HomeVideos

They Should Have Bought A Hard Drive - WAN Show August 14, 2026

Now Playing

They Should Have Bought A Hard Drive - WAN Show August 14, 2026

Transcript

4734 segments

0:01

Hey, what's up everyone and welcome to

0:03

the WAN show. I got myself a birthday

0:07

present today and I'm going to talk

0:11

about it.

0:11

>> What does a guy with a firetruck get

0:13

himself for his birthday?

0:14

>> A robot.

0:15

>> Okay.

0:16

>> In other news,

0:18

>> that makes sense.

0:20

[laughter]

0:22

>> Luke was right again. [clears throat]

0:24

>> Oh no.

0:25

>> A PBS station in St. Louis has lost 70

0:30

years of archives after a cloud storage

0:34

reseller shuts down. Luke warned you and

0:38

we'll be talking about that later as

0:39

well. What else we got?

0:40

>> Did I?

0:41

>> Yeah, you did many times.

0:43

>> Oh, nice. Hell yeah. [laughter]

0:48

Uh I don't know. Where's the other one

0:50

that was good? You took the PBS one that

0:52

I wanted to use. Uh, New Mexico judge

0:54

orders Meta to pay nearly a billion

0:57

dollars. Do we finally have like fees

0:59

for companies that will actually impact

1:01

them uh that aren't coming from the EU?

1:03

Maybe, possibly.

1:05

>> I'm not getting my hopes up.

1:06

>> Um, me neither.

1:09

Which other thing do we want to talk

1:10

about?

1:12

>> You know, you have the entire time

1:14

before the show starts.

1:15

>> I didn't think you were going to take

1:16

the PBS one. I don't know why I thought

1:18

that, but I didn't think you were going

1:19

to.

1:19

>> I always try to pick the one you're

1:20

going to pick.

1:22

Yeah, but I didn't think you'd think I

1:23

would pick that one. [sighs] Man, we're

1:26

getting into some mind games on these

1:27

now. Um, this one's kind of funny.

1:30

Whatever. GeForce Now exploit lets you

1:32

access full Windows desktop through a

1:33

simple file swap, and a modder runs

1:36

local AI models on the ultimate tier

1:38

with 48 gigs of VRAM and no

1:40

restrictions, which is actually just

1:42

kind of funny in my opinion.

1:44

>> Beast

2:04

The show is brought to you today by

2:06

Vessie along with Whisper Flow, X

2:09

Square, and Hrefs with of course a rap

2:13

courtesy of our rap partner Dbrand. a

2:16

chair courtesy of our chair partner. Oh

2:18

goodness, I have a t-shirt in front of

2:19

it, Razer. And our laptops, also

2:22

courtesy of Razer. Why don't we jump

2:25

right into our headline topic of today,

2:27

WHICH IS THAT LUKE WARNED [screaming]

2:29

ALL OF YOU WHEN THE CLOUD WAS HAPPENING.

2:33

LUKE SAID HE CAME DOWN FROM ON HIGH and

2:36

he spaketh such that [cough]

2:39

>> on prem will always have a place.

2:42

>> Yeah.

2:42

>> And a St. Louis public broadcaster

2:46

learned this the hard way. Um, they lost

2:50

access to 50 terabytes of archival

2:54

material after their cloud storage

2:55

provider OSS shut down without notice.

3:00

After losing access to their data, 9PBS

3:03

discovered that the OSS website was

3:05

offline and the company was listed as

3:07

delinquent by the Colorado Secretary of

3:10

State. 9PBS learned that the defunct

3:14

storage provider was reselling Iron

3:16

Mountain services and 9PBS has sent a

3:19

demand letter to preserve its data and

3:22

to pay any reasonable costs associated

3:25

with its demand. But as 9PBS wasn't Iron

3:29

Mountain's customer, they actually, it

3:32

turns out,

3:33

>> can't just hand over the data.

3:36

>> That makes sense. That's so annoying,

3:38

but that makes sense. The full story is

3:40

a convoluted mess. Dan, can we put in

3:43

the link to Gizmodo in the chat cuz you

3:45

guys can read all about it there. It's a

3:47

fascinating rabbit hole. But really, I'm

3:50

only bringing this up because what I

3:52

want to talk about was how much Luke and

3:55

actually I also, but it's just, you

3:57

know, nicer to point out, I guess, your

3:59

co-host. I think that's why I was

4:01

confused. I was like, "What do you

4:02

mean?" [laughter]

4:04

>> How both of us were talking about this

4:07

for so many years

4:08

>> like from the beginning

4:09

>> that while there are good things about

4:11

the cloud,

4:12

>> there are also massive problems because

4:15

the second you don't own your

4:17

infrastructure,

4:19

>> you can lose it. It can be just gone.

4:23

And the

4:24

>> the cloud is just someone else's

4:25

computer. The craziest part of this that

4:28

kills me the most is it was just 50

4:32

terabytes.

4:35

>> You like they could back up 50 terabytes

4:38

these days with literally like four

4:42

drives in a consumer little QAP Naz or

4:46

something. Like actually though

4:49

>> it's actually impressing how solvable

4:52

that is.

4:53

>> You don't need an IT department. But but

4:55

you know what the funny thing is? I I

4:57

had um was it dinner? Yeah. Yeah. I had

5:00

dinner with Pancratz a couple of weeks

5:02

ago just cuz like I feel like he's

5:04

worked here like a year and I had never

5:06

actually like had a meal with him. Um

5:08

>> has it only been a year?

5:10

>> I know, right? He feels like he's been a

5:11

fixture. But anyway,

5:13

>> what

5:13

>> I had dinner with Pancrats and I was

5:15

talking about his past life and like

5:18

like you, he used to uh work providing

5:21

uh IT solutions for like small local

5:24

businesses. Um his was like through a

5:26

company whereas I think you were just

5:28

like freelance, right? Yeah. But um but

5:30

he was talking about how

5:33

how different that world is.

5:34

>> Oh, absolutely. compared to the

5:36

high-tech world. And you know, we he was

5:38

he was talking about how like there was

5:41

uh there was like an auto mechanic or

5:43

something that that he was doing work

5:44

for and the guy was just kind of kind of

5:47

shooting the breeze with him and they

5:49

were going through a bill and Buddy's

5:51

basically just like, "Yeah, so I guess I

5:53

just like have to pay this because I

5:56

don't understand any of it, which is

5:58

really uncomfortable for me because when

6:00

it comes to something mechanical, I can

6:02

look at every O-ring, every washer,

6:04

every nut, every bolt, and I know what

6:06

all of it is, and I understand the

6:08

necessity of it for my business. But

6:10

when I look at these IT services, I

6:13

literally don't know what any of this

6:15

is, and I don't understand if I'm being

6:17

charged a fair rate. So, hey, don't

6:21

me. Basically, right? And a guy like

6:23

Pancratz, I'm sure he never wrote an

6:26

unfair bill in his life, even if he

6:29

could be a

6:29

>> very standup dude.

6:30

>> But there's a but there's a lot of

6:31

opportunity there. And so, you know, you

6:34

and I look at this and we go, "Wow, look

6:36

how solvable that is." They could have

6:37

just bought four hard drives and,

6:41

you know, just set up a

6:42

>> one of the ways

6:44

solvable in my opinion is through things

6:47

like what you just described. But that

6:49

relies on them finding someone

6:51

trustworthy who isn't going to oversell

6:53

them something they don't need. So why

6:55

do people end up going with the cloud?

6:57

>> Because there's transparent pricing. It

7:00

looks credible. It's like a it's low

7:04

fees, like low monthly fees, which is a

7:07

lot of a lot of businesses. And you know

7:09

what? Let's face it, like public

7:10

broadcasters don't have unlimited money.

7:13

So even though they're, you know, not

7:15

small business in the sense that you

7:18

might think of as, you know, a mom and

7:19

pop candy store or something, uh, they

7:21

have a they have a finite budget. So

7:23

just buying infrastructure upfront all

7:26

at once is not necessarily feasible.

7:28

>> Yeah. The the the capex versus opex

7:30

[clears throat] thing, the capex is

7:31

like, oh, I have to buy Okay, with your

7:33

example, I have to buy this QAP thing

7:35

and I have to buy three right now very

7:39

expensive hard drives. Yeah,

7:41

>> that's still a decent amount of cash,

7:43

especially if you're, you know, as far

7:44

as my understanding goes, these like

7:46

smaller kind of local TV stations aren't

7:49

making a crazy amount of money right

7:50

now.

7:51

>> Yeah. Or I mean

7:52

>> or anything. [snorts] Yeah.

7:53

>> Um so it's it's it's a little bit of a

7:55

tough world. Uh but you can try to

7:58

communicate the the the potential

8:01

savings of that. We're going through

8:03

this right now where like I [snorts] was

8:05

actually

8:06

we

8:07

>> should just talk about it. Why don't we

8:08

just talk about it? We can a little bit.

8:10

Yeah, sure. We're becoming an ISP

8:12

apparently, which Luke has given me no

8:15

further clarity on. So,

8:16

>> no, this is an AJ project. Okay.

8:18

>> Uh, so it'll be it'll be most of the

8:20

clarity would come from talking to him

8:22

cuz we're still kind of mapping it out.

8:23

I'm still honestly understanding even

8:26

what's happening. Uh, but there there's

8:28

a lot of stuff going on. You you guys

8:30

are aware of you know RAM price

8:32

increases, everything hardware storage,

8:36

uh hard drive prices increasing,

8:38

>> the price of almost everything is going

8:40

up largely because almost every company

8:43

involved in the manufacturing of stuff

8:46

that you would use in your computer can

8:48

also be involved in the manufacturing

8:50

for things in a data center.

8:52

>> Yeah. And it's even affecting, we talked

8:54

about this last week on WAN show, it's

8:56

even affecting things like the timeline

8:58

for our battery bank product. Uh the the

9:01

IC's that we need, I think particularly

9:03

the buck.

9:05

>> Um

9:05

>> you can correct things I say wrong.

9:09

>> They are they're they're backordered by

9:11

something like 44 weeks or something

9:12

like that is the current lead time for

9:14

us to order the like buck boost

9:16

converters or whatever the there's a

9:18

particular component that has a super

9:20

long lead time on it. So, uh, yeah. So,

9:23

>> so anyways, there there's a lot of

9:24

different things going on. Like one, we

9:26

we got news earlier this week, uh, that

9:28

one of the services that we use wants to

9:30

increase our bill for the next year. And

9:32

this is the whole company, not a flow

9:35

plane thing at all.

9:36

>> Yeah. Okay. So, you're talking at a

9:37

scale that like hurts

9:39

>> 20% increase of bill.

9:42

>> Uh, there's lots of other things going

9:44

up right now because what's what's

9:46

basically happening is most of these

9:47

services

9:49

>> like you you feel the pain. your thing

9:52

you're interested in computers the price

9:54

of it is going up at a company level you

9:57

know laptops your your device that you

9:59

use at the company cost of that probably

10:02

going up um any servers infrastructure

10:06

etc things that they use at a company

10:08

level probably going up in cost

10:10

therefore their service cost is going up

10:12

as well so the amount that we pay as a

10:15

company for these external services is

10:17

going up the reason why I attach that to

10:19

this

10:20

>> is because a cloud provider is likely

10:23

going to have cost increases. OVH has

10:25

been talking about um cost increases.

10:29

Let's see here if I can bring this up

10:31

>> for context. OVH is a a pretty cool um

10:36

hosting option that has been very

10:38

affordable over the years and actually

10:40

pretty cool for us to work with the

10:42

>> good partner for us to work with

10:43

>> and

10:46

[clears throat] 87% price hikes.

10:49

>> Yeah. And to be clear, I don't even

10:51

necessarily blame OVH for this.

10:53

>> No, you can't see this happen

10:55

everywhere.

10:55

>> OVH didn't decide decades ago to have

11:00

Homer Simpson work at a nuclear power

11:02

plant, which apparently did contribute

11:05

significantly to people's perception of

11:07

the safety of nuclear power. Um,

11:10

>> because he couldn't screw it up

11:12

>> and cuz all No, cuz all the problems and

11:15

the way Mr. Burns runs it and stuff. I I

11:18

I have I have read that incredible

11:21

places which is like

11:25

>> wild.

11:27

Um and so that's part of why nuclear

11:30

power did not just because like it's

11:32

just kind of obvious.

11:34

>> That's why it didn't just go to the

11:36

moon.

11:37

>> That's nuts. Anyways, um anyway,

11:39

>> cost everywhere.

11:40

>> Sorry. Sorry, Arion.

11:43

>> It's pronounced nuclear.

11:45

>> Nuclear. Nice. Nuclear.

11:47

>> Uh, different places have different ways

11:48

of pronouncing things.

11:49

>> No, no, it's a joke from the Simpsons.

11:51

>> Okay, sounds good.

11:51

>> He corrects Lisa's pronunciation. Honey,

11:53

it's pronounced nuclear.

11:56

>> Anyway, carry on. Uh costs everywhere

11:59

are going up and costs at places like

12:01

OVH especially things are going up

12:03

because not only are their costs

12:07

skyrocketing but if you think about it

12:10

the people willing to rent from places

12:12

like that

12:15

in some cases are also going up in what

12:17

they will be willing to rent

12:18

>> because their unlimited money faucet

12:21

from investors has to be spent on

12:23

something because if you don't spend

12:25

enough of it you don't look like you're

12:27

trying hard enough to win the AI race

12:28

>> and you know places like OVH might not

12:30

be getting as much pressure from that as

12:32

some other potential places

12:33

>> but when the pressure is coming on them

12:35

from both ends when their costs are

12:37

going up and the amount on offer from

12:41

these customers who are responsible for

12:43

the cost going up

12:44

>> it can be it can even be indirect like

12:47

maybe there's another data center which

12:48

is like more like GPU focused or

12:51

something and the lower-end clients are

12:53

being pushed out of there

12:55

>> so they're now going to OVH so there's

12:56

more competition

12:57

>> or regional power bills are just going

12:59

up which just affects everybody.

13:01

>> So there there's a lot of different

13:03

things at play. I don't necessarily

13:04

blame OVH. OVH has been pretty great to

13:06

work with. That being said, oh my god,

13:08

87% potential increase in bill. U now

13:11

that doesn't mean all of OVH's services

13:14

are going up by 87%. And this was

13:16

largely a warning by them. We're

13:18

actually lacking a lot of information on

13:19

it to be clear. Um

13:23

one second.

13:26

What?

13:26

>> We didn't record.

13:29

>> Oh,

13:30

>> well then

13:31

>> everyone was on time.

13:32

>> Wan was not late because of our fault.

13:35

Uh getting No, you know what? Wan was

13:37

late because of Sammy. I think Sammy

13:39

loses a point.

13:40

>> Yeah. Point point against Sammy. Okay,

13:42

carry on.

13:42

>> Yeah. So, all these prices are going up.

13:45

OVH, don't blame OVH. So, we are looking

13:47

at potentially bizouncing. We're not

13:50

going to plan on bouncing for

13:52

everything. We're largely going to keep

13:53

our like omega large data store there

13:56

because that cost increase is not

13:59

presumably going to be nearly as high as

14:01

the other ones and the capex or buy a

14:04

bunch of hardware cost of setting up a

14:06

massive data store somewhere else very

14:09

high.

14:10

>> Mhm.

14:10

>> The capex of the other things that we

14:12

need to do relatively low.

14:14

>> Right.

14:14

>> So you want to do the thing that has the

14:16

lower upfront impact

14:18

um and the biggest potential benefit of

14:20

dodging the price increase. What does

14:22

that have to do with setting up an ISP?

14:25

>> So you're talking about moving some

14:27

equipment on prem. So maybe like

14:29

whatever you know AI server is going

14:32

through and doing our automated

14:33

subtitling like that would just move on

14:36

prem because it can just be it can turn

14:37

in the background. Is that kind of what

14:38

you're talking about?

14:39

>> Sure.

14:40

>> Transcoding potentially some of it.

14:42

>> Yeah. It's

14:43

>> and then we can just steer the jobs to

14:45

like use as much on prem as we can but

14:47

we keep OVH as a backup. But by onrem we

14:51

mean a collocation data center

14:53

>> that is like Vancouverish.

14:55

>> Got it.

14:56

>> So it will like latency wise it will

14:58

almost like very very very close feel

15:01

like to actually onrem. Uh because

15:03

there's fiber lines that go like

15:04

straight there

15:05

>> right.

15:05

>> Uh but it but it won't be it won't be

15:07

technically onrem. Although the term

15:10

onrem gets really confusing depending on

15:12

in what circles you use it in. Yeah.

15:14

Don't worry about it. Um

15:16

>> but yeah so we'd be putting stuff in

15:18

there. We would also be putting in these

15:19

like insane networking routing switching

15:23

boxes.

15:24

>> Yes.

15:25

>> Which blow my brain cuz they they have

15:28

ports that do 400 gig.

15:32

[clears throat]

15:33

[laughter]

15:34

>> They are sick. They're That's where most

15:36

the capex is, by the way. Right. Um is

15:39

getting those.

15:39

>> Who are these from?

15:41

>> Uh Edge Edge Core.

15:44

>> Okay.

15:45

>> So a sick name.

15:46

>> Yeah. Yeah.

15:48

>> Yeah. So I

15:49

>> So I never quite get to experience the

15:53

core. Just

15:54

>> just right on the edge of the core.

15:55

Yeah.

15:56

>> Yeah. Okay.

15:56

>> Yeah. Yeah. Yeah.

15:57

>> Um and then something something

15:59

something. We bought our own IPs. Uh so

16:03

we bought a range of IPv4. We did

16:06

>> IP. When was this?

16:07

>> A little bit ago we did that. Uh, we

16:10

almost bought them from Canada Post,

16:12

which was going to be really funny, and

16:14

then AJ found a cheaper uh, set, so we

16:16

bought those, but we almost bought them

16:18

from Canada Post.

16:18

>> I love that IPv4

16:20

has been and continues to be one of like

16:23

the soundest investments. This is not

16:25

financial advice.

16:26

>> That's one of the reasons why we bought

16:27

it. Uh, it was because we were like we

16:29

were like, we're pretty sure we're going

16:30

to do this thing, but then AJ AJ found

16:33

this like bank of IPs that we would we

16:35

could buy, and he was like, uh, these

16:37

are up for auction. we should like

16:38

probably buy these. And the reason why

16:40

I'm saying this now is cuz this is a

16:42

decent price and even if we don't do the

16:44

plan, you can probably just resell them

16:46

and make money. So, I think you should

16:48

buy these. I was like, okay.

16:49

>> No asks, wasn't there an IPv4 shortage

16:51

at one point?

16:52

>> Yeah,

16:52

>> there's an IPv4 shortage that is ongoing

16:55

and will never end as far as I can tell

16:57

because it looks like we're just

16:59

>> never going to switch to IPv6. I And to

17:02

be clear, that is not a prediction.

17:04

Hopefully, theoretically at some point,

17:06

IPv4 addresses and they're they're

17:09

they're limited quantities we have of

17:11

them won't matter, but for now it does.

17:13

>> Yeah, AJ included this and it's just as

17:15

like a footnote because who cares, but

17:17

we got some free IPv6, too.

17:20

>> Sounds good. What we cared about was

17:23

buying the the actual IPv4. But now that

17:25

we have that,

17:28

we could provide

17:32

internet for ourselves

17:34

from the colo center.

17:36

>> Holy [laughter]

17:39

>> Which would mean that you could have

17:43

Lionus ISP tips as your internet service

17:47

provider at the office.

17:49

>> Linus Lisp tips. [laughter]

17:54

>> [gasps]

17:55

>> entirely possible. You could be your own

17:57

ISP with one customer, which is

18:00

>> my own ISP.

18:02

>> I mean, tech technically there's more

18:03

than one customer because the the ISP

18:06

>> would have the customer of like float

18:08

plane and also the the offices,

18:13

>> right?

18:13

>> So, float plane's a reseller now. So

18:16

okay

18:16

>> kind of because float plane services

18:19

sauce plus and float plane yeah and does

18:22

some other little things.

18:23

>> So Black Raven asks what stops an

18:26

individual home user from being their

18:27

own ISP at an individual home level. Uh

18:31

the economics don't make any sense but

18:33

there are towns that have had really

18:36

poor service from the major ISPs and

18:39

have just pulled together their tax

18:41

revenues and made their own ISP. that

18:44

that is a thing because if you make a

18:48

connection to an exchange

18:52

then if you own the network addressing

18:55

that you need

18:56

>> and you can pay to go over other

18:58

people's lines which is how a lot of

19:00

small ISPs work

19:02

>> the internet is just a series of tubes

19:05

so if you have the equipment and you

19:07

have the lines or you lease access to

19:09

them

19:11

you can you can join the the network

19:14

that we call the internet. Um,

19:18

are we sure this makes sense?

19:20

>> What we're definitely sure of

19:22

>> is it saves a lot of money.

19:24

>> Okay.

19:25

>> So much money.

19:27

>> Does it really save so much money?

19:29

>> So much money. Like 400 grand over 5

19:34

years.

19:35

>> Oh yeah. That's everything combined.

19:38

That's having the servers in the

19:39

collocation center. That's being our own

19:41

ISP. That's basically AJ's been sitting

19:44

the back cooking up a crazy witch's brew

19:47

and it's going to save you 400 grand

19:48

over five years.

19:50

>> While we're at it, I mean, I've been

19:51

thinking about it some more and I feel

19:53

like we should just do the VPN product,

19:55

too.

19:57

>> I think we drew a very arbitrary line

19:59

before.

20:01

>> I think we should just do it. We'll

20:03

become an ISP. We'll become a VPN

20:05

provider. It'll be great.

20:06

>> We're both of them.

20:08

Shoot, we're going to know what your

20:10

traffic is doing no matter what. Nice.

20:12

[laughter] You put on our network and

20:14

you VPN it, we're going to know. Um I I

20:18

love it. Um

20:20

yeah, I don't know. So, we're we're

20:21

we're moving, we're shaking, we're

20:23

trying to make things happen. Uh

20:26

okay. Right. So, because we have a rack

20:30

at the exchange.

20:32

[snorts] So, how does our how does our

20:35

access from here to the exchange work,

20:37

though? What do you mean? Sorry.

20:39

>> Like that fiber line?

20:40

>> Yeah. So, we will need to lease that

20:43

fiber line. But that's the same way.

20:44

That's what I was kind of getting into

20:45

is like there's tons of like small like

20:48

super regional ISPs.

20:49

>> Yeah. Which I know because we use one.

20:51

>> Oh yeah.

20:52

>> So then what's their role in any of

20:54

this?

20:55

>> The the super regional one is nothing.

20:57

>> Is nothing. No. So we just whatever

20:59

>> they are renting lines from someone

21:00

else. So we will just skip that and rent

21:03

lines for the people that they're

21:04

renting lines from.

21:05

>> Got it.

21:05

>> Yeah.

21:06

>> Okay. Yeah.

21:08

>> And then would we still have

21:09

>> which is another dodge in costs,

21:11

>> right?

21:12

>> So it kind of layers. It's like an onion

21:14

of cost.

21:15

>> It's the Shrek of the internet.

