1 00:00:02,400 --> 00:00:03,520 Speaker 1: Also media. 2 00:00:05,040 --> 00:00:07,200 Speaker 2: Hello and welcome to Better Offline. I'm your host ed 3 00:00:07,280 --> 00:00:09,320 Speaker 2: Zeitron and this week we're joined again by the wonderful 4 00:00:09,320 --> 00:00:23,280 Speaker 2: Steve Burke of Gamers. Next us. How are you doing, Steve? 5 00:00:23,520 --> 00:00:24,079 Speaker 3: Doing well? 6 00:00:24,120 --> 00:00:27,080 Speaker 2: How about you doing great? See you? We're just at 7 00:00:27,080 --> 00:00:28,120 Speaker 2: computext right. 8 00:00:28,560 --> 00:00:30,680 Speaker 3: Yes, yeah, we're over there for a couple of weeks. 9 00:00:31,120 --> 00:00:33,199 Speaker 2: So for the listeners who don't know, because this is 10 00:00:33,200 --> 00:00:35,559 Speaker 2: a fair amount, you don't know every conference. So what 11 00:00:35,800 --> 00:00:37,720 Speaker 2: is computext and what makes it so important? 12 00:00:38,280 --> 00:00:41,360 Speaker 3: Copy text? Is the I would I think it's the 13 00:00:41,520 --> 00:00:46,440 Speaker 3: largest computer hardware and computer technology trade show. So that's 14 00:00:46,440 --> 00:00:50,040 Speaker 3: hosted in Taipei every year, and it's it's like a 15 00:00:50,159 --> 00:00:54,800 Speaker 3: much better CES. People may know CS. 16 00:00:54,440 --> 00:00:57,360 Speaker 2: And is it more focused on enterprise stuff or consumer stuff? 17 00:00:58,200 --> 00:01:04,200 Speaker 3: This year, weirdly, they AI commandeered the entire main hall, 18 00:01:04,480 --> 00:01:07,119 Speaker 3: so there were actually there are a lot of pissed 19 00:01:07,160 --> 00:01:11,520 Speaker 3: off consumer hardware manufacturers who were sort of evicted and 20 00:01:11,600 --> 00:01:13,680 Speaker 3: driven out of the main hall and put into the 21 00:01:14,240 --> 00:01:15,840 Speaker 3: lesser one with consumer hardware. 22 00:01:16,040 --> 00:01:19,440 Speaker 2: What were they even showing with the AI stuff as well? 23 00:01:19,640 --> 00:01:23,000 Speaker 3: I mean one company had an AI computer case? 24 00:01:23,600 --> 00:01:26,080 Speaker 2: Oh how did that? What's the AI do? 25 00:01:26,800 --> 00:01:29,160 Speaker 3: Absolutely no idea sick. 26 00:01:29,800 --> 00:01:32,520 Speaker 2: It's so cool that this is absorbed, so it's also 27 00:01:32,680 --> 00:01:37,600 Speaker 2: kind of brutal, that's literally absorbing space. Yes, did you 28 00:01:37,640 --> 00:01:39,640 Speaker 2: see anything cool that though, anything you're excited about. 29 00:01:39,760 --> 00:01:43,600 Speaker 3: Yeah, there's a lot of cool stuff the I mean 30 00:01:43,760 --> 00:01:46,440 Speaker 3: there was some immersion cooling, which is always nice to see, 31 00:01:46,560 --> 00:01:49,400 Speaker 3: like where they actually literally immersed the system, you know, 32 00:01:49,480 --> 00:01:53,760 Speaker 3: in fluid for cooling. There was a thermociphon from Nocto, 33 00:01:53,920 --> 00:01:56,640 Speaker 3: which is just a really cool use of it's another 34 00:01:56,680 --> 00:02:02,840 Speaker 3: cooling technology. Lots of computer case and cooling solution development. 35 00:02:03,040 --> 00:02:06,360 Speaker 3: And then on the GPU side, I mean N Video 36 00:02:06,520 --> 00:02:09,640 Speaker 3: sort of buried its fifty sixty launch and then AMD 37 00:02:10,480 --> 00:02:13,960 Speaker 3: more formally announced the information for its ninety sixty XT. 38 00:02:14,080 --> 00:02:16,480 Speaker 3: And both of those are sort of the slightly more 39 00:02:16,520 --> 00:02:19,440 Speaker 3: affordable consumer class GPUs. 40 00:02:19,520 --> 00:02:22,840 Speaker 2: Yet so in Video also was doing some funny business 41 00:02:22,880 --> 00:02:26,480 Speaker 2: with you recently you did that wonderful video. So walk 42 00:02:26,560 --> 00:02:28,880 Speaker 2: us through what happened that because they tried to big 43 00:02:28,919 --> 00:02:30,239 Speaker 2: dog you so to speak. 44 00:02:30,639 --> 00:02:34,160 Speaker 3: Yeah, So, I mean the problem we had with them 45 00:02:34,320 --> 00:02:37,919 Speaker 3: we talked about was for several months there had been 46 00:02:37,960 --> 00:02:42,320 Speaker 3: this I felt sort of pressure from Nvideo with where 47 00:02:42,760 --> 00:02:45,080 Speaker 3: so N Video for contact with people, and Video sends 48 00:02:45,080 --> 00:02:48,960 Speaker 3: out GP review samples. We and other reviewers are not 49 00:02:49,120 --> 00:02:51,280 Speaker 3: entitled to them. That's fine, we can buy them ourselves 50 00:02:51,280 --> 00:02:54,400 Speaker 3: as well, but they send out samples for product reviews. 51 00:02:55,040 --> 00:02:58,720 Speaker 3: And in the past, and Video has been in hot 52 00:02:58,760 --> 00:03:03,400 Speaker 3: water for asking another outlet called the hard Run Box 53 00:03:03,520 --> 00:03:06,560 Speaker 3: to I believe the quote was, or the paraphrase was, 54 00:03:07,280 --> 00:03:11,000 Speaker 3: change your editorial direction. Yeah, and so that was several 55 00:03:11,120 --> 00:03:11,560 Speaker 3: years ago. 56 00:03:11,960 --> 00:03:13,480 Speaker 2: And what did they what did they ask them to 57 00:03:13,520 --> 00:03:14,359 Speaker 2: do back then? 58 00:03:14,840 --> 00:03:19,399 Speaker 3: Back then it was seeking more coverage of real time 59 00:03:19,480 --> 00:03:23,480 Speaker 3: rate tracing in video games, which at the time was 60 00:03:23,560 --> 00:03:26,840 Speaker 3: more or less only found in any kind of meaningful 61 00:03:26,840 --> 00:03:29,720 Speaker 3: performance on in video hardware. But also and the reason 62 00:03:31,480 --> 00:03:35,480 Speaker 3: at least we weren't ready to fully do testing for 63 00:03:35,520 --> 00:03:38,760 Speaker 3: real time rate tracing yet was because it just it 64 00:03:38,800 --> 00:03:40,640 Speaker 3: wasn't ready yet. I mean, there weren't that many games 65 00:03:40,640 --> 00:03:40,880 Speaker 3: with it. 66 00:03:40,920 --> 00:03:43,160 Speaker 2: You couldn't really test it because you couldn't use it. 67 00:03:43,360 --> 00:03:45,280 Speaker 3: Yeah, I mean they launched and if I remember correctly, 68 00:03:45,280 --> 00:03:49,360 Speaker 3: there were actually zero games that supported this technology and 69 00:03:49,440 --> 00:03:52,160 Speaker 3: the only thing available were sort of tech demos. And 70 00:03:52,240 --> 00:03:55,200 Speaker 3: so anyway that that was what they wanted, then they 71 00:03:55,280 --> 00:03:57,600 Speaker 3: kind of went away, didn't That's at least as far 72 00:03:57,600 --> 00:04:01,240 Speaker 3: as we know, publicly didn't try to push reviewers in 73 00:04:01,280 --> 00:04:05,040 Speaker 3: a particular direction with their narrative or their editorial style. 74 00:04:06,440 --> 00:04:08,640 Speaker 3: And then over the last several months we were getting 75 00:04:08,640 --> 00:04:16,080 Speaker 3: this pressure of basically so we do a bunch of 76 00:04:16,080 --> 00:04:20,719 Speaker 3: interviews with engineers and technical people at Nvidia, and I 77 00:04:20,760 --> 00:04:23,440 Speaker 3: felt they were starting to use that as a lever 78 00:04:23,760 --> 00:04:26,000 Speaker 3: to be like, oh, it sure seems like you like 79 00:04:26,200 --> 00:04:28,200 Speaker 3: talking to the thermal engineers. It'd be a shame of 80 00:04:28,279 --> 00:04:29,000 Speaker 3: something happened. 81 00:04:29,160 --> 00:04:31,599 Speaker 2: And you did that very long one over within video, right, 82 00:04:31,640 --> 00:04:33,479 Speaker 2: you went and kind of looked at how some with 83 00:04:33,520 --> 00:04:34,120 Speaker 2: stuff worked. 84 00:04:34,360 --> 00:04:37,279 Speaker 3: Yeah, yeah, we did. I mean they're great educational content. 85 00:04:37,360 --> 00:04:40,040 Speaker 3: The engineers are awesome, and I've spent years working with 86 00:04:40,080 --> 00:04:43,560 Speaker 3: mostly people who are in different departments or don't work 87 00:04:43,560 --> 00:04:48,080 Speaker 3: there anymore to get in video on the same page 88 00:04:48,120 --> 00:04:51,040 Speaker 3: where I'm like, look, we're not looking for a marketing 89 00:04:51,160 --> 00:04:53,960 Speaker 3: puff piece. We're looking for actual engineers who just want 90 00:04:54,000 --> 00:04:56,359 Speaker 3: to talk about the stuff they make and people can 91 00:04:56,480 --> 00:04:59,000 Speaker 3: learn about it. But it's not like, obviously, yes, it 92 00:04:59,040 --> 00:05:02,560 Speaker 3: can support the problem act by nature of if it's 93 00:05:02,600 --> 00:05:05,120 Speaker 3: good and the engineers explain it well, then people will 94 00:05:05,200 --> 00:05:08,479 Speaker 3: like it. But the objective is not just like come 95 00:05:08,520 --> 00:05:10,120 Speaker 3: on our show and use us as a platform to 96 00:05:10,160 --> 00:05:13,279 Speaker 3: sell things. We want engineers, not marketers. Yeah, and so 97 00:05:13,320 --> 00:05:18,080 Speaker 3: that was great. And the like I said, lever that 98 00:05:18,160 --> 00:05:21,560 Speaker 3: it became was they wanted more in video on more 99 00:05:21,640 --> 00:05:25,040 Speaker 3: coverage of this thing called MFG, which is multi frame generation, 100 00:05:26,760 --> 00:05:30,559 Speaker 3: which is something that Jensen Huan, the CEO, talks about 101 00:05:30,560 --> 00:05:34,360 Speaker 3: on stage as being part of their technology suite that 102 00:05:35,480 --> 00:05:42,120 Speaker 3: is a big surprise AI accelerated that effectively multiplies the 103 00:05:42,160 --> 00:05:47,000 Speaker 3: frame rate to be used as a smoothing technology. 