well, and it's hard

1:06:43

to predict exactly how things will pan out because,

1:06:45

uh, the way a lot

1:06:47

of these policies are written, even

1:06:49

experts aren't, don't seem to understand

1:06:51

what they say half the time.

1:06:53

Yeah. Like, you know, I just

1:06:56

had Hogue on, uh, broken

1:06:58

silicon who's a. a

1:07:00

business attorney that often deals with

1:07:02

disputes, having to do a tariffs

1:07:04

and trade. And I would quote

1:07:06

him here again. I really recommend this episode. I think it's

1:07:08

one of the better ones over the past couple of

1:07:10

years. The way he put it is

1:07:13

I could, it is my job to do this.

1:07:15

Sometimes I could sit down right now, Tom

1:07:17

and tell you exactly how much it would cost

1:07:19

to switch versus now. I could go through every

1:07:21

component and every little thing and where it comes

1:07:23

from. And I could genuinely like that is my

1:07:25

job. Sometimes I could genuinely tell you what it

1:07:27

would cost, but I'm not going to spend

1:07:29

billable hours doing that for customers because I know

1:07:31

that we'll change the next day and that would

1:07:33

be a waste of their money. You

1:07:36

know, so yeah, I mean, that's what it

1:07:38

all comes down to. These things are changing

1:07:40

back and forth. But at the end of

1:07:42

the day, I guess the end of my

1:07:44

major point too is at the end of

1:07:46

the day, even if you see tomorrow Trump

1:07:48

say he's lifted this and that and that

1:07:50

tariff, a lot of companies are still going

1:07:52

to suspend trade with the US and move

1:07:54

production out of the US or away from

1:07:57

having to depend on the US, I would

1:07:59

say. because they don't know if that's

1:08:01

going to change in a week. And they're

1:08:03

going to give it 90 days. So I

1:08:05

think there's going to be at least three months

1:08:07

of like, just people don't want to bother

1:08:09

talking to the US. Well, they wait to see,

1:08:12

you know, like a, like a wild

1:08:15

cat has just emerged at like a dinner party

1:08:17

and everyone's like, there's a tiger walking around.

1:08:19

Let's just not move. And then wait to make

1:08:21

sure the tiger's gone before we start talking

1:08:23

again. That's what's going to happen. Yeah,

1:08:26

techno writes in with the US economy being

1:08:28

made weaker by tariffs and other political situations. He

1:08:30

copped. Could we actually see the return to

1:08:32

normal prices in the rest of the globe? Since

1:08:34

Nvidia and AMD based their card MSRPs on

1:08:37

the US dollar, then convert for elsewhere. Presumably the

1:08:39

dollar continues to weaken at this rate for

1:08:41

the next few months. Yeah. The US dollar is

1:08:43

weakening. We could see prices drop by around

1:08:45

10 to 15 % in global markets by the

1:08:47

summer, meaning that after inflation, the 58 would have

1:08:49

a similar MSRP to that of the 10

1:08:51

80 in many European countries. I

1:08:53

don't say 10 80. But the

1:08:56

overall premise is plausible. Yes. If the

1:08:58

US dollar weakens and more supply

1:09:00

is sent to countries that are not

1:09:02

the US compared to previous generations

1:09:04

traditionally. Yeah. You're already seeing it.

1:09:06

You are seeing prices for RTX go

1:09:08

down in Europe where they go up

1:09:10

in the US. Yes. Again, this is

1:09:12

us not predicting. It's already happened. It's

1:09:15

already happened. Yeah. I mean, that

1:09:17

also depends on, well,

1:09:19

not, I mean, you could see relative

1:09:21

MSRPs go down, but I wouldn't. Bet

1:09:24

that that will reliably make

1:09:26

prices go down permanently. They

1:09:28

if supply goes down you

1:09:30

might just see Prices go

1:09:32

up like they are in

1:09:34

the US globally to although

1:09:36

probably not as fairly because

1:09:38

we're not dealing With if

1:09:40

that GPU was assembled in

1:09:42

China, you're not dealing with

1:09:44

whatever the hell we're at

1:09:46

185 percent again, it's it's

1:09:49

effectively a blockade so All

1:09:51

right now let us move

1:09:53

on to the wrap up.

