The Interview: Steve Burke of GamersNexus

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Alz Media.

Speaker 2 (00:05):
Hell and welcome to this week's Better Offline. I'm, of
course your host ed Zittron and for this week's episode,
I flew out to North Carolina to do a longer
form interview, the first of it's kind on the show,
a little slower, a little more fun, and I did
it with Steve Burke, founder and host of gamers Nexus,
an incredible hardware YouTube channel that's been going since two
thousand and eight and has over two and a half
million subscribers. We talked about the history of gamers Nexus,

(00:27):
the state of the hardware industry, tech journalism, gamers Nexus
is incredibly scientific approach to hardware testing, and of course
a little bit about AI. I think you're going to
really like it enjoy. So how long has the channel

(00:50):
been running? How long have you been doing this? Now?

Speaker 1 (00:53):
About seventeen years and five months.

Speaker 2 (00:55):
And we started off just doing gaming content. I did
go back in the files and found like a Modern
Warfare trailer I think, and then the Office Space parody,
which is awesome.

Speaker 1 (01:06):
Yeah. Yeah.

Speaker 3 (01:06):
So it started as game reviews and trailer analysis videos,
which was the Modern Warfare one and battlefield trailer analysis
that was like a whole subgenre. Back then you got
the trailer, you kind of picked apart what's going to
be in this game? And you know, then I did
game interviews. We spotlighted a bunch of indie games through

(01:27):
Steam's green Light program, and it kind of slowly got
into hardware.

Speaker 2 (01:31):
What was Greenlight? Is that? Like? Is that kind of
the early early access.

Speaker 3 (01:35):
Yeah, that was Steam's thing where they gave indie developers
who are unestablished kind of a way to try and
get onto Steam.

Speaker 1 (01:42):
And so we did a lot of early coverage of games.

Speaker 2 (01:44):
Yeah, I saw you doing a lot of convention interviews.
How did you grow from there to where you are today?
Which is said huge studio and massive testing facilities and such.

Speaker 3 (01:54):
The interviews and the stuff at the conventions was kind
of the turning point for me because I was up
until maybe twenty twelve or something like that plus and
minus year, I was still technically in college. And then
I went to I went to pas Penny Arcade XPO.
I think I went to the twenty ten to one,
and after I got home from that, I decided, this

(02:16):
is kind of the only thing I want to do
go to stuff like this, And so it took me
two years but you know, eventually just dropped out and
then yeah, then just started doing more of those, and
it was it was the best way. First of all,
they're just fun, but then secondly it was sort of
how I started actually meeting the people we would end
up working with in the hardware industry.

Speaker 2 (02:36):
Right, and did Well, how'd you get into testing? Because
I found one of you, I think your first viral clips,
which was the testing a power supply with a paper clip,
I believe.

Speaker 3 (02:45):
Oh that was the that was to jumpstart a power
supply to yeah, test it, yeah, to make sure it works.

Speaker 2 (02:51):
So how did you move into that? Because it it
seemed like it took a minute to get there.

Speaker 3 (02:56):
Yeah, so I think around So I started the site
in two thousand and eight, I think officially, and then
got into publishing PC build guides in around twenty ten
or so. So that's when the hardware really started. And
I'd already been building computers but hadn't actually really published
much about it, So start publishing build guides. I think

(03:18):
it was just as simple as those were the first
thing that had any kind of like readership because it
was all articles back then, right, and so they were
so you started out writing yeah, exclusively, and then the
channel got added in two thousand and nine, but it
wasn't treating you know, YouTuber was not that's not a
job in two thousand and nine like that. Yeah, it's

(03:39):
a little before you like you had maybe Casey Nystadt
was at the front edge of that.

Speaker 2 (03:42):
Yeah, I think I think it was like maybe Vine
was it was, it was still around when did Vine
pop up? Christ it was still asl but no back then,
so you were just on the fringe of this new content.

Speaker 1 (03:53):
Yeah, and it was.

Speaker 3 (03:54):
I did not start a YouTube channel with the intent
of making it a job, because I don't. If you
go on Internet archive, I didn't actually even remember this
until recently. And you look at the gamers Excess website
and you go to the about page from like two
thousand and eight or nine or ten, you'll see that
somewhere in those really early years, I had said that

(04:14):
we were running it without any form of banner ads,
which is actually true today too. We got rid of
the ads when we reintroduced the website recently. Right, But
like point being, you don't do that having no revenue
unless you aren't really thinking about trying to make it
a sustainable thing.

Speaker 1 (04:35):
Right, it was purely for fun.

Speaker 2 (04:37):
And how did it become a sustainable How did it
become your main job?

Speaker 1 (04:41):
Yeah?

Speaker 3 (04:42):
I think as YouTube grew basically, so we were kind
of in the early We weren't the first sort of
i'll call it generation of YouTubers who were able to
successfully make it a job, but I would say we were
maybe in the next generation right after that. And so yeah,
I started publishing more videos alongside the articles. So it

(05:06):
really was with case reviews computer case reviews, where we
would originally just focus on only writing a case review.
At some point, I'm saying we a lot. For a
while it was just me and then every now and
then it'd be soone to.

Speaker 1 (05:19):
Help, but we would do case reviews.

Speaker 3 (05:22):
And then I realized this would do well with video
because there's depth to it and there's like just a
lot of mechanical stuff, and those videos started doing pretty okay.

Speaker 2 (05:34):
And did you move away? When did you move away
from the website, because I know you've come back to it.

Speaker 3 (05:38):
Yeah, we probably around twenty I think around twenty eighteen
maybe is when I kind of like mothballed it basically
for a couple of years. And that was just because
at that point, we had just moved out of the
house and into an office and then into the first office,

(06:01):
and there was just too much content flow on the
video side. And up until that point, I was the
only one who was capable of maintaining the site and
publishing to it, not because it was a special skill,
but because the website was so completely fucked up and
cobbled together because.

Speaker 2 (06:16):
I didn't have a CMS or anything it did, but
I built it like, oh so right, okay, yes.

Speaker 3 (06:22):
So I had a CMS, but I had stuff bolted
onto it over twelve.

Speaker 2 (06:25):
Years classic CMS. Shit.

Speaker 1 (06:27):
Yeah, and I'm not a professional web developer.

Speaker 2 (06:30):
And don't worry, most CMS developers on either.

Speaker 1 (06:33):
Yes, that was something I came to realize. And so anyway,
it's just more and more, you know.

Speaker 3 (06:40):
All the pieces were falling off the car while I
was driving it, and eventually I was like, I can't,
I can't, Like I have to just do one thing
and YouTube is not a platform I need to maintain,
which is good and bad. Yes, but yeah, we kind
of put the website on ice for a while and
then it wasn't until Wendell for Level one Text eventually

(07:01):
approached me and he was like, hey, I'm a web
developer actually, and he really wanted us to get our article,
our content script preserved in article form again, right, because
he's worried about just the loss of information you know
through video, right, And so he set us up to
where now people on the team like Jimmy are able
to maintain the site.

Speaker 2 (07:22):
That's really cool. Yeah, and you've brought it back though
you're somewhat more fun Like is it do you ever
see the website growing into more of a media outlet
or is it just and is it just you writing
it as well?

Speaker 3 (07:35):
Yeah, it's just us, you know, it's the same team.
It's the it's basically the video scripts that the team
rights converted into an article and so sometimes like it's not.

Speaker 1 (07:44):
Going to read as naturally as a pure article.

Speaker 3 (07:47):
Might, but we try to adapt it and if there's
any right now to me, it's like sort of a
community resource where I know for us, at least internally,
it's way more useful to have words, right, you.

Speaker 1 (08:03):
Know, just skin through and control app than a video.

Speaker 3 (08:07):
But also just like preservation wise, it's easier to preserve
articles and videos.

Speaker 2 (08:12):
Are you how is YouTube as a platform though, is it?
Are you a slave to the algorithm so much do
you try and appeal to it? Do you just make content?

Speaker 3 (08:22):
I think their YouTube has a platform. There's a lot
of ways to feel about it. For me, it's it's
almost like asking me how I feel about water at
this point, right, you know, it's like, how do you
feel about water?

Speaker 1 (08:36):
It's like, well, I mean, if I don't drink it,
I die.

Speaker 3 (08:40):
But like YouTube, it's it's not something I try to
game or play to.

Speaker 1 (08:47):
It's basically just a fact of life, right, you know.

Speaker 3 (08:49):
It's like it's it's there, and I have to think
about it, and every now and then I'll probably complain
about it. But realistically, I personally, I think it's kind
of a fool's errand to chase quote unquote the algorithm
too much because it kind of stifles and takes focus
away from the thing that actually matters, which is the content,

(09:12):
you know, and so like, yeah, there's things you can do,
like you can play with the thumbnail or the title,
but what I really wouldn't want to do is change
the actual type of content that's been produced just for
algorithmic purposes, because I think that that also creates kind
of like a almost like a black hole of creativity
where you stop producing the content for the purpose it

(09:34):
was intended and you start producing the content to be
a farm. To use your phrase, at that point, become
basically a slave to the algorithm where you can't escape it.

Speaker 1 (09:44):
Once you start doing that, I think.

Speaker 2 (09:45):
Yeah, at that point, you'll coverage is just dancing for
somebody else who changes their mind constantly.

Speaker 3 (09:50):
Right exactly, because it's you no longer know who the
viewer is. And it's impossible to know what YouTube wants
because YouTube doesn't know what it wants. You could probably
ask a YouTube bene near a pointed specific question right
about the algorithm and they would probably tell you they
don't know.

Speaker 2 (10:05):
So and has that has your experience with them being
chaotic or is it?

Speaker 4 (10:10):
Uh?

Speaker 1 (10:11):
No, because believe it or not, they don't talk to us.

Speaker 2 (10:16):
So, I mean, I've heard content creators who do argue
with them. I've heard ones that don't talk to them
at all. It's interesting to hear, especially your your two
point six million this I think so yeah, that they
don't interact.

Speaker 3 (10:27):
Yeah, So there was a period where they would assign
basically a channel rep to you, and so I would
get on a call, I don't know, somewhere around maybe
it wasn't the high hundreds of thousands of subs, And
that was kind of cool because like, Okay, if something
catastrophic happens, like let's just say, I don't know whatever,

(10:48):
I get locked out of my account because I fudge
the password too many times, something stupid like that, Right,
like I at least have someone I can talk to.

Speaker 1 (10:55):
Those people were rotated through.

Speaker 3 (10:58):
Revolving door basically every three or four months, so right
when you start to know the person and they can
help you, they're gone. And that was by design, I think,
as far as I understood it, that program got killed.
We don't have a REP now, which is normal, and
and the best I have is a a liaison. He's
called and he's a great guy. He's very nice. It's

(11:20):
not really his job, you know, to help us, but
he does it because he's a nice guy, and so
he's like the guy for like dozens of creators.

Speaker 2 (11:29):
Yeah, it's just so bizoved to me that there isn't
a specific rep with with other people. But like a
rep whose job it is to kate, you would think
that your entertainment talent on some level, you'd.

Speaker 3 (11:42):
Think they'd want to be available to help, and I'm
sure if you asked Google they would say that they
do want to be available to help, and they are
and you can tweet at them, you know.

Speaker 2 (11:52):
Yeah, great. Yeah, but that's the weird thing about these platforms.
It does feel like they're like they benefit as much
as possible while providing.

Speaker 3 (12:00):
So the revenue share is public, you know, for ad
Sense YouTube ad Sense, so tho's the ads before after,
in the middle of videos whatever. I think it's forty
five fifty five, with them getting forty five percent.

Speaker 1 (12:13):
For perspective something like Steam.

Speaker 3 (12:15):
I'm not an expert on this, but the last nieber
I saw was something like thirty percent for Steam, and
I think it's very able dependent on them for them, yeah,
for them, So you know, YouTube certainly benefits.

Speaker 1 (12:28):
They to argue in their favor, Hosting videos is unbelievably expensive,
so I get it.

Speaker 2 (12:33):
I didn't know that.

Speaker 1 (12:35):
Yeah, yeah, well the split's pretty high.

Speaker 3 (12:37):
I'm okay with it, Like I actually, I'm fine with
it with them getting forty five percent or whatever. It
is a bad sense because they don't impose restrictions on
things like us selling our own ads, right, So it's
like I don't really care, you know, and I don't
care if people block ads on our channel or whatever.

