1 00:00:00,000 --> 00:00:02,160 Okay, so we're diving into some pretty intense stuff today, 2 00:00:02,160 --> 00:00:03,440 this whole deep seek thing. 3 00:00:03,440 --> 00:00:06,040 Yeah, it's wild, really. 4 00:00:06,040 --> 00:00:07,840 This company is making some big waves 5 00:00:07,840 --> 00:00:10,520 and they might just turn the whole AI world upside down. 6 00:00:10,520 --> 00:00:12,080 Absolutely, and the potential impact 7 00:00:12,080 --> 00:00:13,760 goes way beyond just software. 8 00:00:13,760 --> 00:00:15,560 We're talking about the chip industry even, 9 00:00:15,560 --> 00:00:17,320 our phones, our cars, everything. 10 00:00:17,320 --> 00:00:19,320 Right, the article we're looking at today. 11 00:00:19,320 --> 00:00:21,280 It's called, How Deep Seek Will Destroy 12 00:00:21,280 --> 00:00:22,680 the Entire Chip Industry. 13 00:00:22,680 --> 00:00:24,760 Catchy title, a little dramatic maybe, 14 00:00:24,760 --> 00:00:26,680 but it definitely grabs your attention, right? 15 00:00:26,680 --> 00:00:28,880 For sure, I mean, chips are in everything these days. 16 00:00:28,880 --> 00:00:31,120 So if this company's really about to disrupt 17 00:00:31,120 --> 00:00:33,240 the whole industry, we need to understand what's going on. 18 00:00:33,240 --> 00:00:35,000 Yeah, no doubt, and what's fascinating is 19 00:00:35,000 --> 00:00:37,240 this isn't some distant future we're talking about. 20 00:00:37,240 --> 00:00:38,480 This could have a real impact, 21 00:00:38,480 --> 00:00:40,000 like way sooner than we think. 22 00:00:40,000 --> 00:00:41,080 Right, exactly. 23 00:00:41,080 --> 00:00:43,200 And it all started with these whispers 24 00:00:43,200 --> 00:00:47,200 about this mysterious Chinese AI model, Deep Seek R1. 25 00:00:47,200 --> 00:00:48,080 Oh yeah, R1. 26 00:00:48,080 --> 00:00:50,120 I remember when those benchmarks first leaked, 27 00:00:50,120 --> 00:00:51,240 everyone was buzzing about it. 28 00:00:51,240 --> 00:00:53,680 It seemed to be right up there with like open AIs, 29 00:00:53,680 --> 00:00:55,920 GPT-4 in terms of performance. 30 00:00:55,920 --> 00:00:56,840 It was kind of shocking, right? 31 00:00:56,840 --> 00:00:58,840 Out of nowhere, this unknown player comes in 32 00:00:58,840 --> 00:01:00,200 and shakes things up. 33 00:01:00,200 --> 00:01:02,360 Even Sam Altman, the CEO of Open AI, 34 00:01:02,360 --> 00:01:04,760 seemed to acknowledge, even if suddenly, 35 00:01:04,760 --> 00:01:06,160 that something big was happening. 36 00:01:06,160 --> 00:01:07,760 Yeah, there was definitely a shift in the air, 37 00:01:07,760 --> 00:01:10,000 but honestly, what really set Deep Seek apart 38 00:01:10,000 --> 00:01:12,480 wasn't just the performance, it was how they achieved it. 39 00:01:12,480 --> 00:01:13,560 Right, their efficiency. 40 00:01:13,560 --> 00:01:16,080 Exactly, the article highlights how they managed 41 00:01:16,080 --> 00:01:17,760 to get that level of performance 42 00:01:17,760 --> 00:01:20,720 using way less computing power than their rivals. 43 00:01:20,720 --> 00:01:22,320 Like significantly less. 44 00:01:22,320 --> 00:01:23,880 Which has huge implications 45 00:01:23,880 --> 00:01:25,480 for accessibility and affordability. 46 00:01:25,480 --> 00:01:26,320 Oh, absolutely. 