1 00:00:15,356 --> 00:00:15,796 Speaker 1: Pushkin. 2 00:00:20,236 --> 00:00:21,836 Speaker 2: Just a quick note at the start of the show. 3 00:00:21,916 --> 00:00:24,396 Speaker 2: We are very eager to hear what you think of 4 00:00:24,396 --> 00:00:26,796 Speaker 2: the show and how we can make it better, how 5 00:00:26,836 --> 00:00:29,316 Speaker 2: we can make a show that you like more than 6 00:00:29,476 --> 00:00:30,236 Speaker 2: the show we're making. 7 00:00:30,236 --> 00:00:30,436 Speaker 1: Now. 8 00:00:30,516 --> 00:00:32,236 Speaker 2: Tell us what we should do, more of what we 9 00:00:32,276 --> 00:00:35,596 Speaker 2: should do, less of what we should do, wildly differently. 10 00:00:35,956 --> 00:00:41,196 Speaker 2: Email us at problem at pushkin dot fm. Again problem 11 00:00:41,436 --> 00:00:45,916 Speaker 2: at pushkin dot fm. I'm looking forward to reading every email. 12 00:00:47,476 --> 00:00:53,716 Speaker 2: Artificial intelligence feels very abstract, feels very ephemeral, feels more 13 00:00:53,796 --> 00:00:57,556 Speaker 2: like an idea than a thing. But there is in 14 00:00:57,596 --> 00:01:01,956 Speaker 2: fact a thing there. AI is rooted in the physical world. 15 00:01:02,476 --> 00:01:05,996 Speaker 2: The thing or things are these little pieces of silicon 16 00:01:06,156 --> 00:01:12,716 Speaker 2: and metal called graphics processing units GPUs. These GPUs are expensive. 17 00:01:12,996 --> 00:01:15,836 Speaker 2: They cost thousands of dollars each. If you want to 18 00:01:15,836 --> 00:01:17,876 Speaker 2: build a state of the art AI model, you have 19 00:01:17,916 --> 00:01:22,116 Speaker 2: to buy tens of thousands of these GPUs, and most 20 00:01:22,156 --> 00:01:26,236 Speaker 2: of the GPUs come from a single company, in Vidia, 21 00:01:26,676 --> 00:01:30,236 Speaker 2: which is why in just the past few years Nvidia 22 00:01:30,316 --> 00:01:34,556 Speaker 2: has become one of the most valuable, most important companies. 23 00:01:34,236 --> 00:01:34,796 Speaker 1: In the world. 24 00:01:40,996 --> 00:01:43,796 Speaker 2: I'm Jacob Goldstein, and this is what's your problem. My 25 00:01:43,876 --> 00:01:47,236 Speaker 2: guest today is take him. He's a staff writer at 26 00:01:47,276 --> 00:01:50,116 Speaker 2: Baron's and the author of a new book about in Nvidia. 27 00:01:50,156 --> 00:01:54,156 Speaker 2: The books called The Nvidia Way. The book, of course, 28 00:01:54,196 --> 00:01:57,076 Speaker 2: points up being a lot about Jensen Wong, who co 29 00:01:57,156 --> 00:02:00,596 Speaker 2: founded the company back in nineteen ninety three and who 30 00:02:00,636 --> 00:02:04,716 Speaker 2: is still the CEO and who is really a wild 31 00:02:04,836 --> 00:02:08,436 Speaker 2: kind of terrifying, brilliant figure. And Tay and I talk 32 00:02:08,516 --> 00:02:12,076 Speaker 2: a lot about Jensen later in the interview, but we 33 00:02:12,116 --> 00:02:16,236 Speaker 2: started with this moment in two thousand and one that 34 00:02:16,316 --> 00:02:19,836 Speaker 2: led from Nvidia being a company that made graphics cards 35 00:02:19,876 --> 00:02:23,516 Speaker 2: that let people play video games on computers to becoming 36 00:02:23,556 --> 00:02:29,556 Speaker 2: this key company at the center of the AI revolution today. 37 00:02:30,716 --> 00:02:33,276 Speaker 2: So you write about this moment in two thousand and 38 00:02:33,356 --> 00:02:37,476 Speaker 2: one when a researcher, right, an academic at the University 39 00:02:37,476 --> 00:02:41,516 Speaker 2: of North Carolina, realizes he can basically like hack these 40 00:02:41,556 --> 00:02:44,316 Speaker 2: graphics cards, right, this hardware that's just made to make 41 00:02:44,356 --> 00:02:48,036 Speaker 2: pretty graphics on the computer to help in his research, 42 00:02:48,076 --> 00:02:50,516 Speaker 2: which is like not at all about graphics, right, It's 43 00:02:50,556 --> 00:02:54,396 Speaker 2: like modeling the weather or something. And it seems like 44 00:02:54,636 --> 00:02:58,556 Speaker 2: this moment is a big winds up being a big 45 00:02:58,636 --> 00:03:02,076 Speaker 2: turning point in the history of technology. Really tell me 46 00:03:02,116 --> 00:03:02,836 Speaker 2: about that moment. 47 00:03:03,236 --> 00:03:05,756 Speaker 1: The key moment was with Mark Harris, where he was 48 00:03:05,756 --> 00:03:09,796 Speaker 1: a researcher at the University of North Carolina, and him 49 00:03:09,876 --> 00:03:12,756 Speaker 1: and a lot a bunch of other academics realized there's 50 00:03:12,796 --> 00:03:16,436 Speaker 1: all this computing power that you can hack the algorithm 51 00:03:16,836 --> 00:03:21,156 Speaker 1: to use that computing power to model thermodynamics of fluids 52 00:03:21,196 --> 00:03:23,076 Speaker 1: inside clouds, which was his thesis. 53 00:03:23,436 --> 00:03:26,556 Speaker 2: And he realized that worked better than the CPU than 54 00:03:26,596 --> 00:03:28,876 Speaker 2: just sort of doing it the traditional way through the computes. 55 00:03:28,996 --> 00:03:32,716 Speaker 1: Yeah, and a GPU has all these different cores that 56 00:03:32,836 --> 00:03:36,156 Speaker 1: run in parallel, and the CPU only typically has four 57 00:03:36,196 --> 00:03:39,556 Speaker 1: to eight cores, and GPUs have hundreds of thousands of cores. 58 00:03:39,996 --> 00:03:43,356 Speaker 2: What's a CPU better at than a GPU? Like, why 59 00:03:43,396 --> 00:03:45,756 Speaker 2: would you have four to eight when you could have thousands? 60 00:03:46,196 --> 00:03:50,436 Speaker 1: Well, it could run more complicated pieces of instructions and software, 61 00:03:50,676 --> 00:03:54,356 Speaker 1: and GPUs typically break down into much simpler tasks. But 62 00:03:54,396 --> 00:03:57,476 Speaker 1: the difference is you could run all those tasks across 63 00:03:57,476 --> 00:04:02,916 Speaker 1: a thousands So for certain workloads like AI, it's so 64 00:04:03,036 --> 00:04:06,396 Speaker 1: much faster than a CPU, which has to do things 65 00:04:06,556 --> 00:04:09,836 Speaker 1: kind of one after another serially. And Mark Harris and 66 00:04:09,836 --> 00:04:14,036 Speaker 1: all these academics started hacking into the graphics algorithms and 67 00:04:14,876 --> 00:04:18,876 Speaker 1: using that power to do all this scientific high performance computing. 68 00:04:19,316 --> 00:04:23,236 Speaker 1: And Mark Harris started a website and congregated all these 69 00:04:23,516 --> 00:04:26,676 Speaker 1: researchers and everyone's sharing the knowledge, sharing the tricks that 70 00:04:26,796 --> 00:04:30,356 Speaker 1: actually hack into this. So Nvidia sees this, they're seeing 71 00:04:30,436 --> 00:04:33,836 Speaker 1: people use this to do model stock options, the model 72 00:04:33,916 --> 00:04:36,876 Speaker 1: weather and saying, wow, this is actually really interesting. 73 00:04:37,436 --> 00:04:40,156 Speaker 2: Yeah, that's like your dream if you're a company. Right, 74 00:04:40,196 --> 00:04:42,236 Speaker 2: It's like, oh, there's all these other things you could 75 00:04:42,276 --> 00:04:44,196 Speaker 2: do with this thing we're already. 76 00:04:43,796 --> 00:04:47,236 Speaker 1: Making exactly so, and Jensen kind of had the foresight 77 00:04:47,356 --> 00:04:49,236 Speaker 1: to see this, Wow, this is a really big deal. 78 00:04:49,476 --> 00:04:51,956 Speaker 1: We should invest in this. So they wind up hiring 79 00:04:51,996 --> 00:04:56,156 Speaker 1: Mark Harris and look at all the things that all 80 00:04:56,196 --> 00:04:59,196 Speaker 1: these researchers are doing. And this is the beginning of 81 00:04:59,276 --> 00:05:03,756 Speaker 1: development Kuda, which is a programming platform for general purpose 82 00:05:04,076 --> 00:05:05,036 Speaker 1: GPU computing. 83 00:05:05,476 --> 00:05:07,436 Speaker 2: And just to be clear, Kuda is like, it's a 84 00:05:07,436 --> 00:05:11,436 Speaker 2: programming language, but in this case they're programming what used 85 00:05:11,476 --> 00:05:13,276 Speaker 2: to be just the thing for graphics, but it turns 86 00:05:13,276 --> 00:05:15,396 Speaker 2: out to be good for these other things. And that 87 00:05:15,956 --> 00:05:18,956 Speaker 2: language it's in video's own language. Right, This is I 88 00:05:18,956 --> 00:05:20,676 Speaker 2: feel like, going to be important in the sort of 89 00:05:20,716 --> 00:05:23,876 Speaker 2: business story for why and video is so dominant today, right, 90 00:05:23,956 --> 00:05:27,716 Speaker 2: Like they write this language, they own the language, and 91 00:05:27,756 --> 00:05:30,116 Speaker 2: it's written just to work with their chips, is that right? 92 00:05:30,276 --> 00:05:33,876 Speaker 1: Yes, no one else can use this. It's extensions on 93 00:05:33,876 --> 00:05:37,636 Speaker 1: a programming language that makes parallel computing easier for programmers 94 00:05:37,676 --> 00:05:43,156 Speaker 1: to make. So Jensen just invests in this, and he 95 00:05:43,676 --> 00:05:46,916 Speaker 1: actually thinks longer term than the average CEO. So most 96 00:05:46,956 --> 00:05:49,716 Speaker 1: CEOs looking out maybe the next quarter, maybe a year 97 00:05:49,836 --> 00:05:53,996 Speaker 1: or two. Jensen thinks in five, ten, fifteen year increments. 