Speaker 1 (12:53):
I don't give a shit.

Speaker 3 (12:54):
But like the I think it's it's only if if
they ever overstep and they start restricting what you're allowed
to put in your content in a way that beyond
like I don't know, whatever they have their own rules
about like hate.

Speaker 1 (13:09):
Speech and stuff like that.

Speaker 3 (13:10):
Right, it is fam Yeah, but like in terms of
if they start restricting, let's just say they decide, yeah,
you're not allowed to sell merch without giving us a cut.

Speaker 1 (13:19):
No, then I'd have a big.

Speaker 2 (13:21):
Propool style is the YouTube way you make most of the.

Speaker 3 (13:24):
Revenue in one way or another. Yeah, I mean YouTube
is the reason it's possible to do any of this
because I was doing it as articles only, and YouTube
didn't really become a focus until I said, like twenty
eighteen or so as a primary source.

Speaker 1 (13:41):
And so yeah, through.

Speaker 3 (13:44):
Mostly merchandise sales on our store, through Patreon support, which is,
you know, the monthly donations from viewers, and then sort
of after that it's like ads we sell.

Speaker 2 (13:55):
And then ad sense right and wait, so you spent
like a decade mostly just writing pretty much.

Speaker 3 (14:02):
Yeah, probably probably till I would say, like till twenty fifteen,
very seriously focused, not only writing.

Speaker 2 (14:09):
Yeah, so was that your full time thing? Did you
have other jobs as well?

Speaker 3 (14:13):
I did several side quests? Yes, I would collect side
quests from local business owners, you know, like Greeting's adventure.

Speaker 1 (14:23):
I need a local website. Can you build that for me?

Speaker 2 (14:26):
It's a living, as they say in the Flintstones, it's
and let's actually get really not necessarily super specific. But
what's your history, like, without getting to biographical, like how
did you come in to do? Are you? Were you
just naturally interested in this? Was this just a childhood thing?

Speaker 1 (14:44):
I guess depends how far back we go.

Speaker 3 (14:46):
But gaming was always an interest and continues to be
an interest, of course, I don't know. I mean the
first game I played was probably Lemmons.

Speaker 2 (14:57):
Hell yes, yeah, hell yes Humans for me? Yeah yeah,
like first Lemmings game, yeah yeah, hell yes. Now I'm
so you were primarily a PC gamer though, yeah.

Speaker 1 (15:08):
For sure?

Speaker 3 (15:09):
Yeah so Lemmons, and then there are some other ones
back then, and uh I played a lot of nes
S nes N sixty four game, keep all that stuff,
a lot of Nintendo eventually.

Speaker 1 (15:20):
Got a PS two as well.

Speaker 3 (15:21):
But but after the N sixty four era, so like
early two.

Speaker 1 (15:25):
Thousands was when I built my first computer.

Speaker 3 (15:28):
Right before that, I was playing games like Command and Conquer,
Red Alert, yeah, and things things of that nature on
what you know. This was back when every family had
a quote unquote computer room, right, yeah, like there was
a room for the computer.

Speaker 2 (15:47):
Or a nook in Mike England's so small.

Speaker 3 (15:50):
Right yeah, but like point being where it's just like
this is the this is the dedicated computer area for
the computer, this household chair.

Speaker 2 (15:57):
Yes.

Speaker 3 (15:58):
So I played games on that built computer, you know,
early twenty penny and four and remained interested, and I
guess I started really getting more interested in the coverage
when we don't feel free to ask a file up
if you're interested in this.

Speaker 1 (16:15):
But basically I was running like a gaming guild like
group of friends, you.

Speaker 2 (16:20):
Know, nice So well, what were you gaming together? Is that? Yeah?

Speaker 3 (16:24):
Counter Strike, Source Age, Evampires three ever, request, things like
that you played?

Speaker 2 (16:30):
Request?

Speaker 1 (16:30):
Yeah?

Speaker 2 (16:31):
Whatsoever?

Speaker 3 (16:33):
I think I was on Kurana, I was on the Raith,
I was moving. I think Karana merged into the race.

Speaker 2 (16:40):
I went on storm Hammer as well. I paid the
extra It was not worth it, man, oh, good to know. Yeah,
same scars.

Speaker 1 (16:46):
Then I interviewed Brad McQuaid.

Speaker 2 (16:49):
I have a long and storied history of being a
problem for Sony online. Statement asking like the one British
journalist who asked any questions about this game, I got
in lots of trouble.

Speaker 1 (17:00):
Ever, ever quest I remember buying it.

Speaker 3 (17:02):
When I bought it was when it was sort of
demonized in some media as yeah, and also though specifically
as like the don't there was like some kind of
devil worshiping, like whatever, kind of like D and D. Right, Yeah,
D and D got the same treatment.

Speaker 2 (17:19):
They didn't, but they didn't seem to have a problem
with the fact that getting to level fifty was a job.
Right you had to work eight hours a day. Yeah,
Oh god, so you had this guilt?

Speaker 1 (17:31):
Sorry?

Speaker 3 (17:31):
Yeah, And long story short. The website I had built
for that group of friends with a forum that had
all these gaming guides on it, we worked.

Speaker 1 (17:41):
Very hard on. As you know, we were pretty at
that point.

Speaker 3 (17:44):
We were mostly high school early college age at the
at the highest end, and everyone put a lot of
effort into making these guides. And then at one point
the website was hacked by just some common CMS breach, right,
and I didn't know how to to restore from backup
and didn't really know what I was doing, so I
lost all the data. And so you could imagine for

(18:05):
a group of like teenagers, losing gaming guides is like devastating.

Speaker 1 (18:10):
Yeah, it's like here, like my life's work. A guy
for how to rush and Asian Vampires.

Speaker 2 (18:15):
Yeah, oh my, well, I mean that was game e
fa ques there, I guess, but yeah, if you were
making your own, they can't have been that good on
game efic hues.

Speaker 1 (18:23):
Yeah.

Speaker 2 (18:23):
So the Library of Alexandria that yeah, sorry.

Speaker 3 (18:26):
Yeah, yeah, so we lost that and that's when I
made a new website and that was the Gamer's next
to site.

Speaker 2 (18:32):
So so you're all self taught then pretty much?

Speaker 1 (18:36):
Yeah.

Speaker 3 (18:36):
I mean self taught in the sense that there's not
a lot of formal education, not self taught in the
sense that the people I've worked with over the years,
especially engineers in the industry, are the ones who actually
taught me a lot. Yeah, right, but nothing formal, I guess.

Speaker 1 (18:53):
Yeah.

Speaker 2 (18:54):
And it seems that something I've noticed with your coverage
and actually the surrounding channels is it seems like people
in hodware are relatively generous with that time, like there
are some people who want but like a lot of
the scientific people seem very key to it and want
to help and make sure there's understanding.

Speaker 1 (19:09):
Yeah, I think so.

Speaker 3 (19:10):
I think people like the person who comes to mind
immediately is Tom Peterson, who currently works at Intel, used
to working on video.

Speaker 1 (19:18):
But he's the type of guy where.

Speaker 3 (19:20):
When you talk to him you can tell he's not
in it because he's trying to sell Intel's or previously
on videos product. He's in it because he likes the technology, right,
And so people like that, I think they tend to
be happy to just share yeah.

Speaker 2 (19:34):
Which is really cool. And other YouTube channels as well
seem like Lewis Rossman around Box, they seem like very
they want to collaborate, which is really cool.

Speaker 1 (19:41):
Yeah. Yeah Rossman and Steve from hard run Box. I've
done a lot of videos with Yeah.

Speaker 2 (19:46):
You're having fun, yes, because there's a lot of cynicism
and kind of depression in media because of the job environment,
but there's also even above that seems like just a
depression around the work, but it doesn't. It seems more
fun what you're doing, and they indeed throughout the hogware people.

Speaker 3 (20:02):
I think it's one of the things I kind of
really actively spend a lot of time managing, is trying
to make sure there's a cadence to the content where
we actually just did this recently. I kind of look
at it and it's like, all right, this is We've
had a lot of heavy stuff recently. We had tariffs
into black Market, into Bloomberg, you know whatever, and so

(20:25):
then we switched to publishing more folks on reviews methodology.

Speaker 1 (20:29):
I ran a review of a toy story.

Speaker 2 (20:31):
Computer which I've seen and it's insanely cool. Yeah, I
actually kind of I don't know how good it is inside,
but I love the look of it.

Speaker 1 (20:37):
Fine, I had problems with shipping.

Speaker 3 (20:38):
But but the point being, you know, we we do
try to, like I try to manage what the tone
of the content is and for how long, because you
just you don't want to lose the fun of it,
you know, And so yeah, I would say I would
say also, even when it is a story that might

(21:00):
be more maybe categorized as depressing, so like if you're
covering some kind of corruption or something in the industry,
there is still.

Speaker 1 (21:10):
Fun to the job.

Speaker 3 (21:13):
Of covering that thing where like the fun part is
not necessarily the topic. It's the trying to piece together
something that's like really complex and figure out how to
explain it to anybody.

Speaker 2 (21:24):
I mean, the GPU tariffs, the tarrist one and the
GPU black market stuff was very seemed like it probably
sucked a little of your soul out, but that looked
like fun. It looked like a fun ben.

Speaker 3 (21:35):
Is I mean, the tarist one, you know, it was
It's a story that there were people we spoke to
who are very negatively affected, and you feel for those people.
And that part is sad if you, you know, put
yourself in that position. But then on the the kind
of keeping it fun for yourself covering it side, if
you really just kind of step back and look at

(21:56):
it's like how fortunate can you be to for your
job be like I'm going to get on a plane
in twelve hours, you know, and fly and then I'm
not going to know where I'm going next until I
get there. Like that on its own is pretty fun. Yeah,
And that's it's very privileged to be able to do
that and know that at the end of it it's
going to be fine, you know.

Speaker 2 (22:17):
Right, And so those and those trips were kind of
semi chaotic.

Speaker 1 (22:21):
Then they're very chaotic. But that's like, that's why it's
fun because you.

Speaker 2 (22:25):
Got to was it Hong Kong and you immediately like
got a price sheet for GPU.

Speaker 1 (22:30):
Yeah.

Speaker 3 (22:31):
Yeah, we were speaking with some suppliers and so we
had a price sheet. Yeah, we went to we went
to China for that trip. We met a guy we
weren't planning to meet. He was mister five in the
video and he was awesome.

Speaker 2 (22:45):
And where do you find these? Is it people come
to you? Are they connections? I realized YOUNI laterally you
probably can't onnset.

Speaker 3 (22:51):
But normally I normally I have a source for kind
of the first link in the chain, right, and then
it just kind of develops. You meet a person, you
hopefully leave a good impression with them, and then they
might say, you know, hey, I know someone who might
be interested in talking to you, and you kind of
go from there. And I think the biggest thing is like,

(23:13):
and this is kind of challenging sometimes, but learning to
to kind of roll with it where I operate at
a relatively high level of anxiety in terms of preparedness, yeah,
you know, and so to have a wrench thrown in
where it's like, hey, you might be interested in meeting
this guy. I'm like, well, like my schedule is really
I have figured it out. You have to be willing

(23:35):
to deviate from it.

Speaker 2 (23:36):
Yes, that's the only way to there's kind of any
good broadcast work.

Speaker 1 (23:40):
Yeah, and you pulled it off.

Speaker 2 (23:41):
It's good. So you've got quite an operation here as
well around how many people work.

Speaker 1 (23:46):
With you day to day.

Speaker 3 (23:49):
It's five to five to ten total day to day,
so it just depends you know who's doing what each.

Speaker 2 (23:57):
Day, right, And then mostly around editing the videos and stuff.

Speaker 1 (24:00):
A couple editors slash camera operators.

Speaker 3 (24:03):
We have a couple of writers slash testers, so kind
of like if you're testing the product, you're probably the
one who's going to write the review that often makes
the most sense, not always, but and then editors often
will shoot the b roll they need because they kind
of they hit a clip, they realize they need something.

Speaker 1 (24:23):
Yeah. And then we've got.