47 00:01:26,320 --> 00:01:27,160 I mean, think about it. 48 00:01:27,160 --> 00:01:29,640 Right now, companies like Open AL and Google, 49 00:01:29,640 --> 00:01:32,880 they control access to these super powerful AI models 50 00:01:32,880 --> 00:01:34,200 through their cloud services, 51 00:01:34,200 --> 00:01:36,440 and those services are expensive. 52 00:01:36,440 --> 00:01:38,520 And then Deep Seek drops this bombshell. 53 00:01:38,520 --> 00:01:40,000 R1 was gonna be open source. 54 00:01:40,000 --> 00:01:41,320 Boom, game changer. 55 00:01:41,320 --> 00:01:43,440 Totally, anyone can use it, develop it, 56 00:01:43,440 --> 00:01:45,400 without paying any licensing fees. 57 00:01:45,400 --> 00:01:46,680 It's a completely different model 58 00:01:46,680 --> 00:01:47,960 than what we've seen before. 59 00:01:47,960 --> 00:01:49,400 It's like they flip the script. 60 00:01:49,400 --> 00:01:51,560 Imagine suddenly startups, researchers, 61 00:01:51,560 --> 00:01:53,600 heck, even individual developers, 62 00:01:53,600 --> 00:01:56,040 they all have the power to compete with the tech giants. 63 00:01:56,040 --> 00:01:59,040 It's a huge potential shift in the whole AI landscape. 64 00:01:59,040 --> 00:02:00,920 And it all comes back to their background. 65 00:02:00,920 --> 00:02:03,320 Deep Seek didn't come from a traditional AI lab. 66 00:02:03,320 --> 00:02:06,320 They came from the world of quantitative trading. 67 00:02:06,320 --> 00:02:08,000 Yeah, that's a key piece of the puzzle. 68 00:02:08,000 --> 00:02:09,080 High stakes finance. 69 00:02:09,080 --> 00:02:11,080 Right, fractions of a second matter in that world. 70 00:02:11,080 --> 00:02:14,200 They're used to dealing with massive amounts of data, 71 00:02:14,200 --> 00:02:15,600 running complex algorithms, 72 00:02:15,600 --> 00:02:18,600 and optimizing for speed and efficiency, like constantly. 73 00:02:18,600 --> 00:02:20,160 It's like they were training for the AI race 74 00:02:20,160 --> 00:02:21,120 without even knowing it. 75 00:02:21,120 --> 00:02:23,360 Totally, they had this preexisting expertise 76 00:02:23,360 --> 00:02:26,440 in financial modeling and lightning fast calculations. 77 00:02:26,440 --> 00:02:28,400 And it turns out that was the perfect background 78 00:02:28,400 --> 00:02:29,480 for AI development. 79 00:02:29,480 --> 00:02:31,200 Yeah, they basically took that expertise 80 00:02:31,200 --> 00:02:33,000 and applied it to a whole new domain. 81 00:02:33,000 --> 00:02:35,120 And the result is Deep Seek R1, 82 00:02:35,120 --> 00:02:37,160 a model that's both incredibly powerful 83 00:02:37,160 --> 00:02:38,920 and remarkably efficient. 84 00:02:38,920 --> 00:02:41,840 Okay, so we've got this mysterious Chinese AI model 85 00:02:41,840 --> 00:02:43,440 developed by a former trading firm 86 00:02:43,440 --> 00:02:44,800 that's more efficient and open source 87 00:02:44,800 --> 00:02:46,560 than anything we've seen before. 88 00:02:46,560 --> 00:02:49,120 But how does the chip industry fit into all of this? 89 00:02:49,120 --> 00:02:51,120 Because the article title is pretty dire. 90 00:02:51,120 --> 00:02:54,000 They're claiming Deep Seek will destroy the entire chip industry. 91 00:02:54,000 --> 00:02:55,520 That's the million dollar question, 92 00:02:55,520 --> 00:02:58,520 and it all circles back to this idea of efficiency. 