1 00:00:10,800 --> 00:00:14,920 Speaker 1: Hello, and welcome to another episode of the Odd Lots Podcast. 2 00:00:14,960 --> 00:00:20,040 Speaker 1: I'm Joe Wisnthal and I'm Tracy Halloway. Tracy, I had 3 00:00:20,079 --> 00:00:22,680 Speaker 1: a thought, and it might be a little controversial, but 4 00:00:23,320 --> 00:00:27,200 Speaker 1: what if um you with a controversial thought? Never No, 5 00:00:27,400 --> 00:00:29,160 Speaker 1: it's like, I know, I'm just gonna throw this out 6 00:00:29,200 --> 00:00:31,240 Speaker 1: there and tell me what you think. What if we 7 00:00:31,680 --> 00:00:35,960 Speaker 1: made that Odd Lots wizon about market and finance anymore, 8 00:00:36,000 --> 00:00:41,200 Speaker 1: but just about the semiconductor history? Could we maybe could 9 00:00:41,200 --> 00:00:45,480 Speaker 1: we maybe start with a semiconductor series before we completely 10 00:00:45,560 --> 00:00:49,920 Speaker 1: rebrand the podcast? All right, we'll we'll revisit the question 11 00:00:50,000 --> 00:00:52,440 Speaker 1: of a total rebrand. But I don't know about you, 12 00:00:52,479 --> 00:00:55,800 Speaker 1: but I found our episode a couple of weeks ago 13 00:00:56,360 --> 00:01:00,800 Speaker 1: with Stacy Raskin from Bernstein about the kind of Intel 14 00:01:00,840 --> 00:01:03,120 Speaker 1: to be fascinating, And after that, I was like, I 15 00:01:03,160 --> 00:01:05,440 Speaker 1: just wanted to talk more chips, Like I just that 16 00:01:05,640 --> 00:01:08,479 Speaker 1: like that really sort of like it really went my appetite. 17 00:01:09,040 --> 00:01:11,960 Speaker 1: It was definitely a lot more interesting than I had 18 00:01:12,000 --> 00:01:15,240 Speaker 1: been expecting. I think I mentioned when we recorded it 19 00:01:15,400 --> 00:01:18,440 Speaker 1: that I hadn't been following all the ins and outs 20 00:01:18,600 --> 00:01:22,640 Speaker 1: of the semiconductor industry, except of course, as it pertained 21 00:01:22,680 --> 00:01:26,399 Speaker 1: to the US China trade tension. So I wasn't aware 22 00:01:27,319 --> 00:01:31,200 Speaker 1: of some of the competition between the big makers, like 23 00:01:31,319 --> 00:01:34,399 Speaker 1: just how bad things had gotten for Intel, and I 24 00:01:34,440 --> 00:01:37,160 Speaker 1: wasn't aware of some of the strategic decisions that we're 25 00:01:37,200 --> 00:01:41,040 Speaker 1: going into the making of their chips and the way 26 00:01:41,080 --> 00:01:44,679 Speaker 1: that some other chip makers have proceeded in a very 27 00:01:44,920 --> 00:01:49,520 Speaker 1: uh different direction. Yeah, right, it's that different direction that's 28 00:01:49,560 --> 00:01:52,760 Speaker 1: interesting because, I mean, we talked with Stacy a lot 29 00:01:52,800 --> 00:01:56,880 Speaker 1: about the sort of manufacturing issues that Intel has been 30 00:01:56,880 --> 00:02:01,279 Speaker 1: facing as it tries to unveil it's seven animator chips. 31 00:02:01,680 --> 00:02:05,840 Speaker 1: But there's more to the story than just faster and 32 00:02:05,960 --> 00:02:11,080 Speaker 1: smaller chips. There's also a fundamental ship happening in what 33 00:02:11,200 --> 00:02:13,960 Speaker 1: chips are and how they're used and how they're designed. 34 00:02:14,320 --> 00:02:19,040 Speaker 1: It goes beyond just sort of manufacturing problems. Yeah, and 35 00:02:19,160 --> 00:02:22,680 Speaker 1: I guess the big news on that front is Apple 36 00:02:22,840 --> 00:02:27,120 Speaker 1: and its new series of Mac processors. Uh. That news 37 00:02:27,160 --> 00:02:29,480 Speaker 1: came out this month, and everyone, as far as I 38 00:02:29,480 --> 00:02:33,280 Speaker 1: can tell, all the technologists that have YouTube channels are 39 00:02:33,400 --> 00:02:37,640 Speaker 1: very very excited about what exactly this means. I don't 40 00:02:37,720 --> 00:02:41,800 Speaker 1: completely understand the technology, but I am interested in learning 41 00:02:41,800 --> 00:02:44,680 Speaker 1: why everyone else um is so excited and why a 42 00:02:44,680 --> 00:02:46,600 Speaker 1: lot of people are saying this is a game changer. 43 00:02:46,639 --> 00:02:49,280 Speaker 1: I heard one there's one guy on YouTube who was 44 00:02:49,320 --> 00:02:53,200 Speaker 1: talking about how this is going to make Apple computers 45 00:02:53,200 --> 00:02:57,119 Speaker 1: basically on a different level to other types of computers. 46 00:02:57,160 --> 00:02:59,680 Speaker 1: So in the way that we don't necessarily compare Apple 47 00:02:59,720 --> 00:03:03,800 Speaker 1: phone with androids anymore, we will no longer be comparing 48 00:03:04,200 --> 00:03:07,480 Speaker 1: Max with you know, your average PC. It's just like 49 00:03:07,760 --> 00:03:11,440 Speaker 1: two different types of equipment. Yeah, there was like a 50 00:03:11,520 --> 00:03:14,480 Speaker 1: medium article about the Apple chip then went viral on 51 00:03:14,560 --> 00:03:20,960 Speaker 1: like articles about chips don't typically go viral, but that 52 00:03:21,120 --> 00:03:23,839 Speaker 1: I think sort of gives you an indication of the enthusiasm, 53 00:03:23,960 --> 00:03:26,840 Speaker 1: the excitement that people have for all this stuff. So 54 00:03:27,040 --> 00:03:28,880 Speaker 1: definitely a lot to talk about. But I have to 55 00:03:28,880 --> 00:03:31,440 Speaker 1: admit I still don't like really get it. Like I'm 56 00:03:31,480 --> 00:03:33,440 Speaker 1: trying to read about this stuff and understand it, but 57 00:03:33,480 --> 00:03:37,560 Speaker 1: I don't totally get Yeah, I'm in that same place. 58 00:03:37,920 --> 00:03:41,600 Speaker 1: So today we are going to be talking more as 59 00:03:41,720 --> 00:03:44,760 Speaker 1: anyone could have guessed about chips. And I'm very excited 60 00:03:44,800 --> 00:03:48,040 Speaker 1: because in addition to being a sequel of sorts to 61 00:03:48,200 --> 00:03:52,240 Speaker 1: that other chip episode, um our guest who is who 62 00:03:52,280 --> 00:03:55,040 Speaker 1: writes on this stuff, and he's prominent on Twitter, etcetera. 63 00:03:55,320 --> 00:03:57,760 Speaker 1: We're going to do something we've never done before. He's 64 00:03:57,760 --> 00:04:01,000 Speaker 1: going to dos himself on this episod. So he's been 65 00:04:01,000 --> 00:04:05,200 Speaker 1: writing under and tweeting under a pseudonyms, agreed to come 66 00:04:05,240 --> 00:04:08,800 Speaker 1: on and with this episode actually tell us who he is. 67 00:04:08,840 --> 00:04:11,320 Speaker 1: So this is a a totally new thing for us. 68 00:04:11,320 --> 00:04:14,480 Speaker 1: I don't think we've done this before. No, we definitely haven't, 69 00:04:14,520 --> 00:04:18,279 Speaker 1: and we should just stress this is a voluntary self 70 00:04:18,520 --> 00:04:21,080 Speaker 1: docs scene. We haven't forced him into it or anything 71 00:04:21,120 --> 00:04:24,359 Speaker 1: like that. That's exactly right. So we're going to be 72 00:04:24,440 --> 00:04:29,000 Speaker 1: speaking with the author of the newsletter called Mules musing. 73 00:04:29,680 --> 00:04:32,920 Speaker 1: He tweets under the handle at full all the time, 74 00:04:33,520 --> 00:04:37,080 Speaker 1: great stuff on the matter of chips. But we're going 75 00:04:37,080 --> 00:04:39,520 Speaker 1: to find out now who he is. So U Mule, 76 00:04:40,560 --> 00:04:43,960 Speaker 1: thank you for joining the Outlaws podcast? Who are you 77 00:04:44,120 --> 00:04:48,360 Speaker 1: and why are we listening to? Hey? I'm My name 78 00:04:48,520 --> 00:04:51,960 Speaker 1: is Doug O Offlin, and I am just a huge 79 00:04:52,000 --> 00:04:56,200 Speaker 1: nerd with an investing background. I guess I previously worked 80 00:04:56,200 --> 00:04:59,240 Speaker 1: at a byeside firm in Dallas and kind of just 81 00:04:59,279 --> 00:05:02,560 Speaker 1: taking a break tween things started writing a newsletter. Um, 82 00:05:02,640 --> 00:05:04,920 Speaker 1: and clearly there's been a lot of interest in semi coonnucters, 83 00:05:04,960 --> 00:05:07,560 Speaker 1: not just from you guys, but from the investing community 84 00:05:07,560 --> 00:05:09,240 Speaker 1: at large, and it kind of just hit a note. 85 00:05:09,400 --> 00:05:13,040 Speaker 1: So I'm I'm really obsessed with the whole you know, 86 00:05:13,080 --> 00:05:15,599 Speaker 1: the whole revolution all the YouTubers are going crazy about. 