Speaker 3 (24:25):
A remote researcher as well who's been contributing to some
of the new stories coming up.

Speaker 2 (24:30):
Very cool. Yeah, and how long does the video take?
Putting aside the obvious ones like the tarist one, which
I think a little bit different. How long does it
take to get a review together, or like any particular video.

Speaker 3 (24:41):
A review of I can give you actual numbers. A
CPU cooler requires forty hours of testing work where there's
some kind of manual involvement from the technician, so that
if we have one cooler, that's going to be someone's
job for one week basically, right and so right now,
that'll be Mike typically running those tests, and then the

(25:04):
actual So if I write that review, it probably takes
me two hours to write it, you know, it takes
me the runtime plus ten or twenty minutes to film it,
and then because you obviously have whatever mistakes that you know,
you do retakes right, so runtime plus hundred twenty minutes,
and then the editors I would say they typically take

(25:24):
about eight hours to complete a relatively simple review edit
plus some camera work.

Speaker 1 (25:30):
So a cooler review you might be.

Speaker 3 (25:32):
Somewhere in the range of fifty to max maybe sixty hours.

Speaker 1 (25:36):
And then something.

Speaker 3 (25:38):
Like like the black Market video. If you don't count
my time, so if my time is zero, I don't
remember exactly how many hours we had in it.

Speaker 1 (25:48):
It was.

Speaker 3 (25:50):
I know if you count my time it was over
three hundred hours. Jesus, it might have been more like
four hundred. But and then the ASRock motherboard video we
just did yeah on CPU failures and Azrok boards. That
one was like with editing and filming time, I think
that was two hundred and forty hours. So like, that's
an example of a content piece that we will lose

(26:12):
money on, but it's like it's subsidized. First of all,
I don't care because I want to do it. Secondly,
you do have to pay for it somehow, and so
for us, we basically we raised so much money from
the black market video, I'm able to go, Okay, that's paid.

Speaker 2 (26:27):
For, you know, moove me through that situation.

Speaker 1 (26:29):
Yeah, which the Azrok thing.

Speaker 3 (26:32):
So the Azrok thing is they have some kind of
yet unknown issue that is resulting in the death of
CPUs expensive ones, and people don't know exactly what it is.
They haven't been entirely forthcoming about it. It seems like they
don't know what the fuck's going on. And so we
got a viewer's motherboard that had killed the CPU and

(26:53):
did a bunch of diagnostics. This was one of the
instances where we couldn't come to a conclusion, and we
decided let's just like, let's just publish everything and maybe
someone can use it.

Speaker 2 (27:02):
As the company being communicative of they been trying to
fix things.

Speaker 3 (27:05):
They appear to be trying to fix things, just not
very successfully. I would not say they've been communicative like
they haven't. I don't think they've done a good job
at telling their customers what's going on.

Speaker 2 (27:17):
Right. So, how is your relationship in general with the
hardware manufacturers you mentioned the Intel. It sounds like you
have some people who are friendly and yeah, others.

Speaker 1 (27:26):
Yeah, it depends, you know.

Speaker 3 (27:27):
I'm sure you've worked with enough people where the people
are often different from the company in terms of the stance,
and it really depends person to person. But generally speaking,
the companies are able to maintain a fairly open line
of communication and be relatively mature about even criticisms because

(27:49):
normally the people actually in between us and their bosses
are pretty good at their job, right, you know, And
so like there's a guy at who he's incredible at
what he does because and I really don't think consumers,
I think a lot of people don't know this job
role exists, but it's really important and the role is

(28:11):
effectively to be like the translator between outside criticism and
internal action.

Speaker 2 (28:17):
Right right, So not quite a PR role, but like
a developer. Not quite developer really.

Speaker 3 (28:22):
It's almost like like tech marketing maybe yeah, but like
they're not really marketing in the traditional sense. But so
the guy at AMD, he does a really good job
because he's told me how if we have a criticism.
I asked me to what happens internally, and he said, well, normally,
like marketing might go to him and kind of be like,
what the fuck, Like why did Steve say this? Or

(28:43):
hardburn boxed or whoever, which is covered by Steve, I guess,
but they might ask why did the Steve say this?
And it's his job to figure that out. And he
was telling me he normally just asked them, well is
it true? Right, And if they say it's true, then
he says, well, then make it not true by fixing it.

Speaker 2 (29:01):
See, this is the thing I've run a p off of.
And it's like when clients come to White they say that,
I'm like, and many times said why did you do that?

Speaker 1 (29:08):
Right?

Speaker 2 (29:09):
And they say, well, it's not a fair I'm like,
how is it unfair? Because if you can explain to me,
I can go and get this fixed, right, you have
to explain to me for and it is interesting that
that role has to exist.

Speaker 3 (29:18):
But yeah, I mean, if it's just a communication problem, right,
if they're like, well, it's it's true, but we didn't
we don't like it. Yeah, right, so it's got to
be a real reason. But yeah, most of the companies,
the relationship is fine.

Speaker 2 (29:32):
When they generally take criticism.

Speaker 3 (29:34):
Well, the people who interface with us take the criticism well,
because it's just a job. The companies don't always.

Speaker 2 (29:42):
Do you have any executive exposure? Do you know if
any executives watch?

Speaker 3 (29:46):
I know that Johnson Juan once watched at least part
of one of our videos, at least because it was
communicated to me by someone it works.

Speaker 2 (29:56):
With him, all bold, all caps.

Speaker 1 (29:59):
Ye, I was told that he was not thrilled.

Speaker 2 (30:02):
Oh that doesn't that doesn't sound like Jensen. He's usually
such a cool head. But do you do you know
if I don't even need names, but do you know
if there's like a good amount.

Speaker 1 (30:10):
Of them, Uh?

Speaker 3 (30:13):
Sometimes, yeah, I mean we AMD we had a video
where we were basically like begging them to not fuck
up the launch of their nine thousand series GPS because
their competition like no one had showed up for the
consumer Intel. They were kind of there, but they're like
not super viable yet. And anyway, I know after that

(30:34):
video they had emailed us and this wasn't a thread
I told them was on record, but they'd emailed us
and said, uh, you know, basically their sort of executive
marketing team had watched it and they did talk about
the problems we raised, and that's cool. It seemed like
they addressed some of them.

Speaker 2 (30:51):
Yeah, do you ever try and get interviews with them?

Speaker 3 (30:54):
Yeah, occasionally, I mean, especially if there's like the Asus
situation with their warranty as context for people has had
an ongoing problem with customer support, where people who need
a warranty field often end up posting online saying they
got screwed.

Speaker 2 (31:12):
Is this across the board or with specific things?

Speaker 1 (31:14):
Definitely Motherboard. I'm not sure about other categories because I
really liked the.

Speaker 2 (31:19):
Rock Gangs ally, but have had a few listeners say yeah,
we've got some warranty issues and it's actually I kind
of wanted to address that with you in the know.

Speaker 1 (31:26):
Well that one.

Speaker 3 (31:26):
I know a lot about the ally specifically yeah, So
to like close the loop on the executive question, the
Asus thing, I we published a series about the warranty problems.
We proposed a number of fixes, and eventually I really
pushed them to let us speak to an executive and
customer support. They put a director of marketing in front

(31:50):
of us, very very nice guy, yeah, but by his
job right, And so I kept pushing back in him
where I was like, look, man, nothing against you. You know,
you're not the guy. And so they eventually got us
to the to the right guy, which is something I
try to remember too.

Speaker 1 (32:09):
We really try.

Speaker 3 (32:10):
To go to executive levels if possible, because if it's
just some dude who was told to do with in,
I can't really press him on the decision.

Speaker 2 (32:18):
You can't punish him. It's not really his decision making.

Speaker 3 (32:20):
It's company, yeah, and he's not paid enough to like
deal with it, you know. So so we try to
go above that person. And we did the same with
New Egg. They sat a room.

Speaker 1 (32:32):
Full of like four, I think through four executives with
us a couple of years ago.

Speaker 3 (32:36):
Yeah, so they a lot of times they'll play ball,
which like credit to them, you know. But the asues
ally thing, the long story short on that is, we
had a defect. I want to say it was like
the joystick or something. We had some kind of defect.
We sent it in for actual repair on our unit

(32:57):
anonymously or you know, as a pseudonym, and they sent
us a photo of the xtue of the chassis where
there was a tiny nick, like a tiny tiny crater
that's basically the size of a pinpoint in the end
the edge of the chassis that is purely cosmetic. And

(33:18):
this is something that like we literally put it under
a microscope to see what they were talking about.

Speaker 2 (33:23):
And they said it wasn't covered.

Speaker 3 (33:24):
And they said that they would have to charge us.
I forget how much it was. It was like ninety
or one hundred and eighty or something dollars goodness to
repair a problem that was unrelated to this cosmetic thing
and was their fault, and they were trying to use
it to charge us for the repair. Yeah, so we
ran that as a story, and you know, I mean

(33:46):
eventually it was handled, but I don't count it as
being done correctly if they find out who we are
and then they fixed it.

Speaker 2 (33:55):
That was the thing. Because you said it was anonymous,
like a dummy thing, right, and I assume you it
with a different card with it, Like.

Speaker 3 (34:02):
I think I bought it from like a retail physical story, right,
so there's no.

Speaker 2 (34:05):
Trace because I like that thing. But now I'm like regretting.

Speaker 1 (34:09):
That's a good device.

Speaker 2 (34:09):
But it's like, hopefully it doesn't break.

Speaker 4 (34:11):
Yeah, it's a great device as long as it never
has a problem.

Speaker 2 (34:25):
So okay, let's change. Let's change tag to the industry
at large. And I want to talk to you about
electronic arts because I know you just did a video.
My whole thing was, how does he get worse? Because
there's a company they've been dog shit for a while.
And that's a personal opinion.

Speaker 1 (34:42):
I think that's an objective.

Speaker 2 (34:43):
Yeah, I mean, look at the Maden franchise.

Speaker 1 (34:46):
I think I don't know. EA is like bizarre to me.
I don't get it.

Speaker 3 (34:53):
I know you can look at their financial reports and whatever,
but the powers at play and the EA acquisition, it's
multiple governments through one connection or another. And I think
for a lot of people, regardless of what they think
about social issues, it is just, if you really think

(35:14):
about it, weird for government or government connected entities to
start acquiring video game companies and I personally, like, I'm
very skeptical of it because the US government in particular
has dragged video games at every opportunity they get for decades.
It's always like the video games cause violence, you know,

(35:37):
and it's from people who've probably never played an actual
game in their lives.

Speaker 2 (35:41):
And from one instead of this deal, it's mostly it's
like Saudi money and PE money.

Speaker 3 (35:45):
Yes, So it's yes, So there's private equity. As he said,
there's Saudi money through the PAF and then on the
US side Affinity Partners, which is helmed by Jared Kushner
right to former senior advisor to the President, and I
think who he just sent over to some kind of
peace discussion or something, so he's still involved somehow. But anyway,

(36:06):
so that's the US side where you've got PE money.
You have somehow indirectly the US government but through a
former US government official and then his firm, the Affinity Partners,
has received somewhere around two billion dollars of initial investment
from the Saudi pif that was a New York Times

(36:27):
report previous.

Speaker 2 (36:27):
So is funded by the other people invested.

Speaker 3 (36:32):
Yeah, And I think if I remember the numbers correct,
it was close to ninety million dollars out of there,
one hundred and sixty or so million in management fees
for the last filing was from managing Saudi money. And
then you've got the PIF from Saudi Arabia working with
this firm to buy EA Games. One of the things

(36:52):
we didn't talk about in the video that I just
I hadn't really thought too much about it, and I
saw some excellent comments that made me think more about it,
But is that companies like EA Games control a huge
amount of data that seems inconsequential on the surface.

Speaker 1 (37:10):
It's video game data, Like, right, what is okay?

Speaker 3 (37:13):
So what Let's just pretend there's a clean path for
the data to exit EA Games and go to the
US government or go to the Chinese or the Saudi
rating whatever. Right, Let's just pretend there's a path there.
They can't really do anything useful with your save game file.
But one of the things that does interest me, and
I just want to maybe caution here that this isn't.

Speaker 1 (37:38):
I don't think there's anything going on here right now, but.