21:17

[laughter]

21:18

>> Shreknet.

21:20

>> You just keep saving.

21:21

>> Can we can do we need to spin up a new

21:22

company for this? Because if we do, I

21:24

want it to be called Shreket. [laughter]

21:30

Um, and AJ's also pointing out we you

21:32

could technically use this as like a

21:34

weird form of redundancy as well because

21:36

we have the crazy switch at the colo.

21:38

Yeah.

21:38

>> If like Telus went down,

21:41

>> as far as my understanding goes, and I'm

21:42

still learning a lot about this, so AJ

21:43

can correct me, but if Telus went down,

21:46

>> if we had our contracts arranged in the

21:48

right way, which this might be stupid,

21:49

but I think you could if Telus went

21:51

down, I think you could literally just

21:52

be like, "Well, I guess we're on Rogers

21:54

now." and just not lose internet.

21:58

I think

21:59

>> as long as it's not somebody putting a

22:00

shovel through like our fiber.

22:03

>> That'd be a little rough. Yeah.

22:04

>> Got it. Okay. And then

22:05

>> I think that works. AJ can can correct

22:07

me.

22:07

>> While we're at it, float plane chat

22:09

asks, "Okay, why not cellular service

22:12

then?" Um, there's a dead spot right by

22:15

my house that causes me no amount of

22:19

frustration because I cannot open and

22:21

close my garage doors with my with my

22:23

phone because it uses your Wi-Fi.

22:25

>> Yeah, Adrian said yes. for CarPlay and

22:27

Android Auto [snorts] and my cellular

22:30

service doesn't work right next to my

22:32

house. So, it's really annoying.

22:35

>> Oh, sorry.

22:36

>> So, if I could just put a tower there,

22:39

that would be swell. I would be super

22:42

down.

22:43

>> Yeah, I have no idea about that at all.

22:45

>> Okay,

22:45

>> that's maybe a later later thing to look

22:47

into. Something else he was describing

22:49

to me is because the colo center, we can

22:51

make like a LAN ring.

22:55

So, like if our network from the studio

22:58

to the lab went down for some reason.

23:00

Yeah, that's okay because the studio can

23:02

talk to the lab through the colo center.

23:06

>> Okay. [laughter]

23:08

>> Which is like sick.

23:10

>> Okay.

23:11

>> So, there's like there's a lot of really

23:13

kind of

23:15

there's been a lot going on, so I

23:17

haven't had like a crazy amount of time

23:18

to talk to AJ about it, but every time I

23:19

talk to AJ about it, I'm just like,

23:21

"What?" [laughter]

23:23

It sounds very fun. Um,

23:25

>> all right. Okay, cool. Um, yes, I still

23:30

have the cell repeaters in my house, but

23:32

I I ran out of of interior antenna slots

23:37

on my repeater, and all of those are

23:39

used inside the house, so they're not

23:42

useful in front of the garage, which is

23:44

where I have my problem right now. Very,

23:46

very annoying. And I even just switched

23:48

ISPs to make it or not ISPs, um, mobile

23:52

providers so that I can do conference

23:54

calls on the way to and from work

23:56

because the one that I was on before had

23:58

a couple of dead zones that were too

23:59

dead. And then the new one, I'm able to

24:02

make calls to and from the office, but

24:04

still cannot open my garage door. So

24:07

part of the problem is the noise

24:08

blocking glass in my car. It turns out

24:10

it's also cellular signal blocking

24:12

glass. So sometimes I'll I'll be that

24:15

guy where I'm like, I really need to

24:17

close my garage door right now. Uh

24:19

>> Oh, yeah.

24:20

>> Oh, can you hear me now? Can you hear me

24:21

now? Boop. Okay. Yeah. Put

24:23

>> Yeah.

24:23

>> Ridiculous. Add a stop sign.

24:25

>> So AJ added the the unfortunate part in

24:28

full plane chat.

24:29

>> Oh,

24:30

>> the two switches will total

24:34

not not each about 65 grand.

24:46

which will equate like all the costs in

24:50

is roughly

24:53

what we would have been paying if we

24:54

didn't do anything over the course of 12

24:57

months.

24:58

>> Okay.

24:58

>> And then you'll be saving money after

25:00

that.

25:00

>> Uhhuh.

25:03

[sighs and gasps]

25:07

>> And we need two of these switches. Okay.

25:09

Is the second one just for redundancy?

25:11

>> Yeah, but it's total.

25:12

>> Can I just have one?

25:13

>> No.

25:14

>> And not have redundancy?

25:15

>> No.

25:15

>> Why?

25:16

>> Because if the whole float plane service

25:19

goes down, it's your fault. [sighs]

25:26

>> But it's not it's not 65k twice.

25:29

>> No, I got that.

25:29

>> Okay. All right.

25:30

>> 65k is okay. Is it at least Canadian

25:32

dollars or is it like real money?

25:36

>> You don't know. You're just waiting for

25:38

>> I don't remember. I think it's CAD. I

25:39

think it's CAD.

25:40

>> Okay. Well, that would help. I don't

25:41

remember.

25:41

>> That would help a lot.

25:42

>> I don't remember.

25:43

>> It's in the dock. It's all in the dock.

25:46

>> Yeah, everything's in the dock.

25:47

>> Yep.

25:48

>> AJ's totally going to reply.

25:50

>> He'll say,

25:51

>> "Man, oh man, I just

25:53

>> a break even of 12 months sounds pretty

25:55

good." It is pretty good. It doesn't

25:56

change the fact that it's still 65K.

25:58

>> And [laughter] this is But this is why.

26:00

>> Yeah, it's cat.

26:01

>> This is Okay, that helps a lot. So, it's

26:03

like 40,000 real dollars. like 45. Um,

26:06

but th this is why businesses have a

26:08

hard time

26:09

>> all the way back

26:10

>> making these kinds of decisions because

26:13

paying a few grand every month is

26:16

something that you can plan for. You

26:18

could you just it smooths out your your

26:22

your your expenses and your income

26:25

because I don't I don't just suddenly

26:27

have necessarily

26:30

$65,000 of excess income right now. I I

26:33

it's hard to plan around that. I mean,

26:35

think about it for like your personal

26:36

life as well. Why do so many people

26:38

finance a car? You're paying way more.

26:41

>> Why do you buy the okay thing that's

26:43

going to break down pretty soon instead

26:45

of the the really expensive thing that

26:47

might last you your whole life? Cuz it's

26:49

really brutal to buy the really

26:50

expensive thing.

26:52

>> And so I just

26:54

>> super normal. [sighs]

26:55

>> I mean, realistically, we're probably

26:56

going to do it. It'll make a great

26:58

episode of Holy on the channel.

27:00

>> Dude, the switches are nuts. Okay,

27:03

[laughter]

27:04

AJ linked me the like product page for

27:06

them. My brain exploded. I was crazy.

27:09

[laughter]

27:10

They look so wild. I've never even

27:13

considered something like like I knew

27:17

that type of infrastructure would exist

27:19

somewhere, you know,

27:21

>> but I've never really considered like

27:23

what it would be. Tell me this like at

27:26

at what point like would I would it make

27:29

sense if I wanted like super low latency

27:33

like the best quality internet at my

27:35

house would I should I should I lease

27:39

the line to there like like where where

27:42

does it end? We could look into how much

27:44

that would cost.

27:45

>> Like what would it cost to be our own

27:46

ISP for like everyone who works here?

27:48

>> As far as my Well, okay. Yeah.

27:51

>> Like if you're

27:52

>> That gets really messy.

27:54

>> What I'm asking is is what what how does

27:56

the lease work? Is it by the like the

27:59

line where it goes? I wouldn't think so.

28:01

It probably wouldn't even be by

28:02

wavelength either because a lot of that

28:04

would be shared. So is it just by total

28:05

network utilization? So as long as we

28:08

told if we told the staff here like as

28:11

long as you have you know Telus fiber to

28:14

your neighborhood then we can offer you

28:17

lisp for

28:22

[panting]

28:23

[laughter and snorts]

28:26

uh for a low for a little low low price.

28:30

>> I knew it was going to end up being

28:32

Lion's internet service provider

28:33

somehow. I knew it. It's It's in my like

28:35

Jo We have a all hands on Monday. It's

28:37

in my like jokey notes for the all hands

28:39

is Linus [laughter] internet service.

28:42

>> Oh man.

28:42

>> Like could it be a better price than

28:45

buying from Telus or whatever?

28:47

>> Okay. Apparently no no apparently AJ

28:50

just said we can actually gain access to

28:52

Telus last mile gon service which I

28:54

think is AJ Ford. Yes.

28:59

>> I don't AJ that's so complicated. I

29:02

[laughter] don't want to be an internal

29:04

ISP for all the employees. That's going

29:06

to be so much work. But it wouldn't make

29:08

sense for us since we're power users.

29:11

>> What does that mean?

29:12

>> That doesn't mean anything.

29:13

>> That means nothing, AJ. And I think that

29:15

was intentional.

29:16

>> Maybe [laughter]

29:17

maybe

29:19

I think we need to not

29:22

do this.

29:23

>> We might use the availability that we

29:25

have.

29:25

>> Yeah. And we're better off we're better

29:28

off sharing quota with grandma who

29:31

>> sends like three emails. Yeah. Okay,

29:34

>> that actually makes sense.

29:36

>> Sorry, AJ. That was

29:37

>> That makes sense.

29:38

>> Reasonable. I don't think we should be

29:41

>> an ISP for all of LMG employees. I don't

29:44

think that's a good idea.

29:45

>> Do it. Do it for Do it for fun.

29:47

>> It'd be pretty badass though.

29:49

>> Oh my god.

29:50

>> Be so cool.

29:50

>> Okay, buy a bunch more fat switches.

29:53

>> Oh no.

29:54

>> Yeah, there you go. No,

29:56

>> no, [laughter]

29:58

>> no.

30:02

>> Got him.

30:03

>> So, how many I how many IPv4 addresses

30:06

do we own then?

30:07

>> 256, I think.

30:08

>> Wow, that's not bad.

30:09

>> Yeah.

30:09

>> Okay.

30:10

>> Pretty sure. [laughter]

30:14

>> What do we pay for those?

30:15

>> Seven grand.

30:16

>> Oh, okay.

30:17

>> Again, we could resell.

30:18

>> Yeah. Yeah. Yeah. Yeah.

30:19

>> That's like actually an asset that I was

30:20

like fairly unconcerned about about

30:23

acquiring.

30:24

>> That's crazy. The switches,

30:26

>> I don't know how many people are going

30:27

to be looking for them on Facebook

30:29

Marketplace. Um, so those might just be

30:33

>> they're gone

30:34

>> a cost. Yeah.

30:34

>> Yeah.

30:35

>> Yeah.

30:35

>> And now we have those switches forever.

30:38

>> Yeah.

30:39

>> Neat.

30:39

>> Yeah.

30:40

>> I mean, a lot of networking equipment is

30:41

like that. No life outside of the data

30:43

center. Like some home labber might

30:46

tinker around with them here and there,

30:47

but the the the size of the market for

30:50

these things, the secondary market is

30:52

like negligible.

30:56

having a 400 gig port, dude.

30:59

>> And how many does it have?

31:00

>> My god, several. As far as my

31:02

understanding goes.

31:02

>> No, don't look them up now. We'll save

31:04

it for

31:05

>> the total switching capacity has to be

31:07

obscene.

31:08

>> It's it's uh Yeah, it's uh it's kind of

31:12

mind-blowing. Um

31:13

>> Okay.

31:14

>> And we're not we're not getting like

31:16

bougie ones. We're also not getting like

31:19

ratttastic ones either. Okay. Um,

31:22

>> AJ, send me the specs. I want to look at

31:24

them when I go home.

31:27

[laughter]

31:28

>> That would have its own potential

31:30

problems, you know, but we we got like

31:32

fairly scrappy ones, but but ones that

31:34

seem good.

31:35

>> Okay.

31:35

>> Um,

31:36

>> yeah. I mean, I'm potentially open to it

31:38

as long as we can make content out of

31:39

it.

31:40

>> The problem is if we don't do it.

31:42

>> Yeah.

31:43

>> The costs

31:45

again, doing this saves 400 grand over 5

31:48

years. not doing this costs 400 grand

31:51

over five years, right? So, it's like

31:53

65k now,

31:54

>> right? Penny saved, penny earned.

31:56

>> Yeah.

31:57

>> So, I don't know. I mean, you it's it's

31:59

technically in your guys' hands. I've

32:00

also talked to Taran about it, but I

32:01

know that an expense this high is going

32:03

to have to go up. So, I'm going both

32:06

angles right now cuz if you guys do

32:08

decide to say no,

32:10

>> I would just question

32:12

what the strategy is. Uh because the the

32:17

okay the 400k savings over 5 years as

32:18

far as my understanding goes also is

32:20

just from the OVH payments.

32:22

>> Well, that also just assumes that

32:23

everything's going to stay flat for 5

32:24

years.

32:25

>> That too which it like very likely will

32:27

go up more.

32:28

>> Yeah, I know.

32:29

>> So it's like

32:31

I I understand it's tough.

32:33

>> Yeah, it is.

32:35

>> It's a lot of money.

32:36

>> It's very tough. It's maybe a lot of

32:38

money when

32:39

>> it's a step above tough.

32:41

>> It's a lot of money when things are when

32:42

things are rough, too. But think about

32:44

it, man. We can potentially get

32:45

enterprise

32:47

>> hardware and sponsors and stuff on the

32:49

videos. We can make videos about it.

32:51

>> Yeah.

32:51

>> Do you want to make videos about being

32:52

your own ISP?

32:53

>> Yeah, I do.

32:54

>> That's kind of fun.

32:55

>> That's pretty cool.

32:55

>> A lot of this project sounds really

32:57

freaking sick.

32:58

>> Can we shoot at the Colo?

32:59

>> I don't 100% know yet.

33:01

>> Okay. Cuz that would be amazing.

33:02

>> Being able to shoot at the Colo would be

33:03

sick.

33:04

>> As far as my understanding goes, we

33:06

reached out. I don't think they just

33:07

like said no.

33:08

>> Okay.

33:09

>> I think there was they were like,

33:11

>> "What?" Um, but it wasn't an immediate

33:14

no. There's some issues because of other

33:16

people's stuff.

33:17

>> So, what I was going to say, I know it's

33:18

a thing where you can maybe it's for a

33:20

fee or something, but you can have like

33:22

blanked out covers. So, maybe we just

33:25

have covers on ours so we can do

33:27

whatever we want inside ours. Maybe we

33:29

could suggest that

33:30

>> so that we can't see anybody else's cage

33:32

from inside our cage.

33:33

>> Yeah, something like that. Maybe like I

33:35

I'm going to try to play ball with them

33:37

because I would like to

33:39

>> AJ AJ says they said most likely.

33:41

>> Nice. Okay. I didn't hear that update,

33:42

but that's great.

33:43

>> Yeah, that's that's terrific because

33:44

that makes a huge difference to us

33:46

because all of a sudden like

33:47

>> that's been crazy. I'm behind on emails.

33:49

>> We could we could go to someone like uh

33:51

you know, I don't want to I don't want

33:52

to put any words in anyone's mouth, but

33:54

we could we could go to companies in

33:56

this space and they have such an

33:58

appetite for oh, you're doing something

34:00

that's like actually kind of sexy and

34:02

cool with like boring business iron.

34:07

Yeah, let's do it. Um they tend to be

34:10

pretty enthusiastic about it. So that

34:11

would be really exciting.

34:12

>> Yeah, it's just there's a lot there's a

34:14

lot of cool um that comes from it.

34:18

There's a lot of like and to be

34:19

transparent, one of the reasons why the

34:22

capex is actually so low because that

34:24

capex for those switches is like almost

34:27

all of it.

34:28

>> Okay,

34:28

>> there's very little outside of that.

34:30

Like genuinely very little.

34:31

>> All right, that's good.

34:32

>> Um and the reason for that is cuz we

34:34

just have a lot of stuff,

34:36

>> right? Oh, that's what Thomas was doing.

34:42

>> Probably.

34:42

>> I saw him dumpster diving and just like

34:45

he had just like every server from the

34:48

warehouse and he was like going through

34:50

them.

34:51

>> We're trying to make stuff happen,

34:52

Brett.

34:52

>> Ah, [groaning]

34:55

I like it. I like it.

34:58

>> The inside of the rack is already

34:59

isolated because of physical access,

35:01

right? Okay. So, that makes sense. So,

35:02

you have to be able to go there.

35:04

>> Yeah. And because it's a colo center, so

35:06

you have to be able to go there to

35:07

service your own stuff,

35:08

>> but not all of them necessarily have

35:10

like complete like blanket coverage.

35:12

>> Yeah. So, we just got to we just got to

35:14

make sure. We just got to make sure

35:16

>> it it Yeah, it might even I I'm just I'm

35:19

not Yeah, I'm not sure what the

35:20

situation necessarily is, but it's it's

35:22

the whole thing is very interesting. Um,

35:25

so yeah, again, a lot of the big bulk

35:27

storage would still be over in OVH in

35:29

Quebec cuz that pricing is expected to

35:31

not go up as much. And buying 10

35:34

trillion hard drives right now sounds

35:36

terrible. Uh,

35:37

>> it really does.

35:38

>> So, we'll just keep doing that. That

35:39

also kind of helps because my 321 backup

35:42

brain goes, "Wait a second.

35:45

>> Earthquake."

35:46

>> Yeah. So,

35:48

>> we live on a major fault line and

35:50

>> that is way overdue.

35:51

>> My entire life. Yeah,

35:54

>> I've been told the big one's coming.

35:56

[laughter]

35:57

>> Not here yet.

35:58

>> Any minute now. The big one's coming.

36:00

I'm like, all right.

36:01

>> And as far as our understanding goes,

36:02

the big one here, if it does happen,

36:05

will be

36:06

>> the longer it goes, the more

36:10

pressure there is, apparently. I don't

36:12

know. I'm not a geologist.

36:14

>> Yeah, I don't know either.

36:15

>> I don't know how fault lines work,

36:16

>> but apparently this one's going to be

36:17

very brutal.

36:19

>> Yeah, there you go. Nice. Both of us.

36:21

Um, so yeah, the data would still be

36:23

kept over there. Yeah.

36:24

>> Um, the compute stuff, Jonathan's done

36:26

some just insanely attractive things

36:30

with Kubernetes. So like if it did all

36:32

blow up, you could spin it up somewhere

36:34

else really quickly. Um, so we're not

36:36

super worried about that part. Um, so

36:40

>> all right. Yeah,

36:40

>> good luck everybody.

36:41

>> Yeah, cuz I know there was there was a

36:43

question of like we moved uh we moved

36:46

some things I'm going to be slightly

36:48

vague. I think you know about this onto

36:50

float plane infrastructure recently

36:52

which saved like 17 grand a year and

36:54

then there was some questions that came

36:55

up with when I've been kind of talking

36:57

around about this project internally

36:59

where people were like well what

37:00

>> what's happening with that cuz you like

37:02

just talked about that it's like well

37:03

that's still happening and literally

37:06

where it was still happening like we're

37:07

not even moving that part.

37:08

>> Yeah

37:09

>> we're moving a lot of the compute and

37:10

everything. Um,

37:12

>> do we need to look at like uh what is

37:14

it? Next cloud.

37:16

>> Like we've had all these conversations

37:18

about like teams and

37:20

>> Oh, I was like, how would that help us

37:22

with this? [laughter] Okay, other

37:24

things.

37:24

>> I'm just talking self-hosting in

37:26

general.

37:27

>> Uh,

37:27

>> is it is it time?

37:31

[snorts]

37:31

>> It's time to consider some of it. Um, I

37:34

know like you and I have talked about

37:37

Okay, so we have we have Slack for some

37:40

of the labs people and all of the flow

37:41

plane people. We have Google Workspace

37:44

for everyone.

37:44

>> Yeah.

37:45

>> And we have

37:46

>> Teams for

37:47

>> Office 365 for like everyone minus

37:51

maybe a couple.

37:53

>> Yeah,

37:53

>> it's basically everybody.

37:55

>> Um,

37:57

do we do we consolidate these services?

38:01

It's gonna

38:02

>> or do we get rid of all of them?

38:03

>> Feel like ripping my own heart out, but

38:05

we're almost certainly going to get rid

38:06

of Slack.

38:07

>> So, that's that's one that's just going

38:09

to go. It's way too expensive per user.

38:11

I love it compared to Teams. Teams is a

38:13

rotten pile of poo. U I was in a meeting

38:16

earlier today and I really needed to

38:17

look something up that was in Teams

38:19

messages and they just wouldn't load.

38:20

It's like cool, nice. Um Discord would

38:23

have loaded just fine. I wish Discord

38:24

had a corporate edition, but they don't.

38:26

Um,

38:28

I know one of the opponents to leaving

38:32

Office 365 because I've brought up like

38:34

how painful would it be to lose Google

38:36

Drive and you amongst many others

38:38

including myself have been like that's

38:40

not really an option.

38:41

>> Um,

38:43

so the easier one for me would be

38:46

Microsoft. One of the opponents to

38:48

losing Office 365 is infra because

38:54

the like authentication stuff,

38:56

>> yeah,

38:57

>> is

38:57

>> single sign on, baby.

38:58

>> Is more ubiquitous. It's not necessarily

39:01

better. I think it's worse, but it's

39:02

more ubiquitous on the Microsoft side.

39:04

>> Yeah.

39:05

>> So, it's like

39:07

>> Jonathan did pipe up in here. I don't

39:09

know if you noticed him in float plane

39:10

chat. He says, "We have Nexcloud with

39:12

high performance backend running in

39:13

Kubernetes with custom Helmchart." Yeah,

39:15

that was a project we did internally to

39:17

try to validate if we can use it.

39:18

>> Oh, okay. And

39:20

>> not for everything.

39:22

>> Maybe for some stuff though.

39:23

>> I know they're moving pretty quick and

39:25

they are like

39:25

>> they do seem to be moving pretty quick.

39:27

>> They are,

39:29

no offense you guys, but kind of overly

39:32

attached girlfriend about working with

39:33

us, too. [laughter] Like they email me

39:35

probably once every quarter or so going,

39:38

"Hey."

39:39

>> The second you mentioned them on the

39:40

show, I was like, "There's an email.

39:42

[laughter]

39:45

It's like, "All right, incoming." Uh, we

39:47

have not forgotten about you. I'll say

39:49

that much.

39:50

>> Yeah, we're we're aware and they do seem

39:53

to be moving pretty fast.

39:54

>> I think they're moving fast. And I get

39:56

the feeling that if we were to move on

40:00

to them,

40:03

it would be a benefit to them having an

40:08

organization like us give feedback to

40:11

help guide bug fixes and and short-term,

40:16

you know, fires that need to be put out.

40:19

I suspect they would be very motivated

40:20

to do that, I guess, is where I'm I

40:22

think there'd be a bit of a give and

40:23

take there. Yeah, there's there's so not

40:26

to be mean xcloud I think it's a really

40:28

fantastic project. Uh VDKA in full chat

40:31

said too bad nextcloud breaks their own

40:32

stuff every other update. My notes app

40:34

has been dead for weeks because of a

40:36

right lock bug they have not fixed

40:41

Google Drive is going to work.