98 00:05:54,636 --> 00:05:57,916 Speaker 1: So he's always looking out, what's the next big computing shift, 99 00:05:57,956 --> 00:06:01,636 Speaker 1: what's the next big technology phase. So he saw KUDA 100 00:06:01,756 --> 00:06:04,476 Speaker 1: as the main thing where eventually his GPUs, you have 101 00:06:04,556 --> 00:06:07,676 Speaker 1: all this latent, enormous computing power will be able to 102 00:06:08,076 --> 00:06:13,636 Speaker 1: be used for science research, all these other simulation things. 103 00:06:14,196 --> 00:06:16,636 Speaker 1: And he didn't give up even when Wall Street wents 104 00:06:16,676 --> 00:06:18,796 Speaker 1: down on his next day, Why you wasting all this 105 00:06:18,916 --> 00:06:22,636 Speaker 1: die space is crushing your gross profit margins? He said, no, 106 00:06:22,796 --> 00:06:25,196 Speaker 1: this is the future computing. I'm going to invest in 107 00:06:25,236 --> 00:06:28,196 Speaker 1: it and it will work out someday. 108 00:06:28,596 --> 00:06:30,596 Speaker 2: And when you say wasting all this die space, like 109 00:06:30,676 --> 00:06:33,036 Speaker 2: de space is like space on the chip, right, Like 110 00:06:33,076 --> 00:06:37,356 Speaker 2: he's making a physical thing and to commit to Kuda 111 00:06:37,436 --> 00:06:40,196 Speaker 2: to have this dedicated programming language, you actually have to 112 00:06:40,316 --> 00:06:43,196 Speaker 2: give up space on the chip. So that's it's costly, right, 113 00:06:43,236 --> 00:06:47,636 Speaker 2: Their margins go down. They're not optimizing for profits at 114 00:06:47,636 --> 00:06:48,156 Speaker 2: that moment. 115 00:06:48,276 --> 00:06:51,716 Speaker 1: It's literal circuits called Kuda cores. Their hardware circuits are 116 00:06:51,756 --> 00:06:56,996 Speaker 1: optimized to run the Kuda language. So even internally executives 117 00:06:56,996 --> 00:07:00,996 Speaker 1: are like, why are we spending all this money allocating 118 00:07:01,036 --> 00:07:03,876 Speaker 1: dye space for something that people really aren't using that much, 119 00:07:03,996 --> 00:07:07,516 Speaker 1: that isn't generating revenue. But Jensen saw this as a future. 120 00:07:07,716 --> 00:07:10,676 Speaker 1: I actually kind of bring up the analogy of read 121 00:07:10,716 --> 00:07:14,676 Speaker 1: Hastings and Netflix. So when he started Netflix, you know, 122 00:07:14,676 --> 00:07:18,196 Speaker 1: the technology wasn't ready, consumers didn't really have broadband, but 123 00:07:18,276 --> 00:07:22,156 Speaker 1: he had this intuitive sense that someday video will be 124 00:07:22,156 --> 00:07:24,476 Speaker 1: streamed over the Internet and that was the future, and 125 00:07:24,556 --> 00:07:27,116 Speaker 1: it makes it's so obvious now, but back then it wasn't. 126 00:07:27,796 --> 00:07:32,396 Speaker 1: So Hastings positioned Netflix. He made money with DVD rentals 127 00:07:32,396 --> 00:07:35,436 Speaker 1: for a while and just stayed on top of the technology, 128 00:07:35,516 --> 00:07:38,196 Speaker 1: kept on investing in investing, and when the technology was 129 00:07:38,236 --> 00:07:42,836 Speaker 1: good enough, that's when he really pivoted Netflix to dominate 130 00:07:42,836 --> 00:07:49,196 Speaker 1: Internet stream Jensen did this multiple times with three D graphics, 131 00:07:49,236 --> 00:07:52,116 Speaker 1: video game graphics. He knew that someday PC video games 132 00:07:52,156 --> 00:07:56,756 Speaker 1: would be a big market programmable GPUs, Kuda and later 133 00:07:56,876 --> 00:08:00,676 Speaker 1: with AI on these full stack data center AI servers. 134 00:08:00,916 --> 00:08:03,036 Speaker 1: He just sees the future and is willing to keep 135 00:08:03,076 --> 00:08:06,756 Speaker 1: investing even if it's five ten years out. And Kuda 136 00:08:06,956 --> 00:08:10,036 Speaker 1: started in two thousand and six. I mean, things got 137 00:08:10,036 --> 00:08:12,836 Speaker 1: incrementally better for the next ten years, but it didn't 138 00:08:12,876 --> 00:08:15,156 Speaker 1: really take off till twenty twenty two. 139 00:08:16,236 --> 00:08:19,436 Speaker 2: Till chat GPT. Basically twenty twenty two is chatchipt That's 140 00:08:19,476 --> 00:08:21,436 Speaker 2: when in video goes totally bonker. 141 00:08:21,556 --> 00:08:23,836 Speaker 1: And that was the power of the large language model 142 00:08:23,876 --> 00:08:27,036 Speaker 1: of AI and all that stuff. And he's investing throughout 143 00:08:27,036 --> 00:08:28,956 Speaker 1: this whole thing for ten fifteen years. 144 00:08:29,036 --> 00:08:31,316 Speaker 2: Yeah, so there is a moment. There's a moment in 145 00:08:31,356 --> 00:08:33,116 Speaker 2: the middle there. I do obviously want to get to 146 00:08:33,156 --> 00:08:35,156 Speaker 2: the chat GPT moment and the present, but there is 147 00:08:35,196 --> 00:08:38,116 Speaker 2: one more moment I feel like in the in the middle, 148 00:08:38,156 --> 00:08:40,116 Speaker 2: that is where in video is in the center of it. 149 00:08:40,156 --> 00:08:43,076 Speaker 2: And that's twenty twelve, right, So you have early you know, 150 00:08:43,076 --> 00:08:46,276 Speaker 2: two thousand and one, people realize, oh, you can pack 151 00:08:46,316 --> 00:08:48,996 Speaker 2: the GPU to do other things. Andybody's like, oh, that's interesting, 152 00:08:49,076 --> 00:08:51,276 Speaker 2: let's build a whole programming language. So people didn't do that, 153 00:08:51,716 --> 00:08:56,356 Speaker 2: And then in twenty twelve, you have this moment that 154 00:08:56,516 --> 00:09:01,476 Speaker 2: really seems like the birth of modern AI. Right, this 155 00:09:01,516 --> 00:09:06,516 Speaker 2: moment when when this AI model called alex net sort 156 00:09:06,556 --> 00:09:09,836 Speaker 2: of emerges into the world and everybody in the AI 157 00:09:09,916 --> 00:09:14,116 Speaker 2: world is like, oh my god, AI is here, right, 158 00:09:14,556 --> 00:09:15,676 Speaker 2: tell me about that moment. 159 00:09:16,236 --> 00:09:19,836 Speaker 1: So alex Nett was a program that was created by 160 00:09:19,876 --> 00:09:24,516 Speaker 1: two researchers at the University of Toronto and they competed 161 00:09:24,516 --> 00:09:28,796 Speaker 1: in this competition called image net, which basically fed images 162 00:09:28,836 --> 00:09:31,396 Speaker 1: into the model, and they were able to recognize and 163 00:09:31,436 --> 00:09:35,436 Speaker 1: category rise image much more effectively than any other model 164 00:09:35,436 --> 00:09:38,876 Speaker 1: in the past. And the breakthrough was they used GPUs 165 00:09:38,916 --> 00:09:40,116 Speaker 1: for the first time and. 166 00:09:40,116 --> 00:09:42,036 Speaker 2: It was like just a few, right, Like they were 167 00:09:42,076 --> 00:09:44,676 Speaker 2: grad students and they got their hands on like a 168 00:09:44,716 --> 00:09:47,396 Speaker 2: few in video GPUs and had them like running on 169 00:09:47,476 --> 00:09:48,836 Speaker 2: servers in the hallway or something. 170 00:09:48,916 --> 00:09:52,396 Speaker 1: Yes, right, and these are video game GPUs. To be clear, 171 00:09:52,636 --> 00:09:56,076 Speaker 1: this isn't some enterprise complicated. They literally went to a 172 00:09:56,116 --> 00:09:59,116 Speaker 1: store and bought a bunch of video game GPUs and 173 00:09:59,236 --> 00:10:02,036 Speaker 1: it turned out to be very effective where they could 174 00:10:02,076 --> 00:10:06,676 Speaker 1: do the power and effectiveness of two thousand CPUs on 175 00:10:06,676 --> 00:10:07,916 Speaker 1: only twelve GPUs. 176 00:10:08,276 --> 00:10:11,116 Speaker 2: And to be clear or that kind of vision AI, 177 00:10:11,996 --> 00:10:16,436 Speaker 2: it's machine learning. It's the same basic technology as is 178 00:10:16,516 --> 00:10:19,036 Speaker 2: used in large language models. Right, it's it's modern AI. 179 00:10:19,316 --> 00:10:23,436 Speaker 2: And so that moment is this moment when it's like, oh, 180 00:10:23,516 --> 00:10:29,756 Speaker 2: holy shit, GPUs are one hundred x better than CPUs 181 00:10:29,956 --> 00:10:33,196 Speaker 2: for AI. Yes, that's interesting, that's useful to know. 182 00:10:33,956 --> 00:10:38,196 Speaker 1: So it was a combination of GPUs data and algorithms 183 00:10:38,276 --> 00:10:41,796 Speaker 1: that kind of combusted at that perfect moment. And then 184 00:10:41,916 --> 00:10:44,556 Speaker 1: Jensen saw this and it's like this is a big 185 00:10:44,596 --> 00:10:48,076 Speaker 1: deal and someday AI is going to be huge and 186 00:10:48,156 --> 00:10:50,476 Speaker 1: we need to invest big in this. So on top 187 00:10:50,516 --> 00:10:54,316 Speaker 1: of KUDA, he invests in all these AI libraries that 188 00:10:54,836 --> 00:10:59,476 Speaker 1: effectively managed the GPUs to do AI workloads the most 189 00:10:59,516 --> 00:11:02,636 Speaker 1: in the most effective manner. And he invested in libraries, 190 00:11:02,796 --> 00:11:07,116 Speaker 1: invested in software, he invested in researchers, and allocated a 191 00:11:07,116 --> 00:11:10,876 Speaker 1: lot of employees to this project starting in around twenty thirteen. 