Speaker 3 (37:41):
I think there is the opportunity to abuse anti cheat systems,
which are like kernel level software in most cases that
run on the computer. So if you wanted to deploy
like some kind of root kit with very low level
access to computers, you could do it through anti cheat solutions.

Speaker 1 (38:04):
Now, I don't think this is.

Speaker 3 (38:05):
Happening right, right, and so I want to make sure
people know that because you don't want it sound like
it's some crazy, wild conspiracy theory.

Speaker 1 (38:11):
I'm just saying it's possible for it to happen.

Speaker 2 (38:13):
But I mean there's probably a trophy of dated with
the EA play or whatever horrible cloud service.

Speaker 3 (38:18):
For sure.

Speaker 1 (38:19):
Maybe they have chat logs that could be useful to elements. Right.

Speaker 3 (38:23):
The US government has talked about wanting to bring Gabe
Newell in for a testimony of some kind to talk
about radicalization of teens who play video games.

Speaker 2 (38:35):
Right, and they're doing that with the Discord CEO. I
think or they are already did that.

Speaker 3 (38:39):
Yeah, And I mean, if you want to go way back, right,
this has been the circus has been had before.

Speaker 1 (38:44):
It was just last time was with rock music.

Speaker 2 (38:47):
Right, So god, I just as just said, it's like,
I don't know how EA gets worse though, because you
were sports I don't know if you're a sports fan
or anything.

Speaker 1 (38:56):
I'm aware of sports.

Speaker 2 (38:57):
Yeah, I said, this is the one of the pigs
that buys Madden every year. It has got worse while
also staying the same. And I don't know, I mean,
the only way for them to make it worse is
to just have it steal your money at this point,
just actually go through your Walter. I just don't have
they spoken of their plans for the company other than using.

Speaker 3 (39:16):
AI Yeah, great, this is why you should film to expression.

Speaker 1 (39:23):
Yeah, just the the revulsions like you melt it. Yeah.

Speaker 2 (39:28):
And that most of the debt, it looks like he's
going to it's sort of most of the revenue is
going into the debt as well.

Speaker 3 (39:33):
Yeah, I think it's I think it was like a
twenty billion dollar debt or something that they have to
pay down relatively fast. I don't remember how much pre
you're off the top of my head, but it's it's
enough where not sure exactly what the game plan is.
But the current CEO, who's remaining in the CEO at
least right now, I think his name is Andrew Wilson,
I want to say, has stated that they have plans

(39:56):
for use of AI EEA, and I think the as
a whole have stated use plans for agents AI agents.

Speaker 2 (40:06):
The very real thing that exists, great stuff fucking EA.
So that's how they do it. They become a AI
and it sucks even harder. But let's change topics to
another weird acquisition investment. Intel, how do you feel about
this in Nvidia Intel situation?

Speaker 3 (40:24):
I think the Nvidia Intel US government situation, Yes, it's
really weird. I am not I mean I don't like it.
I think I understand the US government's interest. I'm sure
they look at it as absolutely a national security asset.
It's our only way we can even dream of making

(40:47):
relevant competitive chips in America. And so if for some
reason you needed a local resource to do that, Intel
is it. So I get it, But at the same time,
the way it came about is kind of bizarre to me.
Where the timeline of events was. Trump says Intel CEO

(41:12):
lip Putin is quote highly conflicted, and that the only
resolution to this problem of being highly conflicted would be
for him to resign. The reasoning it seemed for that
belief of being highly conflicted is pasted or current investments
by Intel's current CEO in Chinese companies, including some which

(41:34):
I have Chinese military ties. So that's my understanding of
that event. Immediately following that, within days, well within one day,
Intel respond publicly. Intel CEO within days gets on a plane,
meets Trump. Now they're friends. Trump says it, calls him
a success. And then shortly after this, you know, you're

(41:55):
on this this roller coaster, Intel stock the whole time
plunges when he says he needs to resign.

Speaker 1 (42:00):
It's skyrockets, you know.

Speaker 3 (42:02):
And then shortly after all that, the US government is
effectively acquiring ten percent of Intel, right and yeah, I mean,
and on then video side, after a couple of weeks
after this, Stop and Video is acquiring five billion dollars
worth of Intel, which I think is around four percent
or something according to Reuters.

Speaker 1 (42:22):
So I just I don't really know.

Speaker 3 (42:26):
I guess my answer is, ed, I don't know what's
happening anymore.

Speaker 2 (42:31):
And that's kind of where I've come down in this whoop,
because it's not clear what happens next. They're going to
do something together. Do you think they're going to keep
making the elk GPUs or is that because they haven't
said they're going to stop making them.

Speaker 1 (42:41):
I think they're at risk. I that was a that
was an unintentional pun.

Speaker 3 (42:47):
Very few people will get no, I loved it nice. Yeah,
I think there is some risk there. So, like the
stated plan is that they're going to work together on
X eighty six, which is an ISA and instruction you know,
architecture that is used in CPUs, and they want to
work together on x eighty six solutions, which there's not
a ton OF's. There's a lot of ARM there's Intel,

(43:09):
and there's AMD for x eighty six for the most part,
some via but and they also are planning to work
together on envy Link integration, which was Nvidia's actually effective
solution was that.

Speaker 2 (43:23):
They go through Mellano's acquisition.

Speaker 3 (43:25):
Mellanox is their networking infrastructure, right and envy Link is
there sort of on board, although now it's expanded, but
basically PCIe alternatives. Oh okay, so so PCI Express wasn't
doing it, friend VideA. This is actually one of the
areas where it's not just all marketing bullshit, Like envy
Link serves a real purpose. It does a real thing

(43:46):
and and they need it. And so they're going to
work with Intel to integrate this protocol into other products
which will make end videos. GPUs more lolliable on CPUs
like x eighty six CPUs.

Speaker 2 (44:04):
Right, do you think there could be a good thing
about this? Could that actually be a positive?

Speaker 3 (44:08):
It'll definitely be better for Nvidia's solutions. I mean, if
you want to look at it purely in a vacuum
of ignore literally all business and all competition aspects and
look at only the product level, their product should be
better as a result of it. In theory, I think
there's risk to both and the end to Intel here.

(44:32):
So in particular, something that's interesting that a lot of
people don't know is the mobile the laptop side of
the business where this is another stated goal of NVIDI
and Intel is they want to make laptop hardware, so
they want to make silicon with RTX chiplets, which is
a tiny piece of silicon for the the sc or
the CPU. And on the laptop side of the business,

(44:54):
the entire industry aligns to end Video's schedule. So if
Nvidia and Intel previous we're both launching a CPU and
this just happened at about the same time they launched
a cp in a GPU between the two of them,
the vendors will align to Nvidia's GPU, meaning all of
the marketing effort, the money, the sampling, the review guidance

(45:14):
where you know they will get in touch.

Speaker 1 (45:16):
With reviewers and educate them on the new architecture.

Speaker 3 (45:20):
All of that basically happens around Nvidia's timing because Nvidia
sells units, right and so because of that, if Nvidia
is suddenly, you know, if they launch a new GPU
for mobile right now, Intel and A, it doesn't matter
which CPU is in it, MSI or Asus or hpuor

(45:41):
the Nobo or Dell. They' launch their notebook with whatever
CPU doesn't matter. The Nvidio GPUs the part that matters
to them. If Nvidia is working with Intel, then suddenly
it would seem that there's motive for them to leverage
allocation of GPU Silicon with the intent being to get

(46:04):
more Intel Nvidia co branded.

Speaker 1 (46:07):
CPUs deployed in notebooks. You see where I'm going with this.

Speaker 2 (46:10):
Yeah, because this is kind of a long tail thing.
But it's like the old ms DOS situation, Yeah, where
OEMs just ultimately went with dot because could Nvidia push
down prices?

Speaker 1 (46:20):
I su moor, Well, I mean they they historically have
been I don't know.

Speaker 3 (46:26):
I guess you'd say maybe revolted against an EVJ situation.
But we've covered plenty of vendors in the past. I'll
just name them now because it's been long enough. But
like a sus, MSI, Gigabyte, and EVJ have all told
us about times where allocation, which is the it's the
pot of gold, right, it's it's how many chips they get,

(46:47):
is withheld or is modulated based on their willingness to
comply with whatever the current sort of requirements are.

Speaker 2 (46:56):
And what would those requirements be.

Speaker 3 (46:58):
In the past, it's been stuff like well, I mean,
on an EVGA side of things, they had always told
us about being restricted in their ability to design certain
high end boards that might have like overclocking solutions, engineering solutions,

(47:18):
and so there was there was kind of like a
trade behind.

Speaker 2 (47:20):
The those restrictions happened. They do they need something of
an nvidious.

Speaker 3 (47:24):
Side or well sometimes yeah, but it can be like
we need you to not do that, or we need
you to sell a certain amount of this priced class
of card. So as an example, there was a time
where we'd reported when multiple sources at EVGA informed us
that the MSRP cards were not actually real, like they

(47:48):
were going to exist for a short period for launch
to comply with Nvideo's requirement, and as soon as it
was no longer a hard requirement to get the allocation,
they were going to kill the product because they didn't
have margin.

Speaker 2 (48:01):
And so they effectively had to release something that made
the money.

Speaker 1 (48:04):
Yeah, basically.

Speaker 3 (48:05):
Yeah, there was one I think we reported on where
they were making like four box or something, and that's
on a that was I think that was like a
three hundred dollars something Boar Christ Yeah.

Speaker 2 (48:14):
Yeah, so this is fairly typical for nvidio.

Speaker 3 (48:17):
This kind of allocation is is their leverage. Yeah, it's
also how they create internal competition between the board partners.
Now AMD does this too, and so does Intel. I
think with Intel the differences they don't really have anything that.

Speaker 2 (48:32):
A ton of leverage at the moment, not really.

Speaker 3 (48:34):
Yeah, and AMD, you know, they they also play games
with allocation. I think the difference is nvideo is just
the difference is if you make nvideo products and you
make am D products like GPUs, you can lose the
AMD ones or ten percent of the supply or whatever
and still be in business. But if you lose the
nvideo ones, you're fucked, right Like business is over.

Speaker 2 (48:58):
So they could use this with CPUs. There could be
a scenario.

Speaker 1 (49:01):
That is my concern.

Speaker 3 (49:03):
Yeah, my concern is if they're in mobile with CPU
and GPU. Now it's not just they're aligning to Nvidya's GPUs.
There's again a there's no evidence they're planning to do this,
but I think I've seen enough historical context to be
concerned about a possibility where they say, Hey, we'd really

(49:25):
love it if you would put more of the Intel
RTX CPUs in your notebooks so that you can get
enough fifty ninety GPUs.

Speaker 2 (49:32):
Oh so they will, They'll use one, So okay, now
I understand. So it's the leverage with other cards that
they'll use the force of this. Yes, potentially potentially. I
know that this is and this they've done this for years, like.

Speaker 1 (49:44):
Yeah, yeah, allocation is a is a big lever.

Speaker 2 (49:47):
Do you think they do similar things with the AI GPUs.

Speaker 1 (49:50):
Uh, I don't know.

Speaker 3 (49:52):
I don't really talk to because that would be probably
more like Dell HP like those big enterprise deployers. I
don't really talk to people there. I mean to me,
it just seems like this is kind of a company
culture thing, and this is all this is all me
speaking from what we've reported on. You know, we have
some facts for some of the stuff where we've we've

(50:13):
covered it, like with the EVJ situation, but yeah, enterprise,
I'm not sure how exactly that side works.

Speaker 2 (50:21):
So when it comes to AIGPUS, is there is it
a limitation of your testing as the way you don't
look into them because it felt like it's the one
thing I'm like surprised that you haven't dug into more
with like a one hundreds h one hundreds, Like.

Speaker 3 (50:34):
Yeah, So there's sort of there's like two sides to it.
There's testing and then there's just reporting on whatever the
news is. And so the reporting part we did that
with the black market video. The testing side, I guess
there's there's like a couple of things that are nested
with in testing. So we've done a little bit of

(50:56):
quote unquote AI testing and that was on the RTX
or six thousand Blackwell GPU. This is something that is
still new. It's not one hundred percent clear to me
what is a reproducible reliable test, right It's part of
the problem is if you're testing LLLMS or generative AI
or whatever, something that makes images by nature of the application,

(51:20):
it's almost semi it's different every time.

