¶ Intro / Opening
Also media.
Hello and welcome to Better Offline. I'm your host ed Zeitron and this week we're joined again by the wonderful Steve Burke of Gamers. Next us. How are you doing, Steve?
Doing well?
How about you doing great? See you? We're just at computext right.
Yes, yeah, we're over there for a couple of weeks.
So for the listeners who don't know, because this is a fair amount, you don't know every conference. So what is computext and what makes it so important?
Copy text? Is the I would I think it's the largest computer hardware and computer technology trade show. So that's hosted in Taipei every year, and it's it's like a much better CES. People may know CS.
And is it more focused on enterprise stuff or consumer stuff?
This year, weirdly, they AI commandeered the entire main hall, so there were actually there are a lot of pissed off consumer hardware manufacturers who were sort of evicted and driven out of the main hall and put into the lesser one with consumer hardware.
What were they even showing with the AI stuff as well?
I mean one company had an AI computer case?
Oh how did that? What's the AI do?
Absolutely no idea sick.
It's so cool that this is absorbed, so it's also kind of brutal, that's literally absorbing space. Yes, did you see anything cool that though, anything you're excited about.
Yeah, there's a lot of cool stuff the I mean there was some immersion cooling, which is always nice to see, like where they actually literally immersed the system, you know, in fluid for cooling. There was a thermociphon from Nocto, which is just a really cool use of it's another cooling technology. Lots of computer case and cooling solution development.
And then on the GPU side, I mean N Video sort of buried its fifty sixty launch and then AMD more formally announced the information for its ninety sixty XT. And both of those are sort of the slightly more affordable consumer class GPUs.
Yet so in Video also was doing some funny business with you recently you did that wonderful video. So walk us through what happened that because they tried to big dog you so to speak.
Yeah, So, I mean the problem we had with them
¶ Computex: AI Takes Over
we talked about was for several months there had been this I felt sort of pressure from Nvideo with where so N Video for contact with people, and Video sends out GP review samples. We and other reviewers are not entitled to them. That's fine, we can buy them ourselves as well, but they send out samples for product reviews. And in the past, and Video has been in hot water for asking another outlet called the hard Run Box to I believe the quote was, or the paraphrase was,
change your editorial direction. Yeah, and so that was several years ago.
And what did they what did they ask them to do back then?
Back then it was seeking more coverage of real time rate tracing in video games, which at the time was more or less only found in any kind of meaningful performance on in video hardware. But also and the reason at least we weren't ready to fully do testing for real time rate tracing yet was because it just it wasn't ready yet. I mean, there weren't that many games with it.
You couldn't really test it because you couldn't use it.
Yeah, I mean they launched and if I remember correctly, there were actually zero games that supported this technology and the only thing available were sort of tech demos. And so anyway that that was what they wanted, then they kind of went away, didn't That's at least as far as we know, publicly didn't try to push reviewers in a particular direction with their narrative or their editorial style.
And then over the last several months we were getting this pressure of basically so we do a bunch of interviews with engineers and technical people at Nvidia, and I felt they were starting to use that as a lever to be like, oh, it sure seems like you like talking to the thermal engineers. It'd be a shame of something happened.
And you did that very long one over within video, right, you went and kind of looked at how some with
¶ NVIDIA's Pressure on Reviewers
stuff worked.
Yeah, yeah, we did. I mean they're great educational content. The engineers are awesome, and I've spent years working with mostly people who are in different departments or don't work there anymore to get in video on the same page where I'm like, look, we're not looking for a marketing puff piece. We're looking for actual engineers who just want to talk about the stuff they make and people can
learn about it. But it's not like, obviously, yes, it can support the problem act by nature of if it's good and the engineers explain it well, then people will like it. But the objective is not just like come on our show and use us as a platform to sell things. We want engineers, not marketers. Yeah, and so
that was great. And the like I said, lever that it became was they wanted more in video on more coverage of this thing called MFG, which is multi frame generation, which is something that Jensen Huan, the CEO, talks about on stage as being part of their technology suite that is a big surprise AI accelerated that effectively multiplies the frame rate to be used as a smoothing technology.