93 00:02:58,520 --> 00:03:01,320 Deep Seek R1 doesn't need as much computing power 94 00:03:01,320 --> 00:03:03,440 so you can run on less powerful hardware. 95 00:03:03,440 --> 00:03:06,120 It doesn't need those super expensive high-end chips 96 00:03:06,120 --> 00:03:07,960 that companies like NVIDIA have been relying on. 97 00:03:07,960 --> 00:03:09,920 Exactly, and that's where the potential 98 00:03:09,920 --> 00:03:11,280 for disruption comes in. 99 00:03:11,280 --> 00:03:13,720 If you don't need to buy the most powerful chips 100 00:03:13,720 --> 00:03:15,400 to run cutting edge AI, 101 00:03:15,400 --> 00:03:18,160 demand for those top tier chips plummets, 102 00:03:18,160 --> 00:03:20,120 that's a big problem for the chip industry. 103 00:03:20,120 --> 00:03:21,880 Especially since they've been betting on the idea 104 00:03:21,880 --> 00:03:24,400 that we'll always need more and more powerful chips, 105 00:03:24,400 --> 00:03:27,160 Deep Seek is challenging that assumption at its core. 106 00:03:27,160 --> 00:03:28,000 Think of it this way. 107 00:03:28,000 --> 00:03:30,200 The semiconductor industry, it's a giant, 108 00:03:30,200 --> 00:03:32,240 multi-trillion dollar giant. 109 00:03:32,240 --> 00:03:34,560 They're built on this idea of constant growth, 110 00:03:34,560 --> 00:03:37,560 this insatiable need for more processing power. 111 00:03:37,560 --> 00:03:38,680 Deep Seek comes along and says, 112 00:03:38,680 --> 00:03:41,160 hold on, what if we can do more with less? 113 00:03:41,160 --> 00:03:43,400 And suddenly the whole foundation of the industry 114 00:03:43,400 --> 00:03:44,240 is in question. 115 00:03:44,240 --> 00:03:46,080 That's the idea the article is exploring 116 00:03:46,080 --> 00:03:47,680 and it's not just some theoretical thing, 117 00:03:47,680 --> 00:03:49,680 it could actually have a real impact on you and me. 118 00:03:49,680 --> 00:03:51,160 Like in our everyday lives. 119 00:03:51,160 --> 00:03:52,920 Yeah, because chips are in everything. 120 00:03:52,920 --> 00:03:54,640 If the chip industry gets disrupted, 121 00:03:54,640 --> 00:03:57,800 we're all gonna feel it, our phones, our cars, everything. 122 00:03:57,800 --> 00:04:00,120 Deep Seek is challenging the status quo 123 00:04:00,120 --> 00:04:02,120 and the ripple effects could be massive. 124 00:04:02,120 --> 00:04:04,640 For sure, it's a wild time to be following AI, 125 00:04:04,640 --> 00:04:05,760 that's for sure. 126 00:04:05,760 --> 00:04:07,240 This whole Deep Seek situation, 127 00:04:07,240 --> 00:04:09,120 it's like watching the tectonic plates 128 00:04:09,120 --> 00:04:10,440 shift beneath our feet. 129 00:04:10,440 --> 00:04:12,080 Definitely, things are changing fast 130 00:04:12,080 --> 00:04:13,720 and Deep Seek is definitely one to watch. 131 00:04:13,720 --> 00:04:14,760 No doubt about it. 132 00:04:14,760 --> 00:04:16,520 We just dove into Deep Seek's background 133 00:04:16,520 --> 00:04:18,320 and this incredible AI model, 134 00:04:18,320 --> 00:04:20,960 Deep Seek R1 that's shaking up the industry. 135 00:04:20,960 --> 00:04:23,440 But I keep coming back to that bold article title, 136 00:04:23,440 --> 00:04:26,760 how Deep Seek will destroy the entire chip industry. 137 00:04:26,760 --> 00:04:29,120 Is it really possible they could have that much impact? 