Speaker 2 (51:24):
We would also need a massive cluster to really emulate
what it is because one A one hundred or eight
one hundred isn't really gonna happen.

Speaker 3 (51:30):
Yeah, and that's kind of like you can run those tests.
But and this would be a valid request from the audience.
I think an audience that really cares about it would
be like okay, cool, So like what about if you
have you know, ten of them or whatever.

Speaker 2 (51:46):
And even then, how do you even practically?

Speaker 4 (51:48):
Yeah?

Speaker 3 (51:48):
And I just I think we can do it, and
I do think we'll probably integrate some kind of long
term testing for single GPU boards or dual GP single boards.
But it's right now, it's so new that to test
it properly requires a lot of research and a lot

(52:13):
of sort of trial and error.

Speaker 2 (52:14):
Well, so how would you even because with something like
GB two hundreds or GB three hundred, they've got their
own distinct cooling. You're not going to get a giant
racking wherever you want.

Speaker 1 (52:24):
Yeah, there's there's definitely a limit.

Speaker 2 (52:27):
And what are you testing for? It's just I've I
have spoke spoke to a few listeners and like, oh
what about AIJ why's you know, and it's kind of
on some level, what would you even be testing?

Speaker 3 (52:36):
Yeah, I think that's a big question for us, And
I think that's where like the most immediate thing that
would maybe make sense for us would be some kind
of consumer level I really hate to call things AI,
but like AI application. So maybe that's someone wants to
just buy a fifty ninety or whatever, use it for gaming,
and then at night they're training something on it, or

(52:59):
they're running some kind of system. That's the I think
that's the most sensible thing we could test anything where
you're getting into like actual data center workloads. I it
seems like the only real source for that stuff is
basically first party at this point.

Speaker 2 (53:15):
And semi analysis to an extent.

Speaker 1 (53:17):
Yeah, you spoke with them at all. They've reached out,
not some analysis, but I know the work.

Speaker 2 (53:22):
Yeah. I mean it's a good news Letterum just cure surprise.
They haven't the depth they go into. But let's talk
about AI. Actually, how do you feel about AI? Like
you you don't do a ton about it? Yeah, which
is fine, it's just how do you feel about it?
And what? Why haven't you done more?

Speaker 5 (53:39):
That?

Speaker 2 (53:39):
Not as a negative.

Speaker 3 (53:40):
Yeah, most of the coverage we've done has been news based,
so it might be reports, you know, it might be
like an investigation or whatever, but but not a lot
of I don't know, like testing, like we're talking about
the biggest problems I have with it right now are

(54:01):
First of all, I think if we go like really
big picture, I don't know to qualify everything. I think
there are use cases for things like LMS. The best
use case I have that I've actually used is translation
where I speak other languages, and I use Google Translate
for most of that. Because Google Translate is a dumb translator.

(54:22):
It translates the words you type into it. Sometimes that
doesn't work for a colloquial phrase, right right, So if
you take any idiom, it's not really going to do
that well through Google Translate. Something like chat GPT makes
it a little easier sometimes to search for like really
specific colloquial or idiomatic ways to express something. So I

(54:46):
would say that's like a real use case. You can
also be deep down there, Yeah, but it is like
the specific thing that they're kind of built around, so
they do it well.

Speaker 1 (55:00):
I think the.

Speaker 3 (55:02):
With that qualifier out of the way, the big picture
concerns I have with AI are weaponization of things like
LMS to propagandas and so this could be for companies
to market products, it could be for governments, it could
be for unknown entities whatever.

Speaker 1 (55:21):
But having seen the bot comments on.

Speaker 3 (55:23):
Videos you know every day, is that the problem? It
is a very consistent problem, and they're getting better.

Speaker 2 (55:29):
So what are they doing?

Speaker 3 (55:30):
So originally it started as the really obvious ones. There
were kind of two kinds. There's the bots that do
financial scam scams where they basically.

Speaker 1 (55:40):
Post something about like some cryptocurrency, right, some scam coin.

Speaker 3 (55:46):
The other one would be the bots that have a
photo of someone's ass, you know, and then three emojis
and some text about whatever. And that's the more like
traditional trying to scam. Some went into interacting with a
fake user that presents themselves as an attractive person.

Speaker 1 (56:05):
So those are the two common types of bots.

Speaker 3 (56:08):
Historically, those have been really easy for a savvy user
to identify because yeah, it's like it's always the same language.
It's like the old old you know, Nigerian prints email
scam right right, you can read it and you're like,
I know this is a scam. But where it's going
now kind of concerns me because there have been times
where I'm not sure if it's a bot or a
real person, right, And there was one recently I just

(56:29):
banned from the channel that originally I thought it was
a user because I forget what the message was. But
it was like, I wonder what Steve and Wendell think
about blah blah blah, And it was clearly polling stuff
from the title and or the transcript of the video
and then create an account.

Speaker 1 (56:49):
And I left it alone for a minute to see
where it would go.

Speaker 3 (56:52):
And there's like replies under the thread where it's a
bot replying to itself through different accounts I guess, uh,
and eventually just tries to you off site to go
you know, get scammed. So I banned all those accounts.
But the thing that was concerning was was that it's
pulling context from the video transcript and forming a sentence

(57:13):
that makes sense, right, and then creating the appearance of
a real dialogue between the appearance of real users about
this subject that.

Speaker 1 (57:23):
Made sense to then scam someone.

Speaker 2 (57:26):
I've seen these people in blue sky for sure. Yeah, sorry,
seeing these boots in blue sky where it's just someone
appearing to have a guy. It's like, wow, I read
that from the thing. Well do you think about this?

Speaker 1 (57:36):
Right?

Speaker 2 (57:36):
I've never seen them get to sending me to a website,
but I imagine that's that's a few down the change.

Speaker 1 (57:41):
The way it normally happens.

Speaker 3 (57:42):
If you go to like CNBC or any finance channel,
and you'll see these on any new video they post.
They all talk so like CNBC is my favorite one
to look at for this because okay, you'll see and
it's not their fault, they're just the target of it.
But you'll see a bot comment that'll say something like
I have hundreds of thousands of dollars and I don't
know how to invest it right, right, And it's always

(58:05):
some ridiculous number.

Speaker 1 (58:07):
Yeah, and then there's a reply to it.

Speaker 3 (58:08):
It's like I had excellent luck with John Smith, and
John Smith is the comments like, yeah, they're comments.

Speaker 2 (58:15):
I've seen these in Korra. Okay, it's like, how do
I invest one hundred and fifty thousand dollars? It's like, well,
I did it on right oscoin or whatever.

Speaker 1 (58:23):
Yeah.

Speaker 3 (58:24):
But these they'll talk about like the name of a
person and have this fake conversation about some financial advisor
who doesn't exist, and then you google that name, which
is fairly unique, right, so that's the only thing that
comes up. And then it's a website that's a scam,
and they you know, these.

Speaker 2 (58:42):
Point new ones compared to the original one.

Speaker 1 (58:45):
It's it's less obvious than.

Speaker 3 (58:49):
By questionable cryptocurrency or nft on website.

Speaker 2 (58:54):
Yeah. I don't know how you even deal with this.
I don't know what the solution is other than just
chutting them all down.

Speaker 1 (59:02):
Yeah, the channel owners have to ban them, but there's
too many.

Speaker 2 (59:05):
You get a lot reddit to them, seeing I don't
know if you're active on your subreddit much. I haven't checked.

Speaker 1 (59:10):
Not too much.

Speaker 2 (59:10):
Yeah, I'm in mind too much. But I do catch
someone occasionally who is just an obvious bar sick host, disgusting. Yeah,
but I just I ignorantly just assumed that they were
being annoying and they were there to sow discord rather
than send me to something to buy.

Speaker 1 (59:26):
Right. I think that's the thing too. I forget who
the report was from.

Speaker 3 (59:31):
There was a recent report about something close actually, I
think I saw it re reported and fact checked by
Kurtzkazak to the channel where it was close to like
fifty percent of Internet traffic is bought.

Speaker 2 (59:44):
Yes, yeah, and that's gonna fuck the ad industry.

Speaker 3 (59:47):
It's gonna be bad for everything because like, fuck the
ad industry. It's gonna be bad for humanity.

Speaker 1 (59:53):
Well.

Speaker 2 (59:53):
Oh, to be clear, I don't care about the idea.
I'm saying that that's how everything's paid for online.

Speaker 1 (59:58):
Yeah, it'll screw that all screw it's this is like
Library of Alexandria is on fire. Yeah. Problem like this
is this is like loss.

Speaker 3 (01:00:07):
Of of knowledge issues. And I think it's because, uh,
you know, there's there's been a consolidation of sites into
media forms of media or media for presentation like videos,
discord things where it's not well preserved.

Speaker 1 (01:00:24):
And as those.

Speaker 3 (01:00:26):
Things collapse, because like you said, the ad industry collapses
or whatever, there's gonna be a loss of information and knowledge.
The death of the internet. Uh, you know, it's gonna
happen from the the fact that you can no longer
tell if you're the only real human in the threat
or not for the conversation, right, and so you're gonna

(01:00:48):
stop interacting at all because you don't know if they're.

Speaker 1 (01:00:51):
Bots or not.

Speaker 2 (01:00:52):
You think that this is gonna happen.

Speaker 1 (01:00:53):
I think it's happening.

Speaker 3 (01:00:55):
What I don't know is are the companies that can
control the large platforms incentivized to stop it.

Speaker 2 (01:01:04):
I wonder because there is a level of, like any engagement,
it's good engagement, but if people don't engage what they're
going to.

Speaker 3 (01:01:09):
Do, there might be a tipping point. Maybe right now
it's something they don't really touch. Like, you know, it
took YouTube and awfully long time to start effectively addressing
the bot problem, right and they still haven't really done it,
and so.

Speaker 2 (01:01:23):
And that was just bot commenters or so views.

Speaker 1 (01:01:25):
Yeah, well there's that too.

Speaker 3 (01:01:27):
Yeah, there's I don't know too much about the view side,
but the bot commentser's side. They it's hard to know.
Are they dragging their feet on this as a counterintelligence operation?
If the bots are intentionally really stupid, then they could
be useful for trying to determine what are YouTube's countermeasures

(01:01:47):
to get rid of those.

Speaker 2 (01:01:48):
Bots so that they can still convent them.

Speaker 3 (01:01:50):
Right, So you maybe if you're playing YouTube side of it,
you know, the thought is we don't want to just
ban these because they're going to figure out our mechanisms
we use for more important one.

Speaker 2 (01:02:00):
It's exactly you know what, I actually buy that third
because that's exactly what they did with s E. With
sh They're just like, well, we couldn't possibly tell you
how this works because someone would just do it.

Speaker 1 (01:02:08):
Yeah yeah with Google or Google.

Speaker 2 (01:02:10):
Yeah, they own YouTube, so it's like Jesus fucking cut.

Speaker 1 (01:02:14):
And the other thing too. You know.

Speaker 3 (01:02:15):
There's that, But there's also if you want to take
the more cynical side, the less YouTube side, it could
be that they want to report high engagement numbers to shareholders.

Speaker 1 (01:02:22):
Yes, and bots do that.

Speaker 4 (01:02:23):
Well.

Speaker 2 (01:02:24):
I mean they did that with by combining Gemini with
Google Assistance and they had three hundred million I think
weekly active views or something, right, Yeah.

Speaker 3 (01:02:31):
So but then the result is there may be a
tipping point maybe at some point human engagement that's like
somehow verified, which is scary for a different reason. But
human engagement becomes almost artisanal. Like the reason you buy
something from Etsy instead of Amazon. You know, some guy
made it in his garage. Yeah, and so.

Speaker 2 (01:02:55):
Etsy full of bots. Now that's ultimate. Do you subscribe
to the whole dead Internet theory? Do you think that?

Speaker 3 (01:03:03):
I think parts of it are starting to look.

Speaker 1 (01:03:09):
Practical, Like Facebook, I feel.

Speaker 2 (01:03:11):
That's my one where it's just like this is like
old people screaming at their TV at this point.

Speaker 3 (01:03:16):
Yeah, and it's filled with bots that either trick them
or that and raise them into a response.