1:09:55

All right. First one here,

1:09:57

Dan, dragon range refresh is

1:09:59

official. I got to say,

1:10:01

I don't have a lot

1:10:03

to say about rebranding Zen

1:10:05

for, for rebranding on strikes

1:10:07

again. mean, I'm sure this

1:10:09

is just leftover Zen for

1:10:11

supply that's been higher because

1:10:13

it is easier to bend

1:10:15

it higher. Um,

1:10:17

I would be surprised if this was

1:10:19

around for a long time, unless

1:10:21

genuinely this is just how they're deciding

1:10:24

to get rid of oversupply. I

1:10:26

mean, I guess they have to keep

1:10:28

making Zen forward to a degree

1:10:30

for their server customers. Presumably

1:10:33

they have some deal

1:10:35

strike in with an OEM

1:10:37

to put these in

1:10:39

laptops. So it'll probably be around

1:10:41

for some amount of time. Yeah. That's true. Yeah.

1:10:43

So, but I don't have anything to say

1:10:45

about this. Zenfors fine,

1:10:48

you're using a Zenfors system. Zenvive

1:10:50

is better. Zenfors fine. Van

1:10:52

Gogh might finally be coming

1:10:55

to do it yourself. I thought

1:10:57

this was interesting. Let

1:10:59

me see here. There's a

1:11:01

rise in AI Z2 stream. Then

1:11:03

there was the Z2A. And a

1:11:05

lot of people weren't sure what

1:11:07

that was, but new data is

1:11:09

emerging that the Z2A APU may

1:11:11

be Van Gogh and we might

1:11:13

be able to finally benchmark. You

1:11:16

know, I, you know, there might be

1:11:18

other handhelds that have the same APU and

1:11:20

the steam deck or who knows, I

1:11:22

think it would be interesting if a mini

1:11:24

PC got busy to a, and then

1:11:26

you could really do some heavy, heavy testing

1:11:28

with different TDP levels and like a

1:11:31

more traditional form factor. I think that'd be,

1:11:33

so I know that's something to watch

1:11:35

out for. Nvidia

1:11:38

supposedly fixes drivers. Although,

1:11:41

you know, let me look here. Dead

1:11:43

AI writes in and says Nvidia can't keep

1:11:45

getting away with this crap. The newest

1:11:47

driver fixes black screens. I was having black

1:11:49

screen issues. It had fixed them by

1:11:51

the way, but broke driver telemetry that's emerging

1:11:53

like clock speed, temp, power monitoring, what

1:11:55

is going, and it's also affecting the 30

1:11:57

of the 40 series. What is wrong

1:11:59

with them? How can they keep effing things

1:12:02

up this badly? And I editorialized that

1:12:04

a little bit, but there's, there's a lot

1:12:06

of not saying effing crap. A lot

1:12:08

of more swearing that I just got to

1:12:10

skip, but I don't know for me,

1:12:12

the new Nvidia drivers seem to fix some

1:12:14

issues that I did have. Um, I

1:12:16

didn't really have any issues, but I just

1:12:18

moved to a 4k OLED to 40

1:12:20

Hertz monitor. So I got that a two

1:12:22

40 Hertz, uh, 1440 P one and

1:12:24

then my old 4k one 21. So I'm

1:12:26

kind of pushing the limits of what

1:12:28

the bandwidth, the 40 90 can supply over

1:12:30

display board. I think that was causing

1:12:32

some black screen issues. They're gone now. They're

1:12:35

all fixed. So I'm happy about that.

1:12:37

And luckily I don't really monitor anything anymore

1:12:39

because I just don't overclock my 40

1:12:41

90 anymore. I just have it in the

1:12:43

low power bio switch and I'm letting

1:12:45

it run. But yeah, I don't know. This

1:12:47

is more evidence that Nvidia may have

1:12:49

fixed one problem and caused another. And I

1:12:51

don't know. It just feels like they

1:12:53

hired the RDNA one driver team. Yeah.