40:43

>> Yeah,

40:44

>> like pretty much no matter what. And

40:48

like I I I trash on cloud stuff all the

40:50

time. This topic is technically us

40:52

trashing on cloud stuff because of the

40:55

uh wherever it went.

40:58

>> Okay,

40:59

>> station losing their data.

41:00

>> Fun fact. Okay, this is me trying to

41:04

open a Google doc earlier today.

41:07

>> Can't open this document. Sorry. Google

41:09

Docs is the reli the reliability of

41:12

Google services

41:14

>> has been moving the wrong direction for

41:16

me for 10 years.

41:18

>> Yeah. No, I agree. And do you know how I

41:20

fixed this?

41:21

>> Search is also useless now. Yeah.

41:22

>> Want to know how I fixed it? Watch this.

41:24

It's the stupidest fix ever.

41:27

>> I already see where it's going.

41:28

>> Make a copy.

41:29

>> Yep.

41:30

>> I made a copy

41:32

>> and then the copy's readable

41:33

>> and immediately it opened.

41:35

>> Yep.

41:36

>> Unfucking believable.

41:38

>> I also have issues where I can't

41:40

freaking find anything in Google Drive.

41:42

Um like very often it's an issue. I

41:44

think the search is really bad now.

41:46

search for Google in general has just

41:48

gone down the hole, which is wild. It's

41:50

weird to even say that. But search on

41:53

everything for them. Man, the fact that

41:54

I've been finding it easier to find

41:56

certain YouTube videos by googling it

41:57

than searching it on YouTube is

41:59

mind-blowing.

41:59

>> Ivonne did that the other day. I was

42:01

like, why aren't you searching on

42:02

YouTube? And she's like, this just works

42:03

better and she found it immediately.

42:05

>> Does Yeah, it like actually does.

42:07

>> And that's with Google search degrading

42:08

the way that it has.

42:09

>> Yeah, which is just wild. The whole

42:12

thing is just very confusing. Um,

42:15

so I don't know. It's it's an

42:17

evaluation. We've been honestly a little

42:19

bit brains sniped by this collocation

42:21

thing. Um, because there are renewal

42:25

dates upcoming.

42:27

>> Oh,

42:27

>> that are like very important.

42:31

So, we're trying to move fast,

42:33

>> right? Is one of them related to the

42:35

email chain that AJ and I were in? I

42:37

see.

42:38

>> But there's like a bunch of them.

42:40

>> I understand.

42:40

>> Yeah. Um, so like

42:42

>> and getting it all aligned so that

42:45

you're not paying for your expensive

42:47

off-rem stuff at the same time as your

42:50

capex on your onrem or like colloccated

42:53

stuff

42:54

>> and there's like shipping times,

42:56

>> right,

42:56

>> and things that we need to be taking.

42:58

>> And availability of anything with an IC

43:00

in it these days is not to be taken for

43:02

granted.

43:03

>> Uh-huh. So it's tough. It's uh

43:07

>> I think that's a habit you picked up

43:09

from me. what they so it's tough.

43:12

>> I say that all the time. Every once in a

43:14

while [laughter] you hang out with

43:16

someone for long enough that you you

43:19

just you pick up mannerisms from each

43:21

other. I I have lucisms too that drive

43:23

me nuts because [laughter]

43:24

I'll say them and I'll be like I can

43:28

hear I can almost hear it in your voice.

43:30

>> You you combing some things with bricks

43:31

or what?

43:34

>> No, the ones that make sense. Um,

43:36

[laughter]

43:37

man, I can't even think of it off the

43:39

top of my head, but they'll just pop out

43:41

of my mouth and I'm just like, "Oh, I

43:43

picked that up from that freaking guy."

43:45

Drive me crazy.

43:46

>> I wonder how many people like actually

43:48

even know actually where that reference

43:51

is from anymore.

43:52

>> Comb with a brick. It was a W show a

43:54

thousand years ago.

43:54

>> It was a W show in the garage.

43:55

>> Yeah, something like that.

43:56

>> Yeah. [snorts]

43:58

Um,

44:00

yeah, it's tough. Oh, yeah. There it is

44:03

again. Um, now I'm not going to be able

44:05

to say [laughter] that without thinking

44:07

about that.

44:09

>> Uh, but it's this and this is happening

44:12

to be clear. Like we've got some really

44:13

exciting things going on. It's

44:15

terrifying for short-term checkbooks,

44:17

really exciting for long-term

44:19

checkbooks. This would save us money

44:21

even if the price increases weren't

44:22

coming.

44:22

>> Noki's got it.

44:23

>> Exciting.

44:24

>> I say that's wild. That's all 100%

44:28

Lucasm that I picked up. [laughter]

44:30

>> Yeah. Thank you, Noki. Yeah, you nailed

44:32

it. cuz I never said that before in my

44:34

life.

44:35

>> That's interesting.

44:35

>> And I picked it up from him 100%.

44:37

[laughter]

44:38

>> Uh but yeah, this is this is happening

44:40

to everyone. Um don't get it twisted. We

44:42

are not the only ones going through this

44:44

problem.

44:44

>> No. And and that's part of the problem

44:46

is that it's happening to everybody

44:47

right now.

44:48

>> Yeah.

44:48

>> This is not isolated. It's like

44:50

industrywide. This is macro forces that

44:53

are just making their way in unexpected

44:57

ways um into the daily lives of just

45:00

like randos running YouTube channels.

45:02

>> Speaking of unexpected ways, I I think

45:04

I've talked to you about this before,

45:05

but like I met the guy a while back. I

45:08

don't know if I don't think he still

45:09

does this anymore, but I met the guy a

45:10

while back who runs like he was the CTO

45:13

of I think it was Kroger International.

45:15

>> Yeah. All right. Turns out they have

45:18

like massive amounts of like servers and

45:21

stuff.

45:21

>> Kroger like the paper company.

45:23

>> No, the grocery chain.

45:24

>> Oh, that one. Sorry. Sorry. Which one's

45:25

Oh, is there

45:26

>> But this Sorry.

45:27

>> I was thinking about this earlier today

45:29

where like I'm looking at this and

45:30

talking about how it's all cyclical or

45:32

not cyclical. It's all uh circular. Oh,

45:35

Krueger, correct?

45:35

>> Krueger is a paper company. No, no,

45:36

you're right.

45:37

>> Oh, okay. Okay. Um

45:40

where like you know [gasps]

45:41

all these other things are making other

45:43

costs go up which make other costs go up

45:45

and all cascades. That's the word I'm

45:46

looking for. Um,

45:48

you're like bananas

45:51

are going to cost more because of AI

45:57

cuz companies have to recoup their

45:59

costs.

45:59

>> That's bananas.

46:01

>> I'm going to go with that now instead of

46:02

that's wild. That'll be my own

46:03

[laughter] thing.

46:05

>> If if they have crazy costs going up

46:07

because of

46:09

uh, you know, hardware increases. If

46:12

they have costs going up because, hey,

46:13

their password manager just is going up

46:16

20%. If they have costs going up because

46:18

of etc, etc., etc. that are cascading

46:20

all over the place. If their machines

46:23

for automated checkout need to go up in

46:26

price, if they're etc., it's all going

46:28

to cascade down onto costs, which are

46:29

going to cascade on down onto you, which

46:31

is going to make your bananas or

46:32

whatever else you're buying potentially

46:34

cost more. So, it's it's impacting like

46:37

everything. And obviously different

46:39

companies are going to get impacted to

46:40

different degrees. For some companies

46:43

it's going to make effectively no

46:44

difference. For other companies it's

46:45

going to make a very large difference.

46:46

For flow plane for instance this is like

46:49

oh god [laughter]

46:52

really bad terrifying solve the problem

46:54

now uh type situation for other

46:57

companies would be less impactful.

46:59

>> Um so

47:01

noi found it using filot.

47:04

You ready Dan?

47:05

>> No.

47:07

>> Nope. Wait, the wall was white.

47:10

>> So the wall

47:11

>> is that the

47:13

>> tech link?

47:14

>> Uh, no. No, this is um this is at the

47:17

Langley house, I believe.

47:19

>> Yeah, but you had a Oh, not Tech Link.

47:21

Sorry, Tech.

47:21

>> This was with the boxes piled up, but

47:23

sometimes they all fell down. [laughter]

47:25

>> So, we just had product boxes behind us

47:28

as a as a set.

47:29

>> I just I thought the bench area had

47:31

pegboard in it really fast.

47:33

>> Uh, no, that was in the garage. This is

47:35

at the Langley house. This isn't in the

47:37

bench den. This is in the bedroom.

47:39

>> Yeah.

47:39

>> Weird. I I had remembered this being in

47:42

the garage. Okay.

47:44

>> Scrutiny. That thing.

47:45

>> Oh, I can't hear. Am I supposed to be

47:47

able to hear

47:49

>> will be combed with like a brick?

47:56

>> That's when you're just [clears throat]

47:56

moving into the house. Okay.

47:58

>> Yeah. I guess we weren't

48:04

>> They'll be able to

48:06

Oh, all right. Well, you just said

48:09

scrutiny. That thing.

48:10

>> Hey, look. After party link. It was a

48:12

separate

48:13

>> people. Was it on the same channel?

48:16

>> What does that mean?

48:17

>> Like everything will be looked at.

48:18

>> It comes with a brick.

48:20

>> Yeah. [laughter]

48:22

>> Your face.

48:23

>> I can tell when I said it. [laughter]

48:30

>> [laughter]

48:33

>> I've never gone back and watched that.

48:35

We've just [laughter] had this reference

48:37

for years.

48:37

>> That is That is the most

48:40

What did you just say that I think I've

48:42

ever

48:43

[laughter]

48:47

>> Oh man, that was pretty good. [laughter]

48:50

I couldn't even hear it. That was pretty

48:51

good. [gasps]

48:53

>> 13 years ago.

48:55

>> Yeah.

48:56

>> Oh, dude. We're so old. especially me.

49:00

Speaking of which, we should probably do

49:01

a new topic. I got myself a birthday

49:04

present.

49:05

>> Yeah.

49:06

>> Not on purpose, mind you. I ordered it a

49:08

year ago. These guys apparently build

49:13

these things to order.

49:14

>> Oh.

49:15

>> And in spite of the fact that they're

49:17

seem like a decentsized company and have

49:19

been around for quite a while, they make

49:22

um tennis training machines. So, you

49:25

know those ones that just like lob balls

49:27

over the net and you can return them?

49:28

They do them for soccer. Apparently,

49:31

that's a thing. You just load it up with

49:33

like

49:34

>> soccer balls and it just kicks them at

49:37

you or something. Uh, volleyball,

49:39

squash, table tennis, uh, reaction

49:43

lights, paddle, pickle ball. They have a

49:46

pickle ball training feeder thing. Well,

49:48

guess what? They also make badminton

49:52

feeding machines. This was a

49:54

recommendation from that uh English pair

49:56

that came over for a couple of collabs.

49:59

And apparently the German national team

50:02

uses

50:04

uh I believe it is

50:09

this guy right here, the 8025A.

50:14

Can this go away please? Badminton. Oh

50:18

my god. Websites are unusable. Okay,

50:22

there we go. [laughter]

50:24

8A hardcore quality

50:28

intelligent equipment. Uh, this arrived

50:32

today.

50:32

>> Have you used it? Did you use it?

50:34

>> I tried it today. I So, I'm going to

50:37

talk about this in a little bit more

50:38

detail at the all hands on Monday, but

50:40

there's going to be a bit of a content

50:42

pivot on LT.

50:44

>> Oh, okay. LTT is going to become

50:47

>> a little bit less of like,

50:50

>> okay, this is the type of video that

50:52

goes on the Linus Tech Tips channel

50:55

versus this is the type of video that

50:57

goes on Techquicki or on Short Circuit

51:01

or on wherever whatever other channels.

51:04

And LTT is going to be way more just

51:09

what I want to make a video about.

51:13

So, sometimes it'll be an unboxing,

51:16

sometimes it might be a review,

51:18

sometimes it might just be like I'm

51:21

building a weird AI server in my

51:23

basement. Um, and I feel like it's been

51:26

a long time coming.

51:27

>> I think this is great, especially right

51:30

now. I think this kind of has made sense

51:32

for a long time. I also think this

51:34

especially makes sense right now because

51:39

I don't know. I I I had a this chat

51:42

quite a few times at

51:44

>> He changed it. I [laughter] tried to

51:46

press I don't know.

51:49

>> That's wild.

51:52

[laughter]

51:55

>> What does mine say, dude? [laughter]

52:01

>> He changed it. THAT'S WHY

52:05

[laughter]

52:07

this uh this paid dividends.

52:10

>> Oh man, little text buttons.

52:12

>> That was [laughter] that was pretty

52:13

good.

52:14

>> All right.

52:14

>> Uh man, what was I even saying? Oh,

52:16

right. At open source, we had a bunch of

52:18

conversations with people about like

52:21

>> what do you even do now? Like I I make

52:24

content in a field that the field is

52:28

toxic right now.

52:29

>> Yeah.

52:30

>> Like it just sucks. No one wants to be

52:32

looking at this stuff right now.

52:34

Computer hardware is way too expensive.

52:36

Even if you can still afford to buy it,

52:39

it's really painful, man.

52:41

>> And it just sucks. And it just feels

52:43

bad.

52:44

>> The whole thing.

52:45

>> And sometimes And sometimes you can find

52:47

deals.

52:48

>> Y

52:48

>> and people don't want to hear that

52:50

either because the deal is not as good

52:51

as it was

52:52

>> and 10 months ago,

52:53

>> they might not be able to find one

52:55

because it I mean I found one. It took

52:57

me months.

52:58

>> Yep.

52:58

>> Like this is it's not super reasonable.

53:00

What if your timeline had to be shorter?

53:02

Like, it's it's it's just kind of icky

53:05

right now. But it's not like when that

53:08

viewership floats away that nerds who

53:12

like nerdy techy things are just gone.

53:15

>> No. If anything, I feel like they're

53:18

hungry for something to engage with.

53:21

>> And we have seen that in some cases

53:23

where there's like tech creators that

53:26

aren't making stuff that's super

53:27

centralized around like performance

53:28

computer hardware. Y

53:30

>> that are actually doing better than they

53:32

were.

53:32

>> I mean,

53:33

>> but most like effectively everybody

53:35

that's in the computer hardware tech

53:36

creator space thing is going down, but

53:39

people that are close to it are going

53:41

up, which is just very interesting.

53:43

>> I mean, even looking at the kind of

53:45

stuff that

53:46

>> dude, you buying a printer

53:48

>> is successful on our channel.

53:50

>> Dang.

53:50

>> I mean, just replacing a battery with my

53:54

daughter almost a million views. That's

53:56

the kind of cont

54:06

meat on this bone. I don't know if

54:08

there's enough learnables. But then I

54:10

was just like, "No, my daughter's phone

54:11

is broken and this is an opportunity for

54:14

her to do something she's never done

54:15

before. And there's a whole new

54:17

generation of viewers that is just like

54:19

her and hasn't seen me replace a battery

54:21

before. Why are we afraid to do this?

54:24

>> It's also not as normal now. That's this

54:26

is

54:26

>> which makes it interesting.

54:28

>> And she had a blast. Actually, you know

54:29

what's really kind of funny, but also

54:33

she like was almost in tears about it is

54:36

on uh on last Friday uh no not last

54:39

Friday, earlier this week um my other

54:43

daughter broke that phone, [sighs]

54:46

dropped it and the screen shattered.

54:49

>> It's kind of Yeah, you're right. It's

54:51

kind of funny, but it's also really sad.

54:53

timing. The timing blows.

54:55

>> She just fixed it.

54:56

>> So, Cat just fixed it. She just had this

55:00

like honestly kind of fun experience of

55:03

like she's invested in it and she has

55:06

like more of an emotional attachment to

55:08

it. It's something we did together and

55:11

I'm looking at it going, "Okay, well,

55:12

like now what path do we take?" Because

55:14

[laughter] it's already a six or

55:16

sevenyear-old phone or something like

55:17

that. And I like was barely even wanting

55:22

to buy a new battery for it, let alone a

55:25

new screen, which is a way more invasive

55:29

replacement.

55:31

I'm like, uh, and more expensive than

55:33

the battery. Do we just do I just get

55:36

her a new phone now? It's pretty close

55:39

to her birthday. Also, it would put her

55:41

on the same cycle as the other two kids

55:43

who both just got new phones. Like

55:45

there's a lot of reasons, but then part

55:46

of me is just like, "No, but we should

55:48

just we should just Frankenstein's

55:51

monster phones it and we should just I

55:54

don't know which which way would you

55:55

go?" She needs a stylus. Not needs, but

55:59

she loves the stylus. So, that really

56:01

limits my options.

56:02

>> Even cool that she likes the stylus,

56:04

too. I would almost see that as a

56:06

benefit. What new phones have a stylus?

56:08

>> Very few. like the top end um the top

56:11

end S series phones do, but they're

56:14

huge. Whereas the Note 10, not the Plus,

56:17

the Note 10 was a very reasonable size

56:20

for a kid stillish.

56:22

Like I meant I miss reasonable sized

56:25

phones.

56:26

>> Yeah.

56:27

>> What the heck?

56:29

>> Is there a route? Oh man, this stuff is

56:33

so complicated cuz it's like how how

56:35

does the kid respond to it? So, I was

56:37

going to say like this is this is more

56:38

expense. Sorry. By the way, $65,000 in

56:42

switches. Uh but uh what if you got her

56:46

to replace the screen and as a reward

56:48

for replacing the screen?

56:50

>> Then she can sell the phone and we'll

56:52

get her new one. [clears throat]

56:52

>> She gets the nice new one.

56:55

>> Yeah. Or something like that. I I could

56:57

see us doing that. Um yeah, I'll I'll

57:00

think about it. How

57:01

>> because like I don't know how much that

57:02

phone's worth, but I think

57:06

Just the process of if if she enjoyed

57:08

the previous one, showing her that there

57:10

can be a reward in fixing things I think

57:13

is is cool. Does that make sense? Even

57:16

[clears throat] if it's not fixing it

57:18

for you.

57:20

>> Dan's in chat. You know, you you're

57:21

right here, Dan. He says, "Make the one

57:24

who broke it fix it."

57:26

>> Yeah. Why not? Why not?

57:29

>> She's still a little young for that

57:31

delicate of an operation.

57:32

>> I don't think so. The phone's already

57:34

broken.

57:35

>> Yeah. And the stakes are lower this time

57:37

because

57:38

>> absolutely

57:38

>> of what happened last time. I did set up

57:40

backup on it. So at least

57:42

>> we're not like

57:44

don't break this thing.

57:46

>> I started when I was like three and I

57:48

turned out normal.

57:52

[laughter]

57:55

>> So

57:56

>> I'm excited for that project next

57:58

weekend.

57:59

>> Me too. [laughter]

58:00

>> See make her fix that screen.

58:03

Do it.

58:07

>> What?

58:08

>> I'll think about it. I'll think about

58:09

it.

58:09

>> Yeah. [clears throat]

58:10

>> I think fixing the screen like if if you

58:13

know I think you did this with 3D

58:14

printing. You you sowed an interest.

58:18

She

58:20

displayed some interest in Thank you. Uh

58:24

in replacing the battery.

58:27

>> Do you just kind of keep that going?

58:29

It's 80 bucks. Kind of sucks. Maybe you

58:32

can recover some of that cost reselling

58:34

the phone. I really don't expect a

58:35

seven-year-old phone to

58:36

>> Worth a couple hundred bucks still.

58:37

>> Really?

58:38

>> It's It's Well, it's kind of a unique

58:39

phone.

58:40

>> I would definitely do it then.

58:41

>> Yeah. And it has a fresh battery

58:43

>> whether you keep it or not.

58:44

>> I would I would replace the screen.

58:46

>> Brand new screen, brand new battery

58:48

version of that phone. It's probably

58:49

worth an okay amount.

58:50

>> Yeah, it's probably worth something.

58:52

>> Yeah.

58:52

>> Um anyway,

58:54

>> video [clears throat]

58:55

uh video coming soon on the the

58:57

badminton machine 8025. Here's a much

59:01

worse version of it, but it just kind of

59:04

craps shuttles at you and then you hit

59:07

them. It's pretty cool

59:10

and it works surprisingly well.

59:13

>> Be rentable at the bon.

59:15

>> I don't know exactly what to do with it.

59:17

I mostly got it because I always wanted

59:20

one and I thought it would be super

59:22

cool, but I wasn't sure what to do with

59:24

it. It could be interesting as like a

59:25

nighttime rentable thing because during

59:28

prime hours, I'm assuming you want more

59:31

more

59:32

feet on courts.

59:34

>> Yeah, I was say butts and seats, but not

59:36

in this case. You want feet on courts.

59:37

Um, but maybe at night time when you

59:40

have like idle empty courts,

59:42

>> you could rent out the the practice

59:45

machine.

59:45

>> Yeah, I could see that. I mean,

59:47

depending on how much people are paying

59:48

to rent the machine because it would

59:50

have to have a pretty considerable cost.

59:52

Yeah,

59:52

>> cuz we have to like set it up and tear

59:54

it down

59:55

>> for you presumably. So, I would assume

59:56

it would be like a hundred bucks an hour

59:58

or something like that.

59:59

>> I have no idea.

60:01

>> And it's it's like a it's like a

60:03

>> it's going to have to be not that cheap.

60:04

>> Yeah, it can't it can't be super super

60:06

cheap.

60:06

>> Maybe it could be a program thing where

60:08

you get like you buy like a bundle of

60:11

sessions with it or something. I could

60:12

see something like that

60:13

>> because I could see someone potentially

60:15

pairing that with um a training program.

60:20

Uh, Magi 1999 says that's too steep. Do

60:24

you have any idea how much a coach to

60:26

sit and feed you shuttles costs? Like

60:29

it's So, first of all, we're talking

60:30

Canadian dollars, so that's something.

60:33

Um, but like coaches are over $100 an

60:38

hour for the ones with like a good

60:39

reputation these days. And from the

60:42

limited time that I've spent with it so

60:44

far, the feeding quality is super

60:46

consistent. How how long does it take to

60:49

set up and take down?

60:50

>> It's pretty quick.

60:53

>> How much is the cost of the thing?

60:55

>> Uh it was like eight or nine grand. Like

60:58

if you're if you're renting, you go to

61:00

like um like uh like United You go to

61:02

United Rentals or something. You rent a

61:04

paint sprayer,

61:05

>> right? You're talking like a $700

61:08

uh or actually they're a little more

61:09

than that these days. I think they're

61:11

like probably closer to a,000 plus. Like

61:13

you rent like a $1,000 paint sprayer,

61:14

it'll easily cost you like a hundred

61:16

plus dollars a day easily. Like the

61:19

economics around renting stuff are kind

61:22

of crazy to me. That's why we bought all

61:23

the tables and chairs for whale land.

61:25

>> Yeah.

61:26

>> Because by the time we ran I think it

61:27

was like three lands. We could literally

61:30

afford to just buy all of our own tables

61:32

and chairs.