104 00:05:47,279 --> 00:05:50,240 Speaker 2: So it's different. Is this different to like DLSS and 105 00:05:50,279 --> 00:05:52,000 Speaker 2: things like that, or is that. 106 00:05:52,480 --> 00:05:55,000 Speaker 3: It is part of the DLSS or the Deep Learned 107 00:05:55,040 --> 00:05:59,440 Speaker 3: Super Sampling Technology suite. So it's sort of it's it's 108 00:05:59,440 --> 00:06:02,520 Speaker 3: almost like a subset of DLSS. 109 00:06:02,160 --> 00:06:05,719 Speaker 2: Right, and so it basically just looks better at higher resolution. 110 00:06:05,839 --> 00:06:08,440 Speaker 2: It fills in the frames. I'm guessing just trying to simplify. 111 00:06:08,720 --> 00:06:13,279 Speaker 3: Yeah, So yes, the simplest version of it is MFG 112 00:06:14,360 --> 00:06:19,320 Speaker 3: sort of interpolates frames, so it's able to add in 113 00:06:19,440 --> 00:06:23,320 Speaker 3: or insert additional effectively artificial or generated frames if you're 114 00:06:23,360 --> 00:06:26,320 Speaker 3: trying to be charitable with it, that are not literally 115 00:06:27,279 --> 00:06:31,000 Speaker 3: built by the game engine, and so the GPU is 116 00:06:31,040 --> 00:06:34,640 Speaker 3: working with its own software to construct these frames and 117 00:06:34,800 --> 00:06:37,760 Speaker 3: construct pixels based on what it thinks is going to 118 00:06:37,839 --> 00:06:40,839 Speaker 3: happen next. And we've actually tested it and there's situations 119 00:06:40,880 --> 00:06:43,320 Speaker 3: where it's not bad. There's situations where it's awful, like 120 00:06:43,360 --> 00:06:49,119 Speaker 3: with any technology. But what it does is on a chart, 121 00:06:49,279 --> 00:06:51,200 Speaker 3: if you just look at the number that comes out 122 00:06:51,960 --> 00:06:54,680 Speaker 3: the so called frame rate or the amount of frames 123 00:06:54,680 --> 00:06:58,800 Speaker 3: in a given period of time, that looks higher with 124 00:06:58,839 --> 00:07:03,839 Speaker 3: this technology, but it's not measuring the same thing as 125 00:07:04,200 --> 00:07:08,000 Speaker 3: something without this technology. So it's effectively, you know, you're 126 00:07:08,040 --> 00:07:12,800 Speaker 3: comparing apples and oranges if you try to put real 127 00:07:12,880 --> 00:07:17,040 Speaker 3: frames that are rendered in the traditional way up against 128 00:07:17,200 --> 00:07:21,080 Speaker 3: these generated AI frames that are effectively guessing at what 129 00:07:21,080 --> 00:07:21,840 Speaker 3: they should look. 130 00:07:21,760 --> 00:07:23,560 Speaker 2: Like, and do they look bad as well? 131 00:07:23,640 --> 00:07:26,320 Speaker 3: Like they can they can look really bad. I mean 132 00:07:26,760 --> 00:07:30,120 Speaker 3: we've done some testing where in some of the games 133 00:07:30,200 --> 00:07:34,080 Speaker 3: it's just like text for example, will will get completely 134 00:07:34,120 --> 00:07:38,480 Speaker 3: garbled text on screen interface interface issues will draw the 135 00:07:38,520 --> 00:07:42,239 Speaker 3: interface in the wrong area. You could have really. 136 00:07:42,600 --> 00:07:46,200 Speaker 2: Sanely bad for like any RPG of any kind or 137 00:07:46,920 --> 00:07:48,880 Speaker 2: any kind of like Twitch I own. 138 00:07:49,240 --> 00:07:50,000 Speaker 3: Yeah, it feels like. 139 00:07:49,960 --> 00:07:52,640 Speaker 2: Anything you'd be doing EVA would be ruined by that. 140 00:07:53,120 --> 00:07:55,680 Speaker 3: Yeah, so there's scenarios where it works, Okay, But the 141 00:07:55,840 --> 00:07:59,240 Speaker 3: RPG example you brought up, so Final Fantasy was one 142 00:07:59,240 --> 00:08:02,960 Speaker 3: of the games we test did where there's just areas. 143 00:08:03,040 --> 00:08:04,960 Speaker 3: I mean, it just I personally would turn it off. 144 00:08:05,000 --> 00:08:07,840 Speaker 3: It looks worse and I'd rather have a lower quote 145 00:08:07,880 --> 00:08:12,360 Speaker 3: unquote frame rate than have this sort of potential blurry 146 00:08:12,400 --> 00:08:13,200 Speaker 3: mess every now and then. 147 00:08:14,160 --> 00:08:17,760 Speaker 2: So strange low qualities. Well, it's just a very strange 148 00:08:17,760 --> 00:08:18,800 Speaker 2: thing to push out the door. 149 00:08:19,520 --> 00:08:24,120 Speaker 3: Yeah, and the you know, I think sort of finished 150 00:08:24,160 --> 00:08:28,440 Speaker 3: the editorial narrative. They know, I mean, it's fine, but 151 00:08:28,560 --> 00:08:30,120 Speaker 3: I'll finish that first and I'll come back to your 152 00:08:30,160 --> 00:08:32,040 Speaker 3: account you just make because that's also interesting. 153 00:08:32,520 --> 00:08:33,000 Speaker 2: But on the. 154 00:08:33,000 --> 00:08:37,160 Speaker 3: Editorial side, they wanted more coverage of MFG, and we 155 00:08:37,360 --> 00:08:41,720 Speaker 3: disagreed from an objective standpoint. You know, I kind of 156 00:08:41,760 --> 00:08:44,000 Speaker 3: made the case of them several times, over several calls, 157 00:08:44,040 --> 00:08:48,600 Speaker 3: over several months of why we think it It just 158 00:08:48,720 --> 00:08:54,040 Speaker 3: doesn't belong on the same charts as normal benchmarking, normal 159 00:08:54,080 --> 00:08:57,920 Speaker 3: objective testing, and so there was, Yeah, there's just a disagreement, 160 00:08:58,240 --> 00:09:02,640 Speaker 3: which is fine. The problem came in when I felt 161 00:09:02,679 --> 00:09:06,720 Speaker 3: repeatedly on various calls with the N Video people that 162 00:09:06,800 --> 00:09:10,560 Speaker 3: they were almost sort of holding hostage the idea of 163 00:09:11,320 --> 00:09:15,520 Speaker 3: working with engineers again, where it's well, the way we're 164 00:09:15,559 --> 00:09:18,400 Speaker 3: able to make this happen is by you covering things 165 00:09:18,440 --> 00:09:21,199 Speaker 3: like MFG. You know, sort of that type of conversation. 166 00:09:21,640 --> 00:09:24,240 Speaker 2: Yeah, that's so. And they mentioned something about like budgets 167 00:09:24,679 --> 00:09:25,480 Speaker 2: in the video. 168 00:09:25,480 --> 00:09:28,040 Speaker 3: I remember, yeah, I mean there was like discussion of 169 00:09:29,600 --> 00:09:34,120 Speaker 3: one of the times specifically, the statement was that this 170 00:09:34,240 --> 00:09:37,800 Speaker 3: is how we can secure budget for these types of things, 171 00:09:38,280 --> 00:09:41,720 Speaker 3: and like to be clear for listeners, we don't take 172 00:09:41,800 --> 00:09:46,320 Speaker 3: money for interviews. Yeah yeah, And in fact, I've paid 173 00:09:46,520 --> 00:09:50,080 Speaker 3: now over five figures to visit N Video on numerous 174 00:09:50,080 --> 00:09:54,079 Speaker 3: occasions for said interviews. Personally, like I'm paid. 175 00:09:53,920 --> 00:09:57,120 Speaker 2: And I can say this from the PR side, and 176 00:09:57,240 --> 00:09:59,840 Speaker 2: not working not work with A video or anyone associated 177 00:10:00,080 --> 00:10:03,120 Speaker 2: five all that. It's like, no one talks about budgets. 178 00:10:03,240 --> 00:10:06,000 Speaker 2: No one is like, oh I can't visit a journalist 179 00:10:06,120 --> 00:10:08,840 Speaker 2: due to budget, and certainly not in Video, a company 180 00:10:08,840 --> 00:10:11,520 Speaker 2: that had thirty nine billion dollars I believe in revenue 181 00:10:11,640 --> 00:10:13,120 Speaker 2: just in the data cent a portion. 182 00:10:14,640 --> 00:10:18,480 Speaker 3: Of classical multi trillion dollar marketing cap. I think if 183 00:10:18,480 --> 00:10:22,960 Speaker 3: I try to take Devil's advocate at maybe there's some 184 00:10:23,840 --> 00:10:28,120 Speaker 3: internal budgeting whatever, like the marketing department needs to budget 185 00:10:28,160 --> 00:10:31,720 Speaker 3: time from the engineering department or something. But we're talking 186 00:10:31,840 --> 00:10:36,160 Speaker 3: like a couple hours of time max normally, you know, yeah, 187 00:10:36,240 --> 00:10:39,200 Speaker 3: especially when I'm traveling to them. So it just seemed 188 00:10:39,240 --> 00:10:42,760 Speaker 3: like I don't know it. I didn't buy it. And 189 00:10:42,840 --> 00:10:49,959 Speaker 3: also that was only one of the numerous UH calls, 190 00:10:50,720 --> 00:10:54,080 Speaker 3: you know, like that the defensible reason for them changed. 191 00:10:54,200 --> 00:10:55,959 Speaker 3: It felt like each time where it's like, oh, let's 192 00:10:55,960 --> 00:10:57,360 Speaker 3: try this angle, let's try this angle. 193 00:10:58,040 --> 00:11:01,199 Speaker 2: So did they end up blocking you? They just how 194 00:11:01,240 --> 00:11:02,840 Speaker 2: did that situation resolve itself? 