1:12:56

Good job. Also Nvidia Zora RTX

1:12:58

demo was provided. I actually meant

1:13:01

to try it out, but I

1:13:03

couldn't before you showed up Dan

1:13:05

for Easter weekend. But

1:13:07

it is a rather large download to

1:13:09

an instructing a 108 like gigabyte zip

1:13:12

was actually kind of annoying. So that's

1:13:14

why I didn't get to it, but

1:13:16

I'll probably be messing with that for

1:13:18

the next week. It'll be interesting. TSMC

1:13:20

moves 30 % of two nanometer to the

1:13:22

U S again, and worth noting that

1:13:24

more that this thing is happening again, to

1:13:26

go to that reader mail about what

1:13:29

is AMD going to do with two nanometer

1:13:31

supply. I don't think they would have

1:13:33

chosen two nanometer air hollow night. unless they

1:13:35

thought they could get a enough supply

1:13:37

and TSMC is rapidly expanding it in the

1:13:39

United States. So maybe it won't be

1:13:41

a big issue. Now this

1:13:43

one's, again, these are not

1:13:45

major stories, but I think this

1:13:47

is worth mentioning. Nvidia, I'm

1:13:49

sorry. I always do that. Nintendo

1:13:51

revised the switch to product

1:13:53

info and removed variable refresh rate

1:13:55

support from official sites. I

1:13:59

don't know what that means. What is going on with

1:14:01

Nintendo? I

1:14:03

don't know. Like we said last

1:14:05

time talking about the switch, something weird

1:14:07

is going on internally at Nintendo. Uh,

1:14:10

we, I know we've been speculating about

1:14:12

that for like a couple of years

1:14:14

now at this point, but there's too

1:14:16

much weird shit at this point to

1:14:18

deny. Like something has to

1:14:21

be going on there. I'm trying to

1:14:23

remember if there was anything equivalent with

1:14:25

like Sony like this. Cause I, to

1:14:27

my memory, right? Sony

1:14:29

eventually added 1440p support.

1:14:32

to the PlayStation five way longer. It took

1:14:34

them to put it on there than

1:14:36

I thought it should have taken, but they

1:14:38

did add it and they didn't claim

1:14:40

it had it and tell it did. And

1:14:42

I don't think they really, and they

1:14:44

added free sync support later. I don't remember

1:14:46

them saying it's like on the box,

1:14:48

free sync and tell it and then removing

1:14:50

it like. They tend to just take

1:14:52

a long time to add stuff, but they eventually

1:14:54

do at PlayStation. Can you

1:14:56

think of anything? Like they removed the

1:14:58

4K, I have to mention this or

1:15:01

the 8K thing from their boxes, I

1:15:03

guess. So I guess

1:15:05

that's comparable, but they didn't

1:15:07

brag about it running 8K a

1:15:09

whole lot. And in Nintendo put

1:15:11

VRR and a lot of official

1:15:13

announcements and they now removed it.

1:15:16

This thing's been done for years. How could they

1:15:18

not know if they're going to have VRR

1:15:20

support? I

1:15:22

don't know. I mean, I'm

1:15:24

holding on and just hoping

1:15:26

this is some weird oversight, like

1:15:28

they omitted it from some

1:15:31

marketing material, but I don't know.

1:15:34

Or yeah,

1:15:37

I really don't know why at the

1:15:39

last minute, they're like, no, no more

1:15:41

VRI support that almost that makes no

1:15:43

sense to me. Yeah. All right.

1:15:46

And the final one here, I just threw

1:15:48

it in because it involves tech. And this

1:15:50

is insane to me. Tesla sued for allegedly

1:15:52

faking odometer readings to avoid warranty repairs. And

1:15:54

it's not the first time someone's claimed

1:15:57

this. I believe this is them. Like they

1:15:59

change the odometer so they don't need

1:16:01

to pay for a warranty at the last

1:16:03

minute or something. So

1:16:05

like you're about to be out of warranty. And then

1:16:07

they speed it up so they don't have to do

1:16:09

it. I'm starting to think

1:16:11

I shouldn't get a Model Y, Tom. Yeah.