192 00:11:11,436 --> 00:11:14,676 Speaker 2: And so that's that's more of the like it's not 193 00:11:15,196 --> 00:11:18,636 Speaker 2: just the chip, right, it's like the chip that is 194 00:11:18,836 --> 00:11:23,396 Speaker 2: optimized to be not only efficient on the hardware level. 195 00:11:23,396 --> 00:11:25,356 Speaker 2: But like there's all this software. So if you are 196 00:11:25,876 --> 00:11:28,516 Speaker 2: if you are building AI, if you're if you're writing 197 00:11:28,556 --> 00:11:31,716 Speaker 2: code for AI, like they make it really easy for 198 00:11:31,756 --> 00:11:34,196 Speaker 2: you to do it on in video GPUs. 199 00:11:34,236 --> 00:11:37,156 Speaker 1: So it's all the math libraries, it's all the AI 200 00:11:37,236 --> 00:11:41,436 Speaker 1: frameworks work the best on in video GPUs. And they 201 00:11:41,476 --> 00:11:44,756 Speaker 1: added these things called tensor cores similar to krudercres. They 202 00:11:44,756 --> 00:11:49,236 Speaker 1: actually added hardware circuitry inside the GPUs that are optimized 203 00:11:49,276 --> 00:11:53,076 Speaker 1: to run AI training and workloads and all the all 204 00:11:53,116 --> 00:11:55,916 Speaker 1: the software that you're running. And this is over again, 205 00:11:56,116 --> 00:12:00,156 Speaker 1: this is over another nine eight years where they're building 206 00:12:00,156 --> 00:12:03,236 Speaker 1: all the software and hardware circuitry to run deep learning 207 00:12:03,276 --> 00:12:06,076 Speaker 1: AI the best. 208 00:12:06,556 --> 00:12:11,236 Speaker 2: Okay, it's time. It's twot twenty two. Twenty twenty two, 209 00:12:12,396 --> 00:12:16,796 Speaker 2: chat GPT comes out. What's that mean for in video? 210 00:12:17,316 --> 00:12:21,676 Speaker 1: So actually check chip takes off and it's it takes 211 00:12:21,676 --> 00:12:25,116 Speaker 1: about two quarters for the whole world to realize this 212 00:12:25,196 --> 00:12:29,356 Speaker 1: is a big deal. This, this model of natural language 213 00:12:29,356 --> 00:12:34,956 Speaker 1: processing where the computer actually understands what you're asking, and 214 00:12:35,076 --> 00:12:40,436 Speaker 1: that ability to draw insights and be effective with all 215 00:12:40,476 --> 00:12:44,996 Speaker 1: that natural language stuff just just blows people away, and 216 00:12:45,196 --> 00:12:48,636 Speaker 1: companies and startups realize we have a new AI boom here, 217 00:12:48,756 --> 00:12:54,476 Speaker 1: because it really unleashes a wave of capabilities that wasn't 218 00:12:55,076 --> 00:12:58,516 Speaker 1: wasn't doable before. So about six months later, that's when 219 00:12:59,196 --> 00:13:01,916 Speaker 1: the big bang I call it for Nvidia happens when 220 00:13:01,916 --> 00:13:06,436 Speaker 1: they say, oh my gosh, we're gonna we're gonna beat 221 00:13:06,516 --> 00:13:09,116 Speaker 1: our numbers that the street expects by four billion up. 222 00:13:10,036 --> 00:13:12,276 Speaker 1: And the stock went up like one hundred and seventy 223 00:13:12,316 --> 00:13:15,396 Speaker 1: billion dollars in value because the world realized. 224 00:13:15,036 --> 00:13:17,556 Speaker 2: In like a day. Yes, right, there was one day, 225 00:13:17,636 --> 00:13:21,516 Speaker 2: and it was spring of twenty twenty three. Way talking 226 00:13:21,556 --> 00:13:21,956 Speaker 2: about and. 227 00:13:21,916 --> 00:13:25,316 Speaker 1: This is three, this is three weeks after I had 228 00:13:25,396 --> 00:13:28,436 Speaker 1: the meeting with the book publisher who asked me if 229 00:13:28,476 --> 00:13:30,996 Speaker 1: I could do a book on video. So literally I started. 230 00:13:30,756 --> 00:13:33,436 Speaker 2: Also good timing for you, good timing for a video, 231 00:13:33,516 --> 00:13:37,356 Speaker 2: Good timing for you, And and right, no, I remember 232 00:13:37,396 --> 00:13:40,876 Speaker 2: that day. And and there was a question then of like, 233 00:13:40,956 --> 00:13:42,956 Speaker 2: oh my god, is this a one off? Are they 234 00:13:42,996 --> 00:13:44,156 Speaker 2: going to keep growing this way? 235 00:13:44,236 --> 00:13:44,396 Speaker 1: Right? 236 00:13:44,436 --> 00:13:48,316 Speaker 2: Basically what they said was we sold way way more 237 00:13:48,436 --> 00:13:51,996 Speaker 2: GPUs than we thought we were gonna. Right, that's that's 238 00:13:52,036 --> 00:13:54,516 Speaker 2: what That's the basic thing they reported in their earnings 239 00:13:54,596 --> 00:13:58,156 Speaker 2: report and the subtext is this is the world and 240 00:13:58,196 --> 00:14:02,036 Speaker 2: in particular, like big tech companies with lots of money realizing, oh, 241 00:14:02,196 --> 00:14:05,596 Speaker 2: we've got to get into this AI game more than 242 00:14:05,636 --> 00:14:07,316 Speaker 2: we have been and to do that, we got to 243 00:14:07,356 --> 00:14:11,436 Speaker 2: buy a lot of really expensive GPUs from video. 244 00:14:11,796 --> 00:14:16,876 Speaker 1: And Jensen is really smart at selling this stuff. Like, 245 00:14:17,396 --> 00:14:22,356 Speaker 1: he's very smart because every company, every company CEO, every startup, 246 00:14:22,956 --> 00:14:26,156 Speaker 1: they saw this risk as the extential for them because 247 00:14:26,476 --> 00:14:30,996 Speaker 1: if your competitor comes out with the AI featured offering 248 00:14:31,236 --> 00:14:34,156 Speaker 1: that's much better than what you have today, that that 249 00:14:34,276 --> 00:14:37,996 Speaker 1: isn't advantaged by the AI models, Yeah, that competitor could 250 00:14:38,876 --> 00:14:41,996 Speaker 1: could drive you out of business. So Jensen was very 251 00:14:41,996 --> 00:14:46,196 Speaker 1: smart as at selling to people that you need to 252 00:14:46,316 --> 00:14:47,316 Speaker 1: get on board. 253 00:14:47,156 --> 00:14:52,196 Speaker 2: At making everybody scared of losing. Yes, well, and in particular, right, 254 00:14:52,276 --> 00:14:55,116 Speaker 2: I mean there is a small number of very large, 255 00:14:55,276 --> 00:15:00,156 Speaker 2: very rich tech companies that are a very big part 256 00:15:00,156 --> 00:15:03,036 Speaker 2: of in videos revenue. Right, it's like whatever the ones 257 00:15:03,116 --> 00:15:09,756 Speaker 2: you quld think of, Google, Meta, Microsoft, Amazon, Yes, I mean, 258 00:15:09,756 --> 00:15:12,356 Speaker 2: I guess open ai sort of is kind of connected 259 00:15:12,396 --> 00:15:15,676 Speaker 2: with Like those companies are buying a huge share of 260 00:15:15,796 --> 00:15:19,556 Speaker 2: these GPUs, right, They account for a big, big chunk 261 00:15:19,596 --> 00:15:22,276 Speaker 2: of in videos revenue. Yes, and those are good companies 262 00:15:22,276 --> 00:15:24,436 Speaker 2: they have as customers because they're super rich, right, they 263 00:15:24,436 --> 00:15:26,356 Speaker 2: can afford to pay you tens of billions of dollars 264 00:15:26,436 --> 00:15:27,156 Speaker 2: for your GPUs. 265 00:15:27,516 --> 00:15:31,636 Speaker 1: But that's being said, they're reselling that GPU power to 266 00:15:32,036 --> 00:15:36,316 Speaker 1: companies and startups, right, so startups are buying that GPU 267 00:15:36,356 --> 00:15:38,636 Speaker 1: competing capacity from these. 268 00:15:38,596 --> 00:15:41,116 Speaker 2: GNT tech companies. Yeah, that's a good point, right. So 269 00:15:41,116 --> 00:15:43,276 Speaker 2: they're in a way not the end customer. They're the 270 00:15:43,316 --> 00:15:45,636 Speaker 2: sort of intermediate kind of service providers. 271 00:15:45,676 --> 00:15:49,516 Speaker 1: But they also use that for their own internal systems too, 272 00:15:49,596 --> 00:15:52,476 Speaker 1: like Meta uses a ton of GP power to make 273 00:15:52,516 --> 00:15:57,516 Speaker 1: their advertising algorithms more effective and to pick the videos 274 00:15:57,516 --> 00:15:58,396 Speaker 1: like tech talk does. 275 00:15:58,796 --> 00:16:01,996 Speaker 2: And obviously Google Search is now becoming more and more 276 00:16:02,076 --> 00:16:04,716 Speaker 2: AI driven, and there's Gemini, which is right, I mean, 277 00:16:04,756 --> 00:16:07,436 Speaker 2: there's there's more direct use of AI by all of 278 00:16:07,476 --> 00:16:12,076 Speaker 2: the big tech companies as well. So there's another piece 279 00:16:12,156 --> 00:16:14,396 Speaker 2: of this which is interesting. I mean, because there's one 280 00:16:14,476 --> 00:16:16,636 Speaker 2: universe wherever it's like, oh, we got to get it 281 00:16:16,676 --> 00:16:19,636 Speaker 2: on AI, and they all buy the GPUs and then 282 00:16:19,676 --> 00:16:21,996 Speaker 2: that's kind of it. It's like a step function where 283 00:16:22,036 --> 00:16:25,476 Speaker 2: like there's this momentary rush where everybody buys the GPUs 284 00:16:25,516 --> 00:16:27,756 Speaker 2: and then whatever. They just upgrade every couple of years, 285 00:16:27,796 --> 00:16:31,796 Speaker 2: and it's more like a regular tech hardware business, which 286 00:16:31,836 --> 00:16:34,436 Speaker 2: is like good but not amazing. But there's another piece 287 00:16:34,476 --> 00:16:37,316 Speaker 2: of it that has been a huge deal for end video, 288 00:16:37,356 --> 00:16:40,476 Speaker 2: and that's the scaling law or the scaling hypothesis. Let's 289 00:16:40,476 --> 00:16:42,796 Speaker 2: talk about that. What's the what's the scaling hypothesis. 