Speaker 2 (01:03:38):
One of my favorite things to do is if you
go on Facebook and you type in Facebook support into
the chat bot ah, and you find the groups that
are people thinking they're posting on Facebook grip and I
know there's one of them that has like ten thousand
people on it, and it's people like boomeras. I hate
to sound out, but it's like people in the seventies
being like, I don't know how to compute work, and

(01:04:00):
three guys from Indonesia responded like, I would love to
help you. His number from Indonesia, which is Facebook support is.

Speaker 3 (01:04:06):
I mean, it's kind of it is the the basically
there's investment comments like on CNBC I was talking about
where it looks like a real thing, you know, in
this case, it maybe looks like actual Facebook support, right,
And the end result is maybe maybe it's someone trying
to help, Maybe they're enthusiasts, you really know face.

Speaker 2 (01:04:27):
Oh it's all Indonesian scammers. Okay, I've looked through. I
may have spent a few hours of what like just
I could be doing literally anything else just looking at them,
and it's entirely guys in the Global South just defrauding people.

Speaker 3 (01:04:41):
Yes, and there's whole channels now that are built around
scam busting, right, Like I forget the names of some
of them, but there's like YouTube channels where they'll they'll
walk through kind of the bot scam link. I think
the AI stuff, the lms, you know, the core question
of like what do I think of AI? All of

(01:05:03):
those can be tools for good things. I think the
bad things right now are very profitable. It's the dumbest
bad thing that's profitable as a scam.

Speaker 2 (01:05:13):
But even then it's not profitable for the AI companies.
It's profitable for the scammers. It's like the only people
making money all Jensen Hwang and scammers.

Speaker 1 (01:05:21):
Right.

Speaker 3 (01:05:22):
Yeah, And maybe there's I don't know, there's probably some
like Fortune five hundred company that thinks they're making money
on it.

Speaker 1 (01:05:31):
I don't know that I've found.

Speaker 2 (01:05:32):
But no, OK, maybe there's one. No, this is like
my one hyper focus of like anyone who mentions the
revenue with AI, I know.

Speaker 3 (01:05:40):
Well there was one, wasn't there some report that said,
what was it like over ninety.

Speaker 2 (01:05:45):
No ROI amazing report. Amazing report because you know it
was good because immediately people said hit piece. The moment
it says hit piece, you know that it's the true
just people immediately think it's a hit piece.

Speaker 1 (01:05:57):
Do you remember who the report was by.

Speaker 2 (01:05:59):
It was by the Nanda lab mit.

Speaker 1 (01:06:02):
Oh wow.

Speaker 2 (01:06:03):
And people got really assy about it because they said, oh,
it's just a bunch of interviews, that's how they did it.
How the fuck do you think surveys work?

Speaker 1 (01:06:09):
Right?

Speaker 2 (01:06:10):
And so they spoke to no, they spoke to a
bunch of Fortune one hundred I think CEOs and people
really miss reddics. They said, oh, it's a learning gap
between people using this and not understanding AI. No, the
paper says it's a learning gap because the ais don't learn,
but no one reads. Okay, but it's it's so strange
as well, because it's everywhere. But it's also nothing. And

(01:06:31):
I mean it feels like there is AI within the hardware.
Well with like I forget, forget the term the upscaling ones, Yes,
do those generally work?

Speaker 1 (01:06:39):
That's actually a really good point. Yeah, so, and.

Speaker 2 (01:06:42):
That's different to large language books different transform based that's
that's really good DLSS and such.

Speaker 3 (01:06:47):
DLSS is like a real thing that works and does
it's good. Broadly it's effective, I mean.

Speaker 2 (01:06:54):
And that's when it fills in the frames.

Speaker 1 (01:06:56):
Right DLS.

Speaker 3 (01:06:57):
So there's kind of like sub technologies they have, but
broadly speaking, DLSS originally started as just it stands for
deep learned super sampling, and it started where they would
take ground truth images, meaning like basically we're saying this
is reality, and I think there were sixteen K resolution.
They're ridiculously high resolution. They would train on all these

(01:07:19):
images and then use that data game by game to
be able to upscale from a lower native render resolution
to a higher i'll call it projected resolution to the viewer,
to the user.

Speaker 1 (01:07:36):
And that was the original implementation and it is pretty
good now.

Speaker 3 (01:07:41):
Like it's actually if you need to run at ten
adp native so that your video card can handle the
game at a good frame rate but make it look
higher resolution, a lot of times it works and actually
two N video's credit, DLSS in some situations can be
better than native, which shouldn't be possible in theory, but
because of the way they've actually integrated the deep learning

(01:08:04):
on it now, which has been rebranded.

Speaker 1 (01:08:05):
These days to AI.

Speaker 2 (01:08:06):
Of course what it was called.

Speaker 3 (01:08:07):
Deep learning because the way they've integrated it, it can
sometimes reconstruct details that should be there but are not,
and we've done some videos on that.

Speaker 1 (01:08:17):
But yeah, that's a use case of deep learning.

Speaker 2 (01:08:19):
Then that's completely different to the world of laws of language.

Speaker 1 (01:08:22):
Yeah, it's not an LM.

Speaker 3 (01:08:23):
Yeah, it's like it has a singular purpose, right, make
the thing look better, or generate frames to insert to
smooth it over, and that's overall not bad.

Speaker 1 (01:08:31):
Also, there's places it's really useless.

Speaker 3 (01:08:33):
Yeah, But I think also though, the difference is when
that technology is useless, it's not really harmful, whereas when
an LM is useless, it is harmful because it's it's
putting bad information.

Speaker 2 (01:08:45):
Out or just even if it's not being used particularly
well and someone's just fucking around with it, it's incredibly environmentally damaging.

Speaker 1 (01:08:52):
That is a great point as well.

Speaker 2 (01:08:53):
Yes, they a fe your stuff being plagiarized for the models,
you know.

Speaker 3 (01:08:56):
For models, Well, I know I did a there was
a great one where I did a Google search and
I searched for the release date of a product, and
you know, Gemini spat out in a.

Speaker 1 (01:09:09):
Year and it was wrong by two years, and I
was like, what the fuck? And I clicked to see
what its sources were, and one of its top source
was us, and I was like, what the fuck? Did
we screw that up?

Speaker 2 (01:09:20):
Yeah? Like what ended up happening?

Speaker 3 (01:09:21):
So I went to our own article and it was
a revisit we had published two years after the original
product came out, and we made that clear in the article.
But it's just looking at the published date and mapping
it to the product name, and it decided this is
the release date two years later, you know.

Speaker 1 (01:09:40):
And and so like.

Speaker 3 (01:09:42):
It learned or took from our content, misrepresented it, misrepresented it,
and then credited the incorrect information to us.

Speaker 1 (01:09:53):
So now it also makes it look like I got it.

Speaker 2 (01:09:55):
Wrong, right, So everyone loses, including the customer. Yeah, it's
a shame though, but I mean I feel like in
the whole AI generitive world, your stuff is more valuable,
the very hands on, very specific work and the very
hardcore testing you though.

Speaker 3 (01:10:11):
It's also we're fortunate that it's kind of at the
front end of like, if we're reviewing products, then.

Speaker 1 (01:10:19):
Someone who's an enthusiast trying.

Speaker 3 (01:10:20):
To decide on a purchase is still coming to us
before it's useful in training.

Speaker 2 (01:10:26):
So I imagine the generative content is kind of antithetical
to the kind of testing you do as well, because
you can't really fake because.

Speaker 3 (01:10:34):
Not really, you can't really like generate a review, you know,
it would be obvious.

Speaker 2 (01:10:37):
And also you have these massive machines for testing stuff here, yeah,
which are insanely cool. And actually that's a good question.
How do you source this kind of stuff? Do you
buy from industrial because you have like pressure tested, Well, actually,
maybe you could speak to some of the machines.

Speaker 1 (01:10:51):
Run through it.

Speaker 3 (01:10:51):
Yeah, So yeah, we have a Hemiana Quick chamber, which
is a sound chamber and acoustic test chamber. We have
a laser scanner that does three D scans of like
of products.

Speaker 1 (01:11:06):
And converted some into three D models.

Speaker 3 (01:11:07):
We have pressure testers, we have fan testers, cooler testing,
and then more traditional power supply tester stuff like that,
and so yeah, generally, the way we go about it
is we identify a problem. Typically, this is kind of
the part of the business I work on the most personally,

(01:11:28):
is like basically test engineering or design. So we identify
a shortcoming, which is like this product is claiming this thing.
We are not able to validate or invalidate their claims
because we don't have the tool to do it. And
then either I just do some research online and find
the tool, or you know, sometimes the real challenge is
knowing the name of a thing and that it exists.

(01:11:49):
And so I went through this with current clamps, which
is just a clamp you put on a wire to
read the current going through it. A long time ago,
when I started this, I didn't know that was a
tool that existed. And it's like magic because you put
a wire through a client a plastic looking clamp, and
then through the magic of electromagnetics, it tells you what
the amperage is. And when I found that, you know

(01:12:13):
early and it's not a new it's been around forever.
But when it was new to me, I was like,
holy shit, this is If I only knew the name
of this thing sooner, I could have bought it.

Speaker 2 (01:12:24):
And all of your testing is effectively new to you.
You've had to kind of build it and learn yourself
with help from right, with.

Speaker 1 (01:12:29):
A lot of expert help outside.

Speaker 3 (01:12:31):
And yeah, I would say sourcing it mostly comes from
either just doing research or we tore a lot of
factories and engineering facilities and make videos on it, and
normally when I'm there, I do my best to talk
to the actual people who use that equipment. And that's
that's kind of how we learned that it exists, you know,

(01:12:52):
Like I did a tour recently of a facility where
they showed us something that they had built for a
client for testing cooling products.

Speaker 1 (01:13:00):
And when they.

Speaker 3 (01:13:01):
Were done explaining it so I could talk about it
in a video, I said to them, can I buy.

Speaker 1 (01:13:06):
One of those?

Speaker 2 (01:13:07):
You know?

Speaker 1 (01:13:08):
But the yeah, yeah, yeah, they sell it, so that's
so cool. Yeah.

Speaker 2 (01:13:13):
And do they come and train you to use it?

Speaker 1 (01:13:15):
Sometimes you can.

Speaker 3 (01:13:16):
You can either pay for training or they'll do it
for free for up to whatever four.

Speaker 1 (01:13:20):
Hours or something.

Speaker 2 (01:13:21):
You know. And do you generally operate these machines or
do they train the.

Speaker 3 (01:13:24):
Crew generally, either I'm the first one to learn how
to use it, or someone specific on the team might
be one of the first ones to learn how to
use it.

Speaker 1 (01:13:34):
So like we had.

Speaker 3 (01:13:37):
For the laser scanner, it was Patrick on the team
who ultimately in that case I had tasked him with
learning how to use it, and he went through he
documented it you know, and now anyone can use the sop.

Speaker 2 (01:13:50):
And the laser's kind of you describing to me earlier.
It's used that to see if there's devotes ye, like
in the in the cooling you explain.

Speaker 1 (01:13:59):
I'm surface, I'm doing it terrible job. Now you're pretty
much right on it.

Speaker 3 (01:14:01):
I mean it's yeah, cooling product has a flat surface
that needs to contact a piece of silicon. And what
we're looking for is, Okay, the performance is really good
or really bad thermally, and we can't look at it
and figure out why. Maybe if we scan this at
a microscopic level, it'll reveal something.

Speaker 1 (01:14:19):
And so sometimes.

Speaker 3 (01:14:20):
It'll reveal that there's these like deep pits in the
surface of the metal that you can't necessarily see by
eye and can really affect the cooling performance. Other times
you might see a curvature to it where maybe externally
it's not obvious, but actually when you install it on
the silicon product, say only two thirds of it or
contact in the metal. And so this laser scanner scans

(01:14:43):
the surface with the laser and then makes a three
D model, and we can use that to inspect the problems.

Speaker 2 (01:14:50):
Do you ever get do companies ever reach out for
any consultancy things.

Speaker 1 (01:14:54):
They do that all the time. We reject all of it.
So because it's too much. There's a lot of reasons.

Speaker 3 (01:15:02):
The core of all of it is it's too much
conflict of interest, right where let's just like, let's just
say it's possible to take a testing job, a private
testing job, and do it with absolutely zero bias towards
future reviews.