195 00:11:03,360 --> 00:11:06,720 Speaker 3: I have not spoken with anyone there since we posted 196 00:11:06,720 --> 00:11:10,000 Speaker 3: our video talking about it. And part of the challenge 197 00:11:10,040 --> 00:11:14,800 Speaker 3: with posting that video is, you know, it's like, the 198 00:11:15,080 --> 00:11:20,240 Speaker 3: first couple of times this came up, I assumed that 199 00:11:20,280 --> 00:11:24,080 Speaker 3: it was just sort of okay, maybe the guy phrased 200 00:11:24,080 --> 00:11:25,800 Speaker 3: it ron you know, I've worked with these people for 201 00:11:25,840 --> 00:11:29,200 Speaker 3: a while and maybe just came out wrong. After a 202 00:11:29,200 --> 00:11:31,080 Speaker 3: couple of times, I did call one of my reps 203 00:11:31,160 --> 00:11:35,280 Speaker 3: after the meeting where they kind of pushed this, you know, 204 00:11:35,920 --> 00:11:39,040 Speaker 3: you testing this technology that you think is not appropriate 205 00:11:39,080 --> 00:11:41,600 Speaker 3: to share on charts with competitors. That's how we get 206 00:11:41,640 --> 00:11:46,120 Speaker 3: you these interviews. I called a rep and said, you know, hey, this, 207 00:11:46,559 --> 00:11:48,280 Speaker 3: if you guys say this to people, this is going 208 00:11:48,320 --> 00:11:50,160 Speaker 3: to be taken the wrong way. And like, I'm still 209 00:11:50,160 --> 00:11:56,800 Speaker 3: assuming it's not meant this way. But but if I 210 00:11:56,880 --> 00:11:59,600 Speaker 3: hear this again, right, like it's I'm did I have 211 00:11:59,679 --> 00:12:03,120 Speaker 3: to start assuming it's it's the way it sounds. And 212 00:12:03,280 --> 00:12:05,920 Speaker 3: it did come up again. And part of the challenge 213 00:12:05,920 --> 00:12:07,440 Speaker 3: with making the video we made was, you know, we 214 00:12:07,480 --> 00:12:09,559 Speaker 3: want to blow the whistle on it in case other 215 00:12:09,559 --> 00:12:12,840 Speaker 3: people are facing that pressure, and also in case consumers 216 00:12:12,880 --> 00:12:16,240 Speaker 3: are watching or reading content which may be influenced without 217 00:12:16,240 --> 00:12:20,080 Speaker 3: their knowledge. The problem is, I think en video may 218 00:12:20,120 --> 00:12:25,079 Speaker 3: now be in a situation where let's just assume everybody 219 00:12:25,120 --> 00:12:29,439 Speaker 3: makes up and it's fine. At some point, I don't 220 00:12:29,480 --> 00:12:32,439 Speaker 3: know that they will ever be able to put an 221 00:12:32,440 --> 00:12:35,319 Speaker 3: engineer on cam RA any time of the near future 222 00:12:35,360 --> 00:12:39,840 Speaker 3: with anybody without the audience questioning it. And for good reason, 223 00:12:40,480 --> 00:12:44,080 Speaker 3: because that was what we felt was being used as 224 00:12:44,120 --> 00:12:47,280 Speaker 3: the lever against us. Yeah, and so it's just kind 225 00:12:47,280 --> 00:12:49,720 Speaker 3: of I don't know. We tried to, you know, I 226 00:12:49,760 --> 00:12:52,040 Speaker 3: explained to them several times why this is a problem 227 00:12:52,400 --> 00:12:54,400 Speaker 3: and why the phrasing is a problem, and the more 228 00:12:54,400 --> 00:12:55,640 Speaker 3: you hear it, the more it's like, Okay, I mean 229 00:12:55,679 --> 00:12:58,680 Speaker 3: at some point, like I think you're saying what it 230 00:12:58,720 --> 00:13:00,720 Speaker 3: sounds like you're saying, I'm going to stop. 231 00:13:01,200 --> 00:13:02,880 Speaker 2: Yeah, it's good that you did it, though, because it's 232 00:13:03,880 --> 00:13:07,120 Speaker 2: GM's got over a million subscribers, so who knows what 233 00:13:07,160 --> 00:13:09,080 Speaker 2: they're doing too much smaller outlets as well. 234 00:13:09,840 --> 00:13:12,240 Speaker 3: Yeah, I mean, and that's what I'm really concerned about 235 00:13:12,400 --> 00:13:16,880 Speaker 3: is I've had conversations with Nvidia in the past that 236 00:13:16,920 --> 00:13:20,160 Speaker 3: will probably get into in a future video, but about 237 00:13:20,200 --> 00:13:24,720 Speaker 3: their strategy for newer creators, and I don't like what 238 00:13:24,760 --> 00:13:27,800 Speaker 3: I've heard. I mean, it sounds very much like we're 239 00:13:27,800 --> 00:13:30,640 Speaker 3: a're gonna sort of try and shape these newer creators. 240 00:13:30,679 --> 00:13:33,560 Speaker 3: And you know, I think the problem is as especially 241 00:13:33,600 --> 00:13:36,800 Speaker 3: old guard media, as in people like OG's before US 242 00:13:37,120 --> 00:13:43,160 Speaker 3: retire or move to other industries, there's openings to change 243 00:13:43,200 --> 00:13:46,000 Speaker 3: how media interfaces with these companies in general, because you've 244 00:13:46,040 --> 00:13:51,319 Speaker 3: got newer people who don't have the historical background with 245 00:13:51,360 --> 00:13:53,760 Speaker 3: these companies to know what to expect, and so it's 246 00:13:53,800 --> 00:13:56,680 Speaker 3: easier to kind of slowly creep that goalpost where it's 247 00:13:57,160 --> 00:14:00,000 Speaker 3: it's less independence and it's more marketing arm of the company. 248 00:14:00,320 --> 00:14:02,199 Speaker 3: And that's what I'm really worried about, is if they 249 00:14:02,240 --> 00:14:05,800 Speaker 3: get in any of these companies with newer creators early 250 00:14:06,679 --> 00:14:13,080 Speaker 3: and are able to change the expectancy of the relationship, 251 00:14:13,160 --> 00:14:17,400 Speaker 3: so change it from independent reviewer analyzing large companies device 252 00:14:17,679 --> 00:14:20,080 Speaker 3: into more of a hey, we're all friends, we're all 253 00:14:20,120 --> 00:14:22,640 Speaker 3: working together to try and benefit the consumer. When if 254 00:14:22,680 --> 00:14:24,920 Speaker 3: you're talking about a multi trillion dollar company that their 255 00:14:25,000 --> 00:14:27,640 Speaker 3: motives are far different from like the guy in his 256 00:14:27,720 --> 00:14:34,160 Speaker 3: bedroom who's trying to do hardware reviews. 257 00:14:39,880 --> 00:14:43,479 Speaker 2: It's just so weird as well, because they're already so successful. 258 00:14:43,720 --> 00:14:46,320 Speaker 2: Is this history? It sounds like historically in Vidia has 259 00:14:46,320 --> 00:14:49,800 Speaker 2: been a bit draconian like this, though not necessarily pressuring 260 00:14:49,840 --> 00:14:50,360 Speaker 2: in this way. 261 00:14:50,800 --> 00:14:53,400 Speaker 3: I mean, I certainly think and Vidia is one of 262 00:14:53,440 --> 00:14:58,240 Speaker 3: the most vindictive companies in the space. They the stories 263 00:14:58,320 --> 00:15:01,440 Speaker 3: we've heard and told in some instance of their partners, 264 00:15:01,520 --> 00:15:06,520 Speaker 3: not reviewers, but actual business partners like board vendors, people 265 00:15:06,520 --> 00:15:09,440 Speaker 3: who make the video cards that the Nvidio GPS go on. 266 00:15:10,240 --> 00:15:15,520 Speaker 3: Those stories would seem to indicate a sort of vindictiveness 267 00:15:16,240 --> 00:15:21,360 Speaker 3: can give an example, Yeah, allocation. So GPU allocation is 268 00:15:22,000 --> 00:15:27,560 Speaker 3: the concept of how many of a high demand GPU 269 00:15:27,760 --> 00:15:30,920 Speaker 3: can you, as a partner get to then attached to 270 00:15:31,160 --> 00:15:34,120 Speaker 3: a product to then sell to a consumer. And there's 271 00:15:34,320 --> 00:15:35,960 Speaker 3: a limited amount of these because they're going to sell 272 00:15:35,960 --> 00:15:41,600 Speaker 3: one hundred percent of them, And allocation is typically measured 273 00:15:41,640 --> 00:15:45,280 Speaker 3: by percent, So company A might get twenty five percent, 274 00:15:45,320 --> 00:15:47,680 Speaker 3: Company B might get five percent of the GPU supply 275 00:15:48,240 --> 00:15:51,840 Speaker 3: whatever that is. And that allocation, we've been told in 276 00:15:51,880 --> 00:15:55,160 Speaker 3: the past, has been used also as sort of a lever. 277 00:15:55,400 --> 00:15:58,480 Speaker 3: And so an example would be back in the EVGA days, 278 00:15:58,480 --> 00:16:01,360 Speaker 3: which Evga sort of famously. 279 00:16:02,480 --> 00:16:04,160 Speaker 2: That was a graphics called company. 280 00:16:04,320 --> 00:16:07,840 Speaker 3: Correct. Yeah, they they were in video's number one partner, 281 00:16:08,440 --> 00:16:12,520 Speaker 3: and they were so Nvidio favored that they only made 282 00:16:12,600 --> 00:16:16,160 Speaker 3: Nvidia video cards. They did not work with any competitors, right, 283 00:16:16,680 --> 00:16:20,560 Speaker 3: and they famously cut that relationship publicly and we ran 284 00:16:20,600 --> 00:16:22,720 Speaker 3: a story about that. But one of the things that 285 00:16:22,720 --> 00:16:27,400 Speaker 3: they were unhappy about was Nvidia slowly tightening the screws, 286 00:16:27,840 --> 00:16:32,120 Speaker 3: which is a direct quote from an Nvidia pm I 287 00:16:32,160 --> 00:16:37,760 Speaker 3: spoke to one. Slowly tightening the tight end the screws 288 00:16:37,840 --> 00:16:42,000 Speaker 3: is the quote, but on what the partners were allowed 289 00:16:42,000 --> 00:16:46,080 Speaker 3: to do with their products, and slowly over time they're 290 00:16:46,080 --> 00:16:47,640 Speaker 3: losing some of this creative freedom. 291 00:16:49,240 --> 00:16:51,160 Speaker 2: What did you say, they lose creative freedom? Like, what 292 00:16:51,200 --> 00:16:53,280 Speaker 2: are they being constrained to do? 293 00:16:53,720 --> 00:16:56,000 Speaker 3: One that's public is we did an interview with a 294 00:16:56,440 --> 00:17:00,800 Speaker 3: famous over clocker who used to work for or EVGA. 295 00:17:01,680 --> 00:17:04,960 Speaker 3: Kingpin is his username and his name is on video 296 00:17:05,000 --> 00:17:08,119 Speaker 3: cards that are world famous. We did a video with 297 00:17:08,200 --> 00:17:12,240 Speaker 3: him and he talked about how he wanted to add 298 00:17:12,280 --> 00:17:15,560 Speaker 3: two power connectors to a card to basically make it 299 00:17:15,600 --> 00:17:18,720 Speaker 3: more fun to overclock with for consumers. And so it'd 300 00:17:18,760 --> 00:17:20,840 Speaker 3: be to be able to pull more power, and it 301 00:17:20,840 --> 00:17:23,520 Speaker 3: would also be able to do so safer because the 302 00:17:23,600 --> 00:17:28,359 Speaker 3: current power connector has this issue of burning and catching 303 00:17:28,400 --> 00:17:30,720 Speaker 3: on fire sometimes, right, so, and so he wanted to 304 00:17:30,720 --> 00:17:32,600 Speaker 3: do two of them to kind of resolve some of that. 305 00:17:33,119 --> 00:17:36,760 Speaker 3: And in the public discussion we had with Kingpin and 306 00:17:36,840 --> 00:17:44,359 Speaker 3: the video he said paraphrasing, but the conversation was he 307 00:17:44,400 --> 00:17:46,639 Speaker 3: said they wouldn't allow it, or they restricted it or 308 00:17:46,680 --> 00:17:48,919 Speaker 3: something like that, And I said who is they? And 309 00:17:48,960 --> 00:17:52,359 Speaker 3: he said they, Yeah, like we we there was only 310 00:17:52,400 --> 00:17:53,560 Speaker 3: one day there could. 