138 00:04:29,120 --> 00:04:30,760 It's a provocative claim, no doubt, 139 00:04:30,760 --> 00:04:33,080 but to understand why it's even being considered, 140 00:04:33,080 --> 00:04:35,400 we need to look closer at the technical innovations 141 00:04:35,400 --> 00:04:36,960 behind Deep Seek R1. 142 00:04:36,960 --> 00:04:39,200 The article mentions model distillation 143 00:04:39,200 --> 00:04:40,880 as key to their efficiency. 144 00:04:40,880 --> 00:04:42,960 It's a concept that's tough to grasp at first, 145 00:04:42,960 --> 00:04:44,640 so let's try to break it down. 146 00:04:44,640 --> 00:04:46,400 Imagine a massive library filled 147 00:04:46,400 --> 00:04:47,760 with every book ever written. 148 00:04:47,760 --> 00:04:50,120 Okay, I can picture that endless shelves, 149 00:04:50,120 --> 00:04:51,360 the smell of old paper. 150 00:04:51,360 --> 00:04:53,160 No, imagine you wanna find the answer 151 00:04:53,160 --> 00:04:54,640 to a specific question. 152 00:04:54,640 --> 00:04:57,120 You could spend years sifting through all those books, right? 153 00:04:57,120 --> 00:04:59,160 That's kinda like how traditional AI models work, 154 00:04:59,160 --> 00:05:01,960 they're huge and they require tons of processing power 155 00:05:01,960 --> 00:05:03,800 to search through all that data. 156 00:05:03,800 --> 00:05:06,760 Model distillation is like having a super librarian 157 00:05:06,760 --> 00:05:09,240 who can pinpoint the exact few books you need, 158 00:05:09,240 --> 00:05:12,320 giving you the answer much faster and with less effort. 159 00:05:12,320 --> 00:05:15,280 So Deep Seek found a way to streamline their AI, 160 00:05:15,280 --> 00:05:17,200 making it smaller and more focused 161 00:05:17,200 --> 00:05:19,880 without losing its ability to deliver powerful results. 162 00:05:19,880 --> 00:05:22,160 Exactly, they've managed to distill the knowledge 163 00:05:22,160 --> 00:05:26,040 of a massive AI model into a smaller, more efficient package. 164 00:05:26,040 --> 00:05:28,600 It's like finding the essence of that vast library, 165 00:05:28,600 --> 00:05:31,320 the core insights, and making them readily accessible. 166 00:05:31,320 --> 00:05:33,640 That makes sense, but there must be some trade-offs, right? 167 00:05:33,640 --> 00:05:36,320 Doesn't shrinking an AI model limit what it can do? 168 00:05:36,320 --> 00:05:38,320 There are always trade-offs in engineering. 169 00:05:38,320 --> 00:05:40,600 Typically, smaller models might be less accurate 170 00:05:40,600 --> 00:05:43,640 or capable of handling very complex tasks, 171 00:05:43,640 --> 00:05:45,600 but Deep Seek seems to have found a way 172 00:05:45,600 --> 00:05:47,480 to mitigate those limitations, 173 00:05:47,480 --> 00:05:49,760 potentially through innovative techniques 174 00:05:49,760 --> 00:05:52,280 like model pruning or quantization. 175 00:05:52,280 --> 00:05:53,560 Can you unpack those a bit for us? 176 00:05:53,560 --> 00:05:55,320 What do those techniques actually involve? 177 00:05:55,320 --> 00:05:57,280 Think of model pruning like removing 178 00:05:57,280 --> 00:05:58,960 the excess branches from a tree. 179 00:05:58,960 --> 00:06:00,680 You get one of the parts in the AI model 180 00:06:00,680 --> 00:06:02,720 that aren't contributing much to the overall results, 181 00:06:02,720 --> 00:06:05,000 streamlining it for better performance. 182 00:06:05,000 --> 00:06:07,640 Quantization is like rounding numbers. 