So it's different. Is this different to like DLSS and things like that, or is that.
It is part of the DLSS or the Deep Learned Super Sampling Technology suite. So it's sort of it's it's almost like a subset of DLSS.
Right, and so it basically just looks better at higher resolution. It fills in the frames. I'm guessing just trying to simplify.
Yeah, So yes, the simplest version of it is MFG sort of interpolates frames, so it's able to add in or insert additional effectively artificial or generated frames if you're trying to be charitable with it, that are not literally built by the game engine, and so the GPU is working with its own software to construct these frames and construct pixels based on what it thinks is going to happen next. And we've actually tested it and there's situations
where it's not bad. There's situations where it's awful, like with any technology. But what it does is on a chart, if you just look at the number that comes out the so called frame rate or the amount of frames in a given period of time, that looks higher with this technology, but it's not measuring the same thing as
something without this technology. So it's effectively, you know, you're comparing apples and oranges if you try to put real frames that are rendered in the traditional way up against these generated AI frames that are effectively guessing at what they should look.
Like, and do they look bad as well?
Like they can they can look really bad. I mean we've done some testing where in some of the games
¶ Analyzing NVIDIA's MFG Tech
it's just like text for example, will will get completely garbled text on screen interface interface issues will draw the interface in the wrong area. You could have really.
Sanely bad for like any RPG of any kind or any kind of like Twitch I own.
Yeah, it feels like.
Anything you'd be doing EVA would be ruined by that.
Yeah, so there's scenarios where it works, Okay, But the RPG example you brought up, so Final Fantasy was one of the games we test did where there's just areas. I mean, it just I personally would turn it off. It looks worse and I'd rather have a lower quote unquote frame rate than have this sort of potential blurry mess every now and then.
So strange low qualities. Well, it's just a very strange thing to push out the door.
Yeah, and the you know, I think sort of finished the editorial narrative. They know, I mean, it's fine, but I'll finish that first and I'll come back to your account you just make because that's also interesting.
But on the.
Editorial side, they wanted more coverage of MFG, and we disagreed from an objective standpoint. You know, I kind of made the case of them several times, over several calls, over several months of why we think it It just doesn't belong on the same charts as normal benchmarking, normal objective testing, and so there was, Yeah, there's just a disagreement,
which is fine. The problem came in when I felt repeatedly on various calls with the N Video people that they were almost sort of holding hostage the idea of working with engineers again, where it's well, the way we're able to make this happen is by you covering things like MFG. You know, sort of that type of conversation.
Yeah, that's so. And they mentioned something about like budgets in the video.
I remember, yeah, I mean there was like discussion of one of the times specifically, the statement was that this is how we can secure budget for these types of things, and like to be clear for listeners, we don't take money for interviews. Yeah yeah, And in fact, I've paid now over five figures to visit N Video on numerous occasions for said interviews. Personally, like I'm paid.
And I can say this from the PR side, and not working not work with A video or anyone associated five all that. It's like, no one talks about budgets. No one is like, oh I can't visit a journalist due to budget, and certainly not in Video, a company that had thirty nine billion dollars I believe in revenue just in the data cent a portion.
Of classical multi trillion dollar marketing cap. I think if I try to take Devil's advocate at maybe there's some internal budgeting whatever, like the marketing department needs to budget time from the engineering department or something. But we're talking like a couple hours of time max normally, you know, yeah, especially when I'm traveling to them. So it just seemed
like I don't know it. I didn't buy it. And also that was only one of the numerous UH calls, you know, like that the defensible reason for them changed. It felt like each time where it's like, oh, let's try this angle, let's try this angle.
So did they end up blocking you? They just how did that situation resolve itself?