1:16:13

I don't know. Besides the fact that it feels

1:16:15

really cheap compared to competing cars. Yeah.

1:16:18

I say that as someone who

1:16:20

tried out a Model Y somewhat recently

1:16:22

and was, man, was I unimpressed

1:16:24

compared to the competition. I

1:16:27

don't what to say. I mean, there's

1:16:29

not really a whole lot to say.

1:16:31

It's not even gaming hardware, but it

1:16:33

involves tech. And it's like, that sounds

1:16:35

highly illegal for multiple reasons. There's

1:16:37

the model. I have Navi 23 in

1:16:39

it or whichever. I don't remember some, some,

1:16:41

some, some tasks to. Okay.

1:16:44

So it's tangentially

1:16:46

related. Sure. Sure. Sure.

1:16:48

All right. Well, that is going

1:16:51

to do it for this

1:16:53

shorter, but I still, I think

1:16:55

fully featured episode of broken

1:16:57

silicon like the model Y. I

1:17:00

argue if it has full, if it's

1:17:02

fully featured, when I have to use a

1:17:04

frickin' touchscreen to control the AC. Just

1:17:08

give me dials on cars, guys. The dials

1:17:10

still work. The dials and the buttons. Give me

1:17:12

the dials and the buttons. Don't give me

1:17:14

a button too for like reverse. Let me move

1:17:16

a stick because then otherwise have to like,

1:17:18

which button do I press the right one? You're

1:17:20

dangerous. dangerous. But

1:17:23

yeah, you know, as

1:17:25

usual, remember support Moore's Law instead on YouTube. I

1:17:28

think like, almost half of you aren't

1:17:30

even subscribed. Please double check that you're subscribed

1:17:32

to the more as lots of YouTube

1:17:34

channel and ring the bell button and tell

1:17:36

your friends about us. Comment below, share

1:17:38

this podcast and consider supporting us on Patreon.

1:17:40

We just put out another over one

1:17:42

hour dye shrink, kind of a bonus broken

1:17:44

silica and going over Nvidia's mind share

1:17:46

and the 50 60 Ti situation. And

1:17:49

we just did some interviews with

1:17:51

developers. There's a ton of just hundreds

1:17:53

of dye shrinks. They are feet. If, you know,

1:17:55

if you just join us at the lowest tier,

1:17:57

you can ask the discord where you can interact

1:17:59

with me, Dan, the team. There's people who work

1:18:01

at Intel and AMD and Nvidia in the Morse

1:18:03

laws at discord where people will ask questions and

1:18:05

they will go, yes, that is how that driver

1:18:07

works. That's why that happened. Like there's thousands of

1:18:09

people in that discord that work in this industry.

1:18:12

And it's probably, I, many people say

1:18:14

it's one the most knowledgeable discords ever

1:18:16

and one of the most civil. discords

1:18:19

ever. I'm very proud of how, uh, not

1:18:22

polite, but how, how civil we've kept

1:18:24

it there. So please join us on

1:18:26

Patreon if you can. That's the best

1:18:28

way to show support. And then I

1:18:30

don't know, Dan, any final words? Happy

1:18:33

Easter. Happy Easter, everybody. And

1:18:35

I will have Gerard edit

1:18:37

this in the beginning. So,

1:18:39

you know why I said

1:18:41

this? I said, holy moly,

1:18:43

because Dan caught Jesse. Messing

1:18:46

with a mole in the yard. And

1:18:48

I never made it clear why I said

1:18:50

Holy moly at the beginning. It

1:18:52

is because Jesse messed with a

1:18:54

mole and she didn't hurt it.

1:18:56

At least not pretty. The mole

1:18:58

of successfully burrowed back underground after

1:19:00

I told her to get away

1:19:02

from it. So we watched it.