61:32

>> I think like like Okay. So how much is a

61:34

coach an hour? You said a little over

61:35

100. So it's like 115

61:37

>> up to like 150 160.

61:39

>> Okay.

61:39

>> Like for the for the really good ones.

61:40

Now, obviously, a machine like this is

61:42

not giving you feedback on your form, so

61:44

there's a trade-off there, but it also

61:48

is able to feed super consistently. You

61:51

will get in more reps with this. So,

61:55

horses for courses, uh, I could see it I

61:57

could see it being I could see it being

61:59

a little lower. Um,

62:00

>> I mean, I have no

62:03

horse in this race.

62:04

>> And also, uh, the shuttles, so you can't

62:07

just I don't think you can just put

62:08

garbage shuttles in it. And we today we

62:12

stacked $2,000 Canadian worth of

62:14

shuttles in it. [snorts]

62:16

So there's a cost there because they're

62:18

hitting them, right? They will wear.

62:20

>> Okay. So it includes the the shuttle

62:23

cost.

62:23

>> I could see that

62:24

>> if you want to soften that blow.

62:27

>> Yeah.

62:28

>> But keep it the same price. I think you

62:31

lower the cost of the machine but make

62:32

someone buy effectively new shuttles

62:34

with it because you don't want bad

62:35

shuttles going through it because it'll

62:36

hurt the machine. I I think people would

62:38

put bad shuttles in it if you don't

62:40

include them. So, I think it's better to

62:42

have a combined rate and we just control

62:44

the quality of of the projectiles.

62:47

>> You take your shuttles home at the end

62:49

of the day when you're done playing with

62:50

the machine.

62:53

>> Um, no one's going to be able to do that

62:55

kind of an upfront. Yeah, I'm going to

62:56

buy two grand worth of shuttles to put

62:58

in it. So, it would have to be provided.

62:59

>> No, no. I'm saying that's like part of

63:01

the benefit.

63:02

>> You take them home.

63:03

>> Yeah.

63:03

>> No, you wouldn't. So they would go back

63:05

in the machine because they can they can

63:07

be reused for quite a while, but it

63:09

would include the wear and tear on them

63:12

that you that you go through in your

63:14

session kind of thing. I could

63:15

>> How do you So So okay, because there's

63:17

$2,000 worth of shuttles, whatever.

63:18

They're never reloading as a you as a

63:20

user.

63:21

>> Uh they might have to, but they'd have

63:23

So they have Do you have to It's And

63:24

that's like the worst part of Admonton

63:26

training is picking up all the stupid

63:27

shuttles and

63:28

>> Yeah. I was thinking because like if I

63:30

this might be a sketchy part of renting

63:33

it out

63:34

>> is what if somebody loads a bad shuttle.

63:37

>> Yeah. Then I I think it shouldn't just

63:39

like be destroyed by a slightly worn out

63:41

one. Like it's not that big of a deal.

63:43

>> Okay.

63:43

>> And the all the hoppers are clear so you

63:46

could see what's in it.

63:47

>> Got it.

63:48

>> So if there's like crap in it then you

63:49

could just I don't know if the

63:51

butttheads put the thing in. You just

63:53

put a new one in.

63:54

>> Anyway, so I've got a video coming on

63:56

that. It was pretty exciting. I

63:57

co-hosted it with my son. That's another

63:59

thing that I think you guys are going to

64:00

see more of is just like

64:02

>> this is something that, you know, me and

64:04

the family are interested in. So, we're

64:06

just going to like make a video about

64:07

it.

64:08

>> Are we doing the thing that I uh

64:13

>> what I retire and I'm replaced by my

64:15

children? Is that what you want? Fine.

64:16

Fine. You know what? No. Fine. No, I'll

64:18

quit. I'll quit right now.

64:20

>> Okay, I'm back.

64:21

>> This What is this?

64:23

>> Uh,

64:25

yeah. I don't see why not. Can I should

64:27

I bring that up?

64:27

>> Sure. Yeah. Let's talk Let's talk about

64:28

it.

64:29

>> Uh my something that I've been talking

64:31

to Lionus about doing is my my dad's

64:33

system. So my dad

64:35

>> It's a crazy build.

64:36

>> It's so wicked. Uh my dad is a plumber

64:39

by trade. He's a plumbing instructor

64:40

now. And he built this case out of

64:43

copper pipe. Uh like fairly big copper

64:47

pipe. Uh whole frame all the way around.

64:49

If you've been to Whaleand, you might

64:50

have seen it. He's brought it a couple

64:51

times. Um, he wants to use the case

64:55

itself as a radiator with an emergency

64:58

out for like particularly hot summer

65:01

days where he can connect it to an

65:03

external rad just in case, but normally

65:05

it should be fine. He did a bunch of

65:06

math on it. He was working with Rod on

65:08

it. Uh, Rod said it's such a cool case,

65:11

which to me just means that it's a fact.

65:14

Um, I So yeah, we're going to do I don't

65:17

think I've even mentioned this to Rod,

65:19

but uh here we go. Oh, that's one of the

65:21

reasons why I wanted to mention it now

65:22

is cuz I knew Rod was in chat. Uh, so

65:24

we're going to try to do that on the

65:25

channel. Uh, which I think like a couple

65:27

years ago when the idea first came up,

65:30

it didn't make as much sense for the

65:32

channel, but with the new vibe that

65:34

you're talking about, it like suddenly

65:35

makes a lot of sense for the channel.

65:37

>> I think so, too.

65:37

>> Um, so there's some plans. I have yet to

65:40

ask for it, but we've been thinking that

65:42

like, oh man, it's such a perfect board

65:45

for this. I hope they say yes. Uh, what

65:47

is it? The Asus 20,

65:48

>> the anniversary one. It's so per It's so

65:51

expensive. It's so perfect for it. Oh

65:53

man,

65:54

>> it's upstairs.

65:54

>> It's so perfect for it.

65:56

>> Should I go get it?

65:57

>> Sure.

65:57

>> I think I know where exactly where it

65:59

is. Maybe just do a do a quick check out

66:01

message while I'm gone.

66:02

>> Impossibly perfect for this build. He's

66:04

going to grab it. But there's a bunch of

66:05

copper accents on it. Uh the heat sinks

66:08

are are like kind of exposed copper heat

66:10

sinks and it has a back plate on the

66:12

motherboard which is just this like huge

66:14

copper back plate inside of

66:16

>> You don't have to describe it.

66:18

>> He's going to go get it. Yeah. Uh,

66:20

inside of a case that is all copper

66:22

tubes. It's going to be freaking sick.

66:25

I'm very excited about it. Rod's going

66:27

to theoretically come up and and help us

66:29

build it. So, it'll be myself, my dad,

66:31

and Rod. Uh, Rod, Rob, and Luke. Uh, and

66:36

we're going to Yeah, we're going to

66:37

build away and have fun. And it should

66:38

be should be really cool. Sorry. What

66:40

were we going to do? What did you want

66:41

me to do?

66:42

>> Maybe a com.

66:43

>> Sure. Do you have one?

66:44

>> Yeah, I got one for you. Uh, if chicken

66:47

was suddenly outlawed, what would you

66:49

eat for protein?

66:51

Uh, if chicken was suddenly outlawed,

66:53

what would you eat for protein?

66:54

>> I was going to say bugs.

66:57

I don't know if you'd be a bug man.

66:59

>> No, not so much. Um, people immediately

67:02

said eggs in full chat and honestly,

67:04

great answer, but I they don't sit well

67:07

with my stomach for whatever reason, so

67:08

I I can't do it. I think eggs would be a

67:10

really good answer, but yeah, I can't do

67:12

it. Um, something I've been enjoying

67:14

lately is extra lean uh ground beef.

67:18

What? That's the motherboard box.

67:22

Are you

67:25

>> What?

67:27

[laughter]

67:28

Why? Seriously, it's so like you could

67:31

ship a computer in this box.

67:33

>> Yeah.

67:37

>> Oh my god.

67:38

>> We're good. We're good.

67:44

I don't I don't know.

67:46

>> You might have to just tip it over.

67:48

>> Yeah,

67:50

it'll be good.

67:55

>> Watch the

67:56

>> Oh, yeah. Oh, yeah, buddy.

67:59

>> Oh my goodness.

68:03

>> Uh uh. What the uh

68:10

>> what? It's in a suitcase.

68:14

>> Why does the motherboard come in a

68:16

suitcase?

68:23

Is there more in there, Dan?

68:29

The heck?

68:32

No wonder this thing costs so much.

68:37

Still not there yet. There we go. Wait.

68:40

Oh, this is the wrong one.

68:42

>> What?

68:42

>> Yeah, this isn't what I'm talking about.

68:43

>> This isn't the one you [laughter] like?

68:44

>> No.

68:49

[laughter]

68:51

You ripped apart off of it already. He

68:53

brought out the anniversary board.

68:54

>> Oh, I thought

68:55

>> which is the new version of it.

68:57

>> I thought that's what you were talking

68:58

about. No, this is a super cool

69:01

motherboard, but not what I was talking

69:04

about.

69:04

>> Oh, all right. Well, get this garbage

69:06

out of here then. This is

69:07

>> Get the Get this out of here. Just throw

69:09

it throw it on the ground.

69:12

>> Um,

69:13

>> what the heck?

69:15

>> Hold on. Hold on. There you go. Okay.

69:17

Thanks, Daniel.

69:18

>> I've got it. This is the one I was

69:20

talking about.

69:21

Oh,

69:24

>> because you see the copper accenting on

69:26

this side, but the back plate, the back

69:29

plate is honestly what really got us

69:31

interested in this. Where can I see the

69:32

back plate? Sick looking motherboard in

69:35

general. Where's the back plate? Show me

69:38

the back plate.

69:39

>> Well, I'm so disappointed. I brought you

69:40

the wrong thing. I failed you. Oh, Dan,

69:42

we may need to adjust that camera again.

69:44

Um,

69:45

>> just a crazy

69:46

>> Oh, there's animations. I don't care

69:47

about that. Where's the back plate?

69:52

These like infinite scroll websites

69:55

drive me insane.

69:57

[cough]

69:59

They don't even show it. Oh my god,

70:02

>> that's crazy.

70:06

>> Um,

70:06

>> of course Noki has it.

70:09

>> Classic Noki.

70:10

>> Okay, here. That's the back plate, dude.

70:13

>> That's insane.

70:14

>> And in his the case that he's designed,

70:16

the back plate will be fully exposed. So

70:19

you'd have this like huge copper 20th

70:22

anniversary ROG back plate exposed

70:25

inside of this like copper framed build.

70:29

So perfect.

70:30

>> The funniest part of honoring this first

70:33

ROG motherboard, like this color scheme

70:35

and everything from 20 years ago is that

70:36

the first ROG motherboard was a piece of

70:39

>> Like our RMA rate on it was, I kid you

70:41

not, 50%.

70:43

>> I remember this. Yeah,

70:44

>> they just all came back.

70:45

>> Yeah. Hopefully this one wouldn't be as

70:47

problematic. Um, hey, we should uh get

70:51

through some sponsors real quick.

70:52

Otherwise, that's going to be

70:53

problematic.

70:54

>> Uh,

70:55

>> the show is brought to you by Vessie.

70:58

Hello, fellow fooths. I have some

71:01

unfortunate news for you. Summer is

71:04

coming to an end, which means rain is

71:06

right around the corner. The good news,

71:09

Vessie's weekend Chelsea shoe has you

71:11

covered. Literally. From your toes to

71:13

your ankles, the Chelsea's will keep

71:15

your feet and socks dry. Vessie says

71:18

their shoes are totally waterproof

71:19

thanks to their highly breathable Datex

71:21

weave that doesn't soak up a ton of

71:23

moisture. And unlike other high top

71:26

shoes, the Weekend Chelsea are easy to

71:28

put on and take off quickly thanks to

71:30

the heel loop. They're built to last

71:32

with grippy rubber outs soles, so your

71:34

new pair of Weekend Chelsea will last

71:35

right through the autumn and into the

71:37

colder and slushier winter. And every

71:39

pair of Vessies also comes with a whole

71:40

year warranty and 30-day hassle-free

71:43

returns. And if your order's over $110,

71:46

they'll ship to you for free as long as

71:47

you're in North America. Your everyday

71:49

escape starts with a little adventure.

71:52

So grab 15% off your first pair of

71:54

Vessie Weekend Chelsea using our link

71:56

down below or the QR code on the screen

71:58

or by going to vessie.com/wanshow.

72:02

The show is also brought to you by

72:03

Whisper Flow. Voiceto text apps and

72:05

devices don't always convey the message

72:07

you're trying to get across. They

72:09

misspell words. Accuracy gets lost in

72:11

longer documents and they may not be

72:13

compatible with the apps you need.

72:16

Whisperflow wants to show you a better

72:18

way. You don't have to accept [laughter]

72:21

You don't have to accept the typos and

72:24

weird formatting to get your point

72:26

across. Whisper Flow is speech to text

72:28

that records what you're actually trying

72:30

to say. It's able to learn in real time.

72:32

And so if you're trying to say the name

72:34

of a co-orker, the app can learn or draw

72:36

from a contacts database and nail the

72:38

correct spelling. No more Luke laugh

72:41

renair in the DMs. A it's built to work

72:45

with any app like Gmail, Slack, notes,

72:47

the LLM of your choice, and so many

72:49

more. And it doesn't matter how fast you

72:51

write or type. Whisper Float just will

72:53

help you work faster. That means more

72:55

yapping and less thumbtapping. It works

72:58

in all the places you do your best

72:59

thinking, like on walks, while

73:00

commuting, even when you have a genius

73:02

idea in the shower. So, don't lose your

73:04

best ideas in translation. Instead, get

73:07

your first month of Whisper Flow Pro,

73:09

totally free. We'll have it linked in

73:11

the video description. The show is also

73:13

brought to you by Xqus

73:16

generalpurpose robots for logistics,

73:18

warehouses, household, and other

73:20

fulfillment operations. Each robot is

73:23

built to continuously assess changing

73:25

parcel piles, determine what to pick and

73:27

how to grasp it, and adjust its actions

73:29

as conditions change. During a recent

73:32

live stream, okay, this was actually

73:34

flipping nuts. Xqu Systems achieved a

73:37

throughput of 1,816

73:39

parcels per hour with an accuracy rate

73:41

of over 98%. [music] Which beat major

73:44

competitors average of 1248 per hour.

73:46

They do this with wall B, [music] their

73:49

proprietary embodied AI foundation

73:51

model. Instead of a traditional

73:52

pre-programmed set of rules that the

73:54

robots follow, X squares robots can

73:56

adjust on the fly. So as parcels are

73:58

added, they can figure out how to grab

74:00

and organize accordingly. The arms

74:03

operate on a six-axis system, which is

74:05

perfect for picking, flipping,

74:06

flattening, repositioning, and handling

74:07

boxes. And even if something is blocking

74:10

a box, the system can automatically

74:12

reassess and retry on its own without

74:15

the need for human intervention. The

74:17

arms can stop what they're doing at

74:18

opportune moments to make sure errors

74:19

are avoided, like a parcel going down

74:21

the wrong chute. And you can learn more

74:22

about X Square using our link in the

74:25

video description. Finally, the show is

74:27

brought to you by Hrefs. When people

74:29

need help with something, a lot of the

74:30

time it just makes sense for them to

74:32

Google it. And if you want to help them

74:34

and make money while you're at it,

74:35

you're going to want to be near the top

74:36

of that Google search. HRES is a free

74:39

toolkit to help you do just that. It's

74:41

for marketing teams, businesses, as well

74:43

as for creators and makers. They offer a

74:45

suite of tools that don't just make your

74:47

brand or business easier to find online,

74:49

but help you understand why you are or

74:51

aren't popping up in the searches you

74:53

want. Their web analytics shows you

74:55

traffic data that helps it. Sorry. Their

74:59

website analytics show you traffic data

75:00

like who's visiting your website and

75:02

where they're coming from. And it can be

75:04

broken down into easyto read graphs.

75:06

Hrefs can even run full SEO audits of

75:09

your website, scanning for over 170

75:12

different technical issues that might be

75:13

affecting your performance. Is one of

75:15

them just being over 40?

75:18

Sorry. They'll show you where your

75:19

website ranks with specific Google

75:21

keywords so you know what strengths to

75:23

lean into more as well. Take your

75:25

digital marketing knowledge to the next

75:26

level for free with Hrefs. Hrefs. We're

75:29

going to have them linked down below. My

75:30

braces are driving me nuts. I had an

75:32

appointment this week and we're so

75:34

close. We're so close, man. They crank

75:36

them on you again.

75:36

>> We just got Oh, yeah. Yeah. Every time.

75:38

But you just got to get this one turned

75:40

a little bit. Just turn that other one a

75:43

little bit and then we're done.

75:44

>> You're free.

75:44

>> Then we're done. Get them off.

75:46

>> Yeah.

75:46

>> I'm really not looking forward to that

75:48

appointment, though. Grinding is the

75:50

worst for me. I I hate the grinding. And

75:54

they have to They have to grind off the

75:56

glue that held all the brackets on. So

75:58

they Oh, and then popping the brackets

75:59

off. They actually had to reposition

76:01

this bracket. So they like they they

76:04

have you like a bite down on a thing and

76:06

they're like

76:08

like pulling it off

76:09

>> right after getting them in place. Just

76:10

yank on them.

76:11

>> It's going to be great, man. I'm

76:12

>> Do you have Do you have your crossbar on

76:14

your bottom teeth?

76:16

>> I They'll probably put that on when I'm

76:18

done. Like I'll ask for a permanent

76:20

retainer there.

76:20

>> Yeah, I got that when I was a kid and I

76:22

asked the guy like, "Oh, when do I get

76:25

this off?" And he said when I was like,

76:27

you know, I don't even know.

76:31

I was like 12, so the age that he told

76:33

me felt like I would be ancient and I

76:35

think it's like roughly now. Uh, but I

76:37

don't think I'm gonna get removed

76:38

anyways because it's just like part of

76:39

me at this point. Um, but I remember him

76:43

telling me that and being like, "Oh, I'm

76:45

I'm doomed forever." But unlike the

76:48

rest, it really it doesn't bother me at

76:50

all.

76:51

>> Yeah. Something on the inside I think

76:52

I'd be fine with. It's just the way they

76:54

shred up my lips.

76:55

>> Yeah. It never shreds anything.

76:57

>> Yeah.

76:57

>> I can scratch my tongue on it. Feels

76:59

great.

77:00

was good.

77:01

>> Nice. So, that's another thing that, you

77:04

know, over the last few years wouldn't

77:06

have been a video on LTS, me just like

77:08

talking about my braces adventure.

77:11

>> Oh, yeah.

77:11

>> Whereas now it will be.

77:13

>> There's technology involved.

77:15

>> There is. You tried 3D printed

77:17

>> 3D printed ceramic brackets.

77:18

>> Oh, they currently are. They are.

77:20

>> Okay.

77:20

>> Yeah. And they suck and so that'll be my

77:22

review

77:23

>> is they are a giant boatload of ass.

77:25

actually help a ton of people,

77:26

>> which legitimately probably will,

77:28

>> including if you especially in including

77:31

if you mentioned how

77:31

>> I'll be naming and shaming them.

77:33

>> Cut your

77:34

uh lips up and stuff.

77:36

>> Oh yeah,

77:36

>> cuz like I had braces when I was never a

77:38

problem. There was problems.

77:39

>> Oh yeah, it's been it's been very suck.

77:42

>> I kissed somebody and our braces got

77:44

connected together.

77:44

>> That's pretty funny.

77:45

>> That was awkward. [laughter]

77:47

>> Yeah,

77:50

>> but like that's the problem. not uh you

77:53

know my my mouth getting cut up.

77:55

>> Yeah. Yeah.

77:56

>> Yeah.

77:58

>> All right.

78:00

What else? Uh oh. Yeah. I Okay. I guess

78:01

we could like maybe do some topics. That

78:03

would be crazy. No, man. Oh, yeah. Okay.

78:07

No. Yeah, we better do a couple topics

78:08

first. YouTube is raising the bar on

78:11

monetization.

78:13

Citing the need to keep pace with the

78:15

growth of the platform. YouTube will be

78:17

increasing the requirements for new

78:19

creators who are applying to the YouTube

78:21

partner program. Potential partners will

78:23

still need 1,000 subscribers, but will

78:26

now need 8,000 qualified watch hours in

78:29

the last year or 20 million qualified

78:33

shorts views in the last 90 days. This

78:35

is up from 4,000 qualified watch hours

78:38

and 10 million qualified shorts views.

78:42

Our discussion question here is what

78:44

advice would you give a creator trying

78:46

to make a content creation a career in

78:48

2026?

78:50

>> The I know one of the main topics that

78:53

I've seen around online. Okay, there's

78:55

been a bunch of stuff. Uh honestly, one

78:57

of the genuinely best resources for

79:00

content creation stuff I've seen in a

79:02

long time is some of Lewig's classes

79:06

from Streamer University. They're like

79:08

genuinely extremely good and I think if

79:10

you're trying to get into this stuff,

79:11

you should go watch it. Um, one of the

79:14

takeaways from that was I think he said

79:16

packaging is like 70% of the deal or

79:18

something and he's not wrong. Uh, so

79:22

>> I hate how important it is.

79:23

>> It is very very frustrating, but it is

79:25

incredibly important. Um, I I think

79:28

we've talked before, but I I've seen

79:30

I've seen view graphs go like 4x from

79:36

having a better thumbnail and title. I

79:38

mean,

79:38

>> and I have also seen retention go up

79:41

with that. So, you have like a more

79:44

accurate title, which everyone says that

79:45

you have to do,

79:47

>> and then you get a title that's more

79:48

clickable, but that doesn't lie, and

79:51

your retention and your views go up.

79:53

>> I mean, we saw this two weeks ago.

79:56

>> Yeah. What?

79:56

>> This WAN show had crazy live numbers.

80:00

Like we peaked over 15,000 or something

80:02

like that, even though we're on a new

80:03

channel that doesn't even have that many

80:05

subscribers. By the way, make sure

80:06

you're subscribed to the Wan Show

80:08

channel. Anywh who, the week before

80:11

peaked at more like 5 or 6,000. The week

80:13

after peaked at more like 6,000, I

80:15

think. And then this week, we're sitting

80:17

at like 5,000 right now, which means

80:19

we're probably tracking somewhere

80:20

between 6,000 6,500. So, the channel's

80:22

like doing good. It's growing. It's

80:24

happening. But this one crushed it.

80:27

>> 3xed out of nowhere.

80:28

>> So I reached out to YouTube. I'm like,

80:29

"Yo, what was up with that?" Like the

80:31

VOD didn't perform particularly out of

80:34

line with everything else. It's a little

80:36

better, but like 3x what is that? And

80:39

they're basically just like, "Oh yeah,

80:41

it seems like it's related to Minecraft

80:43

interest in Minecraft." I'm like,

80:47

I don't think a I don't think that fully

80:50

checks out and b

80:52

um I don't personally love that,

80:57

you know, um a format that people tune

80:59

into, you know, week in week out. You

81:01

guys are deciding whether to serve

81:04

because it got way more impressions.