311 00:17:53,400 --> 00:17:54,800 Speaker 2: Be who could it be? 312 00:17:55,480 --> 00:17:59,399 Speaker 3: Yeah, But there's plenty of examples like that, where the 313 00:18:00,119 --> 00:18:05,720 Speaker 3: another one would be a couple generations ago. The cheaper cards, 314 00:18:05,760 --> 00:18:09,800 Speaker 3: like the two hundred three hundred dollars models they were 315 00:18:10,320 --> 00:18:14,040 Speaker 3: those models were difficult for the partners to actually make 316 00:18:14,200 --> 00:18:16,560 Speaker 3: and still make any money. And so there were times 317 00:18:16,600 --> 00:18:21,440 Speaker 3: where we were told about how this MSRP level video card, 318 00:18:21,720 --> 00:18:24,360 Speaker 3: meaning the number that in video says is the MSRP, 319 00:18:25,000 --> 00:18:27,120 Speaker 3: that would only exist for a fixed period of time 320 00:18:28,640 --> 00:18:31,560 Speaker 3: because it was simply impossible to afford to make it. 321 00:18:32,160 --> 00:18:34,320 Speaker 3: But and Video really wanted them to exist at this 322 00:18:34,400 --> 00:18:37,879 Speaker 3: certain MSRP. It allows them to advertise a lower price, 323 00:18:38,480 --> 00:18:41,560 Speaker 3: and so they're pushed to make these cards that allows 324 00:18:41,560 --> 00:18:44,800 Speaker 3: them to potentially get things like more access to allocation 325 00:18:45,080 --> 00:18:47,800 Speaker 3: or you know whatever privilege they may get. 326 00:18:48,400 --> 00:18:51,760 Speaker 2: So on a much larger level, what do you think 327 00:18:51,840 --> 00:18:53,760 Speaker 2: is going on within Video as a company right now? 328 00:18:53,800 --> 00:18:57,159 Speaker 2: I'm not asking for stock analysis like that, totally not that, 329 00:18:57,200 --> 00:19:00,800 Speaker 2: but just as a company that both makes can zumographics 330 00:19:00,880 --> 00:19:05,800 Speaker 2: cards and undepinds whatever the AI thing is, how like, 331 00:19:05,840 --> 00:19:07,440 Speaker 2: what do you think is going on there right now? 332 00:19:08,359 --> 00:19:15,120 Speaker 3: I think there's definitely a focus on AI to try 333 00:19:15,119 --> 00:19:18,680 Speaker 3: and get money. I mean, there's a Jensen Juan was 334 00:19:18,720 --> 00:19:22,480 Speaker 3: in an interview I don't know last year sometime, I think, 335 00:19:23,240 --> 00:19:26,160 Speaker 3: or earlier this year, and made a coment about how 336 00:19:28,240 --> 00:19:31,800 Speaker 3: they use AI to write their own driver software now 337 00:19:31,880 --> 00:19:35,159 Speaker 3: or their own software or something like that. And you 338 00:19:35,200 --> 00:19:37,800 Speaker 3: know this is this is the number one focus. It's 339 00:19:37,800 --> 00:19:42,000 Speaker 3: what they're making the most I guess investor money on, 340 00:19:42,800 --> 00:19:46,040 Speaker 3: so they're focused on it. The concern I have is 341 00:19:47,560 --> 00:19:50,920 Speaker 3: it's one thing to sort of let the consumer hardware 342 00:19:51,000 --> 00:19:59,000 Speaker 3: side dwindle or get less competitive effort. It's another thing too, 343 00:19:59,400 --> 00:20:01,880 Speaker 3: in my opinion, and sort of drag the entire industry 344 00:20:01,960 --> 00:20:05,879 Speaker 3: down with it as they shift focus to the higher 345 00:20:05,920 --> 00:20:07,760 Speaker 3: revenue AI side. 346 00:20:08,200 --> 00:20:10,280 Speaker 2: Right when you say drag them down, how do you mean? 347 00:20:10,640 --> 00:20:15,359 Speaker 3: I mean things like trying to in my opinion, control 348 00:20:15,440 --> 00:20:18,840 Speaker 3: the media narrative and shape it like that. That to 349 00:20:18,960 --> 00:20:22,800 Speaker 3: me goes beyond Okay, we don't care much about consumer 350 00:20:22,960 --> 00:20:24,960 Speaker 3: We're going to focus on making money with AI. That 351 00:20:25,080 --> 00:20:30,000 Speaker 3: gets into we are going to actively damage the credibility 352 00:20:30,160 --> 00:20:34,640 Speaker 3: of independent media and also potentially harm consumers who are 353 00:20:34,680 --> 00:20:39,439 Speaker 3: influenced by that media, which has been shaped with a 354 00:20:39,440 --> 00:20:42,080 Speaker 3: certain you know, corporate narrative to push. 355 00:20:42,160 --> 00:20:43,480 Speaker 4: And that's my concern. 356 00:20:44,119 --> 00:20:46,440 Speaker 3: So that's that's the That's why I draw the distinction 357 00:20:46,520 --> 00:20:49,040 Speaker 3: between you. As we said in our video, it's if 358 00:20:49,080 --> 00:20:51,040 Speaker 3: you want to just focus on AI and if they 359 00:20:51,040 --> 00:20:54,480 Speaker 3: want to just sell all the GPUs to big enterprise companies, 360 00:20:54,800 --> 00:20:57,679 Speaker 3: go for it, but you know, don't screw all the 361 00:20:57,680 --> 00:20:59,399 Speaker 3: rest of us over at the same time. 362 00:21:00,160 --> 00:21:03,399 Speaker 2: And do you think that they're abandoning consumer or is 363 00:21:03,400 --> 00:21:06,280 Speaker 2: it just they are they are focusing on the shiny 364 00:21:06,320 --> 00:21:07,160 Speaker 2: object of the day. 365 00:21:08,320 --> 00:21:15,359 Speaker 3: I I don't know Jensen Juan personally my read on him, though, 366 00:21:16,760 --> 00:21:20,600 Speaker 3: I feel like it's unlikely he would want to abandon 367 00:21:20,680 --> 00:21:24,480 Speaker 3: something he already has, right, even if it's only out 368 00:21:24,480 --> 00:21:27,840 Speaker 3: of maybe sort of pride, and so I don't think 369 00:21:27,840 --> 00:21:32,199 Speaker 3: they'll abandon it. I do think they're focusing on AI. 370 00:21:32,640 --> 00:21:35,600 Speaker 3: It's just that, uh I don't. I mean, they have 371 00:21:35,720 --> 00:21:39,560 Speaker 3: like ninety two percent market share and consumer GPUs now 372 00:21:39,600 --> 00:21:43,560 Speaker 3: according to John Petty Research. So it's just like they can. 373 00:21:43,960 --> 00:21:46,479 Speaker 3: It's a monopoly. I mean, that's like, it's just like, 374 00:21:46,760 --> 00:21:48,199 Speaker 3: this is what monopolies look like. 375 00:21:48,800 --> 00:21:50,800 Speaker 2: Well, let's it feel I would say there was an 376 00:21:50,840 --> 00:21:54,080 Speaker 2: opportunity for someone to come in. But based on everything 377 00:21:54,119 --> 00:21:56,760 Speaker 2: you've shown with like AMD recently, even on the low end. 378 00:21:56,800 --> 00:21:59,399 Speaker 2: It just it doesn't feel it almost feels like consumer 379 00:21:59,440 --> 00:22:03,560 Speaker 2: graphics up being not treated with the respect they deserve, 380 00:22:03,960 --> 00:22:07,359 Speaker 2: despite PC gaming being huge industry and so on and 381 00:22:07,400 --> 00:22:07,879 Speaker 2: so forth. 382 00:22:08,040 --> 00:22:10,760 Speaker 3: Yeah, it's a massive industry and prosumer too. I mean 383 00:22:10,800 --> 00:22:13,760 Speaker 3: even just non AI use cases include people who edit 384 00:22:13,840 --> 00:22:19,080 Speaker 3: videos or make movies or do three D animations. Yeah, 385 00:22:19,119 --> 00:22:22,720 Speaker 3: I think Intel's a good barometer here because Intel decided 386 00:22:22,760 --> 00:22:25,440 Speaker 3: to get into consumer GPU, and they're new to it, 387 00:22:26,119 --> 00:22:29,080 Speaker 3: and they've had a hell of a time. They're doing 388 00:22:29,160 --> 00:22:33,080 Speaker 3: much better now with this current generation. But one of 389 00:22:33,080 --> 00:22:35,600 Speaker 3: the quotes from an Intel person I spoke with previously 390 00:22:36,240 --> 00:22:39,240 Speaker 3: to us was for their last generation was we should 391 00:22:39,240 --> 00:22:41,760 Speaker 3: be wrapping these with money when we saw them, because 392 00:22:42,000 --> 00:22:46,080 Speaker 3: they were just they were basically subsidizing the buildout of 393 00:22:46,119 --> 00:22:51,119 Speaker 3: this new division, this GPU division, by undercutting prices from 394 00:22:51,200 --> 00:22:53,000 Speaker 3: competitors to just try and get their foot in the 395 00:22:53,040 --> 00:22:56,160 Speaker 3: door with what at the time was a lesser product. 396 00:22:56,680 --> 00:23:00,000 Speaker 3: And so yeah, I think they're a good broader because 397 00:23:01,440 --> 00:23:04,360 Speaker 3: if you want to be first of all, to have 398 00:23:04,600 --> 00:23:09,360 Speaker 3: a fab a fabrication facility costs these days potentially tens 399 00:23:09,400 --> 00:23:12,040 Speaker 3: of billions of dollars in years to build so a 400 00:23:12,080 --> 00:23:16,159 Speaker 3: new vendor would have to be fabless like Nvidia and 401 00:23:16,320 --> 00:23:22,240 Speaker 3: like AMD, unlike Intel, and to be fabulous and make a. 402 00:23:22,840 --> 00:23:25,000 Speaker 2: Sorry about this, this is just for the listeners. What 403 00:23:25,160 --> 00:23:26,439 Speaker 2: is a fab in this case? 404 00:23:26,720 --> 00:23:30,080 Speaker 3: Yeah, so a fab a fabrication plant is basically a 405 00:23:30,080 --> 00:23:35,160 Speaker 3: factory for silicon. It's a highly controlled, clean room environment, 406 00:23:36,119 --> 00:23:40,240 Speaker 3: and that means that there's basically absolutely no dust ingress, 407 00:23:41,680 --> 00:23:45,960 Speaker 3: the small pieces of skin or whatever, you know, just 408 00:23:46,000 --> 00:23:48,920 Speaker 3: working in an environment that's all either captured or controlled. 409 00:23:49,320 --> 00:23:51,919 Speaker 3: And that's because a single particle of dust on a 410 00:23:51,920 --> 00:23:55,840 Speaker 3: wafer can destroy at least part of that silicon supply. 