I have not spoken with anyone there since we posted our video talking about it. And part of the challenge with posting that video is, you know, it's like, the first couple of times this came up, I assumed that it was just sort of okay, maybe the guy phrased it ron you know, I've worked with these people for a while and maybe just came out wrong. After a
¶ NVIDIA's Budgeting Excuses and Fallout
couple of times, I did call one of my reps after the meeting where they kind of pushed this, you know, you testing this technology that you think is not appropriate to share on charts with competitors. That's how we get you these interviews. I called a rep and said, you know, hey, this, if you guys say this to people, this is going to be taken the wrong way. And like, I'm still assuming it's not meant this way. But but if I hear this again, right, like it's I'm did I have
to start assuming it's it's the way it sounds. And it did come up again. And part of the challenge with making the video we made was, you know, we want to blow the whistle on it in case other people are facing that pressure, and also in case consumers are watching or reading content which may be influenced without their knowledge. The problem is, I think en video may now be in a situation where let's just assume everybody
makes up and it's fine. At some point, I don't know that they will ever be able to put an engineer on cam RA any time of the near future with anybody without the audience questioning it. And for good reason, because that was what we felt was being used as the lever against us. Yeah, and so it's just kind
of I don't know. We tried to, you know, I explained to them several times why this is a problem and why the phrasing is a problem, and the more you hear it, the more it's like, Okay, I mean at some point, like I think you're saying what it sounds like you're saying, I'm going to stop.
Yeah, it's good that you did it, though, because it's GM's got over a million subscribers, so who knows what they're doing too much smaller outlets as well.
Yeah, I mean, and that's what I'm really concerned about is I've had conversations with Nvidia in the past that will probably get into in a future video, but about their strategy for newer creators, and I don't like what I've heard. I mean, it sounds very much like we're a're gonna sort of try and shape these newer creators.
And you know, I think the problem is as especially old guard media, as in people like OG's before US retire or move to other industries, there's openings to change how media interfaces with these companies in general, because you've got newer people who don't have the historical background with these companies to know what to expect, and so it's easier to kind of slowly creep that goalpost where it's it's less independence and it's more marketing arm of the company.
And that's what I'm really worried about, is if they get in any of these companies with newer creators early and are able to change the expectancy of the relationship, so change it from independent reviewer analyzing large companies device into more of a hey, we're all friends, we're all
working together to try and benefit the consumer. When if you're talking about a multi trillion dollar company that their motives are far different from like the guy in his bedroom who's trying to do hardware reviews.
It's just so weird as well, because they're already so successful. Is this history? It sounds like historically in Vidia has been a bit draconian like this, though not necessarily pressuring in this way.
I mean, I certainly think and Vidia is one of the most vindictive companies in the space. They the stories we've heard and told in some instance of their partners, not reviewers, but actual business partners like board vendors, people who make the video cards that the Nvidio GPS go on. Those stories would seem to indicate a sort of vindictiveness
¶ Protecting Independent Tech Media
can give an example, Yeah, allocation. So GPU allocation is the concept of how many of a high demand GPU can you, as a partner get to then attached to a product to then sell to a consumer. And there's a limited amount of these because they're going to sell one hundred percent of them, And allocation is typically measured by percent, So company A might get twenty five percent, Company B might get five percent of the GPU supply
whatever that is. And that allocation, we've been told in the past, has been used also as sort of a lever. And so an example would be back in the EVGA days, which Evga sort of famously.
That was a graphics called company.
Correct. Yeah, they they were in video's number one partner, and they were so Nvidio favored that they only made Nvidia video cards. They did not work with any competitors, right, and they famously cut that relationship publicly and we ran a story about that. But one of the things that they were unhappy about was Nvidia slowly tightening the screws, which is a direct quote from an Nvidia pm I
spoke to one. Slowly tightening the tight end the screws is the quote, but on what the partners were allowed to do with their products, and slowly over time they're losing some of this creative freedom.
What did you say, they lose creative freedom? Like, what are they being constrained to do?
One that's public is we did an interview with a famous over clocker who used to work for or EVGA. Kingpin is his username and his name is on video cards that are world famous. We did a video with him and he talked about how he wanted to add two power connectors to a card to basically make it
more fun to overclock with for consumers. And so it'd be to be able to pull more power, and it would also be able to do so safer because the current power connector has this issue of burning and catching on fire sometimes, right, so, and so he wanted to do two of them to kind of resolve some of that. And in the public discussion we had with Kingpin and the video he said paraphrasing, but the conversation was he said they wouldn't allow it, or they restricted it or
something like that, And I said who is they? And he said they, Yeah, like we we there was only one day there could.