87 00:05:15,920 --> 00:05:18,880 Speaker 1: I'm obsessed, Like, it's not just um, it's not just 88 00:05:18,920 --> 00:05:20,960 Speaker 1: all the technology stuff, but there's a lot of business 89 00:05:21,000 --> 00:05:24,400 Speaker 1: and investing um things that could happen, and I think 90 00:05:24,400 --> 00:05:26,720 Speaker 1: it's I think it's truly a revolution. Like it's it's 91 00:05:26,760 --> 00:05:29,120 Speaker 1: about as exciting as it's ever been in semiconductors, which 92 00:05:29,160 --> 00:05:32,520 Speaker 1: is frankly not that exciting, but it's still very exciting 93 00:05:32,560 --> 00:05:35,760 Speaker 1: to me. UM. So yeah, that's that's what I've been 94 00:05:35,800 --> 00:05:39,080 Speaker 1: doing recently. I have to say that while you were 95 00:05:39,400 --> 00:05:42,840 Speaker 1: introducing yourself, I just had a look at your Twitter 96 00:05:42,880 --> 00:05:46,040 Speaker 1: profile page and I saw the binary code on your 97 00:05:46,080 --> 00:05:50,919 Speaker 1: profile and I looked it up and you must be 98 00:05:50,960 --> 00:05:53,000 Speaker 1: the first All Thoughts guests that we've ever had who 99 00:05:53,120 --> 00:05:57,040 Speaker 1: who has basically done like a practical joke in binary code. 100 00:05:57,120 --> 00:05:59,719 Speaker 1: So um, kudos for that, and I guess I give 101 00:05:59,800 --> 00:06:03,840 Speaker 1: some indication of of your interests. H and everyone else 102 00:06:03,880 --> 00:06:06,160 Speaker 1: can go and check it out and translate it into 103 00:06:06,200 --> 00:06:09,120 Speaker 1: English to see what it actually says. But before we begin, 104 00:06:09,200 --> 00:06:13,280 Speaker 1: I'm curious, why did you decide to start the sub 105 00:06:13,320 --> 00:06:17,120 Speaker 1: stack and why were you tweeting under the pseudonym to 106 00:06:17,200 --> 00:06:22,680 Speaker 1: begin with? Well, um, especially in the past past, it's 107 00:06:22,800 --> 00:06:25,480 Speaker 1: it's not you know, I was employed, and it's um, 108 00:06:25,480 --> 00:06:29,200 Speaker 1: it's not you're not supposed to talk publicly about public 109 00:06:29,240 --> 00:06:31,919 Speaker 1: markets if you're in the business. So that was that 110 00:06:32,040 --> 00:06:35,120 Speaker 1: was the big reason for a pseudonym. And frankly, Finfoot 111 00:06:35,160 --> 00:06:37,800 Speaker 1: is an amazing community. There's a lot of um, there's 112 00:06:38,000 --> 00:06:41,240 Speaker 1: more information there than anywhere else, so it's it's an 113 00:06:41,279 --> 00:06:44,320 Speaker 1: awesome place to learn. But obviously you have to be 114 00:06:44,839 --> 00:06:48,480 Speaker 1: extremely private if youtuose to do so. So um that 115 00:06:48,480 --> 00:06:50,359 Speaker 1: that was the reason why I under a pseudonym, and 116 00:06:50,360 --> 00:06:53,520 Speaker 1: then the sub stack. I guess, um, I'm just kind 117 00:06:53,560 --> 00:06:56,440 Speaker 1: of like in between things. Next year, if everything goes right, 118 00:06:56,440 --> 00:06:58,800 Speaker 1: I'll take a little bit of a gap. But you know, 119 00:06:58,920 --> 00:07:03,960 Speaker 1: I'm like born to do or keep researching and doing 120 00:07:04,000 --> 00:07:06,159 Speaker 1: stuff like this, Like this was pretty much what I 121 00:07:06,200 --> 00:07:07,840 Speaker 1: was doing in my free time, and I was like, heck, 122 00:07:08,080 --> 00:07:09,960 Speaker 1: a lot of other people could really learn a lot 123 00:07:10,040 --> 00:07:12,120 Speaker 1: from this. As well, and so then I started writing 124 00:07:12,120 --> 00:07:15,520 Speaker 1: for a broader audience, and um, writing in general has 125 00:07:15,560 --> 00:07:18,680 Speaker 1: been really fruitful. Like you, it's it's a crazy skill 126 00:07:18,720 --> 00:07:21,040 Speaker 1: to have and I'm frankly not I'm getting a lot 127 00:07:21,080 --> 00:07:22,920 Speaker 1: better at it, but it's still um, it's been a 128 00:07:23,000 --> 00:07:25,280 Speaker 1: learning journey. So that's why I started the sub stack 129 00:07:25,440 --> 00:07:27,840 Speaker 1: and I've just been kind of writing the wave. I 130 00:07:27,920 --> 00:07:33,000 Speaker 1: have to say, I feel like a deep, emotional, resonant 131 00:07:33,160 --> 00:07:38,080 Speaker 1: connection with your story. Because I got my start I 132 00:07:38,120 --> 00:07:42,040 Speaker 1: worked for a small portfolio management company. I also started 133 00:07:42,080 --> 00:07:45,360 Speaker 1: I started blogging in two thousand five under a blog 134 00:07:45,480 --> 00:07:48,200 Speaker 1: called The Stalwart, which is now defunct, but that's why 135 00:07:48,200 --> 00:07:49,880 Speaker 1: I got my Twitter handle, and it was kind of 136 00:07:49,920 --> 00:07:53,160 Speaker 1: the same thing. I just wanted to practice writing for 137 00:07:53,320 --> 00:07:55,760 Speaker 1: the skill of writing and keep up to date, and 138 00:07:55,760 --> 00:07:57,800 Speaker 1: it was between jobs, so kind of like you. So 139 00:07:58,440 --> 00:08:00,760 Speaker 1: I'm I'm a big fan of people with your art 140 00:08:00,800 --> 00:08:03,320 Speaker 1: and I think big things. But let's let's move on 141 00:08:03,760 --> 00:08:08,200 Speaker 1: and let's talk chips. I mentioned there's this a viral 142 00:08:08,320 --> 00:08:12,120 Speaker 1: article about the Mac chip that came out, and articles 143 00:08:12,120 --> 00:08:15,640 Speaker 1: about chips don't particularly go viral, but everyone's raving about 144 00:08:15,680 --> 00:08:18,960 Speaker 1: this new chip from Apple, So why is the chip 145 00:08:19,040 --> 00:08:23,160 Speaker 1: going vibral okay, so then yeah, now and I can 146 00:08:23,200 --> 00:08:26,120 Speaker 1: talk ages about this. UM. So the the the M 147 00:08:26,160 --> 00:08:29,720 Speaker 1: one chip is really exciting, so UM. The reason why 148 00:08:29,960 --> 00:08:32,400 Speaker 1: is it's kind of it's kind of like a hybrid 149 00:08:32,520 --> 00:08:35,200 Speaker 1: and we've we've seen we've seen it before, frankly, like 150 00:08:35,240 --> 00:08:37,559 Speaker 1: your iPhone is powered by something that's similar to and 151 00:08:37,760 --> 00:08:39,880 Speaker 1: M one chip, but we've never seen it for a 152 00:08:39,960 --> 00:08:42,960 Speaker 1: Mac or a desktop or any other form factor other 153 00:08:43,000 --> 00:08:46,360 Speaker 1: than iPhone. And part of it is because up until 154 00:08:46,360 --> 00:08:48,440 Speaker 1: this point, all the laptops in the world have been 155 00:08:48,559 --> 00:08:50,920 Speaker 1: mostly running x A D six. So x A D 156 00:08:51,000 --> 00:08:53,760 Speaker 1: six isn't UM is like just think of it as 157 00:08:53,760 --> 00:08:57,600 Speaker 1: like a language of processors. And for a long time 158 00:08:57,600 --> 00:09:00,800 Speaker 1: that was kind of like, you know, the defen ACTO standard. 159 00:09:00,880 --> 00:09:03,360 Speaker 1: But the iPhone getting a lot better has pretty much 160 00:09:03,400 --> 00:09:05,680 Speaker 1: made a second language become a lot more popular, which 161 00:09:05,720 --> 00:09:09,480 Speaker 1: is ARM. And now that it's become a lot more mature, 162 00:09:09,679 --> 00:09:12,640 Speaker 1: it's actually kind of um. In the past, it never 163 00:09:12,760 --> 00:09:16,040 Speaker 1: was fast enough to really compete in a actual computer, 164 00:09:16,559 --> 00:09:19,439 Speaker 1: but now it seems to be that the M one 165 00:09:19,520 --> 00:09:23,520 Speaker 1: chip is like supermature. The ARM infrastructure is ready and 166 00:09:23,600 --> 00:09:27,080 Speaker 1: the laptop looks really compelling, and a big reason for 167 00:09:27,160 --> 00:09:29,880 Speaker 1: what that is is it's not really a CPU like 168 00:09:29,960 --> 00:09:32,720 Speaker 1: it's it's actually a lot of different dies on one 169 00:09:32,800 --> 00:09:35,440 Speaker 1: giant die, and it looks like there's a lot of 170 00:09:35,480 --> 00:09:38,480 Speaker 1: benefits of arm over x eight six that we didn't 171 00:09:38,480 --> 00:09:41,280 Speaker 1: even know where possible up until the M one came out. 172 00:09:41,360 --> 00:09:45,160 Speaker 1: So so, I mean, I think from what I understand, 173 00:09:45,200 --> 00:09:47,559 Speaker 1: this is kind of the heart of it. So normally 174 00:09:47,559 --> 00:09:50,520 Speaker 1: when we talk about chips, you know, certainly when we're 175 00:09:50,520 --> 00:09:53,280 Speaker 1: talking about chips made by Intel or m D or 176 00:09:53,320 --> 00:09:58,320 Speaker 1: something like that, we're talking about CPUs, these central processing units. 177 00:09:58,360 --> 00:10:02,720 Speaker 1: But the Apple hip, the M one is not a CPU. 178 00:10:02,800 --> 00:10:05,560 Speaker 1: It's it's something else, or only one component of it 179 00:10:05,600 --> 00:10:08,000 Speaker 1: is a CPU. Can you can you explain that in 180 00:10:08,000 --> 00:10:10,800 Speaker 1: a bit more detail. Yeah, So it's called a system 181 00:10:10,840 --> 00:10:13,720 Speaker 1: on chip or sock, and what that means is that 182 00:10:14,040 --> 00:10:17,440 Speaker 1: it's it has a traditional die like any other piece 183 00:10:17,480 --> 00:10:20,600 Speaker 1: of silicon, but there are little pieces on the silicon 184 00:10:20,720 --> 00:10:23,839 Speaker 1: that each have different jobs. And so instead of having 185 00:10:24,000 --> 00:10:27,520 Speaker 1: one giant piece of silicon that does everything kind of 186 00:10:27,520 --> 00:10:31,000 Speaker 1: well pretty much, this this silicon each has little parts 187 00:10:31,000 --> 00:10:34,199 Speaker 1: that do their job really well, and then together it 188 00:10:34,240 --> 00:10:36,680 Speaker 1: does a lot better as a system, because it's like, hey, 189 00:10:36,720 --> 00:10:39,079 Speaker 1: instead of having a ton of random you know, instead 190 00:10:39,120 --> 00:10:41,560 Speaker 1: of having a general computer that can do everything kind 191 00:10:41,600 --> 00:10:44,079 Speaker 1: of well, they have like, you know, ten specials in 192 00:10:44,120 --> 00:10:46,560 Speaker 1: the house. And when you look at a laptop, really 193 00:10:46,559 --> 00:10:49,840 Speaker 1: you're only doing ten types of workloads or or however 194 00:10:49,880 --> 00:10:52,600 Speaker 1: many number of workloads. And so they said, instead of 195 00:10:52,640 --> 00:10:55,440 Speaker 1: trying to make just a faster CPU, why don't we 196 00:10:55,440 --> 00:10:58,160 Speaker 1: specialize a little bit, and we'll break down all the 197 00:10:58,200 --> 00:11:01,920 Speaker 1: things that the Mac the laptop has to do, and 198 00:11:01,920 --> 00:11:04,520 Speaker 1: we're gonna make them do it really well and with 199 00:11:04,559 --> 00:11:07,720 Speaker 1: these specialized ships. So that's pretty much what I'm what 200 00:11:08,080 --> 00:11:11,720 Speaker 1: the differences, and so this is kind of in the 201 00:11:11,760 --> 00:11:15,280 Speaker 1: picture of the broader things kind of been a revolution anyways, 202 00:11:15,320 --> 00:11:18,000 Speaker 1: Like getting off the CPU has been happening for a 203 00:11:18,040 --> 00:11:20,800 Speaker 1: long time. And part of the reason why now it's 204 00:11:20,800 --> 00:11:23,080 Speaker 1: become such a big shift is because, you know, in 205 00:11:23,120 --> 00:11:25,480 Speaker 1: the last episode with Stacy, guys talked about more laws 206 00:11:25,480 --> 00:11:27,679 Speaker 1: slowing down or maybe it's even you know, they love 207 00:11:27,760 --> 00:11:29,920 Speaker 1: to say it's dead. It's not quite dead, but it's 208 00:11:30,280 --> 00:11:33,079 Speaker 1: it's slowing down quite a bit. And so now the 209 00:11:33,760 --> 00:11:36,080 Speaker 1: road for it definitely seems to be specialization and so 210 00:11:36,160 --> 00:11:39,000 Speaker 1: the M one did this really well. It's specialized in 211 00:11:39,040 --> 00:11:40,920 Speaker 1: a lot of different ways, put it all onto one 212 00:11:40,960 --> 00:11:43,880 Speaker 1: piece of one piece of silicon, and also coupled the 213 00:11:43,920 --> 00:11:46,120 Speaker 1: memory really close together, which is a whole you know, 214 00:11:46,200 --> 00:11:48,440 Speaker 1: kind of in the same vein but a little different. 