411 00:23:56,320 --> 00:24:00,639 Speaker 3: Cool sorry, Yeah, And so a new vent would have 412 00:24:00,760 --> 00:24:04,639 Speaker 3: to be probably fabulous unless they're already huge like Intel, 413 00:24:05,520 --> 00:24:07,840 Speaker 3: and that means they need to go get supply from 414 00:24:08,600 --> 00:24:14,119 Speaker 3: most likely TSMC, possibly from Intel maybe. I mean the 415 00:24:14,160 --> 00:24:16,400 Speaker 3: past Samsung has been used. There's really not a lot 416 00:24:16,400 --> 00:24:20,560 Speaker 3: of options here, and everybody wants their silicon, so to 417 00:24:20,600 --> 00:24:24,240 Speaker 3: even get allocation of these silicon wafers would be difficult. 418 00:24:25,440 --> 00:24:29,080 Speaker 3: It's just all of the chips are stacked against Someone 419 00:24:29,160 --> 00:24:31,000 Speaker 3: wants to do this, and they would. I think the 420 00:24:31,040 --> 00:24:33,640 Speaker 3: only way you get in is if you're an existing 421 00:24:33,960 --> 00:24:38,720 Speaker 3: multi billion dollar company or have obscene amounts of investment. 422 00:24:39,320 --> 00:24:43,120 Speaker 2: Well that fucking sucks. Yeah, it's actually that's a good 423 00:24:43,200 --> 00:24:46,000 Speaker 2: question though. So this is for the old time is listening. 424 00:24:46,480 --> 00:24:50,000 Speaker 2: I remember having a three FX it's a Voodoo fifty 425 00:24:50,040 --> 00:24:52,760 Speaker 2: five hundred picture of a halo on the on the box. 426 00:24:52,800 --> 00:24:55,879 Speaker 2: Actually don't know what happened there. What happened to all 427 00:24:55,920 --> 00:24:59,280 Speaker 2: the smaller manufacturers. Did they just get kind of absorbed 428 00:24:59,600 --> 00:25:00,919 Speaker 2: or did they just full of punt? 429 00:25:01,240 --> 00:25:06,080 Speaker 3: Yeah, they got bought or yeah, I think three dfxx, 430 00:25:06,080 --> 00:25:12,080 Speaker 3: if I remember correctly, was bought by Nvidia and checked that. 431 00:25:12,200 --> 00:25:16,359 Speaker 3: But some of the others, I mean there's like Via 432 00:25:17,760 --> 00:25:22,240 Speaker 3: Cyrix and some of these CPU companies. Yeah, three dfcs 433 00:25:22,280 --> 00:25:26,240 Speaker 3: bought by Nvidia. SGI was another, so they they. 434 00:25:25,920 --> 00:25:27,199 Speaker 2: Yeah, I forgot about SGI. 435 00:25:27,520 --> 00:25:31,920 Speaker 3: Yeah, So SGI is interesting because they sort of pursued heavily. 436 00:25:32,800 --> 00:25:35,960 Speaker 3: I guess the at the time what enterprise would have been, 437 00:25:36,160 --> 00:25:41,040 Speaker 3: or at least workstation hardware, and they're really invested in workstation. 438 00:25:41,720 --> 00:25:44,480 Speaker 3: The story I got from someone I wasn't covering things 439 00:25:44,520 --> 00:25:48,600 Speaker 3: back then, but was that SGI was less interested in 440 00:25:48,680 --> 00:25:51,719 Speaker 3: consumer and didn't really think consumers could afford video cards. 441 00:25:51,760 --> 00:25:53,600 Speaker 3: Which at the time was sort of true, and that's 442 00:25:53,640 --> 00:25:58,720 Speaker 3: when video came in and you know, started really kicking 443 00:25:58,760 --> 00:26:01,080 Speaker 3: ass in the early days. But a lot of them 444 00:26:01,160 --> 00:26:04,520 Speaker 3: just they either got bought or they sort of exited 445 00:26:04,520 --> 00:26:05,040 Speaker 3: that space. 446 00:26:06,000 --> 00:26:09,040 Speaker 2: So changing text slightly, how do you feel about AI 447 00:26:09,160 --> 00:26:11,960 Speaker 2: in general? I know that I'm not expecting like an 448 00:26:12,000 --> 00:26:15,000 Speaker 2: industry and that I'm just from your perspective, at least 449 00:26:15,000 --> 00:26:17,040 Speaker 2: from what you've told me and watching your videos, it 450 00:26:17,040 --> 00:26:19,800 Speaker 2: feels like AI is something now being stapled onto the 451 00:26:19,840 --> 00:26:23,000 Speaker 2: side of literally everything. Yeah, within your world. Do you 452 00:26:23,000 --> 00:26:25,399 Speaker 2: think there's like do you use it? Do you have 453 00:26:25,440 --> 00:26:26,400 Speaker 2: any thoughts on it? 454 00:26:27,040 --> 00:26:34,719 Speaker 3: There's uses? I mean one thing I am I'm actually 455 00:26:34,720 --> 00:26:39,080 Speaker 3: worried about long term with the AI use cases. And 456 00:26:39,320 --> 00:26:44,520 Speaker 3: the biggest thing I'm worried about is on the media side. 457 00:26:44,560 --> 00:26:47,399 Speaker 3: We obviously get a lot of comments. I probably read 458 00:26:47,680 --> 00:26:49,679 Speaker 3: or at least skim, you know, tens of thousands of 459 00:26:49,680 --> 00:26:54,160 Speaker 3: comments a year, and I'm seeing more and more AI 460 00:26:54,280 --> 00:26:58,359 Speaker 3: comments and not just the really obvious YouTube spambots, but 461 00:26:58,520 --> 00:27:02,160 Speaker 3: somewhere you kind of you and you're like that could 462 00:27:02,200 --> 00:27:06,040 Speaker 3: be real, you know, and I don't like, Yeah, I 463 00:27:06,080 --> 00:27:08,479 Speaker 3: don't want to like remove it. If it's a valid, 464 00:27:08,560 --> 00:27:12,080 Speaker 3: real comment, but then a few comments down the thread 465 00:27:12,359 --> 00:27:15,440 Speaker 3: there will be this discussion amongst what are obvious sort 466 00:27:15,440 --> 00:27:22,679 Speaker 3: of bots about say some new cryptobs or whatever. And 467 00:27:22,720 --> 00:27:25,960 Speaker 3: so I'm actually worried about sort of like that Internet theory. 468 00:27:26,000 --> 00:27:28,919 Speaker 3: I guess, the concept of you're the only human in 469 00:27:28,960 --> 00:27:32,520 Speaker 3: the room and you might not even know it. That's 470 00:27:32,680 --> 00:27:36,560 Speaker 3: concerning to me. There's a loneliness factor there I'm worried about. 471 00:27:37,440 --> 00:27:41,160 Speaker 3: But the big one is just how easily people are 472 00:27:42,119 --> 00:27:47,240 Speaker 3: manipulated by comments that sort of appear to have the 473 00:27:47,320 --> 00:27:51,679 Speaker 3: prevailing opinion just in general, even from real comments. You 474 00:27:51,720 --> 00:27:55,560 Speaker 3: then take that and you apply it to potential AI 475 00:27:55,760 --> 00:28:02,400 Speaker 3: that becomes very convincing in the future. And you then 476 00:28:02,480 --> 00:28:07,000 Speaker 3: think one step further of who owns these lms and 477 00:28:07,400 --> 00:28:10,600 Speaker 3: the companies that sell so called AI products, and all 478 00:28:10,640 --> 00:28:14,440 Speaker 3: of them are these multi billion dollar companies that if 479 00:28:14,480 --> 00:28:18,040 Speaker 3: they want to, could tune it to filter the responses 480 00:28:18,600 --> 00:28:20,879 Speaker 3: in ways that are beneficial to them. And that to 481 00:28:20,960 --> 00:28:25,200 Speaker 3: me is sort of the dystopian old manuals at clouds, 482 00:28:25,240 --> 00:28:42,680 Speaker 3: you know, like concern I have, but it. 483 00:28:42,600 --> 00:28:44,959 Speaker 2: Feels like a more practical one because everyone freaks out 484 00:28:45,000 --> 00:28:47,800 Speaker 2: about AGI, which is fictional and all of these other things. 485 00:28:47,840 --> 00:28:51,000 Speaker 2: But yeah, the idea that comments would be filled with 486 00:28:51,120 --> 00:28:54,800 Speaker 2: people who were just subtly being nudged in different directions. 487 00:28:55,120 --> 00:28:57,880 Speaker 2: And when you talk about corporate narratives as well, who 488 00:28:57,920 --> 00:29:01,200 Speaker 2: knows that. I don't usually with lms in particular, try 489 00:29:01,240 --> 00:29:04,040 Speaker 2: not to lean into anything too conspiratal. But like, this 490 00:29:04,160 --> 00:29:06,560 Speaker 2: is a simple one that you're kind of seeing signs of. 491 00:29:06,680 --> 00:29:09,320 Speaker 2: Even on Blue Sky, I will occasionally get a comment 492 00:29:09,360 --> 00:29:11,520 Speaker 2: from someone it will take me half a second second 493 00:29:11,520 --> 00:29:14,080 Speaker 2: to go, like, what's wrong with you? This doesn't feel 494 00:29:14,120 --> 00:29:17,120 Speaker 2: like a human being saying this. No one speaks like this, 495 00:29:17,480 --> 00:29:20,880 Speaker 2: but you sound human adjacent, right, And I'm and at 496 00:29:20,920 --> 00:29:23,680 Speaker 2: your scale, I imagine that becomes a lot harder to 497 00:29:24,240 --> 00:29:26,360 Speaker 2: moderate slash process. 498 00:29:27,240 --> 00:29:31,080 Speaker 3: Yeah, and it really is just I mean, I'm highly 499 00:29:31,360 --> 00:29:33,600 Speaker 3: my job is to be skeptical, and you try to 500 00:29:33,640 --> 00:29:35,480 Speaker 3: do that in a healthy way. You know, it's easy 501 00:29:35,480 --> 00:29:39,840 Speaker 3: to go way overboard. But this is something where I 502 00:29:40,400 --> 00:29:45,760 Speaker 3: just I think there's so much opportunity to be abused. 503 00:29:45,800 --> 00:29:49,160 Speaker 3: And I was actually looking at there's that meta story 504 00:29:49,160 --> 00:29:54,000 Speaker 3: about meta and Facebook pirate in all of these you know, 505 00:29:54,080 --> 00:29:57,360 Speaker 3: books or taking books from pirated materials and integrating them, 506 00:29:57,360 --> 00:30:00,120 Speaker 3: and there's a lawsuit over it. One of the things 507 00:30:00,160 --> 00:30:01,960 Speaker 3: I was thinking of though, and actually speaking to a 508 00:30:02,040 --> 00:30:06,360 Speaker 3: copyright attorney about, was it really feels like right now, 509 00:30:08,880 --> 00:30:14,200 Speaker 3: the companies that are willing to break the most laws, 510 00:30:14,360 --> 00:30:19,280 Speaker 3: especially intellectual property laws, and steal the most stuff will 511 00:30:19,680 --> 00:30:22,080 Speaker 3: make so much money that it's like, you know what, 512 00:30:22,760 --> 00:30:25,240 Speaker 3: let's just we'll just deal with the consequences in five 513 00:30:25,320 --> 00:30:27,640 Speaker 3: years when it gets through the court, and we'll be 514 00:30:27,680 --> 00:30:30,720 Speaker 3: so far out ahead of anyone who didn't lie, cheat 515 00:30:30,760 --> 00:30:32,960 Speaker 3: and steel, you know, to get here. Yeah, that it 516 00:30:32,960 --> 00:30:37,880 Speaker 3: won't matter, which is just how the big corporations broadly behave. 