81:05

Like it's was served way more

81:07

impressions. So

81:08

>> retention was also good though.

81:09

>> So well, it was Yeah, but it's the same

81:10

as a typical WAN show. It wasn't any

81:12

different.

81:12

>> It was the same which is the crazy.

81:14

>> Yeah. So, why is it getting three times

81:16

the live engagement? And it's basically

81:17

because the packaging. Somehow

81:21

Minecraft resonated with the YouTube

81:24

algorithm in that moment and we got

81:26

three times the viewership. So, that

81:28

gives you some idea just how important

81:29

it is because I don't think there is a

81:32

format that we've ever done that is more

81:34

consistent week to week than the WAN

81:35

show. It's these two guys talking and

81:39

sometimes, okay, he pops in and

81:44

We just

81:45

>> we have intentionally never even changed

81:47

the thumbnails I think from literally

81:49

ever.

81:51

>> I think the thumbnails are just exactly

81:52

the same as they've always been. Uh but

81:54

yeah. Okay. So that's really important

81:56

is that packaging is like literally more

81:58

than half of the equation.

82:00

>> So keep that in mind. Um do a lot of AB

82:03

testing. I talked to I'm not going to

82:05

out them, but I talked to a creator that

82:08

probably almost everyone watching knows

82:10

at open sauce and I was poking around a

82:12

lot of different creators. How many

82:14

revisions do you do of your thumbnail

82:16

and title after the video is uploaded?

82:19

That was the main thing I was trying to

82:20

poke at and someone huge creator

82:24

like 12

82:28

crazy like every time like it's huge

82:32

huge deal. The other one would be the uh

82:34

the like and this is gonna sound really

82:37

negative to say and there's probably a

82:38

way more positive way to say it, but the

82:39

like viewer quality funnel.

82:43

If you get a Tik Tok, a YouTube short,

82:45

an Instagram reel or something like

82:46

that, if you get a view on there, it is

82:49

worth very little.

82:51

>> Yeah,

82:51

>> very very very little. If you can

82:53

offload that to some form of VOD, most

82:56

likely YouTube, long form video content,

82:59

most likely YouTube. there are other

83:00

ones. Um, that value goes way up.

83:03

YouTube's in a pretty good sweet spot

83:05

where you can get a lot of discovery.

83:08

Uh, you can get your channel out there,

83:10

you can get your name out there, you can

83:11

be found pretty easily

83:13

>> through anything. YouTube long form

83:15

content.

83:16

>> I mean, I'd argue long form is getting

83:18

harder, not better.

83:19

>> Way harder, but it's still in a like

83:21

sweet spot range. Um,

83:23

>> better than anyone else, that's for

83:24

sure.

83:24

>> So, it's easier way easier to get views

83:26

than a lot of other platforms. way

83:28

easier to be discovered than a lot of

83:29

other platforms and you can still make

83:30

money because it's a long form YouTube

83:32

video. YouTube is pushing shorts like

83:35

crazy. They are actively killing a lot

83:37

of their long- form creators. There are

83:38

a lot of long form creators that are

83:40

shrinking, laying people off, failing

83:43

entirely, etc. Uh largely in a lot of

83:46

people's opinions due to shorts just

83:48

taking over the watch time of the

83:49

platform. YouTube has kind of saturated

83:52

the world as its watch base. So, it kind

83:55

of shifts its own watch time around on

83:57

the platform and they've shifted stuff

83:59

away from VOD and towards shorts.

84:00

>> You ask YouTube, they'll say, "Well,

84:02

that's what viewers demand." And if we

84:04

weren't giving it to them, Tik Tok would

84:06

or Facebook would.

84:07

>> And if you ask the users and the

84:09

creators, they're like, "Why are you

84:10

shoving it down my throat so much? I'm

84:11

not demanding this at all." Uh, you're

84:14

you're like brain dopamine

84:16

mindontrolling me and just staying on

84:17

these shorts for so long. It's not what

84:18

I actually want. But anyways, the like

84:21

generally seen as highest quality view

84:24

is an active live streamed view.

84:28

So that funnel can be very tough because

84:31

discovery on live stream platforms might

84:35

might as well not even exist.

84:37

>> He was going to do it if I didn't.

84:39

>> Discovery on Twitch is effectively

84:42

useless.

84:43

>> Yeah,

84:43

>> it like might as well not even be there.

84:45

Uh, but having an audience of, you know,

84:51

say 150 people where over the course of

84:54

your stream you're averaging, I don't

84:56

know, like a thousand different people

84:57

coming in and out. Um, and your VOD on

85:00

Twitch gets like 2 3,000 views. That can

85:03

be a successful full-time creator,

85:05

>> right?

85:06

>> That many views on YouTube, not even

85:08

close.

85:08

>> That many views on shorts,

85:11

>> you will make literally nothing.

85:13

probably you probably won't earn enough

85:15

to get a payout. So, it's like it's this

85:17

this ramping thing. It doesn't mean that

85:19

views on short stuff is worthless.

85:21

>> It just means you're going to need like

85:23

millions.

85:24

>> Yeah.

85:24

>> It doesn't mean that views on YouTube

85:25

are worthless. They're absolutely not.

85:27

Views on YouTube are worth a bunch. But

85:28

it also doesn't mean that if you see

85:30

someone on Twitch with a a viewer count

85:32

that doesn't seem super high. That

85:34

doesn't mean that it's not super

85:35

valuable. It kind of slides on this

85:37

chain. And you can you can try to

85:40

promote your your viewers jumping over

85:42

like you can try to promote your viewers

85:43

jumping from shorts to long form. It's

85:46

hard, very difficult. It's tough. Yeah,

85:49

there you go. Um because a lot of people

85:52

are going to watch that short and then

85:53

just scroll on by. But if you can

85:55

convert them, that can be quite

85:56

valuable.

85:57

>> That's some good advice. I have nothing

85:59

to add.

86:01

>> Hell yeah.

86:02

>> Um where did that come from?

86:04

>> It's a merch message, wasn't it?

86:07

>> Excuse me. a

86:10

a com.

86:11

>> Nice.

86:12

>> I don't remember. Checkout message.

86:14

There we go.

86:15

>> Solid.

86:15

>> Totally.

86:16

>> Uh oh, speaking of Twitch, uh Luke

86:18

mentioned him a couple times. You a new

86:21

Twitch feature trains Amazon's

86:24

generative AI on users channels

86:27

>> by default.

86:29

Um, it would be a shame if anybody

86:32

watched this WAN show and I were to say

86:35

if you want to disable that, you go to

86:37

the security and privacy tab of your

86:39

Twitch user settings and look for

86:42

training and generative AI.

86:45

Um, and go ahead and disable that.

86:48

[snorts]

86:49

Obviously,

86:51

Twitch should have made this an optin

86:53

program, but they actually did address

86:56

this. That's what Luke is reacting to.

86:58

Twitch chief product officer Mike Minton

87:01

said um if it was opt in nobody would

87:05

opt in [laughter] so it's going to be on

87:07

by default.

87:12

>> That is

87:14

probably the best example of saying the

87:17

quiet part out loud

87:18

>> that we have seen in at least a solid

87:20

year on the WAN show. Like I can't think

87:22

of a more on the nose

87:26

>> assessment. [snorts]

87:29

>> That made me breathe out of my nose so

87:31

hard that it's itchy now. Um, [laughter]

87:34

wow. That's amazing. Uh, that almost

87:36

feels like it's like begrudgingly said.

87:39

>> Sarin Serendipity says, "Hold on. He has

87:41

a point."

87:42

>> I mean, he really does. He super does.

87:49

[laughter]

87:50

>> Oh man. Meanwhile, McDonald's is making

87:53

dossas on their loyalty program. Our

87:57

notes say, this is great.

88:00

I'm not loving it. Um, Reese Rogers, uh,

88:04

superhero name detected. A Wired

88:06

reporter

88:07

requested his McDonald's loyalty app

88:10

data and received an extensive, this is

88:13

not a typo,

88:16

515

88:18

page dossier that detailed his orders,

88:21

visits, favorite locations, spending

88:23

habits, and this one's the kicker for

88:25

me, predicted behavior.

88:28

McDonald's algorithm even assigned him

88:30

an attrition score of zero, predicting

88:34

that he was essentially never going to

88:37

stop eating Mickey De's.

88:39

McDonald's is now reportedly pooling

88:41

loyalty data from 220 million customers

88:44

across 70 markets into a global AI

88:47

system that is designed to learn from

88:49

billions of transactions. The AI seems

88:52

to be geared towards improving

88:53

suggestion and recommendation algorithms

88:55

to influence customer behaviors. Um, our

88:59

gentleman who prepared this, Nathan, he

89:00

has put in a fought here. They're about

89:02

to make some sort of horrifying

89:04

malevolent clown AI. They'll turn Ronald

89:05

McDonald into the Pennywise of the

89:08

internet age. Everything floats, more

89:09

like everything French fries. Uh, thank

89:11

you for that, Nathan. That was

89:12

definitely worth reading.

89:17

[laughter]

89:22

Uh, he added he did add something worth

89:25

me reading it. Um [laughter]

89:27

uh he actually wrote a techquickie about

89:30

dynamic surveillance pricing recently

89:32

and a former FTC chief technologist told

89:34

told us uh loyalty programs function as

89:37

data harvesting machines and how aware

89:39

are you of that in your day-to-day life?

89:41

Do you think our viewers should be more

89:42

conscious of it? And I I absolutely

89:44

think that's the case. This is our

89:45

discussion question. But I even actually

89:49

it's funny as as recently as yesterday

89:51

when I visited my orthodontist

89:53

they asked me hey are you do you are you

89:56

entering to win our draw do you collect

89:58

our points and I make a conscious

90:01

decision not to collect points pretty

90:04

much anywhere. The one exception is Air

90:07

Miles because we fly so much as a

90:09

company that do you remember that year

90:12

that our like entire

90:15

um uh Christmas party like

90:18

>> all the gift cards

90:19

>> giveaway thing was gift cards. Yeah,

90:21

>> that was that was basically

90:22

>> it wasn't the entire thing. It was just

90:24

a lot.

90:24

>> There was just a lot.

90:25

>> There was a ton.

90:26

>> Well, what happened was a

90:27

miscommunication. Someone asked me,

90:29

"Hey, should I blow all the air miles on

90:32

rewards for the Christmas party?" And I

90:33

was like, "Sure." I didn't know how much

90:36

it was.

90:38

[laughter]

90:40

>> That's a That's a fun way of

90:42

communicating that

90:42

>> stuff happens.

90:45

>> Yeah. So, okay. What I'm learning is I

90:47

just shouldn't have told you the cost of

90:49

the switches. I should have just said,

90:50

"We'll just need to buy a couple

90:52

switches." And then you would have been

90:54

like, "Yeah, sure." And then I could

90:56

have just had my budget approval. Now,

90:58

>> that is not the lesson from this.

90:59

>> I think that's the lesson.

91:00

>> That is nothing to do with this.

91:01

>> I think that's what I just learned. I

91:03

don't think so.

91:04

>> Just be more vague. Suggest it might not

91:06

be as bad as it is or in that case as

91:09

good.

91:10

>> A

91:11

>> that's the lesson.

91:12

>> Yeah, Bent Bob brings up a really good

91:13

point. Loyalty programs can also be

91:15

engines toy. Uh they can dial down disco

91:18

discounts to the lowest level that will

91:20

still appeal to you and then fine-tune

91:22

it so everyone's getting personalized

91:24

discounts that are just enough for them.

91:26

>> Yep.

91:27

>> Oh yeah, 100%.

91:30

>> Um

91:31

>> gross.

91:31

>> Do you want to pick one? Speaking of not

91:33

gross, New Mexico judge orders Meta to

91:36

pay nearly a billion dollars. I have

91:38

been tracking so far since the beginning

91:40

of the show. Have has the world spent

91:42

that much money on AI yet? Since the

91:44

beginning of the show, apparently no,

91:46

but over a tenth of it. So that's that's

91:48

cool. Um, Judge Brian Beadshed

91:54

by

91:55

>> Judge Brian,

91:56

>> Judge Brian [laughter]

91:58

wrote wrote in a judgment that Meta is a

92:01

cause of and has substantially

92:03

contributed to a public nuisance in New

92:05

Mexico and is required to abate that

92:07

public nuisance. Well said. Meta will be

92:10

required to contribute $567 million to a

92:13

fund meant to cover mental health care.

92:15

The remaining 375 million, no I did not

92:19

think 567 was close to a billion. The

92:21

remaining 375 million comes in the form

92:24

of civil penalties. Meta reported 60.8

92:28

billion in revenue and 15.85 billion in

92:32

net income in Q2 2026. So the combined

92:35

penalties work out to just under 6% of

92:37

quarterly income, which is

92:39

>> actually a lot given that this is just

92:41

one state.

92:42

>> A pretty big deal.

92:43

>> Yeah. And fun part, uh, look at this

92:46

from Lemonade Stand. It's like a bunch

92:49

of states trying to do this all right

92:52

now. And this one just worked. So, this

92:54

could be like

92:55

>> a huge deal.

92:56

>> Well, maybe does plan to appeal the

92:59

decision, of course.

93:00

>> Yeah. And I I wouldn't be surprised if

93:02

they get it.

93:02

>> Is it just me or like

93:04

>> But I think it's going to go through.

93:05

Sometimes the appeals process seems like

93:09

really important

93:11

>> and other times it's like

93:12

>> but other times does it just feel like

93:14

like gumming up the works like if you

93:16

want to appeal something

93:17

>> oh that's their goal.

93:18

>> Should it basically just like have to

93:19

happen right away like h how can we how

93:23

can we make

93:23

>> should you should you gain a reputation

93:25

as a company?

93:26

>> Yeah

93:26

>> for your appeals

93:27

>> like like Yeah. Should should it be kind

93:29

of like um uh uh uh yeah I mean okay I'm

93:33

I'm going to bring it back to badminton

93:35

for the second time in a single WAN

93:36

show. So you can actually appeal the

93:40

line judges call.

93:42

>> Okay.

93:42

>> So they they could call it in and you

93:44

could say no I I think that was out. You

93:46

put up your hand. Um but there's

93:48

actually um you can you can only do it a

93:51

limited number of times. And then I

93:53

believe in hockey you can ask for a

93:57

review and then if you fail there's a

94:00

there's like a downside. Yeah.

94:02

>> Hold on. Hockey that could be kind of

94:04

appealing a call.

94:06

>> I also think this should be an uneven

94:10

thing.

94:11

>> Your opponent gets a power play.

94:13

>> Oh wow.

94:14

>> If you're wrong.

94:14

>> I like that. I like that. Like that.

94:17

Like can we can we add a penalty?

94:20

>> If you are a company and you make over a

94:23

certain amount of money, there should be

94:26

penalties for appealing because you know

94:28

that that company's going to have an

94:30

army of freaking lawyers and stuff.

94:31

We're not talking about random

94:32

individuals. I don't necessarily want

94:34

this to be true for like random job

94:36

walking down the street.

94:37

>> Yeah. Like I I wasn't I wasn't there. I

94:39

couldn't possibly have robbed that

94:40

store.

94:41

>> I want to appeal. I don't want this.

94:43

>> Check your cameras.

94:45

>> Yeah. Yeah. I don't want this applying

94:46

to this person. I do want this applying

94:48

to multi-billion dollar mega

94:50

corporations that have armies of

94:52

lawyers. I do want this applying to

94:54

them. So, it's like and there's tons of

94:56

laws that work that way. So, it should

94:57

be a threshold thing. But, I would love

94:58

that. That would be fantastic. But yeah,

95:00

apparently this is like finally and man,

95:03

honestly, like what did it did it take?

95:05

Our generation getting to the point

95:07

where we're like very wellestablished

95:09

famous lawyers or something. I don't

95:11

know what, but like finally these

95:13

companies are being looked at as like,

95:15

oh, they might be causing insanely

95:17

severe amounts of damage to like our

95:20

youth and honestly just everyone and

95:22

maybe something should happen about

95:23

that. Like, thank God.

95:27

Um, horn player says appeals require a

95:30

bond and depending on the amount in

95:31

question that can be expensive enough to

95:33

deter, but I guess that plays perfectly

95:36

into what Luke is saying where if you're

95:38

a large enough company that that's just

95:40

like

95:41

>> who cares

95:41

>> cost of doing business then apparently

95:43

we need something more than that because

95:45

they just appeal

95:47

everything and it takes forever and then

95:49

by the time it's actually finally

95:51

settled you get like what five bucks and

95:53

a coupon for Starbucks. Like what are we

95:55

even talking about here?

95:57

Yeah. [cough] Yeah. I would like it to

96:01

benefit the the groups more as well. So

96:03

like, okay, Meta is required to

96:05

contribute 567 million to a fund to

96:07

cover mental health care. Okay. Well,

96:09

that should just be more. If they appeal

96:11

and fail, you're going to help mental

96:13

health care even more because I'm

96:15

assuming you're taking these groups back

96:16

into court, wasting more of their time,

96:18

wasting more of their money. Um, that is

96:23

you are making life worse for them

96:24

again. You should be on the hook for

96:27

that.

96:27

>> Oh, no. I I know horn player. You

96:29

weren't defending it. Just uh just

96:31

giving an example of an attempt that

96:33

[snorts]

96:34

just is not addressing the underlying

96:36

issue. I got it. I got it.

96:37

>> Yeah. Yeah,

96:38

>> you're good.

96:40

>> Um discussion question. Do you think

96:41

this judgment is large enough to be more

96:42

than just a cost of doing business? Uh

96:46

>> potentially times times many more states

96:48

and regions.

96:49

>> Yes, I think because this is just New

96:51

Mexico, I think it's uh pretty good

96:54

actually. You multiplied that out by all

96:56

the different states. My god.

96:58

>> Yeah. Now we're talking.

96:58

>> Now we're really talking.

96:59

>> Now we're talking maybe don't do that.

97:01

>> Now we're talking more than your income

97:04

for that quarter. It's just a quarter.

97:06

>> Yeah, I know.

97:07

>> My god. Um but but still we're we we're

97:10

start talking about real numbers once

97:12

you start getting up to to all the

97:13

different states that are involved with

97:15

this. Do we have any information on how

97:17

many different states? Cuz it's a ton of

97:18

them. I don't think it's every single

97:19

state.

97:20

>> I do think we have to move on though

97:22

>> because we have quite a few topics left.

97:23

Yeah, good call. I can't find it in

97:25

here. Anyways,

97:27

>> uh this is pretty cool. GeForce Now

97:31

exploit lets you access the full Windows

97:34

desktop through a simple file swap.

97:36

Moder Zortos, working with developer

97:39

D-pad Guy, found a way to escape GeForce

97:42

Now's game streaming sandbox and land on

97:44

the full Windows desktop of Nvidia's

97:46

cloud hardware. So, you launch the

97:49

free-to-play game Trove, click the

97:51

publishers's name to open the built-in

97:53

browser, navigate to the C drive through

97:56

the address bar, and swap the game's

97:58

executable with a replacement file.

98:01

Then, launch the game again, and instead

98:04

of Trove, you get a full Windows 11

98:06

desktop with taskbar, start menu, and

98:07

system tools. The ultimate tier machine

98:10

that Zortos landed on was running a 32

98:13

core AMD Epic, 56 gigs of system RAM,

98:16

and an RTX Pro 6000D with 48 gigs of

98:20

VRAM. So, that's the data center card

98:23

behind the tier that Nvidia Market says

98:25

5080 class performance.

98:28

Once inside, it's basically a rentable

98:31

high-end cloud PC for just [snorts] $20

98:36

a month. Zortos fired up LM Studio and

98:40

loaded a 31 billion parameter local AI

98:43

model [cough and laughter]

98:45

and was able to run it at a very usable

98:48

41 tokens per second. Others have

98:50

installed games outside of GeForce Now's

98:52

catalog like GTA 5 and D-pad Guy built

98:55

an open source tool called Salsa now to

98:58

streamline the whole thing. [laughter]

99:00

>> That's so sick.

99:01

>> Now, this does violate Nvidia's terms of

99:02

service and it could get your account

99:04

banned. It also only works on paid tiers

99:07

and session time limits do still apply.

99:08

Also, the results are inconsistent as

99:10

some users report sessions closing the

99:12

moment the file explorer opens and

99:13

Nvidia will almost certainly patch it

99:14

now that it's everywhere.

99:15

>> Yeah,

99:16

>> but I thought it was pretty funny and

99:19

given that Nvidia sells this exact

99:22

hardware for AI at enterprise prices and

99:24

someone just got it for 20 bucks a month

99:25

with a file swap, I thought it was a

99:27

pretty cool find. Um, good job. It's

99:30

hilarious. It also it also shows us how

99:34

much more we're paying for this stuff

99:37

just because of the AI we're running on

99:39

it versus what it's actually worth

99:43

cuz Nvidia can clearly operate GeForce

99:45

now profitably at $20 a month.

99:48

>> It's crazy crazy.

99:55

>> Speaking of I don't even know. Uh, Intel

99:58

has filed a patent application for some

100:01

space data center concepts that would

100:03

use a few powerful satellites higher up

100:05

to manage thousands of simpler low Wait,

100:08

what? Really?

100:10

Why?

100:12

I mean, isn't that the question?

100:18

Okay. The the higher like literally boss

100:22

satellites. [laughter]

100:24

It's it's like their office is at the

100:26

top of the tower like Okay. What are we

100:28

talking about? Would handle routing

100:30

coordination and other heavy computing.

100:33

Maybe maybe it gives you like more line

100:35

of sight. I any uh reducing reliance on

100:39

groundbased processing and ground

100:41

stations. The goal is cheaper, simpler

100:43

satellites and a more scalable network.

100:45

But this is still a patent concept, not

100:47

a sign Intel is building it. On top of

100:49

that, Intel's latest filing is a

100:51

continuation of a patent application

100:53

that dates back to 2022.

100:55

So, it's not exactly a brand new Intel

100:57

project.

100:59

>> Are we um

101:00

>> are are we still not over the whole the

101:03

whole data centers in space thing?

101:05

>> No.

101:06

>> Okay.

101:07

>> Yeah, that's I think that's going to be

101:08

going on for a little while.

101:10

>> Okay, cool. And then when when are we

101:13

going to get bored of that? I don't uh

101:16

whenever you know the next earnings call

101:18

for some company Elon's involved in is

101:21

and then he has to come up with some

101:22

other new thing that he's doing now

101:23

>> cuz like I like what like like data

101:26

centers in the Kyper belt like what's

101:28

the

101:28

>> I like that like what's

101:29

>> that's pretty good.

101:30

>> Yeah. But like what

101:31

>> data center asteroids?

101:32

>> Yeah. How do we how do we just what's

101:35

the new shiny that we can come up with

101:37

so that we can move on from this

101:43

>> data center asteroid sounds pretty fun.

101:45

>> Yeah, data. I mean data center on moon

101:47

even.

101:47

>> What if we made data moon?

101:49

>> Data moon.

101:50

>> So we have two moons.

101:52

>> Oh, that's no moon. It's a space data

101:56

center [laughter] station.