215 00:11:48,760 --> 00:11:52,000 Speaker 1: And now the M one just rocks. Frankly, it's it's 216 00:11:52,040 --> 00:11:55,280 Speaker 1: not yeah, it's it's a little bit shaky as a 217 00:11:55,320 --> 00:11:57,400 Speaker 1: first product, right like the Apple Watch one or the 218 00:11:57,440 --> 00:12:00,720 Speaker 1: iPhone one wasn't the best product ever, but it's still 219 00:12:01,040 --> 00:12:21,000 Speaker 1: it still is, um a miracle, quite frankly. So you know, 220 00:12:21,080 --> 00:12:24,200 Speaker 1: you talk about how this sort of emerged out of 221 00:12:24,600 --> 00:12:27,720 Speaker 1: Apple's chips for the iPhone and something I think, you know, 222 00:12:27,760 --> 00:12:31,800 Speaker 1: like I buy a new iPhone every few years like 223 00:12:31,920 --> 00:12:34,800 Speaker 1: anyone else. Um, I don't buy every single one that 224 00:12:34,840 --> 00:12:37,160 Speaker 1: comes out, but you know, when my mine feels like 225 00:12:37,160 --> 00:12:39,079 Speaker 1: it's getting old or whatever, I buy a new one. 226 00:12:39,559 --> 00:12:42,280 Speaker 1: But honestly, the only thing I do on my iPhone 227 00:12:42,320 --> 00:12:46,080 Speaker 1: besides tweeting is taking photos. And so even though like 228 00:12:46,120 --> 00:12:49,479 Speaker 1: I have this like incredible computer in my pocket, basically 229 00:12:49,640 --> 00:12:52,920 Speaker 1: is just a camera with Twitter attached or Twitter with 230 00:12:52,960 --> 00:12:56,839 Speaker 1: a camera attached. And so I'm curious like how much 231 00:12:57,360 --> 00:13:01,120 Speaker 1: has this sort of like shift tour words. Uh, These 232 00:13:01,160 --> 00:13:05,040 Speaker 1: sort of focused tasks instead of generalist chips come out 233 00:13:05,080 --> 00:13:07,720 Speaker 1: of this fact that yes, you have I have this 234 00:13:07,760 --> 00:13:11,280 Speaker 1: whole computer. Well, there's one thing that it's really important 235 00:13:11,320 --> 00:13:13,679 Speaker 1: for the iPhone to do for a lot of people, 236 00:13:13,720 --> 00:13:17,040 Speaker 1: and that is take really clear can really uh really 237 00:13:17,120 --> 00:13:20,240 Speaker 1: clear photos, and this sort of recognition that there are 238 00:13:20,280 --> 00:13:23,360 Speaker 1: just a few simple tasks that on a phone you 239 00:13:23,400 --> 00:13:27,280 Speaker 1: have to get done extremely well for most people. Yeah, 240 00:13:27,320 --> 00:13:30,320 Speaker 1: that's actually the perfect that's like the perfect thing to 241 00:13:30,320 --> 00:13:33,240 Speaker 1: bring up because on the Apple phone, right, the A 242 00:13:33,360 --> 00:13:36,160 Speaker 1: fourteen or A fifteen or whatever iteration of the A chip, 243 00:13:36,400 --> 00:13:39,280 Speaker 1: the biggest die, the biggest specialized die is something called 244 00:13:39,280 --> 00:13:43,880 Speaker 1: an IPU or an image process Sorry can you die when? Uh? 245 00:13:44,280 --> 00:13:46,320 Speaker 1: Die is just like a piece of silicon. Sorry, it's 246 00:13:46,320 --> 00:13:49,679 Speaker 1: it's um. Yeah, just just think of it as like 247 00:13:49,720 --> 00:13:52,200 Speaker 1: a unit of silicon, if that makes sense. And the 248 00:13:52,240 --> 00:13:55,200 Speaker 1: biggest piece of like that's been taking share from everything 249 00:13:55,200 --> 00:13:58,640 Speaker 1: else is an IPU, and the CPU has become a 250 00:13:58,679 --> 00:14:01,520 Speaker 1: smaller and smaller part and similar in that same vein right, 251 00:14:01,559 --> 00:14:03,560 Speaker 1: like computers, one of the things that we've been using 252 00:14:03,600 --> 00:14:06,719 Speaker 1: the most is GPUs to play video games. So just 253 00:14:06,800 --> 00:14:09,720 Speaker 1: like the CPUs kind of losing relevance or the you know, 254 00:14:09,880 --> 00:14:13,080 Speaker 1: the CPS losing relevance in your phone to an IPU 255 00:14:13,280 --> 00:14:16,840 Speaker 1: or for your images in in a desktop computer, for example, 256 00:14:16,840 --> 00:14:19,600 Speaker 1: the CPU seems to be losing relevance to a GPU 257 00:14:19,680 --> 00:14:21,560 Speaker 1: for a lot of people who are using it for games. 258 00:14:21,640 --> 00:14:24,480 Speaker 1: So kind of that same that same energy is you know, 259 00:14:24,600 --> 00:14:28,600 Speaker 1: disrupting one and the other. And I think you ask 260 00:14:28,640 --> 00:14:32,400 Speaker 1: a question about like why Apple Apple in the iPhone 261 00:14:32,400 --> 00:14:35,880 Speaker 1: specifically kind of like shifted over to the Mac. What's 262 00:14:35,920 --> 00:14:39,120 Speaker 1: interesting about that is in the iPhone itself, they actually 263 00:14:39,160 --> 00:14:42,200 Speaker 1: had a lot of requirements that were not that you 264 00:14:42,200 --> 00:14:44,680 Speaker 1: didn't really need to use in a in a computer, right, Like, 265 00:14:45,040 --> 00:14:47,080 Speaker 1: it has a lot smaller form factor, it has a 266 00:14:47,080 --> 00:14:49,760 Speaker 1: battery life constraint. It has to also not have a 267 00:14:49,760 --> 00:14:52,760 Speaker 1: fan to actively cool it, and so all these things. 268 00:14:52,960 --> 00:14:54,680 Speaker 1: Um in the beginning, we're just like, you know, a 269 00:14:54,720 --> 00:14:58,040 Speaker 1: pain pain for them to to manufacture around. But as 270 00:14:58,080 --> 00:15:00,000 Speaker 1: they got better and better and better, they pretty much 271 00:15:00,080 --> 00:15:03,000 Speaker 1: got really close to a laptop in an iPhone. And 272 00:15:03,040 --> 00:15:06,120 Speaker 1: that's while having a smaller form factor, while having no 273 00:15:06,600 --> 00:15:09,200 Speaker 1: active fan. And then they're like, wait, why don't we 274 00:15:09,240 --> 00:15:11,520 Speaker 1: just do this for a laptop. And so that's kind 275 00:15:11,520 --> 00:15:13,560 Speaker 1: of what they did. And the M one really is 276 00:15:13,640 --> 00:15:17,160 Speaker 1: like a super souped up um Apple chip. And the 277 00:15:17,160 --> 00:15:19,760 Speaker 1: the announcement that you guys just broke the story on 278 00:15:20,000 --> 00:15:22,040 Speaker 1: is that they're going to actually be making even bigger, 279 00:15:22,080 --> 00:15:24,640 Speaker 1: beefier versions of this because the M one is actually 280 00:15:24,680 --> 00:15:27,240 Speaker 1: made to be put in a laptop with no active cooling. 281 00:15:27,400 --> 00:15:29,360 Speaker 1: Like that's also a big part of it. There's like 282 00:15:29,400 --> 00:15:31,400 Speaker 1: it's supposed to not heat up, it's supposed to not 283 00:15:31,400 --> 00:15:33,440 Speaker 1: have a fan. But what if we what if we 284 00:15:33,520 --> 00:15:36,120 Speaker 1: just like let it run loose and you know, actively 285 00:15:36,120 --> 00:15:38,600 Speaker 1: cool to put a fan. See how hot, how hot 286 00:15:38,640 --> 00:15:40,920 Speaker 1: we can get this thing going? And I think what's 287 00:15:40,920 --> 00:15:43,280 Speaker 1: gonna happen is we're gonna be really shocked by the outcome. 288 00:15:43,720 --> 00:15:47,040 Speaker 1: And so all the manufacturing and the constraints that they 289 00:15:47,040 --> 00:15:49,640 Speaker 1: had to do to make the chip fit into a 290 00:15:49,680 --> 00:15:52,680 Speaker 1: small phone pretty much when you actually scaled it up 291 00:15:52,680 --> 00:15:55,240 Speaker 1: into a laptop, it made it a lot better. And 292 00:15:55,280 --> 00:15:57,160 Speaker 1: that's kind of part of the M one, you know, 293 00:15:57,240 --> 00:16:00,600 Speaker 1: the Apple silicon magic. And Apple isn't the only company 294 00:16:00,640 --> 00:16:04,080 Speaker 1: pursuing this, like I think, you know, almost every company 295 00:16:04,200 --> 00:16:06,600 Speaker 1: is kind of looking at Intel and saying like, okay, 296 00:16:06,640 --> 00:16:09,080 Speaker 1: we need to transition off this vendor. And whether it's 297 00:16:09,080 --> 00:16:11,520 Speaker 1: a data center or whether it's a phone, or whether 298 00:16:11,560 --> 00:16:14,600 Speaker 1: if it's um anything else, Like everyone is kind of 299 00:16:14,640 --> 00:16:17,160 Speaker 1: going this way. And that's what's exciting. Is this isn't 300 00:16:17,280 --> 00:16:20,040 Speaker 1: just an Apple story. Qual Calm mentioned they might do 301 00:16:20,080 --> 00:16:23,680 Speaker 1: heterogeneous chips, which is like one way to call this 302 00:16:23,720 --> 00:16:26,720 Speaker 1: whole thing is heterogeneous, as in many different types of 303 00:16:26,800 --> 00:16:30,120 Speaker 1: chips put together into a larger package. And so you know, 304 00:16:30,240 --> 00:16:33,640 Speaker 1: Android might have a similar and one like chip very soon. 305 00:16:33,920 --> 00:16:36,160 Speaker 1: That's probably you know, a few years away. But the 306 00:16:36,200 --> 00:16:39,320 Speaker 1: writings on the wall that everyone has to shift into 307 00:16:39,640 --> 00:16:42,920 Speaker 1: the this kind of M one style of chip, if 308 00:16:42,920 --> 00:16:46,640 Speaker 1: that makes sense. So it does make sense. Um, this 309 00:16:46,720 --> 00:16:48,200 Speaker 1: is actually what I want to ask you next. So 310 00:16:48,280 --> 00:16:52,400 Speaker 1: I think I understand why Apple sort of struck upon 311 00:16:52,520 --> 00:16:56,680 Speaker 1: this design given its experience with the iPhone, why it 312 00:16:56,840 --> 00:17:02,680 Speaker 1: thought that specialized chips for specialized asks would be more efficient. 