Speaker 1 (01:15:19):
I just we'll just accept that premise for a second.

Speaker 3 (01:15:22):
Even under those conditions, I still don't like it because
now I'm putting a weird spot where when that product launches,
it's gonna be hard for me to review.

Speaker 2 (01:15:31):
It because you, oh, because you've already seen it.

Speaker 3 (01:15:35):
I've seen it, and I might have provided input on
it that they paid for.

Speaker 1 (01:15:39):
If I'm involved.

Speaker 3 (01:15:40):
Right, So, now, like let's say I'm not happy with
their execution, What am I gonna go say?

Speaker 1 (01:15:45):
I told them so, right, Like.

Speaker 2 (01:15:47):
That's not how much can you reveal? And even then
I mentioned the trade secret so brilliant.

Speaker 1 (01:15:51):
You probably end up under an NDA.

Speaker 3 (01:15:53):
The best approach to that would be to recuse yourself
from reviewing it at all. But then the problem is,
like now I can't do my job for consumers.

Speaker 2 (01:16:01):
Yeah, how good is the money on?

Speaker 5 (01:16:02):
I like?

Speaker 1 (01:16:03):
Yeah?

Speaker 3 (01:16:03):
And I also think it just cannibalizes the process where
you're taking things that would make excellent public videos and
public data, and you're presenting it privately to never be
seen by anyone.

Speaker 2 (01:16:15):
Right And by the nature of these agreements, I imagine they
could obfiscate your coverage on some level.

Speaker 1 (01:16:20):
I would think so.

Speaker 3 (01:16:21):
And also it's just, yeah, you're taking the thing that
made you desirable to do the testing to begin with.
You're doing it privately, so it's not public, So the
desire to have you do the private testing reduces, right, right,
Like it seems like it's just a sort of self sabotage.

Speaker 2 (01:16:36):
But how do the hardware companies feel about you having
such high level testing?

Speaker 3 (01:16:40):
Most of the people think it's cool, right, you know,
I think the I'm sure a lot of them don't care.
I know that some of the company representatives have told
me how they'll be more careful with how they market
certain things if they know we're gonna yeah rocks.

Speaker 2 (01:17:01):
I actually genuinely love that.

Speaker 1 (01:17:03):
Yeah.

Speaker 3 (01:17:04):
So, and that's not just us, you know, there's other
reviewers who also do an excellent job. And because the
review community. People kind of specialize in different areas. I
think it just sort of collectively keeps companies somewhat on
their toes. But obviously you know the company is going
to company.

Speaker 2 (01:17:22):
Yeah, they will shit about regardless. What is a kind
of testing you can't do right now that you want
to in the future.

Speaker 3 (01:17:39):
I would say the probably the the Well, so there's
one we can do, but it's not really practical, and
that would be basically more frequent transience testing. So there's
something with power delivery. I know you've been studying lately

(01:17:59):
where where there's transient spikes and so on a GPU
or a CPU.

Speaker 1 (01:18:04):
This would be.

Speaker 3 (01:18:04):
A basically microscopic spike in the current or the power
consumption right for really like normally like one hundred microseconds
or something.

Speaker 1 (01:18:15):
We can test it. We've done in the past.

Speaker 3 (01:18:18):
But because if you're capturing data for seconds out of time,
down to scales of thirty microsecond gaps or whatever, the
amount of data is enormous, it's really hard to process.

Speaker 2 (01:18:29):
It's like a storage and processing problem.

Speaker 1 (01:18:31):
Yeah, it's a major processing problem.

Speaker 2 (01:18:33):
And what would you want to be so is that
if there is, what would those spikes mean?

Speaker 3 (01:18:38):
It would help us to stay on top of especially
GPU manufacturers, but also CPU if if there's sudden spikes
that might take a system offline. So as an example,
this happened with a previous generation of both Nvidia and mdgpus,
but where they would have large spikes and power draw
that deviated from the non old draw. So if you're

(01:19:00):
at four hundred and fifty watts and there's a spike
to eleven hundred watts for one hundred micro seconds or something,
this might be enough to trip over current protection on
the power spot and just shut the computer off. And
then the end user is like, what the fuck? My
power supply is enough to handle four hundred and fifty.

Speaker 1 (01:19:18):
Watts right off?

Speaker 2 (01:19:20):
And this happened that Yeah.

Speaker 3 (01:19:22):
We saw this with I want to say it was
the thirty series, the RTX thirty series, and like it
was a problem that doesn't show up in normal power testing,
but within a silloscope you can see it.

Speaker 1 (01:19:34):
And uh again credit to end video for the problems
we point out.

Speaker 3 (01:19:40):
On the positive side, they did actually fix this problem
in the forty series.

Speaker 2 (01:19:43):
Yeah, so here's here's a weird one, but I know
it's near detO heart Linux Gaming. Yeah, are you actually,
do you think this is a viable alternative to Windows
with Windows ten kind of dying will being killed in
the bag?

Speaker 1 (01:19:57):
Yeah, I think. I think Microsoft is the best marketing
that Linux has ever had.

Speaker 2 (01:20:02):
I go on, Yeah, My.

Speaker 3 (01:20:05):
Biggest concern with Windows is not like the death of
ten ongoing security support. It is the slow intrusion of spyware.

Speaker 1 (01:20:18):
Well like recall.

Speaker 3 (01:20:20):
Yes, you know where recall is marketed as this secure,
locally saved, encrypted reel of your information. But to me,
like capturing screenshots of your desktop while you use it
at all is problematic.

Speaker 1 (01:20:37):
Yes, and we've done some testing.

Speaker 3 (01:20:39):
It's not published yet, but it'll do some dumb filtering
so if it sees the word password on the screen,
it won't take a screen shot, right, But like, you're
not always going to have one of those words on
the screen when it shouldn't take a screen shot.

Speaker 2 (01:20:52):
Right.

Speaker 3 (01:20:53):
Maybe you have some accounting documents open, you know, and
it's information that's sensitive, but recall doesn't know it's sensitive.
And uh, it's stuff like that that concerns me, where
it's breachable and exploitable because it exists not necessarily because
there is an exploit, right and uh, and then Windows

(01:21:13):
eleven is continuing to add telemetry and user data and
engagement monitoring tools.

Speaker 2 (01:21:20):
And that's kind of a big feature of home s,
like all the horrible s operating systems, the cheap ones.

Speaker 3 (01:21:26):
Yeah, it's yeah, everything that. It's just there's so much
data harvesting, you know, Microsoft's This is why I think
Microsoft hasn't really cared too much in recent years.

Speaker 1 (01:21:37):
If you if you don't license Windows. They used to
really care.

Speaker 3 (01:21:40):
About that, yes, because now it's like, yeah, go ahead
and steal it from us, like we're going to just
take all your data and that's how we make the
money anyway.

Speaker 2 (01:21:47):
Horrifying. But do you think Linux is viable as a
gaming plot?

Speaker 3 (01:21:51):
I think it as a gaming platform specifically, I think
it's becoming a lot more viable. So thanks to Valve
and the Steam deck ands and it's pushed for Proton
as a translation layer between the application and.

Speaker 1 (01:22:05):
The operating system.

Speaker 3 (01:22:06):
And what is that just for the Yeah, so Proton
is it's a translation layer, which means it's working to
effectively interpret the code that is that the game is
built with to run with better performance or at all
on a different operating system in this case Linux or
Stemos specifically, and so because of Valve's work on that

(01:22:31):
to make games more compatible, run smoother, you know, deliver
consistent frame rate and pacing of the frames.

Speaker 1 (01:22:39):
Because their work there.

Speaker 3 (01:22:41):
In some instances it actually has better performance than Windows.

Speaker 2 (01:22:45):
Right.

Speaker 3 (01:22:46):
The limiting factor is still sometimes it just simply doesn't work,
but that has become a lot less the case than
what it was, say, ten years ago.

Speaker 1 (01:22:56):
It's not a fix for everybody.

Speaker 3 (01:22:57):
There's still times you're gonna try and run a game
and it's not gonna work, and that's gonna suck.

Speaker 2 (01:23:01):
And how's the driver support?

Speaker 1 (01:23:02):
For example, hit and miss? So we're really early in
this test in but.

Speaker 3 (01:23:09):
As an example, I know, Intel recently laid off a
bunch of their teams that maintains various Linux driversware. Yeah,
and so it's kind of you know, we don't really
know what happens there. Maybe the community picks it up,
I guess, I hope so, but.

Speaker 1 (01:23:25):
Yeah, he will, Yeah, yeah, it depends.

Speaker 2 (01:23:28):
Also, it really feels bad to just be like, who's
gonna take responsibility for this thing we need?

Speaker 3 (01:23:33):
Yeah, and if it's even an open source driver to
begin with. So yeah, it's hit and miss, but I
would say that the biggest limitation of my experience with Linux,
and I'm not an expert in Linux. There's no new
people in your audience who.

Speaker 2 (01:23:45):
Know they email me whenever I say alternative y wes.

Speaker 3 (01:23:48):
Well, that's why we we better not even get into
distributions of Linux.

Speaker 2 (01:23:51):
Or oh no, they will. They will be up my ass.

Speaker 3 (01:23:56):
Arch by the way, But yeah, I think the.

Speaker 1 (01:24:02):
I think.

Speaker 3 (01:24:03):
I think the biggest you know, limiting factors is compatibility
with things like daily applications. So the tools we use
for video editing generally just don't work on Linux. We'd
have to use something else. Yeah, and that's a problem.

Speaker 2 (01:24:17):
But well, adjacent to this Linux conversation, how are you
feel about handheld gaming?

Speaker 1 (01:24:22):
Handhelds are really cool.

Speaker 3 (01:24:23):
I think there's faster innovation happening and handhelds than most
other places.

Speaker 2 (01:24:29):
Are there any really like? Because I know a zeus
because being on the Naughty List a bit. But I
like the steam Deck personally, But like, do you have
at a favorite brand?

Speaker 3 (01:24:38):
The steam Deck and the Ally are both pretty cool devices,
despite the Ally's warranty issues, like it actually is a
cool piece of hardware. I thought the Lenovo Legion Go
the original was a really unique and innovative gimmick.

Speaker 1 (01:24:54):
And I don't use that word.

Speaker 2 (01:24:56):
To like degrade the gimmick.

Speaker 1 (01:24:57):
The gimmick was the controllers detached. I like the switch. Yeah,
and then the steam deck is uh.

Speaker 3 (01:25:07):
It was just kind of the first in the new
round of handhelds. You know, GPD existed before them and
Ioneo and those guys.

Speaker 2 (01:25:14):
I respect GPD. They're weird, but they're trying something.

Speaker 1 (01:25:18):
Yeah. GPD and Ioneo both are weird but trying something.

Speaker 2 (01:25:22):
Yeah. I appreciate you don't have American manufacturers do it.
Actually it's good. Are the American manufacturers trying weird shit
like that or is it predominantly.

Speaker 3 (01:25:32):
It depends like how you how you define American here
because you know design, yes, like who depends on how
you count Lenovo At this point, I don't know if
you count them as American or not.

Speaker 2 (01:25:48):
Right, but they're trying weird shit.

Speaker 1 (01:25:50):
They are doing.

Speaker 3 (01:25:50):
Weird stuff, and and they you know, they certainly have
large US headquarter offices. I don't know where their design happens.
It's probably they have a base I think office. It's
probably there. But Valve is definitely the coolest in terms
of they set the tempo for this to reignite, Like

(01:26:12):
they're the ones who made handheld interesting again, and everyone
else jumped in and started doing stuff and that's pretty cool.
And Valve also said, you know, we're gonna make Stemos
available to other handheld manufacturers if they want to use it.

Speaker 1 (01:26:25):
I don't know if there's a license or not, but
like it's available.

Speaker 2 (01:26:28):
And that's not being picked up. I know you can
kind of sideload it.

Speaker 1 (01:26:32):
Yeah, you can definitely sideload it.

Speaker 3 (01:26:33):
I think officially, I want to say Lenovo might have
been the first one to actually offer it, right zeither Lenovo.
I don't think it was asues, but but yeah, someone's
picked it up.