102:03

Um, our discussion question here is,

102:05

"What concerns do you have about putting

102:07

thousands and thousands more satellites

102:09

into orbit?" Space junk is already a

102:13

significant problem. And the more space

102:16

junk there is, the more pieces of

102:18

shrapnel can go zinging through

102:22

something and cause more fragments,

102:26

which is going to cause more data center

102:29

junk, which is going to make it even

102:30

harder.

102:31

>> Uh like like Earth's

102:34

I didn't really think about it until I

102:36

was I was out of the city recently for

102:39

some reason or other and I was somewhere

102:40

where you can actually see the sky. Um,

102:44

man, I couldn't believe how many

102:46

satellites there were. Like I It hasn't

102:49

I haven't Look, it's overcast here 160

102:52

days out of the year, okay? It just

102:53

doesn't come up that often. But like

102:55

when I was a kid, I lived in in a rural

102:57

area and I used to just, you know, lie

102:59

out on our trampoline of all things and

103:00

like I would look up at the sky and uh

103:04

and there weren't so many satellites.

103:07

>> Yeah.

103:08

>> Yeah.

103:08

>> Yeah. Sounds about right.

103:10

>> There's a lot now. There's pictures from

103:12

the Aremis 2 mission that people took

103:14

from inside the the Orion capsule back

103:17

towards Earth. And there's there's one

103:18

particular series where he was trying to

103:20

just take a bunch of photos to kind of

103:22

see if he had set things up correctly.

103:24

And you can see kind of a time lapse,

103:26

not because he was intentionally taking

103:28

time-lapse photos, but he spammed with

103:29

the shutter a bit. You can see a time

103:31

lapse and you can see a decent amount of

103:33

the satellites kind of sparkling as they

103:37

move a little bit around the Earth. And

103:38

it's like, oh, that's a that's a lot of

103:41

them.

103:42

>> Yeah. So, uh, float plane chat is

103:44

talking about Kesler syndrome.

103:46

>> So, just all this high-speed crap

103:48

ripping around the Earth at tens of

103:50

thousands of kilometers an hour and just

103:53

shredding anything that tries to pass

103:56

through it. Um, if we want to be an a

103:59

multi-planetary species, then we

104:02

actually do need to not put too much

104:04

garbage in orbit around the Earth.

104:05

>> A bit of a problem. There's also stuff

104:07

where it theoretically [clears throat]

104:08

a lot of it will slowly kind of de-orbit

104:10

itself because it's not gaining any more

104:13

>> eventually.

104:14

>> Yeah, can take a long time. But

104:16

>> e gadget guy says there's so much room

104:18

up there. It's not as bad as you think.

104:20

>> But that's the attitude that got us to

104:24

the point where we have the Great

104:26

Pacific garbage layer. Dude,

104:27

>> there's so much ocean.

104:29

>> The ozone layer is so thick.

104:31

[clears throat]

104:31

>> It's so thick and beautiful. Nothing

104:33

will ever hurt the ozone layer. There's

104:36

so much There's so much uh stuff to heat

104:40

up.

104:41

>> Yeah.

104:41

>> You don't have to worry about global

104:43

warming.

104:43

>> Yeah.

104:44

>> There's so much to absorb it. The Earth

104:46

is huge.

104:47

>> Yeah.

104:48

>> What are you even talking about?

104:49

>> Should be fine.

104:50

>> Yeah.

104:50

>> What's one smoke stack?

104:51

>> Yeah.

104:52

>> Don't worry about it.

104:52

>> Yeah.

104:53

>> So, um yeah, this will

104:55

>> ocean absorbs that stuff anyways.

104:56

>> Yeah.

104:56

>> What are we talking about?

104:58

>> Yeah.

104:59

>> Uh

104:59

>> burn more jet fuel.

105:01

>> Yeah. I mean, one of the huge problems

105:03

with this is like you you think data

105:05

centers are expensive now,

105:07

>> right?

105:07

>> What about when you have to launch them

105:09

into space?

105:09

>> Yeah. What are we like what this is the

105:11

thing that's like killing me about this

105:12

whole idea is what are we even talking

105:16

about?

105:17

>> Well, we're talking about infinite

105:18

budgets for AI. So, if your budget for

105:20

AI is infinite and all of the people on

105:22

Earth is going to ban you from building

105:23

data centers, you're just going to

105:24

launch them into space. I think that's

105:26

like the real argument is you know what

105:28

New York you know what Texas you know

105:30

what everywhere else that is putting

105:32

memorandums on building data centers and

105:34

whatever uh would you go put them in

105:37

space

105:38

>> but then the the but it's even it's even

105:40

stupider than that though Luke because

105:43

you could also put them in Antarctica

105:46

and then all you would need is

105:47

communication satellites

105:49

like just and then that takes care of

105:52

the cooling.

105:52

>> I don't think you can but I understand

105:54

your point. Northern Canada, whatever.

105:56

Yeah, I understand your point.

105:57

>> But it just Yeah. I mean, that's that's

106:00

why uh that's why what's their name?

106:01

That's why Hner has their uh has their

106:04

second like main data center in like the

106:06

the Nordic region cuz they're just like,

106:08

"Oh yeah, we can just air cool

106:10

everything for like most of the year."

106:13

[laughter]

106:15

And I I pretty much guarantee you that

106:18

within reason, anything terrestrial is

106:20

going to be better than sticking it up

106:21

into space. Um, okay. That's that's not

106:25

necessarily true, actually.

106:27

>> Ben 34 said, "I'm surprised you guys

106:29

haven't mentioned latency being a

106:30

problem. It's actually a lot less of a

106:31

problem than you think."

106:32

>> Yeah,

106:33

>> that's why Starlink is like actually

106:34

pretty effective.

106:35

>> Yeah, it it's not the it's the latency

106:39

is I think I think is a really easy

106:42

gotcha, but it's also really easy to

106:46

counter because

106:47

>> tasks

106:48

>> for Yeah, exactly. It's good enough to

106:51

return a response to, you know, how much

106:53

wood could a woodchuck chuck if a

106:54

woodchuck could chuck wood? Like, it's

106:56

fine for that.

106:56

>> Great question.

106:57

>> And I

107:00

to me the bigger the bigger issues are

107:03

cooling. Not that there's no way to cool

107:06

them, but just that we're talking about

107:07

these like colossal like here here's the

107:10

thing.

107:12

out of the two sides of their mouths,

107:16

the same the same AI

107:21

drum beaters will tell you that AI data

107:26

centers in space is a great idea and the

107:29

cooling is a solved problem. And then

107:32

also

107:33

be building data centers that are like

107:38

so with with an atmosphere like with

107:41

actual other molecules to bump up

107:44

against and transfer thermal energy to

107:47

that cooling these things is a huge

107:49

challenge and really costly and really

107:52

energy intensive and and really

107:55

difficult. And it's like okay

107:59

which one is it? because it can't be

108:01

both. If you're talking we're using

108:04

super energyefficient hardware in orbit

108:07

that that we can cool with relative ease

108:10

then that's fine. we, you know, we have

108:12

these like giant radiation arrays or

108:14

radiator arrays, right? If you're

108:17

talking the kind of energy density, like

108:19

I was reading on Serve the Home, uh,

108:21

apparently the guess for AMD's latest

108:25

Instinct uh, GPU, or is it even a GPU?

108:28

AI processor, the estimate that Patrick

108:31

had was 2,000 watts each. What What was

108:35

the total thermal dissipation capacity

108:38

of the ISS?

108:40

It's not much

108:44

um

108:47

according to the AI overview. So take

108:51

make that make of that what you will.

108:53

It's about 70,000 watts.

108:56

>> There's also some really

108:57

>> So what 35 GPUs

109:02

with? Have you seen their Have you seen

109:04

their dissipation array on the ISS? It's

109:06

huge.

109:10

And that doesn't even account for any of

109:12

the supporting infrastructure for those

109:13

GPUs. It's just it's just

109:16

asinine.

109:17

>> I That's actually something I'm not

109:19

super aware of is the dissipation array.

109:21

I know they have the crazy solar panel

109:23

array.

109:23

>> It's big.

109:24

>> Yeah, I believe it.

109:25

>> It's big

109:25

>> because I know I know this one. I I I

109:27

looked this up, but like the the

109:29

temperature

109:30

>> swings of the ISS can be really really

109:32

wild. Yeah. Um, like the the shady side

109:36

can be like negative a huge amount of

109:38

degrees and the the side facing the sun

109:40

for a really long time can be like this

109:42

the surface surface temperature uh can

109:45

be just extremely hot. Um,

109:50

>> yeah, it's wild.

109:51

>> Um,

109:51

>> wild wild.

109:52

>> So,

109:53

>> it's it's a fun uh thought experiment.

109:55

>> Yeah, Magi 999 goes, "You have 2000 watt

109:57

GPUs?" I don't. Uh, but here, hold on.

110:02

What is it? the MI uh 355X. Is that the

110:05

Is that the new one that's just coming

110:07

out? Like just announced?

110:10

No.

110:11

Well, dang it. What's the What's the new

110:13

one? Newest AMD in This was like just

110:17

this week.

110:19

New Instinct announcement. MI455X. Yeah,

110:22

Patrick from Serve the Home estimated

110:24

that it could be as high as 2,000 watts

110:26

per GPU. Like, this is what I'm talking

110:28

about. And there is no sign whatsoever

110:32

that the the energy use of these

110:35

processors is going down. If anything,

110:38

the entire industry is looking at them

110:40

going this number is going up. They're

110:44

talking about they're talking about

110:45

putting micro channels into the dyes to

110:48

get liquid into the dyes to take

110:51

the heat away. And we're talking, oh,

110:53

I'm going to use a solar first heat

110:55

radiation array.

110:58

What the are you talking about?

111:00

[laughter]

111:03

And and to be clear, I'm not saying you

111:05

couldn't do it. I'm not saying you

111:08

couldn't do it. I'm just saying

111:11

obviously it is cheaper to do it on the

111:14

ground.

111:16

Obviously.

111:19

So, okay, good luck with that. Uh WY

111:22

giraffe says it's an engineering

111:23

problem, but it's it's an engineering

111:25

problem that there is no need to solve

111:29

because there's actually a lot of

111:32

terrestrial places that don't have a

111:35

data center yet.

111:38

And like look, I know that one one of

111:40

the arguments is that you've got solar

111:42

power that is unimpeded by the

111:43

inconvenient atmosphere down here,

111:47

right? Solar power solar power is good,

111:49

right?

111:49

>> Solar power is really good. But that's

111:50

yet another giant array that you're

111:53

going to have to have on this thing. And

111:54

like again, literally like how how

111:58

sorry, how big it does the solar power

111:59

need to be to do my to do my uh my 35

112:03

GPUs to do 70,000 watts. So these things

112:06

are like they're basically going to look

112:09

like TIE fighters. You have your little

112:11

you you'll have your little data center.

112:12

You'll have your like dissipation.

112:14

You'll have your power array.

112:17

They just want the lawlessness of space,

112:19

says Crow.

112:21

[laughter]

112:23

I mean, I wouldn't be surprised if

112:24

that's part of it.

112:25

>> Yeah.

112:27

Ah,

112:32

I think I think

112:33

>> slop fighter. [laughter]

112:36

[snorts and gasps]

112:38

Yeah.

112:40

Evacus says, "Why not build a data

112:42

center on a mountain?" Like this is what

112:44

this is exactly what I'm talking about.

112:46

Yes, we could solve the problem of data

112:49

center in space, but why don't we just

112:52

use those resources to solve the problem

112:54

of data center in the Arctic or data

112:55

center on a mountain? Like there's there

112:58

are other places we could go that we do

113:00

not need a rocket ship to get to. They I

113:02

think they just want to fly their rocket

113:04

ships.

113:04

>> Oh yeah. No,

113:06

>> that's why I think it's mostly just a

113:08

fun question.

113:09

>> Yeah. Cuz if if you take it not super

113:11

seriously, but then the more seriously

113:13

you take it, the kind of fun it is at

113:16

the same time. And I think this is a

113:18

very like San Francisco kind of thing

113:21

where like you want to

113:22

>> think even bigger.

113:23

>> Yeah. Accelerationism, whatever. Like

113:25

let's just have fun with this thought

113:26

experiment and then maybe we do do it.

113:28

Who knows? But I I think you know

113:30

running power and internet

113:32

infrastructure to somewhere in like

113:35

northern BC or something or the Yukon or

113:38

whatever um Northwest Territories data

113:42

centers um makes more sense to me than

113:48

flying rockets for huge sets of hardware

113:52

infrastructure and then dealing with um

113:56

oh you know a chunk of some other data

113:58

center just ripped ripped off 12 of your

114:00

solar panels and now your data center

114:02

just shut down. That sucks. Time to

114:05

launch a rocket about it. Um, [laughter]

114:08

like what are we talking about, man?

114:13

What are we talking about? And then

114:15

you've got uh you've got uh Russia

114:17

showing off their new uh their new beam

114:20

thing that can shut down SpaceX

114:22

satellites. So, that'll be convenient.

114:25

>> Yeah. Yeah. Everyone should have known

114:28

things like that were going to be a goal

114:29

the second everyone on the planet saw

114:33

how incredibly important SpaceX is to

114:35

the war in Ukraine. Like this the second

114:38

you saw that the second you saw how

114:40

impactful it was cutting Ukraine off of

114:42

it

114:43

>> both

114:44

>> the second you saw I was just going to

114:46

say that as the second one. Uh the

114:47

second you you saw how impactful cutting

114:50

both sides off from it um was then you

114:54

know that's a massive bat signal to

114:55

Russia that's a massive bat signal to

114:57

China to make sure that they have some

114:59

ability to shut this thing down.

115:03

>> Cool.

115:04

>> I mean even Sputnik

115:06

like the Americans were like oh god what

115:08

is this weapon of war?

115:10

>> Yeah.

115:11

>> Like

115:11

>> and I mean kind of I don't blame them at

115:15

the height of the Cold War. Not at all.

115:17

Yeah.

115:18

>> Bernardo Baroo says political wow. This

115:21

is not political. This is just a fact.

115:23

>> It's a fact that when

115:26

uh a potential threat emerges and is

115:29

very it's very obvious to everyone how

115:31

threatening it is that someone will try

115:34

and figure out how to counter said

115:35

threat. That's just, you know, that's

115:38

just a thing that happens.

115:40

>> There have also been lots of ideas um

115:44

for

115:46

things in space that haven't become a

115:48

thing. Uh even though they're they might

115:50

be an an interesting thought experiment.

115:52

Have you never heard of this before?

115:53

>> I heard Rods from God. [laughter]

115:56

>> Project Thor. Just huge

115:59

dumb I'm saying dumb because there's no

116:02

I as far as my understanding goes,

116:03

there's no electronics on them or

116:04

anything. They're just tungsten rods uh

116:07

they put in space. Then the idea would

116:08

be that you just kind of drop it

116:10

>> and then it's a tungsten rod coming from

116:12

space.

116:14

It just needs to kind of whack into

116:16

something and it's gone. That was that

116:18

was the idea. And

116:20

>> this just turns whatever you don't like

116:21

into a crater.

116:23

>> Yeah.

116:23

>> Okay.

116:24

>> Kinetic bomb. Someone said that in chat.

116:26

That is I think generally what they're

116:27

referred to. The problem is is it's a

116:30

giant tungsten rod.

116:32

>> Well, you just put little pieces at a

116:33

time up there. You [laughter] need

116:35

>> assemble them up there.

116:38

Uh yeah, it was a it was um it's a very

116:42

old idea and it was ditched a long time

116:45

ago because it turns out things like

116:48

ICBMs are actually just a lot more

116:51

reasonable than a rack of massive

116:54

tungsten rods that you just keep in

116:55

space for.

116:56

>> Do we need veritassium to be the one to

116:58

tell us that data centers in space are

116:59

not not economical and will never be

117:01

economical?

117:02

>> They won't listen to him.

117:03

>> I I know

117:04

>> they're not going to listen to anybody.

117:06

[laughter]

117:07

I can I throw a weird take out there?

117:10

Sure.

117:10

>> I think someone's going to do it to some

117:12

degree.

117:12

>> Oh, yeah. Oh, there's no question.

117:13

>> I would even be surprised if there was

117:15

like uh little practice versions.

117:19

>> Yep.

117:20

>> Like a like a Jeff Gearling style. I'm

117:23

going to make my data center, but it's

117:24

going to be out of like cube stats.

117:26

Cube. Yeah. Cube stats. There we go.

117:28

>> There are even ways that it could make

117:30

sense. Like that's something I'd like to

117:32

make very clear right now is I'm not

117:34

even anti- progress or like like I I

117:38

hate the idea like um a perfect example

117:41

of something that it could be useful for

117:43

is having some data storage sitting on a

117:46

Starlink satellite. Uh if I was someone

117:48

like a Netflix for instance, could there

117:51

could there be a benefit to putting my

117:54

data closer to the customer? Yeah. I

117:57

mean, remember uh remember

117:58

>> be dude the the the some of this makes

118:00

fun pivot is some of this makes sense

118:03

pivot is going to be a really fun part

118:04

of this conversation

118:05

>> because like uh do you do you remember

118:07

when Netflix first started rolling out

118:10

uh boxes with their entire library

118:12

closer to the home?

118:13

>> Yeah. Yeah. Yeah. They're really cool.

118:15

>> Yeah, they're they're super cool. And

118:17

the Netflix film library is actually not

118:19

very big.

118:20

>> Yeah. So, you know, if every Starlink

118:24

satellite had, you know, a dozen hard

118:27

drive bays on it and someone wanted to

118:31

uh someone wanted to cache important

118:33

data closer to these rural customers, I

118:36

could see something like that making a

118:38

ton of sense. Maybe not actually for

118:40

Netflix because um the latency doesn't

118:44

really matter. So, cutting down on, you

118:46

know, one trip might not be, but I I

118:49

don't know. maybe it would make sense

118:50

whatever is a high super high percentage

118:53

of your traffic is going to make sense

118:55

to minimize how many hops it takes and

118:57

so yeah or or even having a small amount

119:01

of compute there like we're we're going

119:03

to I hope we get to the point where I

119:06

can just have my AI be on device but if

119:08

we don't get there and if you know

119:11

satellite cellular phones become more of

119:15

a thing which I'd be super into because

119:17

better coverage why not I can open my

119:19

freaking garage door. That would be

119:21

great, right? [laughter]

119:22

Um,

119:24

so if if you had to put a little bit of

119:25

compute on the satellite instead of, you

119:29

know, on the data center that it then

119:30

has to like hop over to.

119:32

>> That's not a data center though.

119:34

>> That's not a data center. But that could

119:37

help. So you have a little bit of I

119:39

think you could make the argument. I

119:40

think what's going to happen is they're

119:42

going to move the goalposts

119:44

to well every Starlink satellite has

119:47

some storage and some compute on it.

119:49

>> We've had a data center in space this

119:51

whole time.

119:51

>> Now we're data center in space company.

119:54

Now we're worth two trillion

119:55

>> when it when it comes to space data

119:58

centers they don't have to all be

120:00

physically linked together.

120:02

>> I mean Starlink sat

120:04

>> optical layer is more effective in

120:07

that's is what I'm saying. Optical layer

120:08

is more effective in space anyways. I

120:10

could see that goalpost move happening.

120:11

Yeah. And honestly, that makes a lot

120:13

more sense. It was mind-blowing to me

120:15

that, and I think I've talked to you

120:16

about this before, but when the ISS was

120:17

talking to the Orion shuttle, so Artemis

120:19

2, basically everyone watched it. When

120:21

when they were moving around, they had a

120:24

call with the ISS.

120:27

I just assumed because I know they have

120:30

optical networking that this would be

120:32

like pointtooint to both of them.

120:34

>> Have I read the same article you did? I

120:37

think

120:37

>> I didn't read an article on this. Oh, I

120:38

I came across something that talked

120:40

about this though.

120:41

>> So, I assumed it was going to be really

120:43

fast. It wasn't. There was like 6

120:44

seconds of delay, but I was trying to

120:46

figure out, okay, what's happening here?

120:48

Cuz where is 6 seconds coming from? Cuz

120:51

if you look at the like light time to

120:53

get to Earth, so okay, let's assume it's

120:55

going to Earth and then back up. If you

120:57

look at the light time involved with

120:59

that, still shouldn't be 6 seconds. 6

121:02

seconds is a lot and it was a very, very

121:05

big delay. Um, turns out they're still

121:07

using like stuff from uh I'm going top

121:12

of dome right now, so I'm not really

121:13

going to remember, but it's like the 70s

121:14

or 80s. Like it's deep space network

121:17

from like a bajillion years ago, and

121:20

it's like mxing in processing time is

121:22

like a massive part of that equation. So

121:25

I could see,

121:26

>> you know, okay, we want a freaking base

121:28

on the moon, bros. We might need to, you

121:31

know, update some of our infrastructure

121:33

maybe for once. Maybe the NASA might

121:35

need more than4% of the US budget. Um,

121:39

and they they update things like the

121:41

deep space network so that we can maybe

121:43

communicate within space and then bounce

121:46

things back down, etc. Uh, and that

121:49

might need more of these types of

121:51

things, but I still don't

121:53

feel like I'm looking at

121:55

>> data center,

121:56

>> what most people think of as a data

121:57

center, and then I'm just copy and

122:00

pasting that, but the the layer is now

122:02

space instead of ground. You won't hear

122:03

an argument from either of us that there

122:06

should be no compute in space.

122:08

>> Yeah, there already is.

122:09

>> There already is and there should be

122:10

more.

122:11

>> Probably.

122:13

>> But

122:14

data center has a very specific

122:19

maybe not even meaning, but it

122:20

definitely has a very specific

122:22

connotation.

122:23

>> And some of the pictures and stuff that

122:24

people have released from some of these

122:25

brainstorming sessions are what you're

122:27

thinking of, which is like massive

122:29

physical things.

122:30

>> Very stupid.

122:31

>> Yeah. Um,

122:33

>> but some distributed compute network

122:34

that is effectively like Starlink 2 or

122:37

something like that.

122:38

>> Sure,

122:38

>> could totally make sense.

122:40

>> Um,

122:44

>> they did say they wanted to build

122:45

satellite internet over the moon before

122:47

we sent people to live there. Oh yeah,

122:49

there's tons of stages to Artemis and

122:50

stuff. I just

122:53

Yeah, I just be talking.

122:55

>> All right. Tour, a nonprofit

122:57

organization focused on protecting

122:59

people's privacy and helping them bypass

123:00

internet censorship, has a new Android

123:02

app called Snowflake Volunteer.

123:06

This is pretty cool. It's kind of like

123:10

the uh the news recently where major TV

123:14

brands got busted um for having crappy

123:19

malware apps in their stores that were

123:22

basically turning your TV into an

123:25

endpoint for like a a never do well um

123:29

like proxy network

123:31

>> except the good version. This is the the

123:35

Jedi version of that monstrosity.

123:39

>> So, it's called Snowflake Volunteer, and

123:41

it lets you share your internet

123:43

connection to help users in censored

123:46

countries get online.