313 00:17:03,280 --> 00:17:07,879 Speaker 1: But why was it able to actually produce the chip 314 00:17:08,320 --> 00:17:10,639 Speaker 1: ahead of time? Or maybe another way of saying this 315 00:17:10,760 --> 00:17:14,680 Speaker 1: question is, Um, why was Apple able to do this when, 316 00:17:14,720 --> 00:17:18,040 Speaker 1: as you just mentioned, the entire industry seems to, you know, 317 00:17:18,240 --> 00:17:22,880 Speaker 1: maybe at different speeds be moving in this direction. So 318 00:17:23,160 --> 00:17:27,639 Speaker 1: Apple silicon has always been a leader. They've they've been 319 00:17:27,680 --> 00:17:30,080 Speaker 1: they've been crushing it for a long time. But in 320 00:17:30,119 --> 00:17:32,640 Speaker 1: particular this this kind of goes back to the last 321 00:17:32,640 --> 00:17:35,439 Speaker 1: episode you guys had. T SMC really enabled them to 322 00:17:35,520 --> 00:17:38,840 Speaker 1: do this, and Apple has been buying and building silicon 323 00:17:38,920 --> 00:17:41,959 Speaker 1: expertise in house for years. This is not an overnight 324 00:17:42,000 --> 00:17:45,639 Speaker 1: success story. The the original chip, like the A seven 325 00:17:45,720 --> 00:17:48,560 Speaker 1: or whatever, was when they first started to um to 326 00:17:48,720 --> 00:17:51,639 Speaker 1: really make this more heterogeneous, and they've been every year 327 00:17:52,440 --> 00:17:55,480 Speaker 1: it adds more specialization, and it's gone to this point 328 00:17:55,480 --> 00:17:57,959 Speaker 1: where now the CPU is like one of the not 329 00:17:58,040 --> 00:18:00,560 Speaker 1: the least important, but it's just one aspect of the 330 00:18:00,640 --> 00:18:03,520 Speaker 1: chip in the phone. And so Apple has been doing 331 00:18:03,520 --> 00:18:06,320 Speaker 1: this for a long time. T s MC has been 332 00:18:06,359 --> 00:18:09,480 Speaker 1: a huge part of it enabling this, and then other 333 00:18:09,560 --> 00:18:12,360 Speaker 1: companies kind of look to Apple and are seeing like, okay, 334 00:18:12,400 --> 00:18:15,200 Speaker 1: we can do this too. So I hope that kind 335 00:18:15,200 --> 00:18:18,159 Speaker 1: of answers the question when it comes to manufacturing, and 336 00:18:18,160 --> 00:18:20,560 Speaker 1: of course that was the big thing we talked about 337 00:18:20,640 --> 00:18:24,120 Speaker 1: on our last chip episode, and this idea that it's 338 00:18:24,119 --> 00:18:27,200 Speaker 1: like getting more and more difficult to sort of shrink 339 00:18:27,280 --> 00:18:31,720 Speaker 1: these chips have the smaller an animator form factor. Is 340 00:18:31,760 --> 00:18:34,960 Speaker 1: that an issue here? Or does sort of rethinking the 341 00:18:35,000 --> 00:18:39,800 Speaker 1: fundamental architecture of the chips such that it's it is 342 00:18:40,080 --> 00:18:45,159 Speaker 1: h heterogeneous chip. Does that sort of sidestep some of 343 00:18:45,200 --> 00:18:49,760 Speaker 1: these issues? If that's does that that questions? Yeah? Yeah? 344 00:18:50,000 --> 00:18:52,359 Speaker 1: Um actually, I think that's the huge driver of why 345 00:18:52,400 --> 00:18:55,480 Speaker 1: this whole thing exists, right, Moore's law slowing down and 346 00:18:55,520 --> 00:18:58,440 Speaker 1: the struggle to to shrink smaller and smaller pretty much 347 00:18:58,480 --> 00:19:02,359 Speaker 1: made them think, um, hey, we have to improve this somehow. 348 00:19:02,600 --> 00:19:04,280 Speaker 1: We can't do it through the ways we used to 349 00:19:04,280 --> 00:19:06,280 Speaker 1: do it. Why don't we specialize a little bit. So 350 00:19:06,320 --> 00:19:08,359 Speaker 1: it's kind of like a mix of both that's happening 351 00:19:08,359 --> 00:19:10,960 Speaker 1: at the same time. I definitely think that the desire 352 00:19:11,080 --> 00:19:16,399 Speaker 1: for better, faster chips was what helped the heterogeneous revolution begin. 353 00:19:16,960 --> 00:19:20,640 Speaker 1: But um t SMC being there to essentially make it 354 00:19:21,040 --> 00:19:24,120 Speaker 1: so that anyone who isn't because in the past they 355 00:19:24,119 --> 00:19:27,679 Speaker 1: all until also owned the method to get smaller. Right, 356 00:19:27,720 --> 00:19:29,800 Speaker 1: So it's like, you know, you can't compete against some 357 00:19:29,920 --> 00:19:32,960 Speaker 1: a general chip that's getting smaller, getting faster, better, faster 358 00:19:33,040 --> 00:19:35,760 Speaker 1: than you could specialize your chips faster. But now that 359 00:19:35,800 --> 00:19:38,480 Speaker 1: t SMC is clearly in the lead and anyone can 360 00:19:38,640 --> 00:19:41,520 Speaker 1: can reach out to them and work with them, now 361 00:19:41,560 --> 00:19:43,360 Speaker 1: everyone's like, well, why don't we just make our own 362 00:19:43,400 --> 00:19:46,760 Speaker 1: special little chips? And so that's really been um kind 363 00:19:46,800 --> 00:19:48,920 Speaker 1: of all the things that's made this happen at this 364 00:19:49,040 --> 00:19:51,680 Speaker 1: right moment right now. I mean, on that note, how 365 00:19:51,880 --> 00:19:56,000 Speaker 1: easy is it going to be for competitors to make 366 00:19:56,119 --> 00:19:58,520 Speaker 1: these specialized chips? Then are they going to be able 367 00:19:58,600 --> 00:20:00,919 Speaker 1: to do the same thing that apples do? And in particular, 368 00:20:01,600 --> 00:20:04,480 Speaker 1: is Intel going to be able to do it? Given that, 369 00:20:04,600 --> 00:20:08,080 Speaker 1: as we discussed in the last episode, their business seems 370 00:20:08,160 --> 00:20:11,800 Speaker 1: to be almost entirely based on this idea of selling 371 00:20:12,040 --> 00:20:16,760 Speaker 1: you know, general purpose CPU which then can go into 372 00:20:16,800 --> 00:20:21,760 Speaker 1: a large variety of desktop PCs or whatever. Yeah, so 373 00:20:21,920 --> 00:20:24,840 Speaker 1: Intel is already pursuing this. I mean, the people in 374 00:20:24,920 --> 00:20:27,200 Speaker 1: Intel are smart and they know what they're doing. Um, 375 00:20:27,200 --> 00:20:30,160 Speaker 1: but they're obviously tied to a very large platform. It's 376 00:20:30,200 --> 00:20:33,439 Speaker 1: something called Phobos, which is like kind of their method 377 00:20:33,520 --> 00:20:36,159 Speaker 1: of stacking the chips together. They have this um, this 378 00:20:36,240 --> 00:20:38,920 Speaker 1: new platform and their Intel architecture. Day They're like, okay, 379 00:20:38,960 --> 00:20:40,919 Speaker 1: we're gonna go all in on chiplets, like we're going 380 00:20:40,960 --> 00:20:43,160 Speaker 1: to do the same thing that everyone else is doing. 381 00:20:43,480 --> 00:20:45,880 Speaker 1: But that kind of brings it back to the point 382 00:20:45,920 --> 00:20:47,480 Speaker 1: like Intel is going to be doing the same thing 383 00:20:47,520 --> 00:20:49,320 Speaker 1: that everyone else is doing. And if they don't have 384 00:20:49,359 --> 00:20:53,280 Speaker 1: the manufacturing advantage like they used to, pretty much everyone's 385 00:20:53,280 --> 00:20:56,439 Speaker 1: on the same level playing field and it's it's for 386 00:20:56,520 --> 00:20:58,879 Speaker 1: real right now. Am D for example, does have like 387 00:20:59,160 --> 00:21:01,920 Speaker 1: a big, big reason why they won was kind of 388 00:21:01,960 --> 00:21:04,800 Speaker 1: like a process heater geneous thing where it's like, Okay, 389 00:21:04,880 --> 00:21:06,639 Speaker 1: we don't have to be the fastest, you know, we 390 00:21:06,640 --> 00:21:09,080 Speaker 1: don't have to have the smallest um chip. But what 391 00:21:09,119 --> 00:21:10,760 Speaker 1: we're gonna do is we're just going to make sure 392 00:21:11,200 --> 00:21:14,280 Speaker 1: this really important part is is super small and that way, 393 00:21:14,440 --> 00:21:17,320 Speaker 1: you know, it's like getting across across the finish line 394 00:21:17,600 --> 00:21:19,520 Speaker 1: um by having the best of the little pieces that 395 00:21:19,560 --> 00:21:21,959 Speaker 1: really matter and having a lot of like not as 396 00:21:22,000 --> 00:21:24,760 Speaker 1: good pieces that didn't really matter as much. And so 397 00:21:24,960 --> 00:21:28,720 Speaker 1: they're everyone is kind of going to this like hybrid playbook, 398 00:21:29,080 --> 00:21:32,240 Speaker 1: and every company to a certain extent is pursuing this. 399 00:21:32,440 --> 00:21:35,560 Speaker 1: And yeah, I mean I can talk to individual companies 400 00:21:35,560 --> 00:21:37,159 Speaker 1: in particular, but like I know, it's like I know, 401 00:21:37,240 --> 00:21:40,280 Speaker 1: it's a lot to wrap your mind around sometimes. So 402 00:21:41,000 --> 00:21:44,240 Speaker 1: when we think of Intel, we