517 00:30:38,040 --> 00:30:43,600 Speaker 3: But applying that to a new area like AI, I 518 00:30:43,600 --> 00:30:46,000 Speaker 3: don't know. It's I'm really concerned about it. I think 519 00:30:46,240 --> 00:30:52,720 Speaker 3: there's absolutely good use cases. And I mean, as an example, internally, 520 00:30:52,760 --> 00:30:57,120 Speaker 3: we've used it to concept things, so like I'll have 521 00:30:57,200 --> 00:31:00,880 Speaker 3: an idea in my head for art for a product, 522 00:31:01,360 --> 00:31:03,520 Speaker 3: and I can see it, but I'm a terrible artist, 523 00:31:04,120 --> 00:31:06,800 Speaker 3: and so there's a use case there where I'll sit 524 00:31:06,880 --> 00:31:08,719 Speaker 3: down with my artist and try to explain it as 525 00:31:08,720 --> 00:31:11,280 Speaker 3: best I can, and he'll take a go at it, 526 00:31:11,760 --> 00:31:13,520 Speaker 3: and if I'm like, no, it doesn't really like, look 527 00:31:13,640 --> 00:31:15,680 Speaker 3: how it looks in my head. That's when I might 528 00:31:15,960 --> 00:31:18,280 Speaker 3: try to pull some mid journey thing like okay, here's 529 00:31:18,520 --> 00:31:22,240 Speaker 3: like broadly like the style artistically I'm trying to get, 530 00:31:22,640 --> 00:31:24,280 Speaker 3: you know, and then he can take that and actually 531 00:31:24,360 --> 00:31:24,960 Speaker 3: make something. 532 00:31:25,640 --> 00:31:27,560 Speaker 2: But just to be clear, you would never use any 533 00:31:27,600 --> 00:31:30,479 Speaker 2: of that stuff in the real It is literally just 534 00:31:30,560 --> 00:31:34,360 Speaker 2: a concept tool, correctly, just making sure the listeners have 535 00:31:34,440 --> 00:31:35,160 Speaker 2: a good view on this. 536 00:31:35,520 --> 00:31:38,200 Speaker 3: Yeah, correct the artist so Andrew has been working with 537 00:31:38,200 --> 00:31:41,440 Speaker 3: me forever. He makes it all from scratch, like can 538 00:31:41,440 --> 00:31:43,680 Speaker 3: blender normally and actually flattens it into a two D 539 00:31:43,760 --> 00:31:46,320 Speaker 3: image for the shirts and things cool. So yeah, it 540 00:31:46,360 --> 00:31:49,760 Speaker 3: is purely used. If if in the first conversation he 541 00:31:49,800 --> 00:31:51,920 Speaker 3: can't like quite figure out what I'm trying to explain 542 00:31:51,960 --> 00:31:53,680 Speaker 3: as a non artist, you know, it's like, okay, let 543 00:31:53,680 --> 00:31:56,680 Speaker 3: me go try and figure out, like I don't I 544 00:31:56,680 --> 00:31:58,480 Speaker 3: don't know art words, you know, I don't. 545 00:31:58,280 --> 00:32:02,160 Speaker 2: Know no, no, I know. Yeah, so yeah, yeah, how 546 00:32:02,160 --> 00:32:03,920 Speaker 2: do you deal with the ethical side of that, because 547 00:32:03,920 --> 00:32:07,920 Speaker 2: it's just it is still working on copyright material. But 548 00:32:07,920 --> 00:32:09,640 Speaker 2: I guess if you're not putting it out there, it's 549 00:32:09,720 --> 00:32:12,200 Speaker 2: just like this this shit sucks even like looking at 550 00:32:12,240 --> 00:32:13,000 Speaker 2: it feels bad. 551 00:32:13,240 --> 00:32:15,760 Speaker 3: Yeah, I mean it's hard because it's the first step 552 00:32:15,840 --> 00:32:17,760 Speaker 3: for us. So it's always like a Google search to 553 00:32:17,760 --> 00:32:21,240 Speaker 3: see what's out there already, and so we'll do a 554 00:32:21,240 --> 00:32:23,560 Speaker 3: Google search if I'm trying to explain something I want, 555 00:32:24,360 --> 00:32:29,800 Speaker 3: and normally that's pretty successful. But like now it's the same. 556 00:32:29,960 --> 00:32:34,600 Speaker 3: It's like you're getting all this aiart surface in and 557 00:32:34,680 --> 00:32:37,120 Speaker 3: so at some point it's sort of like the downloading 558 00:32:37,200 --> 00:32:41,640 Speaker 3: and uploading and downloading, uploading a PMNG multiple times, it 559 00:32:41,680 --> 00:32:45,200 Speaker 3: feels like it's just getting like the quality is getting 560 00:32:45,240 --> 00:32:46,440 Speaker 3: lost somewhere in that chain. 561 00:32:47,880 --> 00:32:50,120 Speaker 2: It is funny. We're also like three years into it, 562 00:32:50,160 --> 00:32:52,320 Speaker 2: and the best we've got as far as like from 563 00:32:52,360 --> 00:32:55,400 Speaker 2: you arguably one of the most. Like I don't mean this, 564 00:32:55,680 --> 00:32:57,400 Speaker 2: I mean this actually is in a good way, Like 565 00:32:57,520 --> 00:33:00,360 Speaker 2: you would be excited about something cool in tech coming out, 566 00:33:01,040 --> 00:33:03,280 Speaker 2: like if something cool happened, And the coolest we've got 567 00:33:03,400 --> 00:33:05,920 Speaker 2: is Yes, sometimes I can't work out what I'm thinking, 568 00:33:06,000 --> 00:33:08,320 Speaker 2: so I just make one image and hopefully my artist 569 00:33:08,320 --> 00:33:10,520 Speaker 2: gets in. But yeah, yeah, it's just like that with 570 00:33:10,640 --> 00:33:12,240 Speaker 2: three years in and that's the best we've got. 571 00:33:12,600 --> 00:33:18,640 Speaker 3: Yeah, And I think we early early on with GPT 572 00:33:20,240 --> 00:33:22,680 Speaker 3: when we ran one of the first news stories that 573 00:33:22,680 --> 00:33:24,560 Speaker 3: we ran about it, and I was trying to understand, like, 574 00:33:24,600 --> 00:33:29,200 Speaker 3: what is this thing. I remember feeding it a spreadsheet 575 00:33:29,320 --> 00:33:32,640 Speaker 3: I had made with our own data and asking it 576 00:33:32,680 --> 00:33:38,200 Speaker 3: to analyze the results, and it was just like and 577 00:33:38,240 --> 00:33:40,080 Speaker 3: I keep in mind this was early, but it had 578 00:33:40,160 --> 00:33:46,720 Speaker 3: hallucinated so much that I was just I was disillusioned. 579 00:33:46,720 --> 00:33:49,400 Speaker 3: It was like, Okay, this is literally just made everything up. 580 00:33:49,840 --> 00:33:51,160 Speaker 2: And that was kind of how I was at the 581 00:33:51,200 --> 00:33:54,320 Speaker 2: beginning as well. Because I love my dad's magaizmos. I 582 00:33:54,400 --> 00:33:56,520 Speaker 2: was kind of like, Okay, everyone's talking about this, and 583 00:33:56,560 --> 00:33:58,880 Speaker 2: so I tried to do all of the things. They talked, Oh, 584 00:33:59,160 --> 00:34:02,360 Speaker 2: it could automate work. I'm a spreadsheet heavy guy and 585 00:34:02,360 --> 00:34:04,800 Speaker 2: this crap can't even do that. The funniest one for 586 00:34:04,880 --> 00:34:08,040 Speaker 2: me was that you have a Manus Manus You heard 587 00:34:08,040 --> 00:34:10,640 Speaker 2: of this one now? They claimed to be an AI 588 00:34:10,760 --> 00:34:13,360 Speaker 2: agent company. Okay, and I asked that can you go 589 00:34:13,440 --> 00:34:16,279 Speaker 2: and find all all of the links that mentioned me 590 00:34:16,320 --> 00:34:18,560 Speaker 2: for the last two Years's probably over one hundred of them, 591 00:34:18,760 --> 00:34:22,760 Speaker 2: and it takes eleven minutes or so, and it's writing 592 00:34:22,840 --> 00:34:25,920 Speaker 2: Python for every single step and it comes back with 593 00:34:26,000 --> 00:34:28,960 Speaker 2: eleven links or like nine links. I go, hey, you 594 00:34:28,960 --> 00:34:31,480 Speaker 2: miss some, and it comes back with nine more after 595 00:34:31,560 --> 00:34:34,480 Speaker 2: like another ten minutes. And this is this was about 596 00:34:34,480 --> 00:34:37,120 Speaker 2: a month ago. It's just what it feels like. They're 597 00:34:37,120 --> 00:34:41,799 Speaker 2: pushing them upper hill. It's but back to video for 598 00:34:41,840 --> 00:34:44,080 Speaker 2: a second because this will be the last video one. 599 00:34:44,120 --> 00:34:47,640 Speaker 2: I think you have done a really good job of 600 00:34:47,880 --> 00:34:51,360 Speaker 2: kind of tracking in video as they've jumped towards these movements. 601 00:34:51,800 --> 00:34:54,239 Speaker 2: But they've done this, but they did. How big was 602 00:34:54,239 --> 00:34:57,240 Speaker 2: their move towards Crypto That's actually something I've had trouble 603 00:34:57,320 --> 00:35:01,400 Speaker 2: encapsulating before. Did they really shoot the company that aggressively 604 00:35:01,440 --> 00:35:02,080 Speaker 2: in that direction? 