183 00:06:07,640 --> 00:06:10,400 You slightly reduce the precision of calculations 184 00:06:10,400 --> 00:06:12,920 within the model, saving processing power 185 00:06:12,920 --> 00:06:15,600 without significantly impacting accuracy. 186 00:06:15,600 --> 00:06:17,200 So it's about finding clever ways 187 00:06:17,200 --> 00:06:19,960 to optimize every aspect of the AI model, 188 00:06:19,960 --> 00:06:23,200 squeezing out maximum performance with minimal resources. 189 00:06:23,200 --> 00:06:24,040 Precisely. 190 00:06:24,040 --> 00:06:25,520 And that's where Deep Seek's experience 191 00:06:25,520 --> 00:06:27,680 in quantitative trading really shines. 192 00:06:27,680 --> 00:06:29,880 They're used to finding tiny edges, 193 00:06:29,880 --> 00:06:31,960 optimizing algorithms to the extreme, 194 00:06:31,960 --> 00:06:34,040 because in the world of high-frequency trading, 195 00:06:34,040 --> 00:06:35,280 even milliseconds matter. 196 00:06:35,280 --> 00:06:37,760 They brought that same mindset to AI development, 197 00:06:37,760 --> 00:06:40,800 relentlessly pursuing efficiency at every step. 198 00:06:40,800 --> 00:06:43,400 And the result is Deep Seek R1, 199 00:06:43,400 --> 00:06:45,040 a model that challenges the assumption 200 00:06:45,040 --> 00:06:46,520 that bigger is always better. 201 00:06:46,520 --> 00:06:49,080 This shift toward efficiency could have ripple effects 202 00:06:49,080 --> 00:06:51,000 throughout the entire tech industry. 203 00:06:51,000 --> 00:06:53,080 Imagine a world where even our smartphones 204 00:06:53,080 --> 00:06:56,120 can handle incredibly complex AI tasks, 205 00:06:56,120 --> 00:06:58,120 all without needing a connection to the cloud 206 00:06:58,120 --> 00:06:59,960 or draining our batteries in an hour. 207 00:06:59,960 --> 00:07:02,120 That's a future I can definitely get behind. 208 00:07:02,120 --> 00:07:03,920 But what about the companies that have been profiting 209 00:07:03,920 --> 00:07:05,520 from selling those high-powered chips? 210 00:07:05,520 --> 00:07:06,760 That's where the destroy part 211 00:07:06,760 --> 00:07:08,480 of the article's title comes in. 212 00:07:08,480 --> 00:07:10,760 Deep Seek's efficiency gains directly threaten 213 00:07:10,760 --> 00:07:12,960 the business model of companies like NVIDIA, 214 00:07:12,960 --> 00:07:15,200 whose high-end GPUs have been essential 215 00:07:15,200 --> 00:07:17,480 for running these massive AI models. 216 00:07:17,480 --> 00:07:20,880 If Deep Seek R1 and models like it become the norm, 217 00:07:20,880 --> 00:07:23,160 demand for those expensive chips could plummet. 218 00:07:23,160 --> 00:07:25,160 Suddenly, everyone's scrambling to adapt. 219 00:07:25,160 --> 00:07:28,160 The article argues that the entire semiconductor industry, 220 00:07:28,160 --> 00:07:29,640 which has been built on this assumption 221 00:07:29,640 --> 00:07:32,000 of ever-increasing demand for more powerful chips, 222 00:07:32,000 --> 00:07:33,280 could be caught off guard. 223 00:07:33,280 --> 00:07:35,920 It's a dramatic claim, but not entirely implausible. 224 00:07:35,920 --> 00:07:38,560 And this isn't just about hypothetical future scenarios. 225 00:07:38,560 --> 00:07:41,600 The effects could be felt by everyday consumers very soon. 226 00:07:41,600 --> 00:07:44,560 Imagine your next smartphone being significantly cheaper 227 00:07:44,560 --> 00:07:47,800 because the cost of its AI chip has dropped dramatically. 