1:19:04

She wasn't biting it. I think

1:19:07

she was just kind of poking

1:19:09

it slowly with her paw. Like

1:19:11

what is it? Oh,

1:19:13

and she was doing this. She didn't bite

1:19:15

it at all, but she was like with

1:19:17

her nose, like, Oh, like, like that a

1:19:19

little bit. Like, so

1:19:23

almost trying to play with it. So we

1:19:25

don't think Jesse is a mole killer. I would

1:19:27

not trust her around a rabbit though. She

1:19:29

looks like she wants to get rabbits, but, and,

1:19:31

uh, and squirrels. But yes, that is why

1:19:33

I said Holy moly and Dan said Easter. The

1:19:35

point I was going to make is those

1:19:37

are the two things that were. news for us

1:19:39

today. And it is an hour, over an

1:19:42

hour later that I am correcting this, but we

1:19:44

will try to make Gerard edit in the

1:19:46

video in the beginning as well. Like this is

1:19:48

why Tom said, holy moly, we will say

1:19:50

later, but thank you for watching. Thank you for

1:19:52

listening. Have a good week, everybody. Goodbye.

1:20:03

Tom that guy is me and I

1:20:05

am indeed the creator editor writer

1:20:07

and showrunner of Moore's Laws Dead podcast

1:20:09

videos articles and other media. However,

1:20:12

it's not just me Moore's Laws Dead

1:20:14

is a team with broken silicon

1:20:16

co -hosted by my brother Dan audio

1:20:18

editing by Girog Cortez renders being done

1:20:20

by the industrial designer Jean -Philippe Clermont

1:20:22

and special assistance is also provided

1:20:24

by Carmichael and Kerry Noseau God as

1:20:26

well. Find all of our information

1:20:29

at www .moreslawsdead .com on the about -slash -support

1:20:31

page in the event you do

1:20:33

want to hire me for consulting work,

1:20:35

hire Gerard for audio work, hire

1:20:37

Jean Fleepe for industrial design work, or

1:20:39

you're interested in working with Carbon

1:20:41

Cry or Carino Sugata as well. You

1:20:43

can also find our long -term sponsors

1:20:46

on that page if you want

1:20:48

to show them some love for putting

1:20:50

food on our tables. Or you

1:20:52

can also mail us some love. You

1:20:54

can send letters or hardware donations

1:20:56

to the following address. Moore's Law is

1:20:58

Dead, Peelbox 60632 and Nashville, Tennessee,

1:21:00

zip code 37. Although,

1:21:03

to be honest, the best way

1:21:05

to show Moore's Laws Dead some

1:21:07

love is to support us on

1:21:09

Patreon. Patrons are what makes Moore's

1:21:11

Laws Dead content truly possible. Every

1:21:13

month and really every day, depending

1:21:15

on who you're talking about, me,

1:21:17

Gerard, Dan and John Fleepe are

1:21:19

working tirelessly to provide a steady

1:21:22

stream of content that we could

1:21:24

not keep doing unless we knew

1:21:26

the work was possible without being

1:21:28

reliant on sponsors dictating every little

1:21:30

thing we put out. Don't get

1:21:32

us wrong. We love our sponsors,

1:21:34

but we love directly working for

1:21:36

you, our fans, much more. If

1:21:38

you have any extra money, even

1:21:40

a couple free dollars a month,

1:21:42

consider supporting us directly on Patreon.