123:48

>> Snowflake works by disguising censored

123:51

users tour traffic so it's harder for

123:53

governments or network operators to

123:55

identify and block it. Connections are

123:58

routed through short-lived volunteer

124:00

operated proxies. A broker will

124:03

automatically find another proxy if one

124:05

is blocked. Uh here's an even simpler

124:07

summary. It's pretty much like an

124:09

internet relay system that helps folks

124:11

get on the internet around censorship of

124:14

their connection. The app makes

124:16

volunteering pretty easy. Users can

124:18

leave it running in the background and

124:19

share some bandwidth without having to

124:21

set up complicated infrastructure. Daily

124:24

unique proxy IPs have already risen from

124:26

about 1,300 in May to 1,700 in June, a

124:29

29% increase according to tour. It's

124:32

available on Google Play and Foid and

124:34

its source code is publicly available

124:37

which is pretty cool. Obviously, when

124:40

you're deciding to trust a technology

124:41

like this, um, you know, I can't

124:46

personally say that I've poured over

124:48

every line of the source code, but it

124:50

being open source and nonprofit. And I

124:54

mean, to I mean, I've heard of them, I

124:56

think, contributes to a greater degree

124:58

of trust that I would have in a project

124:59

like this. And it seems like it's a

125:02

pretty uh pretty noble cause compared

125:04

to, you know, using your internet

125:07

connection to collect more more data and

125:09

sell more advertising and just generally

125:13

be crappy and like coordinate bot

125:15

attacks and that sort of thing.

125:16

>> Yeah.

125:17

>> Uh hey, speaking of

125:19

um third party app stores, Google Play

125:23

Store starts distributing thirdparty app

125:25

stores. That's happening after a long

125:27

court battle. Google is opening the Play

125:28

Store to third party app and game stores

125:30

in the US. They will only be accessible

125:33

via links or direct searches. And Google

125:36

says they will enforce their policies

125:38

from the Play Store on these third party

125:40

stores, but will not review every app in

125:43

these stores, nor will they handle

125:45

updates or payments. Installing a store

125:48

is basically like installing any other

125:50

Android app. And the first third party

125:52

store, not an endorsement, haven't even

125:54

looked at it, is called Apptoid Games.

125:59

Okay,

126:00

>> never heard of that.

126:01

>> 95 Google reported that the only

126:03

additional step when using a third party

126:05

app store is accepting the install

126:06

prompt after the app is downloaded.

126:09

These changes come as Google prepares to

126:11

lock down the Play Store to verify

126:13

developers only. And it's not yet clear

126:15

how third party stores will or won't be

126:18

impacted by the new developer rules. You

126:21

know, if there's one thing that Epic did

126:25

that made me happy, it was finally

126:28

addressing the absolute garbage fire of

126:32

App Store and Play Store policies on iOS

126:35

and Android respectively. They were it

126:38

was it was it frustrated me more than it

126:42

probably should have. Um because to me

126:46

it was very very very very very

126:48

painfully obvious that there was

126:52

monopolistic antitrust stuff going on

126:54

with both of them. watching people just

126:59

try so hard to lick the boots of these

127:02

multi- in some case trillion dollar

127:05

corporations that they ended up lodged

127:07

in their rectums. Like I just can't it

127:10

was it was so hard for me to watch

127:12

people defend Apple and Google's

127:14

behavior around this stuff. Um, and it's

127:17

just it's glad it makes me extremely

127:19

glad that some semblance of regulation

127:24

still works a little bit. Um, I'm not

127:27

sure if it actually made it into the

127:29

dock for this week, but um,

127:32

Apple is still being kind of a little

127:36

about their ruling and they're

127:38

they're trying to kind of drag their

127:40

butts and stuff, but but the change is

127:42

happening. The change is happening.

127:44

>> [clears throat]

127:45

>> Uh, speaking about changes, a bit of a

127:47

shorter show. Did we announce these? We

127:49

did, right?

127:50

>> No.

127:51

>> Oh,

127:52

>> I think it's time for those.

127:54

>> We really should.

127:55

>> All right. This week, we've got so many

127:59

new products on LT Store. And if you

128:01

guys want to get a if you guys want to

128:02

get a checkout message in, you're

128:04

probably going to want to do it pretty

128:05

pretty pretty like quick styles here.

128:07

Uh, the foundation collection. These are

128:10

I I actually can't show the store and

128:12

show my notes at the same time. So, you

128:13

want it? Oh, he's on it. Um, basically

128:16

it's a three-piece lineup of everyday

128:18

staples with something for active days,

128:20

something for laidback days, and

128:22

something for those days when you want

128:24

to look a little more put together. If

128:26

you're an active person and need a solid

128:28

foundation piece for your closet, the

128:30

quick dry sport t-shirt is the one to

128:32

grab. Uh, I've been actually using it

128:34

very regularly for my badminton

128:35

sessions. And the moisture wicking,

128:37

quick drying fabric keeps up, while the

128:39

reglin sleeves give your arms more room

128:41

to move, helping reduce underarm

128:43

chafing. Uh, that's a big thing for me

128:45

when I'm moving my arms around. Um, I

128:48

need really good mobility all the way

128:51

up, all the way down, all the way

128:52

forward, all the way back. And there's a

128:54

little bit of stretch to the fabric as

128:55

well. It's really nice.

128:57

>> Yeah. For a more put together look, the

129:00

halfzip sweatshirt is made with our

129:02

comfortable French terry fabric. Same as

129:04

our blank hoodie. It has a half zip, a

129:07

Kangga pocket, and a hidden phone pocket

129:09

because apparently one pocket is not

129:12

enough.

129:13

>> Oh, put together. This guy looks

129:15

>> actually looks pretty looks pretty

129:17

sleek.

129:17

>> Looks pretty put together.

129:18

>> Looks pretty good. The minimal hoodie is

129:21

a more elevated take on the everyday

129:23

hoodie with a sleek, clean look, premium

129:25

feeling fabric, stretch for comfort, and

129:28

a draw cordfree hood. You can check out

129:31

the Foundation collection at

129:32

lmg.gg/foundation.

129:36

But wait, there's more. We've also got

129:39

this month's graphic tee launching, and

129:41

it's a nostalgic one. The can I help

129:44

t-shirt is a little trip back to the

129:46

days when your computer's most

129:48

persistent co-orker was a paperclip that

129:50

offered unsolicited advice. It's printed

129:52

on our classic polyblend te so you get

129:55

the nostalgia without having to dig out

129:57

office XP. You can check it out at

129:59

lmg.gg/

130:00

can I help. This one was really intended

130:03

for the folks out there who

130:08

are asked so often for tech support. I

130:10

just I just I I I imagine I imagine the

130:13

IT guy at his desk just with one with

130:17

one of these. It just it says everything

130:18

you need to say. It's like, "Yeah, I

130:21

don't really want to help you." And

130:22

honestly, I might not end up being that

130:24

helpful, but uh it's my job. So,

130:27

[laughter]

130:29

[snorts]

130:31

>> okay. We had some uh we had some way

130:34

more we had some way darker copy that

130:37

potentially went with like a like a

130:38

clippy design at some point but

130:41

ultimately decided to to go with

130:43

something more helpful.

130:44

>> Yeah. Um,

130:47

like it was I I think we had one that

130:49

was going to be like, "Have you have you

130:50

considered just ending it or something

130:52

like that?"

130:54

>> Like a dark clippy.

130:55

>> Oh, have you considered end task

130:58

>> or Yeah, it's bas basically. Yeah, they

131:00

it was uh or like have you considered

131:02

powering off or like something like

131:04

that. We had we had something along

131:05

those lines.

131:06

>> Alt F4 for yourself. [laughter]

131:08

>> That might have that might have been

131:09

thrown around. We were we were like we

131:11

were in a mood.

131:12

>> We were in a mood. Seems like it. I

131:14

think you guys made the right choice.

131:15

>> Yeah, we went with can I help?

131:17

>> No, they No, they didn't. [laughter]

131:19

>> No, we we I think we made the right

131:21

choice.

131:21

>> Didn't make the right choice for

131:22

>> make both choices.

131:24

>> Um

131:25

>> and then float plane.

131:26

>> Oh, right. Yes, float plane. Uh

131:29

>> which I need to read.

131:30

>> Oh, okay. Loop.

131:32

>> Next week is Lionus's birthday.

131:34

>> It says watch video

131:35

>> and to Yeah. Should we load that up?

131:38

>> I suppose.

131:40

>> Do we watch that first and then I read

131:42

it?

131:43

We need audio.

131:44

>> Oh, ready, Dan?

131:45

>> Yeah.

131:46

>> Why don't we get audio today?

131:48

>> Uh, there was something that happened

131:51

last week as well. The issue with the

131:53

internal routing was left on again.

131:56

>> Okay.

131:56

>> Um, and so the the submix breaks

132:00

unfortunately, but they can hear it.

132:01

>> Good chat. All right.

132:02

>> Okay. Hello.

132:03

>> Next week

132:04

>> is my birthday. Celebrate with me.

132:07

>> Sorry. Sorry. Sorry. This is on me.

132:10

>> Hello. Next week is my birthday.

132:13

Celebrate with me

132:15

>> at lmg.gg/unklinus.

132:19

>> I was wondering what that was.

132:21

[laughter]

132:22

>> Even once I was crossing the border, the

132:24

border guy was like, "Hey, did Lions fix

132:27

your bed?"

132:28

>> You guys really put that on the

132:29

building? Does Ivon know about this?

132:31

>> Yeah, she saw it.

132:32

>> There was a fully nude on Anson in the

132:34

building. Lionus could have gone with me

132:36

but didn't want me to see his dong even

132:39

though I lived in his house and heard

132:40

him do many things for years.

132:42

>> Oh yeah, that poster. Dan called you

132:44

that.

132:45

>> You are 40. Oh, and oh sorry. Yeah.

132:50

>> Oh my goodness. I didn't know you

132:52

already reacted to those videos. But

132:53

anyways, uh next week is Lionus's

132:55

birthday and to celebrate we have four,

132:56

that is right, four videos compared to

132:59

our normal two. Sammy says that this is

133:02

because he's a narcissist. uh the first

133:05

even though it was it was it says it's

133:07

because he's a narcissist, but Sammy

133:08

made the decision. So,

133:10

>> yeah, I had nothing to do with any of

133:12

this

133:12

>> literally at all. So, I I don't know

133:14

what that says about Sammy, but hey, you

133:16

know, uh the first video has been out

133:18

since yesterday. Uh it's the raw video

133:20

of creators sharing memorable moments

133:22

with Lionus. Video submitted by former

133:24

LMG employees, JerryRig Everything,

133:27

MKBHD, and Level One Tech, and more.

133:29

Apparently, we're just saying their

133:30

channel names, not the their names, but

133:33

hey, you know, that works. Um, good job,

133:36

Sammy. Uh, Monday, we'll have Lionus's

133:38

reaction of the video, and there's a bit

133:40

more uh cringe in that than not,

133:43

especially the fact uh that he would not

133:45

show me his dong when we were in Japan.

133:46

That is very cringe. Um,

133:50

>> Wednesday, [laughter] we will showcase

133:52

the LMG rebranding. Uh, what? And end of

133:56

the day. Sam Sammy's notes are so hard

133:58

to read. I make it I make it look

134:00

relatively easy. Better than his team's

134:02

messages.

134:03

>> Uh Wednesday we will showcase the LMG

134:05

rebranding and the day after which is

134:07

Lionus's actual birthday.

134:10

>> Thanks. Uh will be a community gaming

134:13

session with Lionus from 3:30 to 5:30.

134:16

Really? Uh to really thank everyone for

134:18

supporting Lionus and LMG. Not 100%

134:20

locked, but we'll be playing a mix of

134:21

Tape Tape, Halo Infinite, Pico Park, and

134:24

Gang Beast. How the heck are you going

134:25

to handle the lobbies for that?

134:27

>> I don't know. That'll be my problem.

134:29

>> That's what I just said. I don't I had

134:30

nothing to do with any of this.

134:32

>> That's crazy.

134:32

>> I learned about this like yesterday.

134:34

>> Uh anyways, don't miss the event of the

134:37

year at lmg.gg/unklinus.

134:41

Uh and in case anyone is wondering, yes,

134:42

the lionus sign will be out for the

134:45

entire week at the front of the

134:46

building.

134:47

>> Yeah, that is that is actually on the

134:49

building right now.

134:51

>> Yep.

134:52

>> And you couldn't even use a high

134:53

resolution version of the wide linus

134:55

image.

134:55

I'm pretty I'm pretty sure that is

134:57

intentional.

134:59

>> I would definitely try to recompress the

135:02

JPEG to be like,

135:03

>> you know, the sign manufacturer comes

135:04

back and they're like, "This is really

135:05

low quality." And they're like,

135:07

>> "Yep, it's

135:08

>> Can you make it worse?"

135:12

>> Yeah. I mean, knowing Sammy, that might

135:14

have genuinely been intentional.

135:15

>> I believe it.

135:16

>> Yep.

135:16

>> I'm going to give him the benefit of the

135:18

doubt.

135:18

>> I I think I'll give him the the benefit

135:20

of the doubt there for sure.

135:21

>> This is cool. Researchers at LIRMM, a

135:24

French robotics and computer science lab

135:26

in Mont Pelleier, connected some slices

135:29

of post-mortem adult brain tissue to a

135:32

piano playing robot and taught the

135:34

tissue to play notes.

135:36

>> Sick.

135:37

>> The setup is called an OPAB

135:40

>> op,

135:41

>> an organotypic culture of postmortem

135:44

adult brain expplant.

135:45

>> That is way longer of a sentence than a

135:47

four-letter acronym, but okay. It's a

135:49

fancy way of saying a slice of donated

135:51

brain that's kept functioning in a dish

135:54

on a bed of micro electrodes. The robot

135:58

side is an anthropomorphic hand plus a

136:00

microphone acting as an ear. During

136:03

training, the system let the tissue

136:05

associate the robot playing piano keys

136:08

with the sounds that those presses made.

136:11

In testing, a human played notes and the

136:13

brain tissue controlled the robot hand

136:15

to imitate them. The learning was

136:17

unsupervised, meaning nobody programmed

136:19

the association in the neurons formed it

136:22

themselves. This builds on the same

136:24

team's 2024 finding that these

136:26

postmortem brain slices still retain

136:29

synaptic plasticity, the basic mechanism

136:32

behind learning and memory.

136:35

This is a preprint that has not been

136:37

peer-reviewed yet. The tissue is a small

136:40

slice and not a full brain. And there's

136:41

no awareness or consciousness involved,

136:43

just neurons doing what neurons do when

136:46

they're stimulated. The pitch from the

136:48

researchers is an animal-free way to

136:50

study human neural circuits. And it sits

136:51

alongside other biological computing

136:54

work like Australia's CL1 computer which

136:56

was built on lab grown neurons.

137:00

>> Crazy.

137:01

>> The uh the reception to this has been

137:05

>> mixed at best

137:05

>> polarized.

137:07

There's people who are like, "Wow, this

137:10

is great. I can't wait for this future."

137:13

On the one hand, and on the other hand,

137:16

there's, "I hate this. Turn it off now.

137:18

I want stop the ride. I want to get

137:20

off." And there's nothing in between, as

137:23

far as I can tell.

137:26

I will not be taking a position.

137:29

YouTube's AI slop detector has

137:31

incorrectly targeted Kurszig. Kir, I can

137:35

never pronounce. Yes, thank you. As

137:37

other creators fear the same fate. Kurs

137:41

Kurszact,

137:42

that's how you pronounce it, right?

137:43

>> Kurs Gazact.

137:44

>> Thank you. The 23 million subscriber

137:46

science channel, amazing by the way,

137:48

that hand animates videos over months

137:51

with a full team of artists revealed

137:53

that YouTube's automatic AI detection

137:55

tools flagged its work as AI slop and

137:57

quietly throttled the channel. Its Super

138:01

Predator video became the channel's

138:02

worst performing uploads since 2013. But

138:05

the audience data looked strong with

138:07

above average click-through and solid

138:08

watch time. The video simply wasn't

138:10

getting recommended. When Kurszat Kurs I

138:14

can't say this word

138:15

>> go Kurs,

138:17

>> thank you. Contacted YouTube. The

138:18

company confirmed that the AI detector

138:21

was the cause and the channel has since

138:22

pulled the video to rework and re-upload

138:24

it, which is a crazy solution. The

138:27

detector probably misfired because

138:31

>> curs

138:32

>> flat clean vector style is exactly what

138:35

AI slop channels imitate since it's

138:36

cheap to fake. So a detector trained on

138:39

the imitation caught the original that

138:41

inspired it.

138:44

>> Kris Gaz

138:45

>> took it in stride saying they support

138:47

blocking actual slop and calling and

138:50

called their YouTube contacts helpful

138:52

and transparent. Kruz Gazock

138:54

>> had people at YouTube to call, but most

138:57

channels don't. Analytics firm Vid IQ

139:00

highlighted that gap, noting that a

139:02

channel that big can find and correct

139:04

the problem, but smaller ones can't, and

139:07

pointed to a 5 million subscriber

139:08

creator who went through a similar

139:10

unexplained decline. Dozens of creators

139:13

in the replies to

139:14

>> Kris Kazox

139:16

>> noted that their reach was completely

139:18

dead and fear that they've been hit with

139:20

the same flag with no way to confirm or

139:23

fight it.

139:26

I don't really have anything to say

139:27

about that other than that blows.

139:29

>> One of the crazy bits is a a

139:32

knowing a channel that big can find and

139:34

correct it. Winning a smaller example of

139:38

a 5 million subscriber creator is like

139:42

what?

139:42

>> I know.

139:45

[sighs]

139:46

My god.

139:47

>> So, basically, you can't fix it unless

139:49

you're

139:50

>> have to be enormous.

139:51

>> Mega channel

139:52

>> or you're screwed.

139:54

>> Yep. Uh Valve is expanding Steam OS

139:57

support to several gaming handhelds.

139:59

Steam OS 3.8.25 25 beta adds initial

140:02

gamepad support for the Intel powered

140:05

MSI Claw 8X AI Plus, the Iono Pocket S2,

140:08

and the Conquer Fit, plus updated

140:11

controller support for the One Player X

140:13

Apex. The Claw is notable because Steam

140:15

OS has been AMD only territory until

140:17

very recently. And the Pocket S2 is

140:20

notable because it's an ARMbased Android

140:23

handheld, which lines up pretty neatly

140:26

with the ARM compatibility work that

140:28

Valve just shipped with Leptton and FEX

140:31

in preparation for the launch of the

140:32

Steam Frame.

140:33

>> Yeah.

140:35

>> None of this means that everything works

140:36

perfectly, however. Wi-Fi, audio, sleep,

140:38

and drivers can still be rough, and only

140:40

the Steam Deck and Legion Go S have

140:42

fully baked support. The same update

140:45

bundles new Steam Machine firmware,

140:47

fixing a bug where CPU performance

140:48

stayed reduced after waking from sleep.

140:50

Actually, oh, we encountered that uh

140:52

when we were working on our review. So,

140:54

I'm glad to see that they fixed it,

140:55

which we said they probably would, so

140:57

that's great. Along with fixes for HDMI

140:59

audio dropouts and Bluetooth latency.

141:02

Um, discussion question. Valve is slowly

141:05

making Steam OS run on everybody's

141:06

hardware. Ignoring the anti-che elephant

141:08

in the room, is there any reason left to

141:10

keep Windows on a handheld?

141:13

I still run it on mine and it

141:17

>> some some game compatibility is still

141:19

going to be rough.

141:19

>> It does not enhance the experience.

141:21

>> The the AMD based stuff of most

141:25

handhelds will help with that a lot,

141:27

>> but there are still some games that just

141:30

aren't going to be a thing. You're not

141:31

going to be playing Valerant.

141:32

>> I don't know if you want to be playing

141:33

Valerant on a handheld, so maybe it

141:35

doesn't matter much.

141:39

I don't run into this stuff that often.

141:41

The main thing I run into is Nvidia

141:43

problems more than

141:44

>> Linux compatibility problems.

141:46

>> That's just from having an Nvidia GPU.

141:48

>> Yep.

141:49

>> Windows latest finds that Windows 11

141:52

uses five times more memory for its

141:54

adladen weather widget than Mac OS

141:56

equivalent. That's pretty funny. Uh, in

141:58

other Windows news, Justin Marshall

142:00

released a native version of 1990s Word

142:03

1.1A for a Windows 11 on GitHub. And I

142:07

thought this was pretty fun because uh I

142:09

think that was on Techraar. Yeah. Uh

142:12

hold on. It's in here somewhere. How

142:14

much RAM it uses.

142:16

[snorts]

142:17

>> Uh

142:19

hold on. One megabyte of RAM [laughter]

142:22

on Windows 11.

142:26

That's amazing. I wonder how usable it

142:29

is. Honestly, when I'm writing

142:30

documents, I use shockingly few

142:33

features. And this is something that you

142:35

can totally run totally legally.

142:36

Microsoft released the source code for

142:38

this, so Buddy ported it. It's

142:40

apparently a bit of a pain in the butt

142:42

to get it running. Like there's some um

142:47

uh there's there's like some terminal

142:48

involved or something like that. Um you

142:51

need PowerShell. I don't know. I haven't

142:54

actually tried to do it, but it I just I

142:56

just thought that was pretty cool.

142:57

>> I'm surprised there's no pictures in

142:58

here. What does it look like? Uh, what

143:01

does Word 1.1 is going to look pretty

143:04

pretty rough?

143:05

>> Yeah,

143:05

>> here. I'm just going to look up

143:07

>> Word 1.1 interface.

143:13

[sighs]

143:14

>> Dying.

143:14

>> Yeah, you do not actually want to work

143:16

in that right now.

143:17

>> Yeah, maybe. Well,

143:22

>> can you Is it even I mean, it's probably

143:24

not even Can you insert a picture? Maybe

143:27

not. I think we'd probably have to make

143:29

our way to

143:32

I'm trying to think when did they Oh

143:34

man, I I never really thought about

143:36

this. All versions of Microsoft Word.

143:38

Oh, we should switch to After Dark.

143:40

That's it for topics, Dan.

143:41

>> Math co-processor.

143:43

>> That's awesome.

143:47

Ridiculous.

143:48

>> Oh my god,

143:50

dude. This should this should be Labs

143:52

charts now. We should make them look

143:54

like this.

143:56

Yeah.

143:57

>> Oh my god. Word. Perfect.

143:59

>> Actually sick.

144:00

>> It's pretty readable.

144:01

>> I think Word 97 is about as far back as

144:04

I could go and still actually

144:06

>> looks totally fine.

144:08

>> That looks very usable. Lotus notes.

144:10

Let's get some Lotus notes.

144:11

>> I think I could deal I could deal with

144:13

this. No problem.

144:14

>> So really fancy notepad.

144:17

>> Yeah,

144:17

>> that looks like a computer program to

144:19

me.

144:19

>> Yeah, me too.

144:22

Pug boy says, "Fun fact, Word Perfect is

144:24

still around." Nice.

144:27

Darn. All right, hit me, Dan. Hello,

144:29

Lionus, Luke, and Dan. Normal names.