605 00:35:03,280 --> 00:35:08,200 Speaker 3: That's a good question. So from and videos relationship with 606 00:35:08,239 --> 00:35:10,560 Speaker 3: crypto is weird. I remember at one point there was 607 00:35:10,600 --> 00:35:13,719 Speaker 3: some kind of earnings call or something where it was 608 00:35:13,800 --> 00:35:18,080 Speaker 3: during one of the GPU mining sort of apocalypses where 609 00:35:18,080 --> 00:35:20,200 Speaker 3: consumers couldn't get cards because they were all being bought 610 00:35:20,239 --> 00:35:26,919 Speaker 3: by crypto mining. And I remember doing a report on 611 00:35:27,200 --> 00:35:31,480 Speaker 3: in Hardware News on the earnings where it seemed like 612 00:35:31,600 --> 00:35:34,719 Speaker 3: the amount of revenue for mining was sort of obfuscated 613 00:35:34,760 --> 00:35:38,560 Speaker 3: and it was a challenge where the devices are shared 614 00:35:38,600 --> 00:35:41,320 Speaker 3: with gaming and so okay, I could see how it 615 00:35:41,360 --> 00:35:45,799 Speaker 3: would be difficult to decouple these in a revenue chart 616 00:35:45,920 --> 00:35:49,320 Speaker 3: if you sell gaming devices and they're used for multiple things, right, 617 00:35:49,719 --> 00:35:51,360 Speaker 3: so you know, it's like I kind of get it. 618 00:35:51,400 --> 00:35:55,480 Speaker 3: But at the same time, they weren't like hiding from 619 00:35:55,600 --> 00:35:58,719 Speaker 3: making sales to crypto mining. And then later on and 620 00:35:58,840 --> 00:36:01,840 Speaker 3: Video released these l R low hash rate cards. I 621 00:36:01,880 --> 00:36:05,760 Speaker 3: think it was the thirty series, and that was done 622 00:36:05,960 --> 00:36:08,759 Speaker 3: to try and get more card while they say to 623 00:36:08,880 --> 00:36:11,359 Speaker 3: try and get more cards to gamers. Now there were 624 00:36:11,360 --> 00:36:16,040 Speaker 3: bypasses for that, which is a separate story, but it 625 00:36:16,120 --> 00:36:20,319 Speaker 3: was done at a time when the thirty series. It 626 00:36:20,360 --> 00:36:23,680 Speaker 3: was really unfortunate covering it because like the hardware was 627 00:36:23,719 --> 00:36:27,000 Speaker 3: pretty good, the prices at the beginning were good, but 628 00:36:27,040 --> 00:36:31,800 Speaker 3: there's no supply to meet the demand, which was partly 629 00:36:31,800 --> 00:36:34,839 Speaker 3: because of COVID shutdowns where people were building new computers 630 00:36:35,000 --> 00:36:39,000 Speaker 3: for home, and also partly because of mining, and so 631 00:36:39,080 --> 00:36:41,359 Speaker 3: they tried a few things like these low hash rate 632 00:36:41,400 --> 00:36:47,160 Speaker 3: cards to reduce the viability for mining operations. But yeah, 633 00:36:47,160 --> 00:36:49,560 Speaker 3: I mean it definitely they've kind of that pendulum has 634 00:36:49,600 --> 00:36:51,319 Speaker 3: swung back and forth. I think at end Video where 635 00:36:51,280 --> 00:36:52,400 Speaker 3: it's like, Okay, we sell a lot of these and 636 00:36:52,440 --> 00:36:55,000 Speaker 3: make a lot of money to people are really pissed off, 637 00:36:55,040 --> 00:36:57,680 Speaker 3: you know, and we can't get them to consumer users 638 00:36:57,719 --> 00:37:00,080 Speaker 3: who might actually be repeat customers in the future, and 639 00:37:00,360 --> 00:37:04,040 Speaker 3: so let's try and restrict it. So it's definitely don't 640 00:37:04,200 --> 00:37:06,279 Speaker 3: I mean, it's not unexpected behavior, I guess for a 641 00:37:06,280 --> 00:37:10,080 Speaker 3: big corporation where it's just where's the money and if 642 00:37:10,080 --> 00:37:12,360 Speaker 3: the pendulum swings back the other way, you know, we 643 00:37:12,440 --> 00:37:16,600 Speaker 3: need to try and minimize how much both of these 644 00:37:18,160 --> 00:37:22,919 Speaker 3: potential client groups hate us for catering to the other one. 645 00:37:23,080 --> 00:37:28,080 Speaker 2: So so final GPU related question, actually, and forgive me 646 00:37:28,080 --> 00:37:30,080 Speaker 2: if you don't know the answer for this exactly, but 647 00:37:30,800 --> 00:37:33,239 Speaker 2: assuming the AI bubble burst and you don't have to 648 00:37:33,239 --> 00:37:34,719 Speaker 2: commit to whether you think that will happen or not, 649 00:37:35,080 --> 00:37:39,480 Speaker 2: what else can these big enterprise level GPUs actually be 650 00:37:39,600 --> 00:37:39,920 Speaker 2: used for. 651 00:37:40,960 --> 00:37:42,440 Speaker 3: They could be used for a lot. I mean, like 652 00:37:42,800 --> 00:37:48,680 Speaker 3: some of the we're talking to a professor at a 653 00:37:48,760 --> 00:37:54,600 Speaker 3: university for an upcoming piece, and they have some very 654 00:37:54,640 --> 00:37:58,359 Speaker 3: serious like machine learning research use cases, and I've looked 655 00:37:58,400 --> 00:38:00,240 Speaker 3: into some of the other ones, and as you might expect, 656 00:38:00,280 --> 00:38:04,440 Speaker 3: there's really promising use cases in medical and pattern recognition 657 00:38:04,640 --> 00:38:11,760 Speaker 3: and like so called AI being able to do early 658 00:38:11,800 --> 00:38:17,080 Speaker 3: identification of potentially cancerous masses and people that you know. 659 00:38:17,880 --> 00:38:19,520 Speaker 3: It's one of those like if you treat it like 660 00:38:19,600 --> 00:38:24,880 Speaker 3: a tool rather than like the answer. Then I suppose 661 00:38:25,400 --> 00:38:28,440 Speaker 3: in this example, as someone who knows nothing about, you know, 662 00:38:28,520 --> 00:38:32,799 Speaker 3: the medical world, but I suppose a skilled doctor might 663 00:38:32,840 --> 00:38:38,160 Speaker 3: be able to use it to help the process. I 664 00:38:38,239 --> 00:38:42,240 Speaker 3: think the concern I would have as as a skeptic 665 00:38:42,320 --> 00:38:46,680 Speaker 3: would be okay, but can we keep that doctor in 666 00:38:46,719 --> 00:38:49,960 Speaker 3: the mindset of I'm using this to help catch my 667 00:38:50,080 --> 00:38:53,560 Speaker 3: mistakes or help see things I can't, rather than you know, 668 00:38:53,760 --> 00:38:55,560 Speaker 3: I'm overworked and I need to lean on this to 669 00:38:55,600 --> 00:38:58,800 Speaker 3: do my job. Yeah, that would be the concern. 670 00:38:58,920 --> 00:39:01,520 Speaker 2: I'm just my whole thing is just if they and 671 00:39:01,560 --> 00:39:04,080 Speaker 2: I believe they have. This is my personal opinion. I 672 00:39:04,120 --> 00:39:07,520 Speaker 2: think they've massively overbuilt and oversolved these cards, and I 673 00:39:07,560 --> 00:39:11,080 Speaker 2: it's sorry cards is inappropriate to refer to the enterprise 674 00:39:11,160 --> 00:39:14,920 Speaker 2: ones racks, I suppose, But it's what do they do 675 00:39:15,040 --> 00:39:17,640 Speaker 2: next is going to be the question eventually, And I 676 00:39:17,680 --> 00:39:19,959 Speaker 2: mean there's just very clearly not going to be enough 677 00:39:20,200 --> 00:39:22,520 Speaker 2: use cases for them. But I think that's going to 678 00:39:22,600 --> 00:39:25,720 Speaker 2: be really interesting. And now I've got two fun questions 679 00:39:25,760 --> 00:39:27,400 Speaker 2: for you. First of all, did you hear about this 680 00:39:27,480 --> 00:39:32,680 Speaker 2: thirty six hundred dollars keyboard? It's mad the Norbauer Seneca. Okay, 681 00:39:32,719 --> 00:39:34,600 Speaker 2: so I'm going to send you this later. So Nathon 682 00:39:34,719 --> 00:39:37,760 Speaker 2: Edward's over at the Verge reviewed this one, this thirty 683 00:39:37,800 --> 00:39:39,920 Speaker 2: six hundred dollars keyboard, which I need to send you 684 00:39:39,920 --> 00:39:43,040 Speaker 2: because I assume you would Seneca. 685 00:39:43,080 --> 00:39:43,920 Speaker 3: I see the review. 686 00:39:44,239 --> 00:39:48,920 Speaker 2: Yeah, keyboard, it's insane. Listeners, I will drop with everything 687 00:39:48,920 --> 00:39:50,520 Speaker 2: we've been talking about. I'll have links in there. But 688 00:39:50,560 --> 00:39:54,319 Speaker 2: this thing, it's like precision honed and if he gets 689 00:39:54,320 --> 00:39:56,120 Speaker 2: it wrong, he has to start again with it takes 690 00:39:56,200 --> 00:39:59,120 Speaker 2: him like a full day to do, like the things 691 00:39:59,120 --> 00:40:02,160 Speaker 2: that absorb in impact. So cool. 692 00:40:02,760 --> 00:40:07,560 Speaker 3: This is this reminds me of the I don't know, 693 00:40:07,680 --> 00:40:10,680 Speaker 3: probably almost twenty years ago. Now there is the super 694 00:40:10,719 --> 00:40:13,880 Speaker 3: expensive I think it was like an optimistic keyboard or something. 695 00:40:13,680 --> 00:40:16,560 Speaker 2: Like the opt or the dust keyboard where it had 696 00:40:16,600 --> 00:40:17,839 Speaker 2: the little screens on each key. 697 00:40:17,960 --> 00:40:19,919 Speaker 3: Yeah, it was screens for every key and back then, 698 00:40:20,040 --> 00:40:21,840 Speaker 3: you know that was that was before we were putting 699 00:40:21,880 --> 00:40:22,720 Speaker 3: screens on everything. 700 00:40:23,280 --> 00:40:26,239 Speaker 2: Uh huh, I won't. That's the thing. I love that 701 00:40:26,280 --> 00:40:28,600 Speaker 2: this exists because if you look at the Verge story 702 00:40:28,640 --> 00:40:32,080 Speaker 2: with Chell Lnk. It's just a guy who literally sits in. 703 00:40:32,080 --> 00:40:34,600 Speaker 2: This character just builds these things like an artesian. Now, 704 00:40:34,680 --> 00:40:37,520 Speaker 2: I wish we had more weird shit like this. This 705 00:40:37,600 --> 00:40:39,320 Speaker 2: is what I miss about the tech industry. 706 00:40:40,200 --> 00:40:43,239 Speaker 3: Yeah, the the weird stuff and the cool stuff and 707 00:40:43,840 --> 00:40:46,440 Speaker 3: things where that's what I like seeing that computext but 708 00:40:46,520 --> 00:40:50,000 Speaker 3: it's okay, I could this isn't a product yet, but 709 00:40:50,080 --> 00:40:52,280 Speaker 3: I could see it becoming a product, or there's pieces 710 00:40:52,320 --> 00:40:54,920 Speaker 3: of this that could become a you know, an affordable product. 711 00:40:55,280 --> 00:40:56,239 Speaker 3: That's the stuff I like. 712 00:40:56,920 --> 00:40:59,719 Speaker 2: Yeah, did you see anything really bizarre like that there? 