1:21:45

Those couple of monthly dollars will get

1:21:47

you access to the exclusive podcast Dye

1:21:49

Shrink, voting on subjects of future podcast

1:21:51

episodes, the ability to ask guest questions,

1:21:53

and of course, access to the Moore's

1:21:55

Laws Dead Discord full of like -minded people

1:21:58

who I am sure would to meet

1:22:00

you. I am one of them. Additionally,

1:22:02

higher tiers get access to early, ad

1:22:04

-free episodes of Broken Silicon, the ability

1:22:06

to ask questions in all Broken Silicon

1:22:08

episodes and loose ends live streams ahead

1:22:11

of the recording, and the entire back

1:22:13

catalog of Moore's Law is Dead podcasts,

1:22:15

in addition to having thanks in the

1:22:17

credits of videos and podcasts depending on

1:22:19

the tier with other perks available as

1:22:21

well. And hey, if you cannot afford

1:22:23

to support us directly every month, please

1:22:26

do share Moore's Law is Dead videos

1:22:28

and podcasts with friends and family, and

1:22:30

on social media and websites like Reddit

1:22:32

and give Broken Silicon a five star

1:22:34

review on Apple Podcast or your preferred

1:22:36

podcast app of choice. All of this

1:22:39

does really help us so much. But

1:22:41

like I said, this podcast would not

1:22:43

be possible without it, the patrons directly

1:22:45

providing predictable and reliable support every month.

1:22:47

And so now it is time to

1:22:49

give a personal thanks to the greatest

1:22:51

of the fans. The following supporters are

1:22:54

at the 10 gigahertz or higher supported

1:22:56

levels. Brad Medlin, Drita Full,

1:22:58

ZGist, Daniel D. Deek, Nicholas Buckner,

1:23:00

Aaron Close, Jan Rauner, Daniel Hyde,

1:23:02

Mjb1, GZ Ziggy, Toca, SNS Chalmers,

1:23:04

Sharon Ferriere, Valcom Aleph, Jensen Wayne,

1:23:07

Gregorus Hacker, Sarkastro, Chris Trishk, Evan

1:23:09

Dingle, Compress The Eye Blocks, Hal

1:23:11

Booma, Shredbird, Simi Malas, Jordan Simcovic,

1:23:13

The Boss Haas, Christensen Jonathan, Neath

1:23:15

Rizink, Jake223, Jake Martins, Licky, Mead

1:23:18

and Pork, Tim Robb, Ian Clifford,

1:23:20

Travis Gooding, Amers Unite, Mad, Zoot

1:23:22

Suit Taylor, Greg, Michael McGee, Patty

1:23:24

Cake, Stefan Coates, Amy Will Chief,

1:23:26

Mark Mitchell, Tommy, I should James

1:23:28

Anderson, Mark Raidmaker, Cole Attic, Cameron

1:23:31

Henry, Shang, Wesley Sager, Michelle Pell,

1:23:33

Hexapuma, WinWing, Reginald Ari, Teek Autumn,

1:23:35

WinStar, Corey Leonhardt, James I, Raider,

1:23:37

Little Germany, Shay, Milton, Melodic Warrior,

1:23:39

Brett Jones, Chuck Glidden, Mac Daffy,

1:23:41

Steven Dick, Justin Bussell, Austin Hagerty,

1:23:44

Joe Foot, Brian Scanlon, Hardin, Jensen

1:23:46

and Gima, Joseph Kelly, slushed by

1:23:48

Hymns to Gunn, Tails, 2299, Brian

1:23:50

Wright, Sizzifo, 7771, The Forbidden Juice,

1:23:52

Jeffrey Gentlemen, RB Racer, Alex Vega,

1:23:54

Dr. J Matt, Free D, Michael

1:23:57

Cozy, John Swin, Thalo, and five.

1:23:59

Crystal Joe, Elbergum, Keekum, Solarisadis, Brutnik,

1:24:01

Brian P, Houdat, 42, Kornster, 671,

1:24:03

Penta Winta, Hal Richard, Omega Doge,

1:24:05

Mr. Tiger, John Wick, Boomsick, Leo

1:24:07

Goki, and or Keevee, Pegwin's Love,

1:24:10

AI, Porcelain Oath, State Farm, St.

1:24:12

Curry, King 3, Nixie, Usmanias, Akosha,

1:24:14

37 by 73, Ryhill, Chee Bowles,

1:24:16

Sheriff Laffey, Kevin Coots, Smash, Axley,

1:24:18

Armand, Vakatura, Kalan, Raichel,

1:24:20

Jonathan Harris, Kareem, Elma

1:24:22

3, and of course, Angel of Kareem.

1:24:25

of course, of course, thank you to Zahara for the

1:24:27

music.