Be who could it be?
Yeah, But there's plenty of examples like that, where the another one would be a couple generations ago. The cheaper cards, like the two hundred three hundred dollars models they were those models were difficult for the partners to actually make
and still make any money. And so there were times where we were told about how this MSRP level video card, meaning the number that in video says is the MSRP, that would only exist for a fixed period of time because it was simply impossible to afford to make it. But and Video really wanted them to exist at this
certain MSRP. It allows them to advertise a lower price, and so they're pushed to make these cards that allows them to potentially get things like more access to allocation or you know whatever privilege they may get.
So on a much larger level, what do you think is going on within Video as a company right now? I'm not asking for stock analysis like that, totally not that, but just as a company that both makes can zumographics cards and undepinds whatever the AI thing is, how like, what do you think is going on there right now?
I think there's definitely a focus on AI to try and get money. I mean, there's a Jensen Juan was in an interview I don't know last year sometime, I think, or earlier this year, and made a coment about how they use AI to write their own driver software now or their own software or something like that. And you know this is this is the number one focus. It's what they're making the most I guess investor money on,
so they're focused on it. The concern I have is it's one thing to sort of let the consumer hardware side dwindle or get less competitive effort. It's another thing too, in my opinion, and sort of drag the entire industry down with it as they shift focus to the higher revenue AI side.
¶ NVIDIA's Treatment of Partners
Right when you say drag them down, how do you mean?
I mean things like trying to in my opinion, control the media narrative and shape it like that. That to me goes beyond Okay, we don't care much about consumer We're going to focus on making money with AI. That gets into we are going to actively damage the credibility of independent media and also potentially harm consumers who are influenced by that media, which has been shaped with a certain you know, corporate narrative to push.
And that's my concern.
So that's that's the That's why I draw the distinction between you. As we said in our video, it's if you want to just focus on AI and if they want to just sell all the GPUs to big enterprise companies, go for it, but you know, don't screw all the rest of us over at the same time.
And do you think that they're abandoning consumer or is it just they are they are focusing on the shiny object of the day.
I I don't know Jensen Juan personally my read on him, though, I feel like it's unlikely he would want to abandon something he already has, right, even if it's only out of maybe sort of pride, and so I don't think they'll abandon it. I do think they're focusing on AI. It's just that, uh I don't. I mean, they have like ninety two percent market share and consumer GPUs now according to John Petty Research. So it's just like they can.
It's a monopoly. I mean, that's like, it's just like, this is what monopolies look like.
Well, let's it feel I would say there was an opportunity for someone to come in. But based on everything you've shown with like AMD recently, even on the low end. It just it doesn't feel it almost feels like consumer graphics up being not treated with the respect they deserve, despite PC gaming being huge industry and so on and so forth.
Yeah, it's a massive industry and prosumer too. I mean even just non AI use cases include people who edit videos or make movies or do three D animations. Yeah, I think Intel's a good barometer here because Intel decided to get into consumer GPU, and they're new to it, and they've had a hell of a time. They're doing
much better now with this current generation. But one of the quotes from an Intel person I spoke with previously to us was for their last generation was we should be wrapping these with money when we saw them, because they were just they were basically subsidizing the buildout of this new division, this GPU division, by undercutting prices from competitors to just try and get their foot in the door with what at the time was a lesser product.
And so yeah, I think they're a good broader because if you want to be first of all, to have a fab a fabrication facility costs these days potentially tens of billions of dollars in years to build so a new vendor would have to be fabless like Nvidia and like AMD, unlike Intel, and to be fabulous and make a.
Sorry about this, this is just for the listeners. What is a fab in this case?
Yeah, so a fab a fabrication plant is basically a factory for silicon. It's a highly controlled, clean room environment, and that means that there's basically absolutely no dust ingress, the small pieces of skin or whatever, you know, just working in an environment that's all either captured or controlled. And that's because a single particle of dust on a wafer can destroy at least part of that silicon supply.