290 00:16:42,836 --> 00:16:45,276 Speaker 1: So the scaling hypothesis, similar to what I said before, 291 00:16:45,556 --> 00:16:48,876 Speaker 1: is the combination of computing power. The more computing power 292 00:16:48,916 --> 00:16:53,436 Speaker 1: you add, the more data you increase, and the better 293 00:16:53,916 --> 00:16:57,076 Speaker 1: ways you figure out software algorithms for each of these 294 00:16:57,116 --> 00:17:00,516 Speaker 1: three buckets. If you increase it, the AI model becomes 295 00:17:00,556 --> 00:17:02,716 Speaker 1: more capable and more effective. 296 00:17:02,636 --> 00:17:05,516 Speaker 2: And like, just to be clear, like even if the 297 00:17:05,556 --> 00:17:08,676 Speaker 2: algorithms aren't getting that much better, right, Like my understanding is, 298 00:17:08,716 --> 00:17:11,956 Speaker 2: and this is kind of a new idea. It's like, oh, 299 00:17:12,036 --> 00:17:16,196 Speaker 2: if we can just have more data and more computing 300 00:17:16,276 --> 00:17:18,276 Speaker 2: power will get better results. 301 00:17:18,396 --> 00:17:21,436 Speaker 1: Yes, so it's great if you have all three doing 302 00:17:21,556 --> 00:17:25,756 Speaker 1: you know, basically showing up. But for this time period, 303 00:17:26,196 --> 00:17:28,716 Speaker 1: the last few years, the scaling law has really taken 304 00:17:28,796 --> 00:17:34,956 Speaker 1: off and companies are literally ten xing their compute power, 305 00:17:34,996 --> 00:17:40,316 Speaker 1: and these AI clusters going from sixteen thousand GPUs now 306 00:17:40,356 --> 00:17:42,676 Speaker 1: to one hundred thousand gps and now people are saying 307 00:17:42,676 --> 00:17:46,516 Speaker 1: they're going to build one million GPU clusters in the 308 00:17:46,556 --> 00:17:47,836 Speaker 1: next two three years. 309 00:17:47,556 --> 00:17:51,396 Speaker 2: And in videos selling most of those GPUs. 310 00:17:51,476 --> 00:17:53,676 Speaker 1: Yes, So if you go from sixteen thousand to one 311 00:17:53,716 --> 00:17:55,436 Speaker 1: hundred thousand GPUs in a couple of. 312 00:17:55,476 --> 00:17:57,116 Speaker 2: Years, that's a good business to do. 313 00:17:57,556 --> 00:18:00,836 Speaker 1: If you're increasing the scale of the hardware by ten 314 00:18:01,156 --> 00:18:04,196 Speaker 1: ten x every couple of years. And the exciting thing 315 00:18:04,236 --> 00:18:07,516 Speaker 1: for Nvidia is that, like what I argue is that 316 00:18:08,076 --> 00:18:09,996 Speaker 1: this thing is going to this trend is going to 317 00:18:09,996 --> 00:18:13,196 Speaker 1: continue in the next few years because you have the 318 00:18:13,236 --> 00:18:17,156 Speaker 1: scaling laws. Then you have this thing called multimodal where 319 00:18:17,236 --> 00:18:19,796 Speaker 1: they're using this GPUs not just for texts like in 320 00:18:19,876 --> 00:18:24,676 Speaker 1: chat TPT, you're using it for video and images generating 321 00:18:24,716 --> 00:18:28,956 Speaker 1: those things. And now there's these two other two other 322 00:18:29,076 --> 00:18:31,716 Speaker 1: waves of demand that are happening. There's this thing called 323 00:18:31,796 --> 00:18:35,876 Speaker 1: AI agents where these AI models can do multi step 324 00:18:35,956 --> 00:18:36,756 Speaker 1: tasks for you. 325 00:18:37,156 --> 00:18:40,356 Speaker 2: Right, you could be like book me a ticket to 326 00:18:40,436 --> 00:18:44,676 Speaker 2: San Diego some weekend in June whenever it's the best deal, yes, 327 00:18:44,716 --> 00:18:45,756 Speaker 2: and then it does that. 328 00:18:45,876 --> 00:18:49,156 Speaker 1: So that's going to happen in the next twelve months. Yes, 329 00:18:49,596 --> 00:18:52,516 Speaker 1: these AI agents will eventually, in the next year or 330 00:18:52,516 --> 00:18:56,796 Speaker 1: two be able to do all that tedious work automatically 331 00:18:56,916 --> 00:18:59,716 Speaker 1: and probably with less errors than an actual human being. 332 00:19:00,236 --> 00:19:03,636 Speaker 1: So AI agents multimodal. And now there's this thing called 333 00:19:03,676 --> 00:19:07,516 Speaker 1: test time compute, where the AI models, instead of just 334 00:19:07,756 --> 00:19:11,556 Speaker 1: spinning back an instant apply, can actually think about what's 335 00:19:11,596 --> 00:19:14,556 Speaker 1: the best way to respond to your question and spend 336 00:19:14,596 --> 00:19:16,716 Speaker 1: more time thinking about it and then give you a 337 00:19:16,796 --> 00:19:20,556 Speaker 1: higher quality answer. So there's all these things that all 338 00:19:20,596 --> 00:19:25,636 Speaker 1: of these test time compute, AI agents, multi modal. These 339 00:19:25,636 --> 00:19:29,036 Speaker 1: are all things that need more computing power, they need 340 00:19:29,116 --> 00:19:31,436 Speaker 1: more GPUs, and these are all things that are going 341 00:19:31,516 --> 00:19:33,756 Speaker 1: to drive in video's revenue in the next couple of years. 342 00:19:37,076 --> 00:19:38,756 Speaker 2: We'll be back in a minute and we will talk, 343 00:19:38,796 --> 00:19:42,756 Speaker 2: among other things, about Jensen Wong, one of the most 344 00:19:42,836 --> 00:19:46,556 Speaker 2: successful entrepreneurs of the twenty first century and a man 345 00:19:46,596 --> 00:19:49,476 Speaker 2: who once told the colleague that he wakes up every morning, 346 00:19:50,036 --> 00:19:53,396 Speaker 2: looks in the mirror and says to himself, you suck. 347 00:19:59,236 --> 00:20:01,316 Speaker 2: I want to talk about Jensen a little bit. I 348 00:20:01,356 --> 00:20:04,076 Speaker 2: mean your book. You know, he is obviously the main 349 00:20:04,316 --> 00:20:06,516 Speaker 2: character in the history of Nvidio is the main character 350 00:20:06,516 --> 00:20:09,836 Speaker 2: in your book. He's been the CEO of n Video 351 00:20:09,876 --> 00:20:12,276 Speaker 2: for more than thirty years, which is like longer than 352 00:20:12,276 --> 00:20:14,516 Speaker 2: Bill Gates with the CEO of Microsoft, longer than any 353 00:20:14,876 --> 00:20:17,476 Speaker 2: you know, almost anybody in the s and p. Five 354 00:20:17,516 --> 00:20:20,556 Speaker 2: hundred has been a CEO at this point. His childhood 355 00:20:20,596 --> 00:20:22,756 Speaker 2: is quite interesting, right, Like, tell me just a little 356 00:20:22,796 --> 00:20:23,876 Speaker 2: bit about his childhood. 357 00:20:25,276 --> 00:20:28,956 Speaker 1: He was born in Taiwan, and they moved around a 358 00:20:28,956 --> 00:20:33,316 Speaker 1: little bit to Thailand, and his father came to training 359 00:20:33,356 --> 00:20:36,556 Speaker 1: in New York and fell in love with America. This 360 00:20:36,636 --> 00:20:41,596 Speaker 1: is like the great American dream story. And the mother 361 00:20:41,676 --> 00:20:46,636 Speaker 1: and father started teaching Jensen's brother English ten words a day, 362 00:20:47,276 --> 00:20:49,236 Speaker 1: and they sent him to his aunt and uncle when 363 00:20:49,236 --> 00:20:52,076 Speaker 1: he was about age eight or nine, and the aunt 364 00:20:52,196 --> 00:20:55,716 Speaker 1: and uncle and the families. The funny thing is they 365 00:20:55,756 --> 00:20:58,916 Speaker 1: sent him to a reform school in Kentucky by mistake, 366 00:21:00,036 --> 00:21:02,916 Speaker 1: thinking that he will get a great education at this 367 00:21:02,996 --> 00:21:04,676 Speaker 1: boarding school in Kentucky, which. 368 00:21:04,516 --> 00:21:06,676 Speaker 2: He just thought it was a boarding school, yes. 369 00:21:06,676 --> 00:21:09,196 Speaker 1: But it turned out to be a school for troubled. 370 00:21:08,876 --> 00:21:12,956 Speaker 2: Kids, huh, which Jensen was not. He was just like 371 00:21:13,036 --> 00:21:15,636 Speaker 2: a smart kid with ambitious for him parents. 372 00:21:15,756 --> 00:21:18,916 Speaker 1: Yes, But he talks about how his time at Oneita 373 00:21:18,956 --> 00:21:21,836 Speaker 1: Baptist and student Kentucky was formational for him. 374 00:21:21,996 --> 00:21:23,076 Speaker 2: What was it like for him? 375 00:21:23,836 --> 00:21:26,956 Speaker 1: It was hard at the beginning, but he started befriending people. 376 00:21:27,996 --> 00:21:32,196 Speaker 1: He started playing chess with the janitor, and he just 377 00:21:32,276 --> 00:21:35,876 Speaker 1: learned how to deal with other people much better. And 378 00:21:36,436 --> 00:21:39,036 Speaker 1: he talks about how he learned his street fighter. 379 00:21:38,756 --> 00:21:41,356 Speaker 2: Mentality because he was literally fighting. 