713 00:41:00,560 --> 00:41:04,759 Speaker 3: Yeah, I mean the the well actually most recently this 714 00:41:04,880 --> 00:41:07,399 Speaker 3: wasn't computext. But the most sort of bizarre cool thing 715 00:41:07,520 --> 00:41:09,480 Speaker 3: I saw and worked with that I was really happy 716 00:41:09,480 --> 00:41:14,280 Speaker 3: about was the pre built PC made by a small 717 00:41:14,360 --> 00:41:18,480 Speaker 3: company called Cherry Tree. They make in computer cases and 718 00:41:19,160 --> 00:41:22,880 Speaker 3: it's just a computer inside of the husk of a 719 00:41:22,960 --> 00:41:23,680 Speaker 3: video card. 720 00:41:24,320 --> 00:41:27,400 Speaker 2: Okay, that's cool. Yeah, I love that. That's it. That 721 00:41:27,520 --> 00:41:30,080 Speaker 2: reminds me of the corse Air thousand D case, which 722 00:41:30,080 --> 00:41:32,560 Speaker 2: I actually have a home where they have you can 723 00:41:32,560 --> 00:41:35,960 Speaker 2: fit micro micro at X and the regular sized motherboard 724 00:41:36,000 --> 00:41:39,319 Speaker 2: in there. I love that stuff like. It's that's why 725 00:41:39,360 --> 00:41:42,000 Speaker 2: I that's the thing I love about tech, the weirly, 726 00:41:42,120 --> 00:41:45,160 Speaker 2: really dumb, kind of niche stuff. And actually, with that 727 00:41:45,200 --> 00:41:47,719 Speaker 2: in mind, what are you actually excited for as we 728 00:41:47,800 --> 00:41:50,080 Speaker 2: wrap up, Like what the thing's coming out soon that 729 00:41:50,120 --> 00:41:52,399 Speaker 2: you're kind of pumped up to see? 730 00:41:52,920 --> 00:41:56,160 Speaker 3: I mean, the casesn't cool in are really interesting to me. 731 00:41:56,440 --> 00:42:01,720 Speaker 3: I just think in the puter actually, CPUs have gotten 732 00:42:01,760 --> 00:42:06,560 Speaker 3: interesting too, So on the silicon side, AMD has done 733 00:42:06,600 --> 00:42:09,600 Speaker 3: really well to actually get into gear competitively over the 734 00:42:09,680 --> 00:42:14,640 Speaker 3: last several years. And these they call them X three 735 00:42:14,719 --> 00:42:18,080 Speaker 3: D CPUs, which have a bunch of extra cash on them. 736 00:42:18,800 --> 00:42:21,759 Speaker 3: They've been really good for gaming and that's a I 737 00:42:21,800 --> 00:42:28,000 Speaker 3: mean seeing the extra Yeah. So, because you can fit 738 00:42:28,160 --> 00:42:32,799 Speaker 3: more stuff local to the CPU, like like you can 739 00:42:32,840 --> 00:42:37,080 Speaker 3: store more of what the CPU needs to process within 740 00:42:37,200 --> 00:42:40,040 Speaker 3: the CPU silicon itself, it doesn't have to go out 741 00:42:40,080 --> 00:42:42,000 Speaker 3: to system memory. So the way it normally works is 742 00:42:42,280 --> 00:42:44,719 Speaker 3: like if you're playing a game and you need some 743 00:42:44,840 --> 00:42:49,240 Speaker 3: kind of file for drawing something in the game or whatever, 744 00:42:49,320 --> 00:42:51,520 Speaker 3: I mean, go into memory of some kind. So it 745 00:42:51,600 --> 00:42:53,799 Speaker 3: might be GPU memory, might be system memory. But if 746 00:42:53,800 --> 00:42:56,600 Speaker 3: you take the system memory example, something goes into system 747 00:42:56,640 --> 00:43:02,000 Speaker 3: memory you need that file, there'll be a transaction where 748 00:43:02,400 --> 00:43:04,400 Speaker 3: you can imagine these bits going down a bus to 749 00:43:04,440 --> 00:43:06,280 Speaker 3: the memory and coming all the way back to the CPU, 750 00:43:06,280 --> 00:43:09,920 Speaker 3: and that's a long highway to go down. Whereas with 751 00:43:09,960 --> 00:43:12,800 Speaker 3: extra cash you can store more local to the CPU, 752 00:43:12,840 --> 00:43:17,640 Speaker 3: it can transact via basically like a shortcut, and that 753 00:43:17,719 --> 00:43:23,080 Speaker 3: allows in the real world significantly higher performance in gaming 754 00:43:23,200 --> 00:43:24,920 Speaker 3: or in cash heavy applications. 755 00:43:25,640 --> 00:43:30,480 Speaker 2: And so the new AMD ones, yeah, it's. 756 00:43:30,320 --> 00:43:32,320 Speaker 3: The new like the ninety eight hundred x three D 757 00:43:32,440 --> 00:43:36,359 Speaker 3: for example. So those I'm really excited about that. How 758 00:43:36,400 --> 00:43:39,640 Speaker 3: that's gone because that actually you may like this, but 759 00:43:39,920 --> 00:43:42,800 Speaker 3: that was actually a sort of a skunkworks project where 760 00:43:43,040 --> 00:43:46,560 Speaker 3: this team of engineers in AMD and we ran a 761 00:43:46,600 --> 00:43:51,120 Speaker 3: story on this too, but they decided, let's just try 762 00:43:51,160 --> 00:43:55,800 Speaker 3: this idea, and they experimented with it, they got it working, 763 00:43:55,920 --> 00:43:58,719 Speaker 3: and they brought it through management and said, hey, this 764 00:43:58,760 --> 00:44:01,520 Speaker 3: looks viable and uh, you know, and then they were 765 00:44:01,520 --> 00:44:05,680 Speaker 3: able to actually make a product that it came from 766 00:44:06,160 --> 00:44:09,680 Speaker 3: an abnormal path of an underground sort of you know, 767 00:44:09,760 --> 00:44:12,719 Speaker 3: skunk works effort at the in the engineering department to 768 00:44:12,920 --> 00:44:16,760 Speaker 3: actually being the leading gaming product for CPUs on the market. 769 00:44:17,280 --> 00:44:19,840 Speaker 2: That's it I We will have that link in the 770 00:44:19,880 --> 00:44:23,480 Speaker 2: episode notes as well, and we can wrap on the product, 771 00:44:23,760 --> 00:44:27,360 Speaker 2: the AMD skunk works product. You mentioned in a recent 772 00:44:27,440 --> 00:44:30,359 Speaker 2: video that I had not AMD made a mountain bike 773 00:44:30,400 --> 00:44:32,759 Speaker 2: at one point, just want to one. 774 00:44:32,960 --> 00:44:36,600 Speaker 3: Yeah, so yeah, that was not a good moment for AMD. 775 00:44:37,080 --> 00:44:37,200 Speaker 2: Uh. 776 00:44:38,040 --> 00:44:43,000 Speaker 3: Fortunately, fortunately it didn't particularly damage their CPU or gp 777 00:44:43,080 --> 00:44:44,399 Speaker 3: A credibility because it was like so. 778 00:44:44,440 --> 00:44:47,319 Speaker 2: Long imagine so but why why did they do. 779 00:44:47,480 --> 00:44:51,120 Speaker 3: What they It was turning the COVID bike shortages. I 780 00:44:51,160 --> 00:44:53,960 Speaker 3: don't know if you remember, but like all kinds of bicycles, yeah, 781 00:44:54,000 --> 00:44:58,320 Speaker 3: were difficult to get. And uh, yes, they made these 782 00:44:58,440 --> 00:45:02,359 Speaker 3: three hundred dollars so called mountain bikes. I think they're 783 00:45:02,360 --> 00:45:07,560 Speaker 3: like Kent Walmart bike rebrands basically nice. So they went 784 00:45:07,640 --> 00:45:11,520 Speaker 3: real cheap, yeah, and uh yeah it was advertised, you 785 00:45:11,520 --> 00:45:14,400 Speaker 3: know for mountain biking and off road trails and whatever. 786 00:45:14,719 --> 00:45:16,719 Speaker 3: And so we took it down a trail and I 787 00:45:16,760 --> 00:45:19,680 Speaker 3: actually had trouble getting out of the parking lot. That's 788 00:45:19,680 --> 00:45:21,920 Speaker 3: how messed up the bike was. We had to take 789 00:45:21,960 --> 00:45:23,640 Speaker 3: it to a shop before I rode it for a 790 00:45:23,640 --> 00:45:27,080 Speaker 3: safety check because the the used V brakes and they 791 00:45:27,080 --> 00:45:30,120 Speaker 3: were going through the spokes. So they definitely would have 792 00:45:30,120 --> 00:45:30,520 Speaker 3: stopped me. 793 00:45:32,040 --> 00:45:34,080 Speaker 2: Jenson woe sitting there being like we could have had 794 00:45:34,120 --> 00:45:36,439 Speaker 2: a fucking mountain butt man. Yeah we could. 795 00:45:36,840 --> 00:45:38,960 Speaker 3: How do we get Yeah, if we make a bike, 796 00:45:39,000 --> 00:45:39,920 Speaker 3: we can get rid of them. 797 00:45:41,280 --> 00:45:45,279 Speaker 2: They're just like two twenty billion dollars shoved into mountain bike. 798 00:45:45,360 --> 00:45:45,640 Speaker 3: Resa. 799 00:45:45,960 --> 00:45:48,440 Speaker 2: That would be so good, Steve, It's always such a 800 00:45:48,440 --> 00:45:50,760 Speaker 2: pleasure having you on here. Where can people find. 801 00:45:50,560 --> 00:45:52,840 Speaker 3: You and gamers next us on YouTube? 802 00:45:53,400 --> 00:45:55,439 Speaker 2: Thank you so much for being here. I'm of course 803 00:45:55,480 --> 00:45:57,799 Speaker 2: that's itch when you've been listening to Better Offline, have 804 00:45:57,840 --> 00:45:59,719 Speaker 2: a great week, everyone, staves, thanks for coming on. 805 00:46:00,000 --> 00:46:10,160 Speaker 4: Thank you, Thank you for listening to Better Offline. 806 00:46:10,280 --> 00:46:12,719 Speaker 2: The editor and composer of the Better Offline theme song 807 00:46:12,800 --> 00:46:15,439 Speaker 2: is Mattersowski. You can check out more of his music 808 00:46:15,440 --> 00:46:19,120 Speaker 2: and audio projects at Mattasowski dot com, m A T 809 00:46:19,120 --> 00:46:23,560 Speaker 2: T O s O W s ki dot com. You 810 00:46:23,600 --> 00:46:26,120 Speaker 2: can email me at easy at Better offline dot com, 811 00:46:26,200 --> 00:46:28,480 Speaker 2: or visit Better Offline dot com to find more podcast 812 00:46:28,560 --> 00:46:31,880 Speaker 2: links and of course my newsletter. I also really recommend 813 00:46:31,920 --> 00:46:33,840 Speaker 2: you go to chat dot Where's youreed dot at to 814 00:46:33,920 --> 00:46:36,719 Speaker 2: visit the discord, and go to our slash Better Offline 815 00:46:36,760 --> 00:46:40,000 Speaker 2: to check out our Reddit, Thank you so much for listening. 816 00:46:40,840 --> 00:46:43,520 Speaker 1: Better Offline is a production of cool Zone Media. For 817 00:46:43,640 --> 00:46:46,800 Speaker 1: more from cool Zone Media, visit our website cool zonemedia 818 00:46:46,840 --> 00:46:49,719 Speaker 1: dot com, or check us out on the iHeartRadio app, 819 00:46:49,760 --> 00:47:09,960 Speaker 1: Apple Podcasts, or wherever you get your podcasts.