Cool sorry, Yeah, And so a new vent would have to be probably fabulous unless they're already huge like Intel, and that means they need to go get supply from most likely TSMC, possibly from Intel maybe. I mean the past Samsung has been used. There's really not a lot of options here, and everybody wants their silicon, so to
¶ NVIDIA's Monopoly and AI Focus
even get allocation of these silicon wafers would be difficult. It's just all of the chips are stacked against Someone wants to do this, and they would. I think the only way you get in is if you're an existing multi billion dollar company or have obscene amounts of investment.
Well that fucking sucks. Yeah, it's actually that's a good question though. So this is for the old time is listening. I remember having a three FX it's a Voodoo fifty five hundred picture of a halo on the on the box. Actually don't know what happened there. What happened to all the smaller manufacturers. Did they just get kind of absorbed or did they just full of punt?
Yeah, they got bought or yeah, I think three dfxx, if I remember correctly, was bought by Nvidia and checked that. But some of the others, I mean there's like Via Cyrix and some of these CPU companies. Yeah, three dfcs bought by Nvidia. SGI was another, so they they.
Yeah, I forgot about SGI.
Yeah, So SGI is interesting because they sort of pursued heavily. I guess the at the time what enterprise would have been, or at least workstation hardware, and they're really invested in workstation. The story I got from someone I wasn't covering things back then, but was that SGI was less interested in consumer and didn't really think consumers could afford video cards. Which at the time was sort of true, and that's when video came in and you know, started really kicking
ass in the early days. But a lot of them just they either got bought or they sort of exited that space.
So changing text slightly, how do you feel about AI in general? I know that I'm not expecting like an industry and that I'm just from your perspective, at least from what you've told me and watching your videos, it feels like AI is something now being stapled onto the side of literally everything. Yeah, within your world. Do you think there's like do you use it? Do you have any thoughts on it?
There's uses? I mean one thing I am I'm actually worried about long term with the AI use cases. And the biggest thing I'm worried about is on the media side. We obviously get a lot of comments. I probably read or at least skim, you know, tens of thousands of comments a year, and I'm seeing more and more AI comments and not just the really obvious YouTube spambots, but somewhere you kind of you and you're like that could be real, you know, and I don't like, Yeah, I
don't want to like remove it. If it's a valid, real comment, but then a few comments down the thread there will be this discussion amongst what are obvious sort of bots about say some new cryptobs or whatever. And so I'm actually worried about sort of like that Internet theory. I guess, the concept of you're the only human in the room and you might not even know it. That's concerning to me. There's a loneliness factor there I'm worried about.
¶ Challenges for GPU Competitors
But the big one is just how easily people are manipulated by comments that sort of appear to have the prevailing opinion just in general, even from real comments. You then take that and you apply it to potential AI that becomes very convincing in the future. And you then think one step further of who owns these lms and the companies that sell so called AI products, and all of them are these multi billion dollar companies that if they want to, could tune it to filter the responses
in ways that are beneficial to them. And that to me is sort of the dystopian old manuals at clouds, you know, like concern I have, but it.
Feels like a more practical one because everyone freaks out about AGI, which is fictional and all of these other things. But yeah, the idea that comments would be filled with people who were just subtly being nudged in different directions. And when you talk about corporate narratives as well, who knows that. I don't usually with lms in particular, try not to lean into anything too conspiratal. But like, this is a simple one that you're kind of seeing signs of.
Even on Blue Sky, I will occasionally get a comment from someone it will take me half a second second to go, like, what's wrong with you? This doesn't feel like a human being saying this. No one speaks like this, but you sound human adjacent, right, And I'm and at your scale, I imagine that becomes a lot harder to moderate slash process.