380 00:21:42,636 --> 00:21:45,716 Speaker 1: Yes, I mean there were bullies and all that stuff, 381 00:21:45,436 --> 00:21:47,556 Speaker 1: but he just learned how to deal with people and 382 00:21:48,436 --> 00:21:53,076 Speaker 1: the rough and tumble of kids back then. Eventually his 383 00:21:53,196 --> 00:21:58,476 Speaker 1: parents came from abroad and they settled in Oregon, and 384 00:21:58,676 --> 00:22:02,996 Speaker 1: he learned his work ethic from working at Denny's. He 385 00:22:03,076 --> 00:22:06,956 Speaker 1: talks about how he washed more dishes and cleaned more 386 00:22:06,996 --> 00:22:11,596 Speaker 1: bathrooms than any CEO in the history of CEOs. He 387 00:22:11,636 --> 00:22:14,236 Speaker 1: says he Denny's helped them in terms of social skills 388 00:22:14,636 --> 00:22:18,156 Speaker 1: and dealing with time pressure and dealing with customers. But 389 00:22:19,836 --> 00:22:22,796 Speaker 1: it's the work ethic that really sets him apart. So 390 00:22:22,916 --> 00:22:25,916 Speaker 1: even at the beginning, he was working from nine am 391 00:22:26,036 --> 00:22:29,756 Speaker 1: to midnight, and he just set a culture where Nvidia 392 00:22:29,756 --> 00:22:31,956 Speaker 1: employees work really hard. 393 00:22:32,356 --> 00:22:34,956 Speaker 2: And I feel like at the beginning that's common, right, 394 00:22:35,076 --> 00:22:38,276 Speaker 2: Like that is the classic startup story. But the classic 395 00:22:38,276 --> 00:22:41,636 Speaker 2: startup stories you do that for five years and then 396 00:22:42,116 --> 00:22:44,596 Speaker 2: either you get giant and you hire a grown up 397 00:22:44,676 --> 00:22:47,476 Speaker 2: CEO and you you know, go start your blimp company 398 00:22:47,556 --> 00:22:51,596 Speaker 2: or whatever, or you sell or whatever. He's still doing that, 399 00:22:51,876 --> 00:22:55,436 Speaker 2: right He's sixty years old and like wildly rich and 400 00:22:55,836 --> 00:22:57,516 Speaker 2: and he's still working that much. 401 00:22:57,556 --> 00:23:02,196 Speaker 1: He's working all weekend. He's working Saturday Sunday. When he 402 00:23:02,236 --> 00:23:04,916 Speaker 1: talks about when he goes to a movie, he doesn't 403 00:23:04,956 --> 00:23:07,836 Speaker 1: remember the movie because he's thinking about work. He finds 404 00:23:07,916 --> 00:23:11,236 Speaker 1: work relaxing and fund This is what drives him. He 405 00:23:11,436 --> 00:23:12,156 Speaker 1: loves working. 406 00:23:12,596 --> 00:23:15,036 Speaker 2: There's a moment, you know, you don't you don't write 407 00:23:15,076 --> 00:23:17,156 Speaker 2: about his personal life at all. Once he sort of 408 00:23:17,196 --> 00:23:21,276 Speaker 2: grows up and starts in video. You know, reasonably because 409 00:23:21,276 --> 00:23:23,796 Speaker 2: the book's about in video, and so I just assumed 410 00:23:23,796 --> 00:23:25,876 Speaker 2: he just worked all the time and didn't have a family. 411 00:23:25,956 --> 00:23:28,236 Speaker 2: And the I think the only time in the book 412 00:23:28,476 --> 00:23:31,276 Speaker 2: that his family comes up is there's this scene where 413 00:23:31,276 --> 00:23:35,436 Speaker 2: he's on vacation and he's talking to some senior manager 414 00:23:35,436 --> 00:23:36,796 Speaker 2: at a video on the phone and they're like, what 415 00:23:36,836 --> 00:23:38,676 Speaker 2: are you doing. He's like, I'm sitting here on the 416 00:23:38,676 --> 00:23:42,276 Speaker 2: balcony watching my kids play in the sand and writing emails. 417 00:23:42,316 --> 00:23:44,436 Speaker 2: And it's like, like, if it's a movie, the kids 418 00:23:44,476 --> 00:23:47,396 Speaker 2: are never on stage, right, you just hear them off stage. 419 00:23:47,436 --> 00:23:50,436 Speaker 2: At that moment, You're like, oh, he has kids, Like 420 00:23:50,636 --> 00:23:53,516 Speaker 2: that was that was a weird moment to me reading 421 00:23:53,556 --> 00:23:54,596 Speaker 2: that line. 422 00:23:54,796 --> 00:23:58,756 Speaker 1: So both executives told me they hate it when Jensen 423 00:23:59,516 --> 00:24:03,356 Speaker 1: so called goes on vacation because he winds up giving 424 00:24:03,396 --> 00:24:06,676 Speaker 1: them more directives and orders and more stuff to do 425 00:24:06,716 --> 00:24:09,076 Speaker 1: when he's not on vacation because he's emailing in them 426 00:24:09,356 --> 00:24:13,276 Speaker 1: do this, do that, and they they actually yell at him, 427 00:24:13,436 --> 00:24:15,676 Speaker 1: So play with your kids. He's like, no, I could 428 00:24:15,716 --> 00:24:19,316 Speaker 1: get real work done when I'm on vacation by doing 429 00:24:19,356 --> 00:24:23,716 Speaker 1: his emails. I mean, so it shows you his obsessions. 430 00:24:23,756 --> 00:24:28,196 Speaker 1: He's constantly thinking, he's constantly worried about what's happening at video. 431 00:24:28,596 --> 00:24:34,436 Speaker 2: I mean it's interesting, right, Like he's not a balanced person, like, 432 00:24:34,676 --> 00:24:39,436 Speaker 2: which is why to some extent he has built the 433 00:24:39,436 --> 00:24:40,636 Speaker 2: thing that he has built. 434 00:24:40,476 --> 00:24:46,516 Speaker 1: Right, He's completely obsessed with winning and he's extremely competitive. 435 00:24:46,956 --> 00:24:49,916 Speaker 2: Yeah, there's a few other specific moments that really stood 436 00:24:49,916 --> 00:24:54,556 Speaker 2: out to me. There's one where I think this, I 437 00:24:54,596 --> 00:24:57,076 Speaker 2: think a salesperson told you this where they had just 438 00:24:57,156 --> 00:25:00,756 Speaker 2: had a great quarter they sold a ton of GPUs, 439 00:25:01,396 --> 00:25:04,236 Speaker 2: and the sales guy's talking to Jensen about how great 440 00:25:04,276 --> 00:25:08,476 Speaker 2: they're doing, and Jensen says to the sales guy about Jensen, 441 00:25:08,516 --> 00:25:11,636 Speaker 2: about himself. Jensen says, I look in the mirror every 442 00:25:11,676 --> 00:25:13,796 Speaker 2: morning and say you suck. 443 00:25:14,516 --> 00:25:18,476 Speaker 1: So he he's almost like a self psychologist. He knows 444 00:25:18,836 --> 00:25:21,676 Speaker 1: if you start thinking that you're you're the best and 445 00:25:21,756 --> 00:25:24,676 Speaker 1: your your hot stuff, you might get complacent, you start 446 00:25:24,756 --> 00:25:26,956 Speaker 1: resting on your laurels, you might not work as hard enough. 447 00:25:27,116 --> 00:25:29,596 Speaker 1: So he he sees that, oh, we just had a 448 00:25:29,636 --> 00:25:33,236 Speaker 1: blowout quarter. What's the risk here? The risk is me 449 00:25:33,316 --> 00:25:36,156 Speaker 1: gain complacent. So I'm gonna look myself in the mirror 450 00:25:36,196 --> 00:25:37,916 Speaker 1: and say you suck and psych myself out. 451 00:25:38,196 --> 00:25:41,836 Speaker 2: I mean that's that's that's one reading of it. That's 452 00:25:41,876 --> 00:25:44,276 Speaker 2: the like ten dimensional chest reading of it. I mean 453 00:25:44,276 --> 00:25:48,596 Speaker 2: there's another reading of it, which is like he's messed 454 00:25:48,676 --> 00:25:51,396 Speaker 2: up in a way that makes him super driven and successful. 455 00:25:51,916 --> 00:25:52,676 Speaker 1: But it works. 456 00:25:52,956 --> 00:25:55,956 Speaker 2: He actually definitely works. We can agree that it works. 457 00:25:56,076 --> 00:25:59,556 Speaker 1: Like he does the opposite too, So when things are 458 00:25:59,676 --> 00:26:01,916 Speaker 1: really intimidating and he feels like, oh my gosh, how 459 00:26:01,916 --> 00:26:04,236 Speaker 1: am I going to do this, he tells himself, how 460 00:26:04,276 --> 00:26:06,876 Speaker 1: hard can it be? How hard can it be? So 461 00:26:06,916 --> 00:26:09,836 Speaker 1: he does it on both both ends in the spectrum 462 00:26:09,956 --> 00:26:14,356 Speaker 1: where he cites himself out when he feels intimidated, and 463 00:26:14,636 --> 00:26:16,596 Speaker 1: when he feels like he's on top of the world, 464 00:26:16,716 --> 00:26:19,116 Speaker 1: he tries to bring himself down back the earth. 465 00:26:19,996 --> 00:26:23,556 Speaker 2: So the one other big piece of the Jensen Wong 466 00:26:23,596 --> 00:26:25,796 Speaker 2: experience that we haven't talked about is the way he 467 00:26:26,116 --> 00:26:30,636 Speaker 2: treats the people who work for him. Right, I want 468 00:26:30,636 --> 00:26:32,596 Speaker 2: to talk about a particular scene because I think it 469 00:26:32,916 --> 00:26:34,476 Speaker 2: puts a finer point on it. A scene from the 470 00:26:34,516 --> 00:26:37,196 Speaker 2: book where it's like a company wide meeting and it's 471 00:26:37,196 --> 00:26:40,076 Speaker 2: on like zoom or whatever, and he's yelling at a 472 00:26:40,116 --> 00:26:42,556 Speaker 2: guy in the meeting. He's yelling at him, and then 473 00:26:42,596 --> 00:26:45,516 Speaker 2: on top of that, he keeps telling the person filming 474 00:26:45,596 --> 00:26:49,196 Speaker 2: the meeting to zoom in on the guy he's yelling at. 475 00:26:49,516 --> 00:26:52,316 Speaker 1: I had multiple sources tell me that it was the 476 00:26:52,316 --> 00:26:53,956 Speaker 1: most humiliating thing they've ever seen. 477 00:26:53,996 --> 00:26:58,316 Speaker 2: I mean, it's that seems like bullying, like why zoom 478 00:26:58,356 --> 00:27:01,076 Speaker 2: in on the guy? Like what lesson is added by that? 479 00:27:01,436 --> 00:27:05,636 Speaker 1: So he keeps on saying to mister Rayfield, you got 480 00:27:05,636 --> 00:27:08,436 Speaker 1: to get this chip back on track. The chip that 481 00:27:08,476 --> 00:27:11,836 Speaker 1: heels in charge was behind schedule, and he kept us 482 00:27:12,596 --> 00:27:16,076 Speaker 1: pointing to him, zooming in his face and telling him, 483 00:27:16,196 --> 00:27:17,756 Speaker 1: You've got to get this. This is not how you 484 00:27:17,836 --> 00:27:20,036 Speaker 1: run a business. You need to get this chip back 485 00:27:20,076 --> 00:27:23,596 Speaker 1: on track. And this kind of like high standard demanding 486 00:27:23,636 --> 00:27:28,716 Speaker 1: attitude I think is effective. It drives people. If you 487 00:27:28,836 --> 00:27:32,036 Speaker 1: get dressed down by Jensen, the next time, you're going 488 00:27:32,116 --> 00:27:35,396 Speaker 1: to work ten times more to make sure that you 489 00:27:35,436 --> 00:27:36,196 Speaker 1: do a better job. 490 00:27:36,756 --> 00:27:39,036 Speaker 2: Maybe I don't want the world to be that way, 491 00:27:39,076 --> 00:27:41,156 Speaker 2: but it is that way, you know what I mean, Like, 492 00:27:41,196 --> 00:27:44,076 Speaker 2: maybe I want to believe that, like there's a kinder 493 00:27:44,116 --> 00:27:47,796 Speaker 2: world where people could do equally good work. But maybe 494 00:27:47,836 --> 00:27:48,636 Speaker 2: I'm wrong. 