think of this sort 403 00:21:44,280 --> 00:21:47,520 Speaker 1: of like basic business model, which I guess you know 404 00:21:47,600 --> 00:21:50,440 Speaker 1: still exists. Is still primates tons of one a where 405 00:21:50,440 --> 00:21:53,159 Speaker 1: the the one company makes the chips and then another 406 00:21:53,240 --> 00:21:58,040 Speaker 1: company puts together boxes or laptops or servers where they 407 00:21:58,040 --> 00:22:03,000 Speaker 1: assemble various components, taking Intel chips or chips from elsewhere 408 00:22:03,000 --> 00:22:06,240 Speaker 1: and then packaging them and distributing them, etcetera. Apple is 409 00:22:06,280 --> 00:22:08,800 Speaker 1: always kind of the strewed that model, being much more 410 00:22:09,160 --> 00:22:12,600 Speaker 1: vertically or horizontally integrated, I can never remember which is 411 00:22:12,640 --> 00:22:16,320 Speaker 1: which and having a sort of custom set. Is that 412 00:22:16,400 --> 00:22:20,040 Speaker 1: fundamental model going to remain or is there going to 413 00:22:20,080 --> 00:22:24,600 Speaker 1: be pressure on the very premise of having a separate 414 00:22:25,000 --> 00:22:30,800 Speaker 1: computer that's distinct from a chip. So, yeah, that's a 415 00:22:30,840 --> 00:22:34,879 Speaker 1: great question, I think. Um, I mean, man, I can't. 416 00:22:35,119 --> 00:22:37,440 Speaker 1: I think something that's happening in the near term is 417 00:22:37,480 --> 00:22:41,679 Speaker 1: definitely verticalization, like you're seeing it with Apple, and they're 418 00:22:41,720 --> 00:22:45,840 Speaker 1: they're probably furthest down the road. I actually wrote about 419 00:22:45,880 --> 00:22:49,040 Speaker 1: this quite a bit, But like Amazon in some ways 420 00:22:49,160 --> 00:22:52,280 Speaker 1: is making sure that their whole data center is pretty 421 00:22:52,320 --> 00:22:55,200 Speaker 1: much owned by them using arm chips that they license 422 00:22:55,760 --> 00:23:00,119 Speaker 1: manufactured on t SMC and in some ways, they're their 423 00:23:00,160 --> 00:23:02,520 Speaker 1: whole data center will be vertical and they'll start to 424 00:23:03,440 --> 00:23:05,439 Speaker 1: like think of the data center as a giant computer. 425 00:23:05,680 --> 00:23:08,560 Speaker 1: They'll own the entire thing, and it won't be like 426 00:23:08,600 --> 00:23:10,000 Speaker 1: how it used to be where you could build your 427 00:23:10,000 --> 00:23:13,160 Speaker 1: own computer or in this case, a data center. Instead, 428 00:23:13,400 --> 00:23:15,919 Speaker 1: you're going to be forced to be vertical, just like 429 00:23:16,000 --> 00:23:18,320 Speaker 1: Apple is kind of for your phone, if that makes 430 00:23:18,320 --> 00:23:20,960 Speaker 1: any sense. That's a huge that's a huge new trend 431 00:23:21,040 --> 00:23:24,920 Speaker 1: that is like right now happening. The AWS announcements last 432 00:23:25,359 --> 00:23:28,480 Speaker 1: like the last week or so pretty much they announced 433 00:23:28,480 --> 00:23:31,719 Speaker 1: like three pieces of custom silicon, which is which is crazy. 434 00:23:32,040 --> 00:23:34,920 Speaker 1: So so they're they're doing this, but think of think 435 00:23:34,920 --> 00:23:37,040 Speaker 1: of it in a much bigger picture versus Apple out 436 00:23:37,040 --> 00:23:40,000 Speaker 1: a much smaller picture. And same with Microsoft, same with UM, 437 00:23:40,119 --> 00:23:43,760 Speaker 1: same with Google especially, and Facebook has even recently hired 438 00:23:43,840 --> 00:23:47,040 Speaker 1: someone to probably start this this path for them as 439 00:23:47,040 --> 00:23:50,760 Speaker 1: well too. So I have a dumb question kind of 440 00:23:50,800 --> 00:23:54,040 Speaker 1: going in the opposite direction. But if Apple is making, 441 00:23:54,440 --> 00:23:57,119 Speaker 1: you know, this new type of chip that everyone is 442 00:23:57,240 --> 00:24:01,800 Speaker 1: very excited about, could they ever sell it on the 443 00:24:01,840 --> 00:24:05,400 Speaker 1: market in the way that you know, Intel sells its 444 00:24:05,480 --> 00:24:08,040 Speaker 1: chips or does it just not make any sense for 445 00:24:08,119 --> 00:24:11,000 Speaker 1: them to do it. I think I think it makes 446 00:24:11,040 --> 00:24:13,600 Speaker 1: a lot of sense. The YouTube community is freaking out 447 00:24:13,640 --> 00:24:16,160 Speaker 1: right now about if M one can run Windows, because 448 00:24:16,160 --> 00:24:18,879 Speaker 1: if it can, that would be awesome, and so Apple 449 00:24:18,960 --> 00:24:21,479 Speaker 1: could actually, um they could sell it. I don't know 450 00:24:21,520 --> 00:24:24,600 Speaker 1: if that would be kind of like in their ethos right, 451 00:24:24,640 --> 00:24:27,760 Speaker 1: like they make very very defined um in like a 452 00:24:27,880 --> 00:24:33,440 Speaker 1: very closed ecosystem, but um, yes they could, and even 453 00:24:33,560 --> 00:24:35,840 Speaker 1: Microsoft would buy it for their UM for their high 454 00:24:35,920 --> 00:24:38,240 Speaker 1: end surface books. A lot of a lot of the 455 00:24:38,359 --> 00:24:40,960 Speaker 1: laptop O E M s would definitely be an insta buyer. 456 00:24:41,400 --> 00:24:44,040 Speaker 1: So um yeah, if they wanted to, they could instantly 457 00:24:44,080 --> 00:24:46,320 Speaker 1: they could open it up and sell it, and I 458 00:24:46,359 --> 00:24:48,080 Speaker 1: think it would be really good for their business, Like 459 00:24:48,359 --> 00:24:51,240 Speaker 1: it would be gross margin, a creative. Um it's actually 460 00:24:51,240 --> 00:24:54,120 Speaker 1: probably more profitable than selling the phones out right, which 461 00:24:54,160 --> 00:24:57,960 Speaker 1: is crazy to think about. And I have two questions. 462 00:24:58,000 --> 00:25:00,080 Speaker 1: One is very short and one is a little that 463 00:25:00,640 --> 00:25:03,280 Speaker 1: following onto what you just said. Hey, I you've said 464 00:25:03,280 --> 00:25:05,480 Speaker 1: the term ARM or you talked about ARM a few times. 465 00:25:05,600 --> 00:25:08,040 Speaker 1: What does ARM mean or what does it stand for? 466 00:25:08,119 --> 00:25:11,120 Speaker 1: What does it represent and to just there you said, 467 00:25:11,119 --> 00:25:14,040 Speaker 1: the question is can it run Windows? What do we 468 00:25:14,240 --> 00:25:16,600 Speaker 1: really talking about when like, here's this powerful chip, can 469 00:25:16,600 --> 00:25:18,399 Speaker 1: it run a piece of software? What would be the 470 00:25:18,440 --> 00:25:22,040 Speaker 1: tension or the difficulty in running Windows? And how much 471 00:25:22,119 --> 00:25:24,600 Speaker 1: does that sort of add to all this this this 472 00:25:24,800 --> 00:25:30,520 Speaker 1: need to sink the chip with the with the operating system. Okay, 473 00:25:30,560 --> 00:25:34,320 Speaker 1: so arm is UM is kind of one of those languages. 474 00:25:34,440 --> 00:25:37,240 Speaker 1: I don't remember what it stands for off the top 475 00:25:37,240 --> 00:25:40,760 Speaker 1: of my head, but it's it's actually it's it actually 476 00:25:40,800 --> 00:25:43,639 Speaker 1: got sold to Navidio. Well it's in the process of 477 00:25:43,640 --> 00:25:45,399 Speaker 1: the deal to the video, which is like, you know, 478 00:25:45,480 --> 00:25:49,480 Speaker 1: super High States is arm accompany? Yes, Arms arms a company. Yeah, 479 00:25:49,520 --> 00:25:51,679 Speaker 1: I mean I know that arm is a company, but 480 00:25:51,760 --> 00:25:54,040 Speaker 1: it's also a language. Like this is where I always 481 00:25:54,440 --> 00:25:57,119 Speaker 1: Army is a company, and they pretty much they pretty 482 00:25:57,160 --> 00:26:00,959 Speaker 1: much license their software. They're they're tied their language of 483 00:26:01,320 --> 00:26:04,920 Speaker 1: silicon out to anyone who wants to use it. So um, 484 00:26:04,920 --> 00:26:07,960 Speaker 1: whenever ARM licensit, they essentially say like, hey, you can 485 00:26:08,160 --> 00:26:10,240 Speaker 1: you can use this, will even help you tinker with 486 00:26:10,240 --> 00:26:12,200 Speaker 1: it a little bit, make sure it works out, and 487 00:26:12,320 --> 00:26:14,400 Speaker 1: you're gonna have to like two of revenues or something 488 00:26:14,400 --> 00:26:17,320 Speaker 1: like that. It's usually a percentage of and it's it's 489 00:26:17,320 --> 00:26:19,240 Speaker 1: pretty cheap for a lot of companies, like right if 490 00:26:19,280 --> 00:26:21,600 Speaker 1: you if you don't have any silicon expertise in house, 491 00:26:21,680 --> 00:26:23,879 Speaker 1: like building anything from the ground up is you know, 492 00:26:23,920 --> 00:26:27,119 Speaker 1: probably a billion dollars plus enterprise. And so ARM just 493 00:26:27,200 --> 00:26:29,280 Speaker 1: does that all for free. You can essentially just buy 494 00:26:29,400 --> 00:26:33,359 Speaker 1: off the off the shelf ARM you know, cores or 495 00:26:33,480 --> 00:26:36,240 Speaker 1: i P and then you can kind of build build 496 00:26:36,280 --> 00:26:40,560 Speaker 1: your own little piece UM itself and and UM. Oh 497 00:26:40,720 --> 00:26:43,920 Speaker 1: and so the software side of it, so x A 498 00:26:44,040 --> 00:26:46,560 Speaker 1: D six is pretty much what everything for the PC 499 00:26:46,720 --> 00:26:49,720 Speaker 1: has been built on. ARM obviously is everything mobile has 500 00:26:49,720 --> 00:26:52,960 Speaker 1: already been built on UM. And there's kind of been 501 00:26:53,040 --> 00:26:56,160 Speaker 1: this push, this push and pull right like because ARM 502 00:26:56,320 --> 00:27:00,640 Speaker 1: right now definitely has the ecosystem momentum, but