Yeah, and it really is just I mean, I'm highly my job is to be skeptical, and you try to do that in a healthy way. You know, it's easy to go way overboard. But this is something where I just I think there's so much opportunity to be abused. And I was actually looking at there's that meta story about meta and Facebook pirate in all of these you know, books or taking books from pirated materials and integrating them,
and there's a lawsuit over it. One of the things I was thinking of though, and actually speaking to a copyright attorney about, was it really feels like right now, the companies that are willing to break the most laws,
¶ What Happened to 3DFX?
especially intellectual property laws, and steal the most stuff will make so much money that it's like, you know what, let's just we'll just deal with the consequences in five years when it gets through the court, and we'll be so far out ahead of anyone who didn't lie, cheat and steel, you know, to get here. Yeah, that it won't matter, which is just how the big corporations broadly behave. But applying that to a new area like AI, I don't know. It's I'm really concerned about it. I think
there's absolutely good use cases. And I mean, as an example, internally, we've used it to concept things, so like I'll have an idea in my head for art for a product, and I can see it, but I'm a terrible artist, and so there's a use case there where I'll sit down with my artist and try to explain it as best I can, and he'll take a go at it, and if I'm like, no, it doesn't really like, look
how it looks in my head. That's when I might try to pull some mid journey thing like okay, here's like broadly like the style artistically I'm trying to get, you know, and then he can take that and actually make something.
But just to be clear, you would never use any of that stuff in the real It is literally just a concept tool, correctly, just making sure the listeners have
¶ Concerns About AI Generated Content
a good view on this.
Yeah, correct the artist so Andrew has been working with me forever. He makes it all from scratch, like can blender normally and actually flattens it into a two D image for the shirts and things cool. So yeah, it is purely used. If if in the first conversation he can't like quite figure out what I'm trying to explain as a non artist, you know, it's like, okay, let me go try and figure out, like I don't I don't know art words, you know, I don't.
Know no, no, I know. Yeah, so yeah, yeah, how do you deal with the ethical side of that, because it's just it is still working on copyright material. But I guess if you're not putting it out there, it's just like this this shit sucks even like looking at it feels bad.
Yeah, I mean it's hard because it's the first step for us. So it's always like a Google search to see what's out there already, and so we'll do a Google search if I'm trying to explain something I want, and normally that's pretty successful. But like now it's the same.
It's like you're getting all this aiart surface in and so at some point it's sort of like the downloading and uploading and downloading, uploading a PMNG multiple times, it feels like it's just getting like the quality is getting lost somewhere in that chain.
It is funny. We're also like three years into it, and the best we've got as far as like from you arguably one of the most. Like I don't mean this, I mean this actually is in a good way, Like you would be excited about something cool in tech coming out, like if something cool happened, And the coolest we've got is Yes, sometimes I can't work out what I'm thinking, so I just make one image and hopefully my artist
gets in. But yeah, yeah, it's just like that with three years in and that's the best we've got.
Yeah, And I think we early early on with GPT when we ran one of the first news stories that we ran about it, and I was trying to understand, like, what is this thing. I remember feeding it a spreadsheet I had made with our own data and asking it to analyze the results, and it was just like and I keep in mind this was early, but it had hallucinated so much that I was just I was disillusioned. It was like, Okay, this is literally just made everything up.
And that was kind of how I was at the beginning as well. Because I love my dad's magaizmos. I was kind of like, Okay, everyone's talking about this, and so I tried to do all of the things. They talked, Oh, it could automate work. I'm a spreadsheet heavy guy and this crap can't even do that. The funniest one for me was that you have a Manus Manus You heard of this one now? They claimed to be an AI
agent company. Okay, and I asked that can you go and find all all of the links that mentioned me for the last two Years's probably over one hundred of them, and it takes eleven minutes or so, and it's writing Python for every single step and it comes back with eleven links or like nine links. I go, hey, you miss some, and it comes back with nine more after like another ten minutes. And this is this was about a month ago. It's just what it feels like. They're
pushing them upper hill. It's but back to video for a second because this will be the last video one. I think you have done a really good job of kind of tracking in video as they've jumped towards these movements. But they've done this, but they did. How big was their move towards Crypto That's actually something I've had trouble encapsulating before. Did they really shoot the company that aggressively in that direction?