495 00:27:49,276 --> 00:27:51,916 Speaker 1: And Steve Jobs was the same way. Yeah, I mean 496 00:27:52,156 --> 00:27:55,196 Speaker 1: it wasn't all sure, It wasn't all sunshine and rainbows. 497 00:27:55,476 --> 00:27:59,396 Speaker 2: No, no, I mean the Isaacson book was really clear, right, 498 00:27:59,436 --> 00:28:01,956 Speaker 2: the Isaacson biography of Steve Jobs. He did not seem 499 00:28:02,036 --> 00:28:04,236 Speaker 2: like a good person in that book. He seemed like 500 00:28:04,596 --> 00:28:07,876 Speaker 2: really good at building amazing products, but like not like 501 00:28:07,956 --> 00:28:10,356 Speaker 2: a good human being by the sort of standard way 502 00:28:10,436 --> 00:28:11,956 Speaker 2: we think of what makes a good person. 503 00:28:12,156 --> 00:28:14,876 Speaker 1: But I think the key point that we also have 504 00:28:14,956 --> 00:28:17,036 Speaker 1: to think about is people at Nvidia stay. 505 00:28:17,676 --> 00:28:18,956 Speaker 2: Yeah, that's really interesting. 506 00:28:19,276 --> 00:28:21,276 Speaker 1: The turnover is like one of the lowest in the 507 00:28:21,276 --> 00:28:23,756 Speaker 1: industry at only three percent compared to the average of 508 00:28:23,756 --> 00:28:24,796 Speaker 1: thirteen to fifteen percent. 509 00:28:24,876 --> 00:28:27,476 Speaker 2: Yeah, certainly in the last few years. People stay because 510 00:28:27,916 --> 00:28:29,876 Speaker 2: you get rich if you stay, and you lose your 511 00:28:29,916 --> 00:28:31,836 Speaker 2: equity if you leave. But is that number true if 512 00:28:31,836 --> 00:28:34,196 Speaker 2: you go back to when the stock was flat. 513 00:28:34,316 --> 00:28:36,236 Speaker 1: I don't have the numbers, but a lot of the 514 00:28:36,276 --> 00:28:40,436 Speaker 1: senior executives have been there more so than any other company. Yeah, 515 00:28:40,476 --> 00:28:42,236 Speaker 1: fifteen twenty twenty five years. 516 00:28:42,316 --> 00:28:43,236 Speaker 2: Yeah, that's it. 517 00:28:43,676 --> 00:28:46,756 Speaker 1: You don't really hear other than one instance of people 518 00:28:46,916 --> 00:28:50,356 Speaker 1: leaving to become a CEO of a different company. Uh huh. 519 00:28:50,556 --> 00:28:53,476 Speaker 1: And I think part of it is people realize that 520 00:28:53,636 --> 00:28:58,316 Speaker 1: Nvidia or the Gensen way of doing things is super effective, 521 00:28:58,356 --> 00:29:01,796 Speaker 1: so people want to be on the winning team. So 522 00:29:02,076 --> 00:29:05,156 Speaker 1: in one sense, he draws people hard, he dresses them down. 523 00:29:05,436 --> 00:29:08,916 Speaker 1: He's very blunt and direct. But he compensates people. No 524 00:29:08,916 --> 00:29:13,396 Speaker 1: one ever, hardly anyone complains about compensation. If you're effective 525 00:29:13,796 --> 00:29:16,196 Speaker 1: and you do a good job, he will like double 526 00:29:16,236 --> 00:29:20,956 Speaker 1: your stock compensation on the spot. So there's this meritocracy 527 00:29:21,036 --> 00:29:25,596 Speaker 1: that people really adore. And if you see an effective 528 00:29:25,676 --> 00:29:30,556 Speaker 1: leadership a culture that's based on meritocracy, if you're a 529 00:29:30,556 --> 00:29:33,036 Speaker 1: top engineer, you stay at that company. 530 00:29:33,276 --> 00:29:37,996 Speaker 2: Yeah. Yeah, So you interviewed Jenson as as you were 531 00:29:38,036 --> 00:29:41,596 Speaker 2: working on the book. Was it scary to interview him? 532 00:29:42,356 --> 00:29:46,636 Speaker 1: He was intimidating in the meeting, he he at he 533 00:29:46,676 --> 00:29:48,556 Speaker 1: didn't yell at me, but a couple of times he's saying, 534 00:29:48,716 --> 00:29:52,196 Speaker 1: you don't get nvidio in this line of questioning, and 535 00:29:52,876 --> 00:29:55,596 Speaker 1: I it takes a while to get used to it, 536 00:29:55,596 --> 00:29:58,596 Speaker 1: but you appreciate it because you see where the other 537 00:29:58,636 --> 00:30:02,556 Speaker 1: person is. Like when someone is blunt and direct like Jensen, 538 00:30:02,836 --> 00:30:05,556 Speaker 1: you know where that where he is at all times, 539 00:30:05,596 --> 00:30:09,516 Speaker 1: and then you could work at improving yourself. So he 540 00:30:09,596 --> 00:30:12,156 Speaker 1: talks about how what are you optimizing for? Are you 541 00:30:12,196 --> 00:30:16,956 Speaker 1: optimizing for a person's feelings or optimizing for what's good 542 00:30:16,956 --> 00:30:19,076 Speaker 1: for the company. So that's why he doesn't do one 543 00:30:19,116 --> 00:30:22,996 Speaker 1: on one meetings or career coaching. Typically at a large company, 544 00:30:23,276 --> 00:30:26,516 Speaker 1: when someone's doing poorly, the CEO will take them aside 545 00:30:26,516 --> 00:30:28,476 Speaker 1: and say, Bob, you need to do this. 546 00:30:28,556 --> 00:30:31,316 Speaker 2: You need if it's a person's. 547 00:30:30,756 --> 00:30:34,476 Speaker 1: Yeah, Jensen's like, why is Bob the only one that 548 00:30:34,516 --> 00:30:37,556 Speaker 1: gets a learning here? If I am blunt and direct 549 00:30:37,916 --> 00:30:41,476 Speaker 1: and show Bob what he's doing incorrectly, like that person 550 00:30:41,556 --> 00:30:46,556 Speaker 1: at that meeting, everyone in the room can learn. All 551 00:30:46,556 --> 00:30:49,116 Speaker 1: the employees up and down the ladder can learn. So 552 00:30:49,316 --> 00:30:53,076 Speaker 1: that's his philosophy is everyone should learn from their mistakes 553 00:30:53,156 --> 00:30:55,356 Speaker 1: and everyone should know where they are at all times. 554 00:30:56,556 --> 00:30:58,876 Speaker 2: So I mean, so you write in the book that 555 00:31:00,036 --> 00:31:03,876 Speaker 2: in video really is like an extension of Jensen, right. 556 00:31:03,956 --> 00:31:06,196 Speaker 2: I think he used the metaphor of like the formula 557 00:31:06,236 --> 00:31:09,436 Speaker 2: one card that's like optimized for him as the driver. 558 00:31:12,676 --> 00:31:20,516 Speaker 2: So and he's in his sixties, which is not old 559 00:31:20,556 --> 00:31:25,156 Speaker 2: old but definitely not young. Like, is he gonna run 560 00:31:25,156 --> 00:31:27,036 Speaker 2: in video for another twenty years? What's going to happen? 561 00:31:27,356 --> 00:31:29,716 Speaker 2: Like there's no obvious successor, Like where does that go? 562 00:31:31,076 --> 00:31:36,116 Speaker 1: I think there's there's no one else at in video. 563 00:31:36,156 --> 00:31:39,356 Speaker 1: I think that can run in video as effectively as Jensen. 564 00:31:39,756 --> 00:31:42,276 Speaker 1: And he loves the company so much. He loves what 565 00:31:42,316 --> 00:31:46,236 Speaker 1: he's doing. They're having enormous impact with this AI wave, 566 00:31:46,316 --> 00:31:49,956 Speaker 1: and he's so excited about the potential for curing cancer 567 00:31:50,036 --> 00:31:54,156 Speaker 1: and digital biology, the potential for robots, the potential for 568 00:31:54,516 --> 00:31:59,316 Speaker 1: AI to kind of disrupt education and help kids learn better. 569 00:32:00,076 --> 00:32:03,076 Speaker 1: That I don't see him leaving anytime soon. He loves 570 00:32:03,076 --> 00:32:06,596 Speaker 1: his job so much and he can't argue that he's 571 00:32:06,596 --> 00:32:10,956 Speaker 1: being ineffective. So I don't think for the next few 572 00:32:11,036 --> 00:32:13,556 Speaker 1: years there's anyone that's going to take over for him. 573 00:32:13,796 --> 00:32:15,956 Speaker 1: Someday Nvidia is going to have to come up with 574 00:32:15,996 --> 00:32:18,836 Speaker 1: a new CEO or a successor to Jensen, and that's 575 00:32:18,876 --> 00:32:22,956 Speaker 1: going to be a big question mark. It's that next 576 00:32:22,996 --> 00:32:26,396 Speaker 1: person going to be as effective as Jensen in terms 577 00:32:26,436 --> 00:32:28,676 Speaker 1: of being able to have the technical skill and the 578 00:32:28,676 --> 00:32:33,636 Speaker 1: compency to steer in video in the right direction. Well, 579 00:32:33,716 --> 00:32:37,156 Speaker 1: they have his business genius of coming up with all 580 00:32:37,196 --> 00:32:39,756 Speaker 1: these new strategies that he does time and time again. 