228 00:07:47,800 --> 00:07:49,080 Or think about the possibilities 229 00:07:49,080 --> 00:07:50,960 for medical devices in remote areas. 230 00:07:50,960 --> 00:07:53,440 They could now leverage powerful AI capabilities 231 00:07:53,440 --> 00:07:55,840 without needing bulky, expensive hardware. 232 00:07:55,840 --> 00:07:57,280 It's like Deep Seek has thrown a wrench 233 00:07:57,280 --> 00:07:59,440 into the works of the entire tech ecosystem. 234 00:07:59,440 --> 00:08:00,880 And it's fascinating to consider 235 00:08:00,880 --> 00:08:02,400 how different players will react. 236 00:08:02,400 --> 00:08:05,160 Will chip manufacturers try to compete on efficiency? 237 00:08:05,160 --> 00:08:07,800 Or will they pivot to specialized AI hardware? 238 00:08:07,800 --> 00:08:09,320 Will we see a new wave of startups 239 00:08:09,320 --> 00:08:12,200 building AI solutions for low-power devices? 240 00:08:12,200 --> 00:08:14,960 These are all questions that experts are grappling with right now, 241 00:08:14,960 --> 00:08:17,400 and the answers will shape the technological landscape 242 00:08:17,400 --> 00:08:18,800 for years to come. 243 00:08:18,800 --> 00:08:20,280 But one thing is certain. 244 00:08:20,280 --> 00:08:23,520 Deep Seek has ignited a debate about the future of AI, 245 00:08:23,520 --> 00:08:25,600 and the implications are far-reaching. 246 00:08:25,600 --> 00:08:28,520 So it's not just about Deep Seek versus the chip industry. 247 00:08:28,520 --> 00:08:30,840 It's about a fundamental shift 248 00:08:30,840 --> 00:08:32,680 in how we think about AI development, 249 00:08:32,680 --> 00:08:35,080 a shift that could touch every aspect of our lives. 250 00:08:35,080 --> 00:08:37,440 This story goes far beyond just one company 251 00:08:37,440 --> 00:08:38,840 or one AI model. 252 00:08:38,840 --> 00:08:40,840 It's about a new era of AI, 253 00:08:40,840 --> 00:08:44,240 one where efficiency and accessibility are paramount, 254 00:08:44,240 --> 00:08:45,920 and it's unfolding right now. 255 00:08:45,920 --> 00:08:48,520 This global AI race is getting more interesting by the day, 256 00:08:48,520 --> 00:08:51,080 and Deep Seek has definitely shaken things up. 257 00:08:51,080 --> 00:08:52,640 But there's another layer to this story 258 00:08:52,640 --> 00:08:55,440 we need to explore the geopolitical dimension. 259 00:08:55,440 --> 00:08:58,440 China has made it clear that they have ambitions 260 00:08:58,440 --> 00:09:00,320 to lead the world in AI. 261 00:09:00,320 --> 00:09:03,160 How does Deep Seek fit into their larger strategy? 262 00:09:03,160 --> 00:09:05,840 So we're just discussing how Deep Seek's approach to AI 263 00:09:05,840 --> 00:09:08,640 is sending ripples throughout the tech industry. 264 00:09:08,640 --> 00:09:11,920 But this story's bigger than just chips and algorithms. 265 00:09:11,920 --> 00:09:15,520 It's all tangled up with this global race for AI dominance. 266 00:09:15,520 --> 00:09:16,720 And China seems to be playing 267 00:09:16,720 --> 00:09:18,480 a whole different game than the US. 268 00:09:18,480 --> 00:09:20,160 Yeah, it really is a stark contrast 269 00:09:20,160 --> 00:09:22,000 when you look at how the two are approaching this. 270 00:09:22,000 --> 00:09:25,400 The US has often focused on fostering innovation 271 00:09:25,400 --> 00:09:26,760 within its tech giants. 