That's a good question. So from and videos relationship with crypto is weird. I remember at one point there was some kind of earnings call or something where it was during one of the GPU mining sort of apocalypses where consumers couldn't get cards because they were all being bought
by crypto mining. And I remember doing a report on in Hardware News on the earnings where it seemed like the amount of revenue for mining was sort of obfuscated and it was a challenge where the devices are shared with gaming and so okay, I could see how it would be difficult to decouple these in a revenue chart if you sell gaming devices and they're used for multiple things, right, so you know, it's like I kind of get it. But at the same time, they weren't like hiding from
making sales to crypto mining. And then later on and Video released these l R low hash rate cards. I think it was the thirty series, and that was done to try and get more card while they say to try and get more cards to gamers. Now there were bypasses for that, which is a separate story, but it
was done at a time when the thirty series. It was really unfortunate covering it because like the hardware was pretty good, the prices at the beginning were good, but there's no supply to meet the demand, which was partly because of COVID shutdowns where people were building new computers for home, and also partly because of mining, and so they tried a few things like these low hash rate
cards to reduce the viability for mining operations. But yeah, I mean it definitely they've kind of that pendulum has swung back and forth. I think at end Video where it's like, Okay, we sell a lot of these and make a lot of money to people are really pissed off, you know, and we can't get them to consumer users who might actually be repeat customers in the future, and
so let's try and restrict it. So it's definitely don't I mean, it's not unexpected behavior, I guess for a big corporation where it's just where's the money and if the pendulum swings back the other way, you know, we need to try and minimize how much both of these potential client groups hate us for catering to the other one.
So so final GPU related question, actually, and forgive me if you don't know the answer for this exactly, but
¶ Ethics, AI Tools, and Experiences
assuming the AI bubble burst and you don't have to commit to whether you think that will happen or not, what else can these big enterprise level GPUs actually be used for.
They could be used for a lot. I mean, like some of the we're talking to a professor at a university for an upcoming piece, and they have some very serious like machine learning research use cases, and I've looked into some of the other ones, and as you might expect, there's really promising use cases in medical and pattern recognition and like so called AI being able to do early identification of potentially cancerous masses and people that you know. It's one of those like if you treat it like
a tool rather than like the answer. Then I suppose in this example, as someone who knows nothing about, you know, the medical world, but I suppose a skilled doctor might be able to use it to help the process. I think the concern I would have as as a skeptic would be okay, but can we keep that doctor in the mindset of I'm using this to help catch my mistakes or help see things I can't, rather than you know, I'm overworked and I need to lean on this to do my job. Yeah, that would be the concern.
I'm just my whole thing is just if they and I believe they have. This is my personal opinion. I think they've massively overbuilt and oversolved these cards, and I it's sorry cards is inappropriate to refer to the enterprise ones racks, I suppose, But it's what do they do next is going to be the question eventually, And I mean there's just very clearly not going to be enough use cases for them. But I think that's going to be really interesting. And now I've got two fun questions
for you. First of all, did you hear about this thirty six hundred dollars keyboard? It's mad the Norbauer Seneca. Okay, so I'm going to send you this later. So Nathon Edward's over at the Verge reviewed this one, this thirty six hundred dollars keyboard, which I need to send you because I assume you would Seneca.
I see the review.
Yeah, keyboard, it's insane. Listeners, I will drop with everything we've been talking about. I'll have links in there. But this thing, it's like precision honed and if he gets it wrong, he has to start again with it takes him like a full day to do, like the things that absorb in impact. So cool.
This is this reminds me of the I don't know, probably almost twenty years ago. Now there is the super expensive I think it was like an optimistic keyboard or something.
Like the opt or the dust keyboard where it had the little screens on each key.
Yeah, it was screens for every key and back then, you know that was that was before we were putting screens on everything.
Uh huh, I won't. That's the thing. I love that this exists because if you look at the Verge story with Chell Lnk. It's just a guy who literally sits in. This character just builds these things like an artesian. Now, I wish we had more weird shit like this. This is what I miss about the tech industry.
Yeah, the the weird stuff and the cool stuff and things where that's what I like seeing that computext but it's okay, I could this isn't a product yet, but I could see it becoming a product, or there's pieces of this that could become a you know, an affordable product. That's the stuff I like.
Yeah, did you see anything really bizarre like that there?