UM people 503 00:27:00,640 --> 00:27:03,560 Speaker 1: are really nervous. Like even even the laptops, the M one. 504 00:27:03,760 --> 00:27:06,399 Speaker 1: The biggest problem with the new Mac laptops with with 505 00:27:06,440 --> 00:27:08,160 Speaker 1: the M one in it is that some of their 506 00:27:08,400 --> 00:27:10,359 Speaker 1: some of the programs won't work with it and you 507 00:27:10,400 --> 00:27:13,040 Speaker 1: pretty much have to you have to redo it to 508 00:27:13,280 --> 00:27:18,000 Speaker 1: make it into UM an ARM compatible style. However, I 509 00:27:18,040 --> 00:27:20,119 Speaker 1: mean a lot of companies kind of can see what 510 00:27:20,240 --> 00:27:22,239 Speaker 1: the writing on the wall is. So they are. But 511 00:27:22,640 --> 00:27:25,040 Speaker 1: there's also something called an emulator where essentially it's like 512 00:27:25,400 --> 00:27:28,400 Speaker 1: the ARM chip will run a mini you know, fake 513 00:27:28,520 --> 00:27:31,600 Speaker 1: x A six and then it will by using that 514 00:27:31,680 --> 00:27:34,240 Speaker 1: like you know emulation, it can do can move that 515 00:27:34,280 --> 00:27:37,480 Speaker 1: software and use it on that hardware. But um, so 516 00:27:37,560 --> 00:27:39,840 Speaker 1: it is a big deal. So right now it seems 517 00:27:39,840 --> 00:27:42,960 Speaker 1: like Windows um from my understanding, can run on M 518 00:27:43,000 --> 00:27:45,280 Speaker 1: one and but the problem is like when it does, 519 00:27:45,520 --> 00:27:47,960 Speaker 1: it's gonna be buggy, Like have you ever used a 520 00:27:48,040 --> 00:27:50,680 Speaker 1: Linux computer for example, Like you can't play video games 521 00:27:50,680 --> 00:27:53,440 Speaker 1: on the next computer because they're not compatible. And so 522 00:27:53,720 --> 00:27:56,960 Speaker 1: that same, that same kind of like problem with the 523 00:27:57,000 --> 00:27:59,479 Speaker 1: software ecosystems is going to have to be hammered out 524 00:27:59,480 --> 00:28:03,200 Speaker 1: and that's gonna, um a huge growing pain for everyone involved. 525 00:28:04,200 --> 00:28:08,960 Speaker 1: So I'm curious. Everyone is very excited about these specialized 526 00:28:09,119 --> 00:28:12,800 Speaker 1: chips is And we were talking a little bit about 527 00:28:12,880 --> 00:28:16,240 Speaker 1: Moore's law at this idea that like maybe it's not 528 00:28:16,760 --> 00:28:19,440 Speaker 1: completely dead, but people are having to look at new 529 00:28:19,440 --> 00:28:21,919 Speaker 1: ways to make their chips more efficient rather than just 530 00:28:22,000 --> 00:28:25,840 Speaker 1: making them faster. After we do this, after we do 531 00:28:25,880 --> 00:28:28,040 Speaker 1: these integrated chips or whatever, you want to call them. 532 00:28:28,240 --> 00:28:31,520 Speaker 1: What's the next big thing that you see down the 533 00:28:31,600 --> 00:28:34,800 Speaker 1: line for the industry. It's kind of like there's this 534 00:28:34,880 --> 00:28:38,880 Speaker 1: like mix of things going on. We're still getting smaller, right, Um, 535 00:28:38,880 --> 00:28:41,920 Speaker 1: that's that's one thing. That's a huge, a huge thing. Um. 536 00:28:42,120 --> 00:28:44,920 Speaker 1: TSMC just had their five minimeter as we discussed, as 537 00:28:44,960 --> 00:28:47,600 Speaker 1: you guys discussed in your last episode. Um, you know, 538 00:28:47,640 --> 00:28:50,080 Speaker 1: the nanimeters don't actually mean anything anymore. It's still a 539 00:28:50,080 --> 00:28:53,440 Speaker 1: marketing term. But but you do get benefits from getting smaller. 540 00:28:53,560 --> 00:28:56,440 Speaker 1: So that's still going to continue, just at a slower pace. Um, 541 00:28:56,440 --> 00:28:59,400 Speaker 1: the specialization is going to continue until it stops working. 542 00:28:59,800 --> 00:29:02,800 Speaker 1: And people there's kind of like you know, I've been 543 00:29:02,840 --> 00:29:05,360 Speaker 1: I read this paper that there is a theoretical limit 544 00:29:05,680 --> 00:29:07,920 Speaker 1: until when that stops working, but like we're going to 545 00:29:08,040 --> 00:29:11,040 Speaker 1: keep running down that road until it stops working. After that, 546 00:29:11,200 --> 00:29:16,200 Speaker 1: I I really don't know. I mean quantum stuff. You know. 547 00:29:16,280 --> 00:29:19,600 Speaker 1: I keep hearing and reading and researching and trying to 548 00:29:19,600 --> 00:29:22,040 Speaker 1: figure out if this is real. Clearly it's a little 549 00:29:22,040 --> 00:29:24,640 Speaker 1: bit away, but that would still only be really useful 550 00:29:24,680 --> 00:29:29,080 Speaker 1: for a very small subset of problems. And so I 551 00:29:29,120 --> 00:29:32,480 Speaker 1: think the way that we're going down which is specialization 552 00:29:32,840 --> 00:29:36,960 Speaker 1: is probably the the surest surest way forward. And not 553 00:29:37,080 --> 00:29:39,520 Speaker 1: only will it become it will like start from most 554 00:29:39,520 --> 00:29:42,240 Speaker 1: of the hardware, but like the software will probably start 555 00:29:42,280 --> 00:29:45,480 Speaker 1: to have have to be written better to couple with 556 00:29:45,760 --> 00:29:48,360 Speaker 1: the hardware. That's something that's like a big trend where 557 00:29:48,560 --> 00:29:51,000 Speaker 1: UM where the software and the hardware will kind of 558 00:29:51,000 --> 00:29:55,240 Speaker 1: specialized together. Because that's that's something that you can continue forward. Right, 559 00:29:55,240 --> 00:29:57,840 Speaker 1: Like the less flexible it is, the faster it will be, 560 00:29:58,040 --> 00:30:01,160 Speaker 1: versus the more flexible the less faster. And so if 561 00:30:01,160 --> 00:30:03,440 Speaker 1: you can just like kind of make it as at 562 00:30:03,560 --> 00:30:06,440 Speaker 1: the least flexible as possible in theory, you can make 563 00:30:06,480 --> 00:30:09,920 Speaker 1: everything into an ASI, and the speed ups that you 564 00:30:09,960 --> 00:30:14,760 Speaker 1: get between UM specialization and non specialization are our magnitudes 565 00:30:15,320 --> 00:30:17,640 Speaker 1: order faster. So that's the reason why it's so exciting, 566 00:30:17,680 --> 00:30:19,760 Speaker 1: because it's like you could have UM if you could 567 00:30:19,760 --> 00:30:22,360 Speaker 1: if you just made a single program that only had 568 00:30:22,400 --> 00:30:24,360 Speaker 1: to run one type of thing, you could make it 569 00:30:24,400 --> 00:30:26,760 Speaker 1: a hundred times faster by just running it on the 570 00:30:26,800 --> 00:30:30,000 Speaker 1: most specialized chip possible. Now you have to be really 571 00:30:30,040 --> 00:30:32,320 Speaker 1: really sure that all you want to do is just 572 00:30:32,520 --> 00:30:35,400 Speaker 1: this program and UM. A great example of this is 573 00:30:35,520 --> 00:30:39,040 Speaker 1: Google's TPU, which is a tensor processing unit, which is 574 00:30:39,520 --> 00:30:44,080 Speaker 1: their their ecosystem called TensorFlow, which is focused on deep learning. 575 00:30:44,760 --> 00:30:47,880 Speaker 1: So Google is like, hey, we know that we're gonna 576 00:30:47,920 --> 00:30:50,440 Speaker 1: be running a lot of tensor TensorFlow in the future, 577 00:30:50,480 --> 00:30:53,200 Speaker 1: and AI is going to become this huge exponential problem. 578 00:30:53,360 --> 00:30:55,280 Speaker 1: So we're gonna make We're gonna make a chip that 579 00:30:55,360 --> 00:30:58,920 Speaker 1: only runs TensorFlow. And it's really really fast. It's um 580 00:30:58,960 --> 00:31:01,440 Speaker 1: compared to how would be done on a CPU. I 581 00:31:01,480 --> 00:31:04,000 Speaker 1: couldn't really tell you, but it's like ten two hundred, 582 00:31:04,280 --> 00:31:07,280 Speaker 1: maybe even more like the magnitude it's it's a lot faster. 583 00:31:07,520 --> 00:31:10,600 Speaker 1: And so until the specialization stops working, I don't really 584 00:31:10,600 --> 00:31:14,640 Speaker 1: see anything that's going to um gonna stop stop it 585 00:31:14,760 --> 00:31:19,160 Speaker 1: for now. So basically it sounds like sort of wrapping 586 00:31:19,200 --> 00:31:24,000 Speaker 1: some of these thoughts together, is that as chip architecture 587 00:31:24,400 --> 00:31:30,080 Speaker 1: has migrated towards more specialized tasks, more specific tasks you 588 00:31:30,160 --> 00:31:34,080 Speaker 1: just mentioned Google and it's deep learning AI stuff, it 589 00:31:34,200 --> 00:31:38,200 Speaker 1: really becomes more part and parcel with every technology. So 590 00:31:38,320 --> 00:31:43,480 Speaker 1: if you're working on machine learning or AI, you it's 591 00:31:43,480 --> 00:31:45,880 Speaker 1: almost like you have to have a chip component to it, 592 00:31:46,120 --> 00:31:49,240 Speaker 1: or I guess maybe it makes sense to pursue a 593 00:31:49,320 --> 00:31:52,800 Speaker 1: chip component rather than just sort of writing code, writing 594 00:31:52,840 --> 00:31:57,960 Speaker 1: software and then running that generic off the shelf chip 595 00:31:58,040 --> 00:32:01,040 Speaker 1: from some other companies. Yeah, that's the perfect way to 596 00:32:01,040 --> 00:32:04,120 Speaker 1: put it together. And and frankly, it used to be awesome, 597 00:32:04,200 --> 00:32:06,000 Speaker 1: and we used to just do it, like think about 598 00:32:06,040 --> 00:32:07,480 Speaker 1: how awesome would be. It's like, oh, you can just 599 00:32:07,520 --> 00:32:09,719 Speaker 1: write the software and every year it just gets you know, 600 00:32:09,800 --> 00:32:13,640 Speaker 1: every two years it gets faster, like faster. Um, you 601 00:32:13,640 --> 00:32:15,400 Speaker 1: didn't have to think about any of this, and it's 602 00:32:15,440 --> 00:32:17,200 Speaker 1: it's a pain to think about all this stuff where 603 00:32:17,240 --> 00:32:19,440 Speaker 1: it's like, hey, we have to make a custom chip 604 00:32:19,520 --> 00:32:22,440 Speaker 1: platform in order for our software to get faster. You know, 605 00:32:22,480 --> 00:32:24,360 Speaker 1: we missed the old days of more slow for a reason. 