581 00:32:40,836 --> 00:32:42,116 Speaker 1: That's going to be a huge question Marke. 582 00:32:44,796 --> 00:32:49,796 Speaker 2: I mean people also talk about limits to scaling right 583 00:32:51,316 --> 00:32:54,996 Speaker 2: in AI. We were talking earlier about how part of 584 00:32:55,076 --> 00:32:57,996 Speaker 2: in video's wild growth over the last couple of years 585 00:32:57,996 --> 00:33:03,116 Speaker 2: has come from this fact that that you can just 586 00:33:03,316 --> 00:33:07,836 Speaker 2: add more GPUs basically and get better results. More GPUs, 587 00:33:07,836 --> 00:33:13,036 Speaker 2: more data, and you know, people talk about a running 588 00:33:13,036 --> 00:33:17,596 Speaker 2: out of data because like they're basically, as I understand it, 589 00:33:17,996 --> 00:33:20,996 Speaker 2: training on the whole Internet right now, which is like, Okay, 590 00:33:21,036 --> 00:33:25,316 Speaker 2: it's a lot of data. I mean, is that it 591 00:33:25,316 --> 00:33:27,316 Speaker 2: seems like nobody knows, right. It seems like they are 592 00:33:27,356 --> 00:33:29,716 Speaker 2: smart people who say, no, no, we'll have synthetic data 593 00:33:29,796 --> 00:33:32,156 Speaker 2: blah blah blah, we won't hit a sort of scaling wall. 594 00:33:32,196 --> 00:33:34,436 Speaker 2: And then there are people who make the other argument. 595 00:33:34,476 --> 00:33:36,316 Speaker 2: I certainly don't know, but does that seem right, Like 596 00:33:36,436 --> 00:33:38,276 Speaker 2: is that an open question right now? And is that 597 00:33:38,316 --> 00:33:39,796 Speaker 2: a meaningful question for nvideo? 598 00:33:40,516 --> 00:33:43,836 Speaker 1: I think it's an open question. But like I said, 599 00:33:44,156 --> 00:33:49,236 Speaker 1: there's the multimodal stuff, there's AI agents, and then there's 600 00:33:49,316 --> 00:33:51,436 Speaker 1: proprietary data inside companies. 601 00:33:51,756 --> 00:33:52,476 Speaker 2: Uh huh. 602 00:33:52,516 --> 00:33:57,516 Speaker 1: So all these corporations have data going back decades. Companies 603 00:33:57,516 --> 00:34:00,316 Speaker 1: are going to use the power of these AI computing 604 00:34:00,356 --> 00:34:05,116 Speaker 1: systems to go through all their data internally, all their 605 00:34:05,116 --> 00:34:10,276 Speaker 1: proprietary data, and have all that knowledge at employees fingertips. 606 00:34:10,916 --> 00:34:15,196 Speaker 1: So an employee can ask what's the best way to 607 00:34:15,276 --> 00:34:18,196 Speaker 1: do this, and the AI computer is going to be 608 00:34:18,196 --> 00:34:21,436 Speaker 1: able to go back thirty years of data and figure 609 00:34:21,476 --> 00:34:24,116 Speaker 1: out the best piece of insight to help that employee. 610 00:34:24,156 --> 00:34:25,756 Speaker 1: And that's not really being done today. 611 00:34:26,836 --> 00:34:30,836 Speaker 2: So Nvidia designs their chips their gips, but they don't 612 00:34:30,836 --> 00:34:33,876 Speaker 2: actually manufacture them, right Is it right that they're made 613 00:34:33,916 --> 00:34:36,676 Speaker 2: in Taiwan, where most cutting its chips are made now? 614 00:34:36,836 --> 00:34:41,076 Speaker 1: Yes, So TESMC actually makes and manufacturers in videos chips 615 00:34:41,436 --> 00:34:44,636 Speaker 1: ever since the late nineteen nineties, and they're the best 616 00:34:44,676 --> 00:34:48,236 Speaker 1: at doing this all. A lot of the fabulist chip 617 00:34:48,276 --> 00:34:52,916 Speaker 1: designers in California use TSMC, and Video uses them too. 618 00:34:53,516 --> 00:34:57,556 Speaker 2: And so I mean it is really interesting to me 619 00:34:58,036 --> 00:35:02,796 Speaker 2: to a lot of people that Taiwan is this incredibly 620 00:35:03,116 --> 00:35:07,196 Speaker 2: fraught geopolitical place. Right, China says Taiwan is part of China. 621 00:35:07,716 --> 00:35:10,556 Speaker 2: Taiwan says no, we are not part of China, and 622 00:35:10,796 --> 00:35:13,436 Speaker 2: that Taiwan is the only place in the world that 623 00:35:13,596 --> 00:35:17,636 Speaker 2: makes like the most important physical thing in the world 624 00:35:17,716 --> 00:35:21,596 Speaker 2: of technology today. Right, it makes like that's wild, Like 625 00:35:21,636 --> 00:35:23,356 Speaker 2: what do you make of that? And how does that 626 00:35:23,396 --> 00:35:24,636 Speaker 2: fit with the in video story? 627 00:35:25,036 --> 00:35:29,156 Speaker 1: I think eventually Nvidia chips will be made in the 628 00:35:29,276 --> 00:35:32,396 Speaker 1: US TSMC. I mean, that's all part of this Chips 629 00:35:32,436 --> 00:35:35,036 Speaker 1: Act that the Biden administration has posed. 630 00:35:35,156 --> 00:35:39,636 Speaker 2: So this this plant that TSMC is building in Arizona, Like, 631 00:35:39,836 --> 00:35:44,836 Speaker 2: is that advanced enough to make like frontier in Nvidia chips? 632 00:35:45,196 --> 00:35:48,236 Speaker 1: So the CSMC factories in the US are always going 633 00:35:48,276 --> 00:35:52,396 Speaker 1: to be one or two generations behind the factories in Taiwan. 634 00:35:53,316 --> 00:35:57,396 Speaker 1: But that doesn't mean Nvidia can't use the factories in 635 00:35:57,436 --> 00:36:00,316 Speaker 1: the US if they're one or two generations behind for 636 00:36:00,396 --> 00:36:03,796 Speaker 1: their older chips. So I think that will eventually happen. 637 00:36:04,356 --> 00:36:07,556 Speaker 2: It seems wild to me, Like, I mean, it seems 638 00:36:07,716 --> 00:36:11,756 Speaker 2: very possible that China will try and make Taiwan be 639 00:36:11,796 --> 00:36:13,916 Speaker 2: part of China, right, and and like we have all 640 00:36:13,956 --> 00:36:16,676 Speaker 2: these export controls to try and keep China from getting 641 00:36:16,676 --> 00:36:19,956 Speaker 2: cutting edge chips. Like it's a really interesting, complicated dynamic. 642 00:36:21,076 --> 00:36:24,396 Speaker 1: I think people make a big issue at this, But 643 00:36:24,436 --> 00:36:28,036 Speaker 1: I think if it happens, like in Vidio, chips are 644 00:36:28,076 --> 00:36:30,596 Speaker 1: not going to be the main issue. It will cost 645 00:36:30,676 --> 00:36:34,196 Speaker 1: a global calamity where you know, we won't be able 646 00:36:34,236 --> 00:36:40,916 Speaker 1: to upgrade our cars, at laptops, nearly every computing device, 647 00:36:41,316 --> 00:36:45,436 Speaker 1: every appliance will not work if we don't have access 648 00:36:45,476 --> 00:36:47,236 Speaker 1: to Taiwan chips. 649 00:36:47,756 --> 00:36:50,836 Speaker 2: I mean, presumably we could do some like sub Manhattan 650 00:36:50,876 --> 00:36:55,556 Speaker 2: Project scale Manhattan project to get a state of the 651 00:36:55,676 --> 00:36:59,516 Speaker 2: art TSMC factory somewhere that is not Taiwan, right, Like. 652 00:37:00,436 --> 00:37:02,556 Speaker 1: We're doing some of that now, but it's not going 653 00:37:02,596 --> 00:37:05,276 Speaker 1: to be able to it's going to be ten twenty 654 00:37:06,196 --> 00:37:07,676 Speaker 1: of the capacity we need. 655 00:37:08,396 --> 00:37:11,116 Speaker 2: Yeah, the capacity, right, it takes a long time, Like 656 00:37:11,116 --> 00:37:15,636 Speaker 2: the fabs are wildly complicated to build, and they cost 657 00:37:15,716 --> 00:37:17,596 Speaker 2: billions of dollars and they take years, so you couldn't 658 00:37:17,676 --> 00:37:18,116 Speaker 2: just do it. 659 00:37:18,156 --> 00:37:20,396 Speaker 1: They cost ten to twenty billion dollars to build and 660 00:37:20,476 --> 00:37:23,036 Speaker 1: three to four years over three to four years, So 661 00:37:23,076 --> 00:37:27,076 Speaker 1: it's not something that can happen overnight. And if something happens, 662 00:37:27,076 --> 00:37:29,596 Speaker 1: like it's just gonna be so terrible and it'll be 663 00:37:29,636 --> 00:37:33,556 Speaker 1: like a depression. So I just hopefully it doesn't happen, 664 00:37:33,676 --> 00:37:36,836 Speaker 1: and you know it's not gonna The question isn't gonna 665 00:37:36,836 --> 00:37:38,756 Speaker 1: be about in videotips. The question is gonna be about 666 00:37:38,756 --> 00:37:39,756 Speaker 1: the global economy of that. 667 00:37:40,076 --> 00:37:43,156 Speaker 2: So what do you think is the biggest risk for 668 00:37:43,356 --> 00:37:46,156 Speaker 2: video in the like five year timeframe? 669 00:37:48,236 --> 00:37:49,196 Speaker 1: It's really tough to say. 670 00:37:49,236 --> 00:37:49,356 Speaker 2: Now. 671 00:37:49,396 --> 00:37:51,556 Speaker 1: I mean I see the next few years with all 672 00:37:51,596 --> 00:37:55,756 Speaker 1: the different AI innovations and all the progress. But again, 673 00:37:56,156 --> 00:38:00,076 Speaker 1: just like every big computing shift, what's the next big thing? 