Yeah, I mean the the well actually most recently this wasn't computext. But the most sort of bizarre cool thing I saw and worked with that I was really happy about was the pre built PC made by a small company called Cherry Tree. They make in computer cases and it's just a computer inside of the husk of a video card.
Okay, that's cool. Yeah, I love that. That's it. That reminds me of the corse Air thousand D case, which I actually have a home where they have you can fit micro micro at X and the regular sized motherboard in there. I love that stuff like. It's that's why I that's the thing I love about tech, the weirly, really dumb, kind of niche stuff. And actually, with that in mind, what are you actually excited for as we wrap up, Like what the thing's coming out soon that you're kind of pumped up to see?
I mean, the casesn't cool in are really interesting to me. I just think in the puter actually, CPUs have gotten interesting too, So on the silicon side, AMD has done really well to actually get into gear competitively over the last several years. And these they call them X three D CPUs, which have a bunch of extra cash on them. They've been really good for gaming and that's a I
mean seeing the extra Yeah. So, because you can fit more stuff local to the CPU, like like you can store more of what the CPU needs to process within the CPU silicon itself, it doesn't have to go out to system memory. So the way it normally works is like if you're playing a game and you need some kind of file for drawing something in the game or whatever, I mean, go into memory of some kind. So it
might be GPU memory, might be system memory. But if you take the system memory example, something goes into system memory you need that file, there'll be a transaction where you can imagine these bits going down a bus to the memory and coming all the way back to the CPU,
and that's a long highway to go down. Whereas with extra cash you can store more local to the CPU, it can transact via basically like a shortcut, and that allows in the real world significantly higher performance in gaming or in cash heavy applications.
And so the new AMD ones, yeah, it's.
¶ NVIDIA's Crypto Mining Era
The new like the ninety eight hundred x three D
for example. So those I'm really excited about that. How that's gone because that actually you may like this, but that was actually a sort of a skunkworks project where this team of engineers in AMD and we ran a story on this too, but they decided, let's just try this idea, and they experimented with it, they got it working, and they brought it through management and said, hey, this looks viable and uh, you know, and then they were able to actually make a product that it came from
an abnormal path of an underground sort of you know, skunk works effort at the in the engineering department to actually being the leading gaming product for CPUs on the market.
That's it I We will have that link in the episode notes as well, and we can wrap on the product, the AMD skunk works product. You mentioned in a recent video that I had not AMD made a mountain bike at one point, just want to one.
Yeah, so yeah, that was not a good moment for AMD.
Uh.
Fortunately, fortunately it didn't particularly damage their CPU or gp A credibility because it was like so.
Long imagine so but why why did they do.
What they It was turning the COVID bike shortages. I don't know if you remember, but like all kinds of bicycles, yeah, were difficult to get. And uh, yes, they made these three hundred dollars so called mountain bikes. I think they're like Kent Walmart bike rebrands basically nice. So they went real cheap, yeah, and uh yeah it was advertised, you know for mountain biking and off road trails and whatever. And so we took it down a trail and I
actually had trouble getting out of the parking lot. That's how messed up the bike was. We had to take it to a shop before I rode it for a safety check because the the used V brakes and they were going through the spokes. So they definitely would have stopped me.
Jenson woe sitting there being like we could have had a fucking mountain butt man. Yeah we could.
How do we get Yeah, if we make a bike, we can get rid of them.
They're just like two twenty billion dollars shoved into mountain bike.
Resa.
That would be so good, Steve, It's always such a pleasure having you on here. Where can people find.
You and gamers next us on YouTube?
Thank you so much for being here. I'm of course that's itch when you've been listening to Better Offline, have a great week, everyone, staves, thanks for coming on.
Thank you, Thank you for listening to Better Offline.
The editor and composer of the Better Offline theme song is Mattersowski. You can check out more of his music
¶ Future of Enterprise GPUs
and audio projects at Mattasowski 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 you go to chat dot Where's youreed dot at to visit the discord, and go to our slash Better Offline to check out our Reddit, Thank you so much for listening.
Better Offline is a production of cool Zone Media. For more from cool Zone Media, visit our website cool zonemedia dot com, or check us out on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.