144:31

Love listening to the show on my way to

144:33

work. My question is, what hard

144:35

experience at the beginning of building

144:36

LMG made the biggest impact on the

144:38

company for you today? Or you the

144:40

company or you?

144:44

>> At the beginning.

144:47

I ring Luke.

144:50

Aw, that's such a nice answer.

144:51

>> That was a very hard experience.

144:53

>> Yeah, it wasn't hard though. Well,

144:57

[laughter]

144:57

>> Luke, you need more practice.

145:00

>> Right into After Dark.

145:01

>> Oh, yeah.

145:02

>> You know what? I'm going to go with the

145:04

um whether it ultimately was a good

145:07

thing or not, I don't know. But the the

145:10

decision to move out of the garage and

145:13

spend everything that we had made so far

145:16

and what we intended to make for for

145:20

months and months uh on a new space.

145:24

That culture of just like, okay, take

145:27

everything we made and dump it right

145:28

back into the company. Okay, do it

145:30

again. Okay, do it again. Was what

145:33

ultimately led to major successes like

145:37

Backpack and Screwdriver. So that um

145:39

that willingness to just go, "Okay, I've

145:41

got all these chips.

145:43

>> I'm going to play them.

145:44

>> I could keep them.

145:45

>> I'm going to play them all.

145:47

>> But no, I'm going to just bet bet hard.

145:49

>> I'll drive the Civic another year.

145:51

>> Bet hard."

145:51

>> Yeah.

145:52

>> Yeah.

145:52

>> Yeah. I think for me,

145:56

>> I've been doing this 10 years before I

145:58

finally like bought the car, you know?

146:01

>> Like,

146:02

>> hey, it took a minute.

146:03

>> And Luke had to convince you. [laughter]

146:05

>> I don't think Luke tried to convince me.

146:07

I don't think so.

146:08

>> Luke thinks my car is stupid.

146:10

>> Kind [laughter] of.

146:11

>> Yeah, it is stupid. That's fine, though.

146:14

>> I know. I I also think you're entirely

146:16

within your rights to buy a stupid car.

146:18

>> I think I remember you. I'm 40, baby.

146:20

>> Yeah. It's like whatever, dude. It's

146:22

fine.

146:22

>> I don't judge you for it. I just, you

146:24

know, it's not my thing.

146:25

>> Yeah.

146:26

>> But that's okay.

146:27

>> Um I think for me, unironically, it was

146:29

a 700 series launch.

146:32

>> Oh, yeah. I I knew it was going to be

146:34

the GPU.

146:34

>> Yeah.

146:35

>> Yeah. the Titan.

146:36

>> I had to grow a lot as a person in

146:40

learning

146:42

about that stuff because now suddenly it

146:46

like uh you know we took it seriously

146:48

when we were part of NCX but now it was

146:50

like our thing.

146:51

>> Yeah. No, but like actually

146:52

>> now it's like if I if I this up

146:55

>> I think I bleeped that properly. Um

146:57

yeah, we're all just like dead and it's

147:00

like oh wow this is like very serious.

147:01

>> You go crawling back to NCX like what

147:03

are we talking about? There was no

147:04

route. It it was it was you know I I go

147:07

back to school, Ed goes back to school

147:09

and you find a job somewhere else. Like

147:11

it was it was so there was a lot of

147:13

pressure on there.

147:14

>> Do or die, baby.

147:16

>> That was the that was a massive

147:18

expansion in the amount of tests and

147:20

what we tested for a GPU launch compared

147:22

to normal. Um if I remember correctly,

147:25

you like wore a lab coat for that video

147:27

because we were like trying to hit the

147:28

like really hit it home of like this is

147:30

a step.

147:31

>> Yeah.

147:31

>> Um

147:32

>> that was pretty cool. And that was like,

147:35

[clears throat]

147:36

yeah, I don't know if I remember

147:37

correctly, I didn't sleep for like

147:39

a really weird amount of time, like 3

147:41

days or something. And I remember one

147:43

night I was so tired and had to stay

147:45

awake, so I ate through an entire tub of

147:47

peanut butter and a full box of uh wheat

147:51

pins.

147:52

>> I don't

147:52

>> And just like lived off of just

147:54

calories.

147:55

>> I don't think I would have been um happy

147:57

if I'd known you did that. Um

147:59

>> I'm pretty sure I told you afterwards.

148:01

>> Yeah.

148:01

>> Yeah. Yeah.

148:03

>> Yep.

148:06

>> Got her done. [laughter]

148:09

>> That was a mo I had already like, you

148:10

know, locked in multiple times, but that

148:12

was a moment where I was like, I'm just

148:14

going to do whatever to like try to make

148:16

this work.

148:18

>> We got her done.

148:19

>> Yeah,

148:19

>> man. We we released so many videos on

148:21

that card. Remember back when a new GPU

148:23

launch was exciting and people were

148:24

happy about it?

148:25

>> Dude, the 700 series was very exciting,

148:27

though.

148:28

>> Actually, that was a sick release. Um,

148:32

>> I think you're actually thinking of 600

148:34

series though.

148:36

>> Thought it was the

148:37

>> the first one when we got in the garage,

148:38

right?

148:39

>> Was that 600? No, no, no. I think you're

148:41

right. I think it was 700. Wait, was it

148:44

600?

148:45

>> Which one was the first generation to

148:46

have a Titan?

148:48

>> This is our video on 680.

148:50

>> Oh, it's got to be It's got Oh my god.

148:52

Ads

148:53

>> 14 years ago. So 2012. No,

148:57

>> no,

148:58

>> 680.

148:58

>> It would have been

148:59

>> Yeah. 700 then. And yeah, you're right.

149:01

>> 780.

149:03

>> That's it.

149:04

>> New King. This one.

149:05

>> That's it.

149:05

>> 12 years ago.

149:06

>> I think I shot that B-roll because I

149:08

don't think Brandon worked here yet.

149:09

>> Well, I'm sure you did. When I was done

149:10

the testing, I just like died. Um I

149:13

slept for like a day if I hotspot shield

149:16

though. That came later. No, it must

149:18

>> We were in the garage.

149:19

>> It must be Titan. I think you're

149:20

thinking of Titan.

149:22

>> How do How do you search for Titan on

149:24

the channel? It's going to bring up so

149:25

stupid. Why do companies have to call

149:27

everything the same thing?

149:29

>> Go down. Not this one.

149:31

>> Keep going.

149:31

>> It was silver.

149:32

>> Keep going. Keep going.

149:35

>> GTX Titan.

149:36

>> That's the unboxing.

149:37

>> That's the unboxing.

149:38

>> Yeah. Keep going then. You'll find it.

149:40

You'll find it.

149:40

>> Is that little man? That's crazy.

149:42

>> Yeah, I know, right? It kills me.

149:44

>> That's wild.

149:44

>> I know.

149:49

Okay. And a bunch of crap that's totally

149:51

unrelated. Thanks, YouTube. Not what I

149:52

was looking for.

149:54

>> Unbelievable.

149:55

>> It's also crazy to me that the unboxing

149:57

is the only thing that really surfaced.

150:00

It's just completely unrelated things.

150:02

Now,

150:02

>> another fun game is like no matter what

150:04

you search for with my name in it, how

150:06

far do you scroll before Yeah. there.

150:07

Before you come across like a drama a

150:10

drama video.

150:11

>> Yeah, [laughter]

150:12

that's a good point. Um,

150:15

>> okay.

150:16

>> Crazy.

150:17

>> Yeah, it was it was Titan though. It was

150:19

We were just like

150:19

>> That makes sense.

150:20

>> All hands on deck. All six total hands

150:23

plus Ivonne being on Matt leave. I don't

150:26

know doing correctly. We were trying to

150:27

do dual camera shoots at that time and

150:29

stuff, too. Like things were things were

150:31

kind of crazy.

150:32

>> We had ideas.

150:33

>> Yeah.

150:33

>> But you did it.

150:35

>> Yeah.

150:35

>> Yeah. We We're still here, baby.

150:38

>> Hi, Dad. Dad and Dan uh at Lionus.

150:42

Thanks for raising me. It is concerning

150:44

that I see you more than my own dad,

150:47

even though he's the CEO of the company

150:49

I work at. [laughter]

150:52

>> P.S. How's it going with the ADD meds?

150:55

>> [laughter]

150:56

>> I have not actually tried them yet.

150:58

>> What the hell, dude? Uh, my wife was out

151:01

of town, so I got confused about how

151:03

getting medications works, and so I

151:06

didn't get them.

151:07

>> Well, my doctor didn't mention it to me.

151:13

>> If you need drug advice, just let me

151:15

know. My doctor didn't My doctor didn't

151:17

mention it to me because I guess it's

151:19

just like routine, but apparently she

151:22

just faxes the script straight to the

151:25

pharmacy that we like usually go to,

151:27

which I wasn't sure which one it was

151:29

because I don't know where medications

151:31

come from. Um, no, actually I I did

151:34

know. I did know which one cuz sometimes

151:35

I pick up things for other people. I

151:36

haven't like been on many medications,

151:38

but I have kids. They get sick

151:39

sometimes. Um, you know,

151:41

>> but but I didn't I didn't think about

151:43

it. I just I just got off the call. She

151:45

was like, "Yeah, I'll get that for you."

151:46

And I was like, "Okay."

151:48

And then I didn't know like where I

151:52

would get the prescription from. So

151:54

Ivonne came home and was like, "Hey, did

151:56

you ever get that thing from the

151:58

doctor?" And I was like, "No." And she

151:59

was like, "Oh, okay. Would you like me

152:02

to grab it next time I'm in town?" And I

152:04

was like, "Yes, please. Thank you,

152:06

lady." And then she got it for me.

152:09

>> I'm sorry. I know.

152:10

>> I haven't tried it yet, though. I know

152:11

I'm derailing us and bringing us back to

152:12

something we should be moving on from,

152:14

but I found some of them.

152:15

>> Nice. I

152:16

>> surround performance review, 3D vision

152:20

gaming review.

152:21

>> I remember when cards had that.

152:22

>> We just did everything.

152:24

>> 1080p

152:26

like

152:26

>> and they're like 7 minutes, 4 minutes.

152:28

>> Man, the thumbnail is just a graph.

152:30

That's crazy.

152:31

>> I know.

152:33

>> 1600p got a whole video.

152:34

>> P video by itself 10 minutes long.

152:37

>> I don't think 4K was a thing yet. I'm

152:39

pretty sure there was more videos than

152:40

just these as well that they're just not

152:42

surfacing. But yeah, we went really

152:44

hard. [laughter]

152:45

>> Yep.

152:47

>> Anyways, early WAN show means I can

152:49

reach out before Sabbath again. Any

152:52

updates on the tech detox you talked

152:54

about last time? PS, there's lots of

152:56

cool tech for Sabbath, including

152:58

specific home assistant plugins.

153:00

>> Oh, that's interesting.

153:03

So, as long as whatever I do doesn't

153:06

count as work, my home assistant could

153:08

do something for me.

153:12

It doesn't

153:13

>> Trey is interesting.

153:14

>> Wasn't there something that you couldn't

153:16

have somebody else do work for you?

153:18

>> Not sure. I can't remember all the

153:20

rules.

153:20

>> I can't either.

153:21

>> Um,

153:22

>> I never knew them.

153:24

>> Well, I only No, I mean I mean I can't

153:26

remember the rules that I learned the

153:29

last time we talked about the Zwan show.

153:31

I didn't I didn't know them. It was just

153:33

>> people in chat were kind of like

153:35

explaining stuff.

153:36

>> I think you can't manually get the work

153:39

to start.

153:40

>> Yeah, but then you could employ somebody

153:41

to like cook for you and turn on light

153:44

switches and follow you around,

153:47

which I think kind of defeats the

153:48

purpose.

153:49

>> But you're not employing a person.

153:51

>> You're forcing a machine to do that for

153:54

you.

153:54

>> Yeah. Chat says it depends on the rabbi.

153:56

Some of them don't like switches or

153:58

relays closing even. Oh,

154:00

>> cuz you're you're connecting something

154:02

together which is like I don't know.

154:05

>> Yeah.

154:05

>> Uh tech detox.

154:08

>> I have not gotten there yet.

154:10

>> I don't think you ever will.

154:12

>> I did have a week off.

154:13

>> What were you going to do to tech?

154:14

>> We were We were talking about doing like

154:15

a tech detox.

154:17

>> Oh, right. We were both like go away and

154:19

not Yeah.

154:20

>> Yeah. I gave that literally zero thought

154:22

um [laughter] ever since we last talked

154:25

about it. Um but hey I um

154:29

>> it's it is progressively becoming more

154:32

possible for me which is interesting. My

154:34

work is getting more time consuming and

154:36

more allconsuming and it's also getting

154:38

more possible for me to

154:41

step away from it for periods of time

154:42

which is those things happening at the

154:44

same time is very strange but I think

154:46

it's true.

154:47

>> How do I how do I edit this?

154:49

>> People on my teams getting really good

154:53

>> is essentially what that means.

154:54

>> Yeah. which is awesome.

154:56

>> Yeah.

154:58

Okay, I'm gonna send Dan. Okay. How do I

155:01

edit this? How do I edit the stupid

155:02

photo? Yeah, here it is. Okay, perfect.

155:05

Oh, I hate how the image editor in

155:07

Teams, if you drag the corner, it

155:09

assumes you mean just that corner and

155:11

then it's so hard to grab the uh the top

155:14

bar. Okay, so I mean this is my this is

155:17

my version of of detox. My my chickens

155:20

are big now.

155:23

Yep. They're big girls. One, it turns

155:26

out, is a rooster.

155:28

>> Oh, you got dinner in the works.

155:30

>> The heck. Chickens.

155:32

>> There's my chickies.

155:33

>> WHOA.

155:34

>> HEY, BABIES.

155:36

>> They have small crowns.

155:38

>> That's cute.

155:40

>> And then since we're just like putting

155:42

pictures of Lionus's pets on the stream,

155:45

uh here's a shot of my cat that I took

155:46

the other night that I thought was

155:48

pretty funny.

155:58

>> [laughter]

155:58

>> He's just chilling.

155:59

>> Cats are so weird.

156:01

>> He was just sitting there like that. I

156:02

was like, "What the heck, man? You're

156:04

not even like

156:04

>> What are you doing?

156:05

>> You don't even look comfortable."

156:07

[laughter]

156:09

>> Uh, I went to that 1080p performance.

156:12

Oh, no. Not nothing. I'm just looking at

156:14

my cat.

156:15

>> I went to the 1080p performance video

156:16

for GTX Titan cuz I was wondering what

156:18

the comments were like. Yeah.

156:19

>> Um, look who I randomly though. Look who

156:23

I randomly found.

156:24

>> Jack Septic Eye. Nice.

156:27

>> With 13 likes somehow.

156:28

>> Oh my goodness.

156:30

>> That's good. Have fun wasting all that

156:31

money.

156:32

>> That's nice. [laughter] What What's his

156:34

net worth now? I wonder.

156:35

>> Oh, just a crazy amount, I'm sure.

156:37

>> Yeah, that's awesome.

156:39

>> But yeah, wild.

156:44

>> All right, hit me, Dan.

156:45

>> Yeah. What do we got?

156:46

>> Oh, yeah. Sorry, I got distracted by the

156:47

cat. Um,

156:49

2027 Global RAM is completely sold out.

156:52

Luke, since you're good with this, can

156:54

you make a prediction for when this will

156:55

be over? Not because I think you'll be

156:57

right. I just want some hope, please.

156:59

Love the show.

157:00

>> I don't think I don't think you can

157:01

force uh these predictions. They just

157:04

kind of come out of nowhere.

157:05

>> Just give us false hope.

157:07

>> I think it's a while yet. Um,

157:11

I think we're looking If it was before

157:13

this time, 2028, I think I would be

157:16

stunned.

157:17

And my reasoning for that is even if

157:19

things crash, it's going to take a lot

157:22

of time to kind of reset.

157:23

>> And a lot of this is contract pricing.

157:25

>> Yes. So it it will take a hot minute for

157:30

this to correct. I think, you know, I've

157:32

been talking like, oh, maybe China will

157:33

save us in regards to RAM prices. If

157:36

they do, it's going to take a while.

157:38

>> And with that said though, sometimes

157:40

change happens really, really slowly and

157:42

then all at once. a major insolveny or

157:47

like a major like an Enron scale uh

157:50

crash of a of an individual company

157:52

could trigger a chain reaction. We don't

157:54

know.

157:54

>> And there do be some companies with

157:56

Enron type math going on at the moment.

158:00

So it could happen. It's it's very

158:02

difficult to predict right now.

158:04

>> Fascinating.

158:04

>> I think if somebody was like, "Oh, my

158:07

company builds a lot of systems and I'm

158:09

going to buy and hoard a bunch of RAM

158:11

right now." I think it would be pretty

158:13

hard to very confidently argue in either

158:16

direction of whether that was a good

158:18

idea or a bad idea.

158:20

And like this this was true a while ago.

158:23

We were talking internally about how we

158:25

we bought 20 extra laptops in December

158:27

of 2025 for the company.

158:30

>> And I remember at the time that was like

158:32

really risky.

158:33

>> Kind of crazy actually. Yeah.

158:35

>> But we were worried. And

158:38

>> yeah, my kids ah my kids need new

158:40

Chromebooks for back to school

158:41

>> and they're going to suck

158:42

>> and they're Yeah, they're going to

158:44

totally suck or they're going to be

158:45

really expensive. Yep.

158:46

>> Framework doesn't make a Chromebook

158:47

anymore.

158:48

>> As as I was uh as I was like looking at

158:50

the options, I was like, "Oh, I wish I'd

158:52

grabbed a framework last time cuz then I

158:53

could just fix this issue that I have on

158:55

my daughters."

158:56

>> And they don't make one anymore. Anyway,

158:58

sorry. Carry on.

159:00

>> Yeah. Um and you know, luckily that bet

159:03

worked out really well.

159:04

>> Yeah. And it's really easy to look back

159:07

and be like, "Oh, I wish we bought

159:08

more." And it's like, "Yeah, but then

159:10

buying more would have just been

159:11

incurring more risk." And there was no

159:13

way that we could have fully known in

159:15

December of last year what the heck was

159:17

going to happen over these insane eight

159:19

months that we've gone through. Uh, so I

159:22

don't know. It is what it is. I'm happy

159:23

we bought those. That's nice. This has

159:26

saved us a bunch of money already. It's

159:28

effectively bought us multiple laptops,

159:30

which is cool.

159:31

>> Nice. But like, yeah,

159:33

>> Dan, if you're gonna have your battles,

159:35

>> Dan, if you're gonna have my

159:36

there, can you also put my It's

159:38

the black one.

159:40

>> Sure.

159:43

>> Oh, so now I get to see it. Uh,

159:46

>> all you have to do is ask.

159:51

>> Thanks, [clears throat] Dan.

159:53

>> Oh, yeah. Merch messages.

159:57

>> Do you want me to read it out?

159:58

>> Sure.

159:58

>> No, I'm not going to read a merch

159:59

message. I'm going to read a com. Nice.

160:02

>> Solid.

160:03

>> Hey, LDL. I've been shopping for home

160:05

battery backups and sonar panels, solar

160:08

panels, and it seems most of the major

160:10

companies seem to be under the same

160:12

umbrella company. Lionus, how's your DIY

160:14

battery going? Oh, I'm so glad you asked

160:18

because I had a long meeting with uh

160:21

Pancratz and Frey earlier this week and

160:25

we have everything we need to build the

160:28

economical custom packs that are going

160:31

to be way safer than the uh it's just

160:34

the one in the front, the black one.

160:37

>> Oh, sorry.

160:37

>> That's fine. That's good enough.

160:38

>> I can't crop that small. It's far too

160:40

big.

160:40

>> Yeah, the other ones are hens. Nice.

160:43

Yep. All

160:45

right,

160:45

>> that's my

160:48

>> I'll just put that here.

160:50

>> Perfect. Okay. Um, yeah. So, we've got

160:53

everything we need. And then once we do

160:55

phase two, the plan was always to do

160:58

this in phases. So, once we do once we

161:00

do phase two, phase three is going to be

161:02

scaling it up to power the entire house

161:05

for some period of time. Now, obviously,

161:08

I'm not going to charge my car off of it

161:10

or like run the clothes dryer. Um, but

161:13

it could run like all the lights and

161:16

hopefully computers

161:18

um for, you know, at least a while if

161:22

the power goes out.

161:25

And let's see, how many more do I have

161:28

for [laughter] you? I found this in my

161:29

folder. Uh, I'm a professional. Last one

161:33

I've got for you today. You spend most

161:36

of your careers in tech. And while I'm

161:38

only a few years younger than Lionus, I

161:41

don't recall the tech space being like

161:43

it is. Has tech in your experience ever

161:46

seemed this evil? Genuinely?

161:50

No, I don't think so. And that's someone

161:52

like I I used to be really obsessed with

161:55

this uh I think it was Australian thing.

161:58

Yeah. Called the Beast Files. This is

162:01

well before Mr. Beast, I'll tell you

162:02

that much. Uh so was when I was in high

162:04

school. I don't remember how I used to

162:06

watch them online. I think this was

162:08

preyoutube, but maybe it wasn't. Uh

162:11

they're on YouTube now as an archive on

162:13

this like real human stories YouTube

162:15

channel, but I remember this one,

162:16

Google's plan for world domination from

162:19

the Beast Files. This was a fantastic

162:21

video. Somehow it was only 2 minutes and

162:23

45 seconds. Uh but it was it was talking

162:25

back then about people used to talk

162:27

about this all the time like Google got

162:28

rid of their don't be evil thing. Uh,

162:31

and they started doing some stuff that

162:32

was like a little bit spooky.

162:35

Man, if you could sit back then and look

162:38

at the stuff the different tech

162:38

companies were doing back then and

162:40

compare it to now, it would look like a

162:42

freaking joke. All of them. Like,

162:47

yeah, we are in completely unprecedented

162:49

times

162:50

>> because they keep getting away with it

162:52

>> constantly.

162:53

Like, it's it's crazy. And so, yeah, as

162:56

someone who has been in tech for like my

162:58

entire professional life to a certain

163:00

degree, um, no, it's never been even

163:03

close to this crazy.

163:08

Not even sort of.

163:09

>> No, no, don't show them my balls.

163:15

Thanks for tuning in. We'll see you next

163:16

week. Same bad time, same bad channel.

163:19

>> Bye. [laughter]

163:25

>> [snorts]

163:25

>> Oh my god. Oh, that's good.

163:34

The world spent 276.5

163:37

million on AI during the show.

Interactive Summary

The WAN show discusses various tech industry topics, including the loss of 50 terabytes of archives by 9PBS due to a cloud storage provider shutting down, highlighting the risks of not owning one's infrastructure. They detail LMG's shift towards more self-hosted and on-prem solutions, including becoming an ISP and purchasing networking hardware to save costs over five years. The conversation also covers the rise of AI-related costs across the industry, potential content changes on Linus Tech Tips, and interesting anecdotes about hardware and corporate behavior.

Suggested questions

4 ready-made prompts