606 00:32:24,720 --> 00:32:28,479 Speaker 1: But now now that we're slowing down, clearly clearly like 607 00:32:28,480 --> 00:32:31,400 Speaker 1: we're gonna have to do some very specialized, very top 608 00:32:31,400 --> 00:32:35,120 Speaker 1: to bottom structures of software and hardware that work together 609 00:32:35,160 --> 00:32:53,640 Speaker 1: to make the best outcome. So so I just have 610 00:32:53,840 --> 00:32:56,760 Speaker 1: a couple more questions. But one that I'm interested in 611 00:32:56,880 --> 00:32:59,280 Speaker 1: is like, you know, we've been talking about Intel, We've 612 00:32:59,280 --> 00:33:01,560 Speaker 1: been talking about Apple, We've been talking about Google, Amazon, 613 00:33:02,280 --> 00:33:05,360 Speaker 1: in video and Taiwan Semi. But in video has become 614 00:33:05,440 --> 00:33:08,000 Speaker 1: this monster, and I'm curious like their role in it 615 00:33:08,040 --> 00:33:11,080 Speaker 1: because I mostly think of them as a company that 616 00:33:11,120 --> 00:33:13,760 Speaker 1: makes chips for video games, but then I always see 617 00:33:13,800 --> 00:33:17,160 Speaker 1: stuff like about how they're also doing stuff with self 618 00:33:17,280 --> 00:33:21,120 Speaker 1: driving cars and sometimes I see these really cool demos 619 00:33:21,240 --> 00:33:24,560 Speaker 1: or like other sort of machine learning image recognition types 620 00:33:24,600 --> 00:33:27,600 Speaker 1: of thing. What is their role in the ecosystem? What 621 00:33:27,680 --> 00:33:30,240 Speaker 1: is their competitive advantage and how did they emerge to 622 00:33:30,320 --> 00:33:32,720 Speaker 1: be just like such an incredible player in the stock's 623 00:33:32,760 --> 00:33:35,600 Speaker 1: just done? Wild? Yeah, yeah, so the company is wild. 624 00:33:35,680 --> 00:33:39,560 Speaker 1: Jensen Jong is awesome, like he's he's he's an amazing CEO, 625 00:33:39,640 --> 00:33:45,000 Speaker 1: but notoriously um, notoriously stubborn. He's very like Steve Jobs. 626 00:33:45,040 --> 00:33:47,840 Speaker 1: Ask GPUs think about this way that there's a gradient 627 00:33:47,920 --> 00:33:50,440 Speaker 1: of like the most general to the to the least general. 628 00:33:50,480 --> 00:33:54,440 Speaker 1: GPUs are one step past the CPU and where it's 629 00:33:54,480 --> 00:33:56,560 Speaker 1: like it's still pretty flexible and you can, you know, 630 00:33:56,760 --> 00:33:58,400 Speaker 1: you can make it do a lot of things, but 631 00:33:58,440 --> 00:34:00,920 Speaker 1: it's not like so but it's a lot faster, but 632 00:34:00,960 --> 00:34:03,120 Speaker 1: it's not so inflexible that it's like a pain to 633 00:34:03,160 --> 00:34:06,120 Speaker 1: write for. And so that's kind of they're been they're 634 00:34:06,240 --> 00:34:10,600 Speaker 1: huge benefits. So as the original use of a GPU 635 00:34:10,800 --> 00:34:13,680 Speaker 1: is essentially to make all the pixels on your screen 636 00:34:14,120 --> 00:34:16,320 Speaker 1: in parallel, and and that's really hard because you have 637 00:34:16,400 --> 00:34:18,040 Speaker 1: to do it all at the same time. And so 638 00:34:18,280 --> 00:34:22,400 Speaker 1: whenever you hear about paralyzation and that, GPUs are like 639 00:34:22,440 --> 00:34:26,239 Speaker 1: these super parallel engines that are really good at calculating 640 00:34:26,360 --> 00:34:28,560 Speaker 1: everything at the same time. And so the reason why 641 00:34:28,560 --> 00:34:31,480 Speaker 1: that's so important is like a I actually happens to 642 00:34:31,480 --> 00:34:34,319 Speaker 1: be that kind of that same use set almost identically 643 00:34:34,480 --> 00:34:39,399 Speaker 1: UM and so they have this huge ecosystem and they're like, hey, well, 644 00:34:39,400 --> 00:34:42,040 Speaker 1: why don't you use your all your AI things and 645 00:34:42,040 --> 00:34:44,560 Speaker 1: we'll speed up with this GPU. And it's very easy 646 00:34:44,600 --> 00:34:47,680 Speaker 1: to use, and so that's been their huge they're huge 647 00:34:47,960 --> 00:34:50,560 Speaker 1: step is that it's much better than the CPU. It's 648 00:34:50,600 --> 00:34:54,160 Speaker 1: extremely easy to program and and they've they've started to 649 00:34:54,160 --> 00:34:56,040 Speaker 1: add a lot of software demos and and so they're 650 00:34:56,120 --> 00:34:58,799 Speaker 1: kind of in some ways, what what Navidio is doing 651 00:34:59,000 --> 00:35:01,680 Speaker 1: is that they're they're also making their own ecosystem, but 652 00:35:01,719 --> 00:35:04,000 Speaker 1: instead of starting with the software and moving down like 653 00:35:04,080 --> 00:35:06,520 Speaker 1: let's say Apple, they're starting with the hardware and trying 654 00:35:06,560 --> 00:35:09,400 Speaker 1: to move up. And GPUs are, by the way specialized 655 00:35:09,520 --> 00:35:12,520 Speaker 1: they are. They are less. They're less specialized than a 656 00:35:12,600 --> 00:35:14,520 Speaker 1: six or an I P or what you know, all 657 00:35:14,560 --> 00:35:17,120 Speaker 1: the other really really specific things that just do one 658 00:35:17,160 --> 00:35:20,040 Speaker 1: thing really well. They can do multiple things pretty well, 659 00:35:20,400 --> 00:35:24,160 Speaker 1: but um, they are obviously more specialized than a CPU 660 00:35:24,200 --> 00:35:27,520 Speaker 1: that can do everything. You know, kind of Okay, So 661 00:35:27,920 --> 00:35:30,560 Speaker 1: that's that's been their big role, and they've really taken 662 00:35:30,600 --> 00:35:32,520 Speaker 1: a lot of share as as things have started to 663 00:35:32,560 --> 00:35:34,440 Speaker 1: specialized a little bit. They're they're the they're there like 664 00:35:34,480 --> 00:35:37,719 Speaker 1: the easiest first step, if it makes sense. But I 665 00:35:37,760 --> 00:35:41,040 Speaker 1: have one more question. So you know, for a long time, 666 00:35:41,840 --> 00:35:44,319 Speaker 1: I feel like, and correct me if I'm wrong, but 667 00:35:44,480 --> 00:35:48,040 Speaker 1: I feel like the semiconductor industry for like decades was 668 00:35:48,120 --> 00:35:50,719 Speaker 1: sort of like a cyclical industry, and you'd have these 669 00:35:50,760 --> 00:35:54,160 Speaker 1: sort of it's almost like manufacturing or like an industrial thing, 670 00:35:54,200 --> 00:35:56,920 Speaker 1: and you have these up cycles when there was growth, 671 00:35:57,400 --> 00:36:00,560 Speaker 1: and then you'd have like an inventory correction and the 672 00:36:00,560 --> 00:36:03,000 Speaker 1: foundries and the chip companies have made too many chips, 673 00:36:03,000 --> 00:36:05,880 Speaker 1: and then they'd have to cut prices or d stock 674 00:36:05,920 --> 00:36:08,520 Speaker 1: and they'd go into a downturn and so forth. And 675 00:36:08,520 --> 00:36:11,320 Speaker 1: then it feels like in the last several years, maybe 676 00:36:11,360 --> 00:36:14,080 Speaker 1: five years, ten years, I don't know, it has stopped 677 00:36:14,080 --> 00:36:17,879 Speaker 1: being a cyclical industry and become more secular. And and 678 00:36:17,920 --> 00:36:20,399 Speaker 1: that is as chips get put into more and more 679 00:36:20,520 --> 00:36:23,920 Speaker 1: things and phones and all kinds of different stuff. So 680 00:36:23,960 --> 00:36:25,920 Speaker 1: it's like less of an up and down cycle and 681 00:36:26,000 --> 00:36:29,440 Speaker 1: more just like a sort of growth industry. A Is 682 00:36:29,480 --> 00:36:34,120 Speaker 1: that true? Is my perception correct? And be? Is there 683 00:36:34,160 --> 00:36:38,640 Speaker 1: any slow down on the horizon for the world's overall 684 00:36:38,680 --> 00:36:41,160 Speaker 1: demands for chips or do we just keep finding new 685 00:36:41,200 --> 00:36:43,239 Speaker 1: places and new new places to put them in, new 686 00:36:43,280 --> 00:36:47,879 Speaker 1: needs for them. Um? So yes, historically it was a 687 00:36:47,960 --> 00:36:51,480 Speaker 1: crazy cyclical industry. If you guys have ever read um 688 00:36:51,680 --> 00:36:57,359 Speaker 1: the Innovators what I'm dilemma, Yeah, innovatives sylemma. The case 689 00:36:57,360 --> 00:37:00,520 Speaker 1: studies he uses is almost all semi conductor bunnies because 690 00:37:00,520 --> 00:37:03,520 Speaker 1: they're so cyclical and they're they're like constantly you know, 691 00:37:03,920 --> 00:37:06,040 Speaker 1: you know, destroying each other's modes, and they're you know, 692 00:37:06,160 --> 00:37:09,080 Speaker 1: they're they're innovating and then you know, growing up and 693 00:37:09,080 --> 00:37:12,319 Speaker 1: then they're dying like it's it's chaoss um. Things have 694 00:37:12,560 --> 00:37:15,520 Speaker 1: definitely changed a little bit since those days, and so 695 00:37:16,280 --> 00:37:19,200 Speaker 1: one of the largest markets that really dictates the cycklicality 696 00:37:19,320 --> 00:37:23,680 Speaker 1: of all the semiconductor and UM markets is memory, and 697 00:37:23,760 --> 00:37:26,640 Speaker 1: there's been this thesis that UM it's become a lot 698 00:37:26,680 --> 00:37:29,279 Speaker 1: more consolidated, meaning that there's like I think there used 699 00:37:29,280 --> 00:37:31,200 Speaker 1: to be you know, tens of players. There's like three 700 00:37:31,239 --> 00:37:33,560 Speaker 1: for DRAM, which is a type of memory, and five 701 00:37:33,640 --> 00:37:36,040 Speaker 1: for nand five to six or whatever. And so that's 702 00:37:36,160 --> 00:37:40,120 Speaker 1: on one hand, that's that's dampen the cycklicality. And pretty 703 00:37:40,200 --> 00:37:42,560 Speaker 1: much they used to do they used to do the 704 00:37:42,600 --> 00:37:45,759 Speaker 1: races bottom thing where um, whenever things would go down, 705 00:37:45,800 --> 00:37:47,959 Speaker 1: they would just add more supply and then everyone would 706 00:37:48,120 --> 00:37:51,000 Speaker 1: would bankrupt together kind of like shale is to some 707 00:37:51,200 --> 00:37:54,279 Speaker 1: extent or or does. And and they've done less of that, 708 00:37:54,560 --> 00:37:58,240 Speaker 1: and that's been one one thing that's dampened the cycklicality. UM. 709 00:37:58,320 --> 00:38:02,479 Speaker 1: But I think going forward, it seems like there's more 710 00:38:02,560 --> 00:38:05,000 Speaker 1: end markets than ever. In the past. It was really 711 00:38:05,040 --> 00:38:09,080 Speaker 1: just like PCs, right, and then phones became well PCs, 712 00:38:09,160 --> 00:38:12,440 Speaker 1: and and there's some like some industrial and like you know, aerospace, 713 00:38:12,480 --> 00:38:14,560 Speaker 1: and there's always been small markets, but for the most 714 00:38:14,600 --> 00:38:17,360 Speaker 1: part it was just the PC market. And then phones 715 00:38:17,400 --> 00:38:20,080 Speaker 1: became a thing, and then so now became PC and mobile. 716 00:38:20,360 --> 00:38:22,520 Speaker 1: And now it seems like auto is becoming a thing. 717 00:38:22,640 --> 00:38:25,239 Speaker 1: So auto is this new third third market, and so 718 00:38:25,280 --> 00:38:27,920 Speaker 1: it's industrial IoT stuff, which I keep hearing about, but 719 00:38:28,000 --> 00:38:30,520 Speaker 1: I'm always confused about how the semiconductors work into it. 720 00:38:30,800 --> 00:38:33,720 Speaker 1: And then um, the AI market in itself will probably 721 00:38:33,719 --> 00:38:37,800 Speaker 1: become so demand intensive from the cloud, from the cloud 722 00:38:37,840 --> 00:38:41,200 Speaker 1: customers that they'll have their own end cycle. And so 723 00:38:41,400 --> 00:38:44,680 Speaker 1: there's each of these markets have their own cycles to 724 00:38:44,680 --> 00:38:47,840 Speaker 1: to their own extent, but together all the demand is 725 00:38:47,840 --> 00:38:51,560 Speaker 1: starting to like overlap and make it a lot more smooth, 726 00:38:51,600 --> 00:38:54,560 Speaker 1: if that makes sense. To be fair, the market is 727 00:38:54,600 --> 00:38:57,880 Speaker 1: still cyclical, it will always be. It will always probably 728 00:38:57,880 --> 00:39:01,640 Speaker 1: be cyclical, but I think all the demand aspects are 729 00:39:01,719 --> 00:39:05,239 Speaker 1: really pulling things forward, and at least in the near term, 730 00:39:05,360 --> 00:39:09,840 Speaker 1: I think next year seems really well positioned. Like, for example, 731 00:39:10,040 --> 00:39:13,720 Speaker 1: they're saying, like this huge influx of auto demand in China, 732 00:39:13,800 --> 00:39:16,440 Speaker 1: They're like, well, we can't really make it because we're 733 00:39:16,640 --> 00:39:19,239 Speaker 1: we're struggling with semiconductor capacity. That's the headline that came 734 00:39:19,280 --> 00:39:22,280 Speaker 1: out in the last few days. And I think that means, 735 00:39:22,560 --> 00:39:24,839 Speaker 1: you know, that's kind of a sign of of all 736 00:39:24,880 --> 00:39:27,600 Speaker 1: the new things that semiconductors have become a part of, 737 00:39:27,719 --> 00:39:30,400 Speaker 1: not just your PC, but your car, but you're you know, 738 00:39:30,760 --> 00:39:33,399 Speaker 1: your nests, thermostat all the other things in the world 739 00:39:33,400 --> 00:39:35,839 Speaker 1: that you use. So can I just ask one follow 740 00:39:35,920 --> 00:39:38,799 Speaker 1: up on what you just said. You said someone came 741 00:39:38,840 --> 00:39:43,840 Speaker 1: out and was talking about Chinese car production like ramping up. 742 00:39:43,880 --> 00:39:45,680 Speaker 1: Who who was that and that they couldn't make all 743 00:39:45,680 --> 00:39:50,399 Speaker 1: the chips needed for it. It was Volkswagen. Volkswagen. Okay, 744 00:39:50,440 --> 00:39:52,040 Speaker 1: thank you, Sorry, I just wanted to look it up 745 00:39:52,080 --> 00:39:55,799 Speaker 1: after this. Yeah, yeah, I want to say it was 746 00:39:55,800 --> 00:39:58,080 Speaker 1: a Reuter's article, but it pretty much was like, hey, 747 00:39:58,360 --> 00:40:02,759 Speaker 1: semicoucter manufacturers can add capacity because it's like so, I mean, 748 00:40:03,239 --> 00:40:08,239 Speaker 1: that's really interesting, and like like auto constantly is just 749 00:40:08,360 --> 00:40:10,040 Speaker 1: called out as this market that's going to be this 750 00:40:10,160 --> 00:40:13,200 Speaker 1: huge thing, and it is obviously a very big market. 751 00:40:13,280 --> 00:40:15,719 Speaker 1: But like if we're really going down the route of 752 00:40:15,719 --> 00:40:18,640 Speaker 1: EV stuff like hardcore, you know we can we can 753 00:40:18,760 --> 00:40:21,000 Speaker 1: you know, no comment on the ev s back bubble 754 00:40:21,160 --> 00:40:23,640 Speaker 1: or whatever, then there's gonna be a lot of more 755 00:40:23,719 --> 00:40:29,799 Speaker 1: semi connects. Right, It's it's like almost completely digital. Now. 756 00:40:30,120 --> 00:40:34,120 Speaker 1: That was fantastic. You're you're great at explaining this and 757 00:40:34,400 --> 00:40:39,120 Speaker 1: really appreciate you making your public non anonymous debut on 758 00:40:39,280 --> 00:40:41,719 Speaker 1: Odd Lots and I learned a lot to Thanks for 759 00:40:41,719 --> 00:41:07,680 Speaker 1: coming up. Thanks man, I really enjoyed it. So, Tracy, 760 00:41:07,719 --> 00:41:11,120 Speaker 1: are are you cool now to change the focus of 761 00:41:11,120 --> 00:41:15,480 Speaker 1: our podcast on a permanent basis. I don't think I'm 762 00:41:15,560 --> 00:41:20,320 Speaker 1: quite there yet. I admit that semiconductors are a lot interesting. Um, 763 00:41:20,320 --> 00:41:24,880 Speaker 1: then perhaps I gave them credit for before we've recorded 764 00:41:24,920 --> 00:41:27,680 Speaker 1: these two episodes. But yeah, I don't think I want 765 00:41:27,719 --> 00:41:32,080 Speaker 1: to do all thoughts um Comma Semiconductor Specialists just yet. 766 00:41:32,360 --> 00:41:33,840 Speaker 1: Well I was gonna say, maybe, like when we're a 767 00:41:33,880 --> 00:41:36,160 Speaker 1: gigantic media brand in our own right, we could do 768 00:41:36,239 --> 00:41:42,240 Speaker 1: like a spinoff show, The Semiconductor Vertical. Yeah, the semi 769 00:41:42,239 --> 00:41:46,120 Speaker 1: Conductor Vertical excellent. Um. What I was gonna say is 770 00:41:46,600 --> 00:41:50,120 Speaker 1: I'm quite interested in possibly looking at one of those 771 00:41:50,160 --> 00:41:54,680 Speaker 1: Apple UM and one computers at some point. Yeah. I 772 00:41:54,680 --> 00:41:57,520 Speaker 1: don't think we intended for this episode to be a 773 00:41:57,600 --> 00:42:01,400 Speaker 1: giant advertisement for Apple snow chip, but a lot of 774 00:42:01,440 --> 00:42:03,680 Speaker 1: people are excited about it, and it was really interesting 775 00:42:03,880 --> 00:42:07,520 Speaker 1: listening to Doug go into some of the technical details 776 00:42:07,560 --> 00:42:12,320 Speaker 1: about why exactly it's considered such a new thing. Yeah. No, 777 00:42:12,440 --> 00:42:15,400 Speaker 1: I thought that was super interesting as well. Um, I 778 00:42:15,440 --> 00:42:17,880 Speaker 1: feel like the next one we do in our series, 779 00:42:18,360 --> 00:42:23,040 Speaker 1: we should do like at t s MC focus for sure. Yeah, 780 00:42:23,239 --> 00:42:25,680 Speaker 1: let's do that. Like I feel like, you know, like 781 00:42:25,680 --> 00:42:27,880 Speaker 1: we've done until this sort of a sort of a 782 00:42:27,920 --> 00:42:30,240 Speaker 1: little more Apple heavy all the everyone. But it feels 783 00:42:30,280 --> 00:42:33,160 Speaker 1: like t SMC is just like such a fascinating story 784 00:42:33,200 --> 00:42:35,920 Speaker 1: and it's own right, combining their sort of tech prowess 785 00:42:35,960 --> 00:42:39,719 Speaker 1: and geopolitical significance. That that that's that whenever we do 786 00:42:39,760 --> 00:42:43,000 Speaker 1: our next trip episode, let's do it on that agreed, 787 00:42:43,400 --> 00:42:46,799 Speaker 1: all right? Should we leave it there? Yeah, let's leave 788 00:42:46,800 --> 00:42:50,759 Speaker 1: it there. Okay. This has been another episode of the 789 00:42:50,800 --> 00:42:53,440 Speaker 1: All Thoughts podcast. I'm Tracy Alloway. You can follow me 790 00:42:53,520 --> 00:42:57,440 Speaker 1: on Twitter at Tracy Alloway. And I'm Joe Wisenthal. You 791 00:42:57,480 --> 00:43:00,640 Speaker 1: can follow me on Twitter at the Stalwart. Follow our 792 00:43:00,680 --> 00:43:04,359 Speaker 1: guest Doug Laucelin on Twitter. He's under the handle at 793 00:43:04,560 --> 00:43:07,160 Speaker 1: full All the Time. Also be sure to check out 794 00:43:07,200 --> 00:43:12,680 Speaker 1: his newsletter Mules Musing. Follow our producer Laura Carlson at 795 00:43:12,760 --> 00:43:16,480 Speaker 1: Laura M. Carlson. Follow the Bloomberg head of podcast, Francesco 796 00:43:16,600 --> 00:43:19,839 Speaker 1: Levi at Francisco Today, and check out all of our 797 00:43:19,880 --> 00:43:24,200 Speaker 1: podcasts at Bloomberg under the handle at podcast, Thanks for listening,