674 00:38:00,116 --> 00:38:03,756 Speaker 1: Could it be quantum computing? It can be like who knows, five, ten, 675 00:38:04,076 --> 00:38:07,956 Speaker 1: fifteen years from now? And every from the pcach with 676 00:38:08,116 --> 00:38:11,996 Speaker 1: Microsoft and Intel it's called Wintel. And then it went 677 00:38:12,036 --> 00:38:17,196 Speaker 1: to smartphones where Apple dominated with the iPhone, and now 678 00:38:17,436 --> 00:38:21,676 Speaker 1: in Vidia is dominating. In terms of the AI computing shift, 679 00:38:21,916 --> 00:38:25,276 Speaker 1: There's gonna be another shift, and right now we're at 680 00:38:25,316 --> 00:38:29,756 Speaker 1: the early stages of the AI computing movement. In five 681 00:38:29,836 --> 00:38:32,356 Speaker 1: ten years, there might be the next big thing that 682 00:38:32,676 --> 00:38:36,676 Speaker 1: I can't can't even foresee right now. And is Nvidia 683 00:38:37,036 --> 00:38:41,476 Speaker 1: gonna be able to see that coming? If Jensen's around, 684 00:38:41,516 --> 00:38:44,796 Speaker 1: I think he will. But if Jensen's not there, you know, 685 00:38:44,996 --> 00:38:49,076 Speaker 1: then just like every other computer, major computer company in history, 686 00:38:49,276 --> 00:38:53,356 Speaker 1: say IBM, it's very easy to get disrupted in the 687 00:38:53,396 --> 00:38:57,716 Speaker 1: technology industry. 688 00:38:58,716 --> 00:39:05,396 Speaker 2: We'll be back in a minute with the Lightning round. 689 00:39:09,876 --> 00:39:12,276 Speaker 2: I want to finish with the Lightning round, which is 690 00:39:12,316 --> 00:39:15,716 Speaker 2: going to be considerably more random and digressive than the 691 00:39:15,756 --> 00:39:20,676 Speaker 2: conversation to this point. I want to talk about video 692 00:39:20,716 --> 00:39:26,236 Speaker 2: games and your experience with video games. What was the 693 00:39:26,236 --> 00:39:28,876 Speaker 2: first video game you ever loved? 694 00:39:30,196 --> 00:39:32,556 Speaker 1: I think it's the first VAM I've ever tried, which 695 00:39:32,676 --> 00:39:34,956 Speaker 1: was Combat for the Datar twenty six hundred. 696 00:39:35,196 --> 00:39:35,756 Speaker 2: I love that. 697 00:39:36,076 --> 00:39:39,156 Speaker 1: I mean it's literally I remember coming home and my 698 00:39:39,276 --> 00:39:42,556 Speaker 1: dad buying the Autari twenty six hundred connected to a 699 00:39:42,596 --> 00:39:46,276 Speaker 1: black and white TV and being to use that joystick 700 00:39:46,356 --> 00:39:49,876 Speaker 1: to control the little tank going around the screen. That 701 00:39:49,956 --> 00:39:52,676 Speaker 1: was an amazing moment. It was an incredible moment that 702 00:39:52,756 --> 00:39:56,156 Speaker 1: this was possible. At the home, I had. 703 00:39:56,036 --> 00:39:59,756 Speaker 2: An Atari twenty six hundred. I remember the joystick. I 704 00:39:59,836 --> 00:40:02,836 Speaker 2: remember Demon Attack, so kind of second tier game, but 705 00:40:02,876 --> 00:40:07,116 Speaker 2: I got really into it. What's the game you spent 706 00:40:07,156 --> 00:40:08,116 Speaker 2: the most hours on? 707 00:40:11,316 --> 00:40:14,396 Speaker 1: Probably this game called Lemmings. I don't know if you 708 00:40:14,596 --> 00:40:18,756 Speaker 1: it's it's a puzzle strategy game where you control lemmings 709 00:40:19,036 --> 00:40:23,076 Speaker 1: and you guide them across a maze to get. 710 00:40:23,276 --> 00:40:26,236 Speaker 2: The little creatures that are famed for jumping off cliffs, 711 00:40:26,236 --> 00:40:27,756 Speaker 2: which maybe they don't actually do. 712 00:40:27,916 --> 00:40:31,076 Speaker 1: Yes, if you guide them the wrong way, they'll fall 713 00:40:31,076 --> 00:40:33,196 Speaker 1: off a cliff and to their death, and they make 714 00:40:33,276 --> 00:40:37,716 Speaker 1: this cute sound like oh no, so so. I played 715 00:40:37,956 --> 00:40:39,116 Speaker 1: so many hours of. 716 00:40:38,996 --> 00:40:42,556 Speaker 2: That game, one thousand hours. 717 00:40:43,636 --> 00:40:45,716 Speaker 1: Probably hundreds of hours. I want to say a thousand. 718 00:40:45,956 --> 00:40:50,676 Speaker 2: Yeah. What's your most proud video game accomplishment? 719 00:40:52,756 --> 00:40:56,596 Speaker 1: I still remember playing the Legend of Zelda, the first 720 00:40:56,596 --> 00:40:59,836 Speaker 1: one for the Nintendo Entertainment System and beating that game 721 00:41:00,316 --> 00:41:04,916 Speaker 1: and just like pumping my fists in the air. I 722 00:41:05,076 --> 00:41:06,956 Speaker 1: was over young man. Yeah. 723 00:41:07,436 --> 00:41:12,276 Speaker 2: What is the most exciting video game innovation that's going 724 00:41:12,316 --> 00:41:14,556 Speaker 2: to happen in the next few years? 725 00:41:15,796 --> 00:41:18,836 Speaker 1: Well, this is a little technical, but I think DLSS 726 00:41:19,036 --> 00:41:21,956 Speaker 1: from the video is going to get even more powerful, 727 00:41:22,196 --> 00:41:26,436 Speaker 1: which is this upscaling technology the gents actually invented in 728 00:41:26,476 --> 00:41:31,196 Speaker 1: a meeting where they fill in the details using AI, 729 00:41:31,916 --> 00:41:35,836 Speaker 1: so the frame rates are able to go much faster 730 00:41:35,956 --> 00:41:36,356 Speaker 1: and better. 731 00:41:36,836 --> 00:41:39,196 Speaker 2: So it's basically like interpolation. It's sort of doing what 732 00:41:39,276 --> 00:41:42,316 Speaker 2: AI AI always does, yes, fort. 733 00:41:41,716 --> 00:41:44,596 Speaker 1: But that's really effective. People can't tell the difference between 734 00:41:44,636 --> 00:41:47,916 Speaker 1: the real thing and the interpolated AI graphics, so that 735 00:41:48,596 --> 00:41:52,596 Speaker 1: will allow as this technology get better, that will allow 736 00:41:52,796 --> 00:41:56,236 Speaker 1: the graphics to be more for the realistic, and the 737 00:41:56,316 --> 00:41:59,636 Speaker 1: physics and the rate tracing better than ever before. 738 00:42:01,436 --> 00:42:03,356 Speaker 2: What game are you most excited to play in twenty 739 00:42:03,436 --> 00:42:03,956 Speaker 2: twenty five? 740 00:42:05,596 --> 00:42:07,876 Speaker 1: I mean Grand Theft Auto six if it comes out, 741 00:42:07,956 --> 00:42:08,956 Speaker 1: but it might get delayed. 742 00:42:08,956 --> 00:42:13,356 Speaker 2: Next here, Okay, let's do a few non video game questions. 743 00:42:13,916 --> 00:42:19,436 Speaker 2: What's your favorite tech book besides the one you wrote. 744 00:42:22,036 --> 00:42:24,516 Speaker 1: I guess it's still a tech book, The Innovator's Dilemma, 745 00:42:24,556 --> 00:42:26,756 Speaker 1: which is actually one of Jens's here books. It talks 746 00:42:26,796 --> 00:42:32,436 Speaker 1: about how companies get disrupted by startups and people underneath them. 747 00:42:33,036 --> 00:42:36,876 Speaker 1: It really goes into in depth about how the distrived industry. 748 00:42:37,396 --> 00:42:41,196 Speaker 1: Every successive generation that distribes there was a new market 749 00:42:41,276 --> 00:42:45,956 Speaker 1: leader because they couldn't disrupt themselves to the new or 750 00:42:45,996 --> 00:42:46,756 Speaker 1: smaller format. 751 00:42:46,956 --> 00:42:49,596 Speaker 2: I mean, to me, the really interesting insight of that 752 00:42:49,716 --> 00:42:56,716 Speaker 2: book is the innovator is actually making a crappier product, right, 753 00:42:56,796 --> 00:42:59,636 Speaker 2: Like that's the surprise. It's not exactly like they come 754 00:42:59,676 --> 00:43:02,236 Speaker 2: along and do something better. They go to the crappy 755 00:43:02,356 --> 00:43:05,076 Speaker 2: end of the market and they make a cheap product 756 00:43:05,716 --> 00:43:09,036 Speaker 2: and it's not better than the thing the market leaders 757 00:43:09,276 --> 00:43:11,236 Speaker 2: and the market lader' is like, oh that, who cares 758 00:43:11,236 --> 00:43:15,396 Speaker 2: about that? That's just some low market, some low margin thing, 759 00:43:16,036 --> 00:43:19,316 Speaker 2: and the innovator comes up from below. Like that to 760 00:43:19,356 --> 00:43:22,436 Speaker 2: me is the really key insight of that book. 761 00:43:22,516 --> 00:43:25,476 Speaker 1: And they scale the volumes, and once you have the volume, 762 00:43:25,556 --> 00:43:28,796 Speaker 1: it pretty much becomes game over for the incumbent because 763 00:43:28,876 --> 00:43:32,276 Speaker 1: they can't match that scale and the economies of scale 764 00:43:32,276 --> 00:43:33,076 Speaker 1: that comes with that. 765 00:43:36,076 --> 00:43:37,596 Speaker 2: Defend pineapple on pizza. 766 00:43:39,116 --> 00:43:42,116 Speaker 1: It just tastes good. I love it. I don't know 767 00:43:42,156 --> 00:43:44,556 Speaker 1: if you saw my tweets on it. I actually like 768 00:43:44,636 --> 00:43:45,516 Speaker 1: pineapple on pizza. 769 00:43:46,076 --> 00:43:50,996 Speaker 2: Yes, I think it was Instagram. I think it was Instagram. 770 00:43:51,116 --> 00:43:52,756 Speaker 2: What's the best deal you ever got at Costco? 771 00:43:55,436 --> 00:43:57,516 Speaker 1: It's still the hot dog and you can't beat the dollar. 772 00:43:57,516 --> 00:44:08,276 Speaker 2: Fishbee Take Him is the author of the Nvidia Way. 773 00:44:09,316 --> 00:44:12,556 Speaker 2: Today's show was produced by Gabriel Hunter Chang. It was 774 00:44:12,836 --> 00:44:16,276 Speaker 2: edited by Lyddy Jean Kott and engineered by Sarah Bruguer. 775 00:44:16,796 --> 00:44:20,316 Speaker 2: You can email us at problem at Pushkin dot FM. 776 00:44:20,476 --> 00:44:22,796 Speaker 2: I'm Jacob Goldstein and we'll be back next week with 777 00:44:22,876 --> 00:44:32,876 Speaker 2: another episode of What's Your Problem.