Speaker 2 (01:26:45):
How do you feel about the Xbook's Rogue Ally.

Speaker 3 (01:26:49):
I think there's a lot of confusion around it in
the more mainstream audience. I've noticed talment threads where people
think it's going to be required to play what you
would consider an Xbox game, But it's like basically a PC.

Speaker 2 (01:27:03):
It's just a rogue x ally with Xbox branding.

Speaker 1 (01:27:05):
Right basically yeh. And I'm sure there's some software gifts,
you know.

Speaker 2 (01:27:10):
But I wonder if it works because that's not being
I love my rogue ally, but the fucking software that
pops up really gets in the way of gaming.

Speaker 1 (01:27:17):
Yeah.

Speaker 3 (01:27:18):
I don't really know what Microsoft is doing with Xbox.
It just seems like they're they're not sure what to
do with that brand right now.

Speaker 2 (01:27:24):
I have to I was gonna ask, like, what do
you think Microsoft's deal with gaming is right now? They
seem almost like they don't want to do it.

Speaker 3 (01:27:31):
It's weird because they they were so committed to locking
people into the Xbox for like a generation or two,
and then at some point my memory of it is
they kind of realized, like, wait a minute, you know,
as we moved to X eighty six architectures for consoles PC,
gaming kind of works on both these devices, so we

(01:27:53):
don't really care if they buy it on PC or
on Xbox, as long as they use our device.

Speaker 2 (01:27:58):
And they raise the price on game Pass as well.

Speaker 1 (01:28:00):
Yeah, and and games oh.

Speaker 2 (01:28:02):
Of course at the same time, and also the Xbox.

Speaker 3 (01:28:04):
Yeah, and then you know, surprised the Pikachu face when
they raised the price on games and then they raised
the press on games Pass and then everybody cancels.

Speaker 2 (01:28:14):
Well, they lost so much. I read something they lost
like one hundred million dollars or something even more because
they put cool of duty on game Pass.

Speaker 1 (01:28:21):
Yeah, I'm sure it was genies.

Speaker 2 (01:28:23):
Fucking move lets it really. It's a shame as well,
because when you get past all the horrible menus, it
is kind of cool that you can just plug a game.
I don't know, I've been playing PC games long enough
that I was excited when you could just plug controller
and it wasn't some sort of nightmare.

Speaker 1 (01:28:38):
Yeah.

Speaker 2 (01:28:38):
Yeah, though PS five controllers you still have to do
the weird duel shock emulate a thing.

Speaker 3 (01:28:43):
Yeah, it's gotten a lot more accessible, but I think
the handholds are are pretty cool.

Speaker 1 (01:28:52):
There's a lot of focus there. I'm not sure when
Valve is going to do it.

Speaker 3 (01:28:57):
I really want them to bring Stemos to stop properly
because it's not currently officially released for desktops, but they've
done so much optimization work there where. Like the coolest
thing that gets probably the least coverage outside of our
space is they've really tuned the.

Speaker 1 (01:29:14):
Pacing of delivery of frames.

Speaker 3 (01:29:17):
So if you have sixty fps, sixty frames per second
and it's delivered an inconsistent interval, so you have a
frame delivered in sixteen milliseconds and then the next frame
is delivered in one hundred milliseconds and that repeats. That's
going to feel really bad. Despite being sixty fps averaged
over the period, you might have some that are four milliseconds,

(01:29:39):
some that are sixteen, some that are one hundred.

Speaker 2 (01:29:40):
We never thought that fps could be a marketing term,
and now I finally, Oh Jesus, yeah.

Speaker 3 (01:29:45):
And so like the work they've done to optimize for
the frame delivery is great where they sometimes have better
frame time pacing than Windows, and I would like to
see that come to desktop.

Speaker 2 (01:29:57):
So as we wrap up, what are you excited about?
What in teke in hardware and whatever is actually like
kind of bringing your chair.

Speaker 3 (01:30:05):
Yeah, I think on the DIY enthusiast side, there's still
a lot of really cool innovation happening, especially in cases
right now, believe it or not, like computer cases, you
would think that it's like a pretty answered I mean
it's it's all thermodynamics, which is well documented, right, It's

(01:30:26):
not like stilicon.

Speaker 1 (01:30:27):
Engineering, but they're still there. I don't know. There's just
a lot of improvement in.

Speaker 3 (01:30:34):
Getting an affordable, like good looking design that actually has
a lot of function to it. So over the last
six or seven years they've really switched to focus on
coin performance and then coin performance that still looks good
and so that's it's exciting to see that.

Speaker 2 (01:30:48):
And what does that edit? Does it the cases get smaller?

Speaker 1 (01:30:50):
Are they just they can?

Speaker 2 (01:30:52):
Yeah, there's better thermal management.

Speaker 3 (01:30:54):
Better thermals as a result of that, maybe lower noise.
The computer just looks cooler. The smaller boxes have gotten
a lot more viable. So if you want to build
something that you take from I don't know, like if
there's a kid who travels between a college dorm and
home in the summer, you know, having like a small

(01:31:15):
a small mini it x box is pretty appealing, and
those have gotten really good.

Speaker 2 (01:31:19):
Yeah, the mani atxs are always so interesting to me.
I love the idea of kind of like an Apple
TV sized one. That's a pipe dream, but still.

Speaker 1 (01:31:26):
It's not hard to do.

Speaker 3 (01:31:27):
It's yeah, I mean, they've gotten easier. But the other thing,
I think coolers are also pretty exciting right now. A
lot of really neat technological development on the cool end
products GPUs I think are probably the biggest hang ups
for people where the pricing has just gotten kind of

(01:31:48):
stupid for consumer market GPUs, and yeah, I think that
that's the one that's kind of like kind of a bummer.

Speaker 2 (01:31:55):
I guess, yeah. And I mean, do you see any
viable competitor to win video and am D at any point?
Do you think that there's any chance of any smaller competitor.

Speaker 1 (01:32:05):
The only one is Intel, which I never really answered
your question on ARC. I think right now ARC is
going to stick around.

Speaker 3 (01:32:10):
But yeah, right now it only goes up to like
two hundred and fifty dollars and after that it's all
AMD and Nvidia and AMD. You know, part of it's
their fault. They need to really want it and compete,
and they just aren't. Like the products are okay, but
they keep doing the stupidest things with their prices and
their marketing. And I think part of it is they

(01:32:33):
sell every epic CPU they make, so they allocate all
the way for supply to that, and GPUs are kind
of back burnered. But you know, the end result is
we just end up in this stalemate we've been in forever.

Speaker 1 (01:32:44):
Now, what about.

Speaker 2 (01:32:45):
Combined GPUs and CPUs, is that is that something you're
seeing much growth from. Do you think that that's going
to be a focus or is that kind of just the.

Speaker 3 (01:32:53):
CPUs basically, yeah, yeah, yeah, I mean the the entire handheld.

Speaker 1 (01:32:57):
Market is that.

Speaker 3 (01:32:58):
Yeah, So there's a lot of success there and they're
clearly good devices. Now for that, they're not good enough
for like high fidelity, high resolution, high graphics quality gaming
on a desktop, but yeah, mobile devices, they're great for
their faults Upscalers like DLSS and FSR do make it

(01:33:19):
more viable to run an integrated graphics solution and actually
have a decent experience. Can't get close to a dedicated GPU,
but they're good. They're just they're in the same spot
they've been in for fifteen years where it's you know,
it's it's not really a replacement for one hundred dollars.

Speaker 1 (01:33:38):
GPU is going to be better.

Speaker 2 (01:33:39):
Probably probably doesn't make enough money as well from the
sink a bunch of R and D into it either.

Speaker 3 (01:33:44):
Or it becomes a problem of ballooning the die area
because you have to allocate some amount of your silicon
to the GPU and the CPU, and so if you
give more to the GP to make more powerful, you're
making your CPU weaker. And the only really way to
solve that is to put more silicon on the products,
which is expensive.

Speaker 2 (01:34:05):
So yeah, So a couple questions before we wrap up,
who are your favorite creators at the moment, because you
there's a there seems like a good kind of solidarity
between them. But who do you watch?

Speaker 3 (01:34:16):
What do you I think historically, like going back, the
large influences were Tech Report, Scott Wasson specifically, if I
remember correctly, he was involved in the early days of
ours Technico also early early days.

Speaker 2 (01:34:32):
Oh oh sure, I'll have to look out.

Speaker 3 (01:34:34):
And so Scott Wasson did a lot of excellent like
test design work. Ryan Shrout was part of his kind
of era and the two of them with Tom Peterson
from Intel now worked.

Speaker 1 (01:34:47):
On frame time testing.

Speaker 3 (01:34:48):
So so their publications set the framework for modern benchmarking
with frame times and I that was you know, I
still reference those content pieces and in tech I think
as well respected but has been shut down by I
think it was Future bought them.

Speaker 2 (01:35:05):
Yeah yeah, I mean anyone who's worked the Future knows
what happened, which is the Future bought them and then
just didn't do shit with them.

Speaker 1 (01:35:12):
Yeah, But who do you watch right now now? For
I like the work that.

Speaker 3 (01:35:18):
Jared's Tech does on laptops, So he's an excellent laptop reviewer,
does really good work there.

Speaker 1 (01:35:25):
We've helped.

Speaker 3 (01:35:29):
Just Josh Tech, who's another laptop reviewer, with some of
his testing methodology, and I think they've been doing good
things to really try to advance on that. They're still like,
relatively new to testing, but they're really trying. I think
Jay's two cents with his recent changes to testing, He's
really kind of put in a lot of effort to
overhaul it. And then the hard run box guys are

(01:35:52):
also awesome. But yeah, it's there's there's I like, hesitate
to start naming channels because there's a lot of channels
you know that I like or we talk to at shows,
and I hate to leave anyone.

Speaker 2 (01:36:04):
Oh no, I know, I wasn't trying to single anyone else.
It's just I think that, especially in my own work,
I could be quite critical. So it's good to just
focus on the things you actually like. Yeah, so as
we wrap, what's next for game is nextus? What do
you want to do next beyond testing? Is there a
virtually want to move into? I know you've got the
other channel.

Speaker 1 (01:36:22):
Going, yeah, yeah, I have to go to the UK
for that.

Speaker 2 (01:36:25):
I'm sorry.

Speaker 3 (01:36:28):
Recommendations from here leave okay, don't go where the airport is.

Speaker 1 (01:36:35):
Yeah, yeah, I think that we.

Speaker 3 (01:36:38):
Have some relatively large focus on that channel, which is
like looking at some more consumer advocacy topics on the
test inside.

Speaker 1 (01:36:46):
Uh.

Speaker 3 (01:36:47):
The biggest thing we're doing right now is overhauling our
gaming benchmarks. So we're working to introduce some new types
of charts and metrics that don't exist right now in reviews,
cold animation, error, and just trying to basically, we're taking
a moment right now to try and find Yeah, Okay,

(01:37:08):
we've been representing things a certain way in charge forever
as a review community. Is there anything new here we
can do? And that's kind of what we're trying to research.
So that's the immediate focus.

Speaker 2 (01:37:20):
Well, Steve, it's been such a pleasure to come out here.
Thank you so much for your time.

Speaker 5 (01:37:24):
Thanks, thank you for listening to Better Offline.

Speaker 2 (01:37:34):
The editor and composer of the Better Offline theme song
is Matasowski. You can check out more of his music
and audio projects at Matasowski dot com, M A T
T O, s O W s ki dot com. You
can email me at easy at Better Offline dot com,
or visit Better offline dot com to find more podcast
links and of course my newsletter. I also really recommend

(01:37:56):
you go to chat dot Where's Youreed dot at to
visit the discord, and go to aslan.

Speaker 5 (01:38:00):
Bet off lines check out I'll Reddit. Thank you so
much for listening Better. Offline is a production of cool
Zone Media.

Speaker 2 (01:38:07):
For more from cool Zone Media, visit our website Coolzonemedia
dot com, or check us out on the iHeartRadio app,
Apple Podcasts, or wherever you get your podcasts.

All Episodes

Episode Transcript

Host

Ed Zitron

Popular Podcasts

My Favorite Murder with Karen Kilgariff and Georgia Hardstark

24/7 News: The Latest

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}The Interview: Steve Burke of GamersNexus