272 00:09:26,760 --> 00:09:28,880 But China, they've got this more 273 00:09:28,880 --> 00:09:30,920 centralized, strategic approach. 274 00:09:30,920 --> 00:09:31,760 It's very deliberate. 275 00:09:31,760 --> 00:09:33,480 They've set that clear national goal, 276 00:09:33,480 --> 00:09:36,440 becoming the world leader in AI by 2030. 277 00:09:36,440 --> 00:09:38,040 And Deep Seek's actions, 278 00:09:38,040 --> 00:09:40,560 well, they seem to fit right into that ambition. 279 00:09:40,560 --> 00:09:41,400 Oh, absolutely. 280 00:09:41,400 --> 00:09:44,560 And you pointed out how China's been pouring resources 281 00:09:44,560 --> 00:09:46,480 into AI research and development, 282 00:09:46,480 --> 00:09:47,760 creating these national labs, 283 00:09:47,760 --> 00:09:49,200 really pushing STEM education. 284 00:09:49,200 --> 00:09:50,720 It's like they're setting the stage 285 00:09:50,720 --> 00:09:52,920 for a massive AI ecosystem. 286 00:09:52,920 --> 00:09:54,400 Right, and then Deep Seek comes in 287 00:09:54,400 --> 00:09:56,000 with this open source approach. 288 00:09:56,000 --> 00:09:58,080 And it's a perfect fit for that strategy. 289 00:09:58,080 --> 00:10:00,680 By making their technology widely available, 290 00:10:00,680 --> 00:10:02,600 it's like they're throwing the doors open, 291 00:10:02,600 --> 00:10:04,760 attracting talent and collaboration from, 292 00:10:04,760 --> 00:10:06,000 well, everywhere, 293 00:10:06,000 --> 00:10:08,200 could really accelerate China's progress. 294 00:10:08,200 --> 00:10:09,560 It's clever, for sure. 295 00:10:09,560 --> 00:10:12,120 Especially when you contrast that with the more, 296 00:10:12,120 --> 00:10:13,240 I don't know, guarded approach 297 00:10:13,240 --> 00:10:14,920 we've seen from some US companies. 298 00:10:14,920 --> 00:10:15,960 It's almost like they're saying, 299 00:10:15,960 --> 00:10:16,960 hey, come join us. 300 00:10:16,960 --> 00:10:18,920 Let's build the future of AI together. 301 00:10:18,920 --> 00:10:20,400 But at the same time, 302 00:10:20,400 --> 00:10:22,920 they're subtly shifting the whole center of gravity 303 00:10:22,920 --> 00:10:24,440 in the AI world, you know? 304 00:10:24,440 --> 00:10:26,280 Yeah, it's a very subtle power play. 305 00:10:26,280 --> 00:10:27,200 And this approach, 306 00:10:27,200 --> 00:10:30,200 it could really reshape the global AI landscape. 307 00:10:30,200 --> 00:10:31,040 Think about it. 308 00:10:31,040 --> 00:10:33,120 It has the potential to empower those 309 00:10:33,120 --> 00:10:35,760 who've traditionally been on the sidelines of AI development. 310 00:10:35,760 --> 00:10:38,320 Like, we could see new hubs of innovation 311 00:10:38,320 --> 00:10:40,800 popping up in unexpected places. 312 00:10:40,800 --> 00:10:44,400 It could create a more multipolar world. 313 00:10:44,400 --> 00:10:47,240 That's what I find so fascinating about this whole thing. 314 00:10:47,240 --> 00:10:50,200 It's not just about like, who builds the best AI, 315 00:10:50,200 --> 00:10:52,120 but who sets the rules, right? 316 00:10:52,120 --> 00:10:54,320 Who gets to define the ethical standards, 317 00:10:54,320 --> 00:10:56,760 who controls the flow of data and expertise, 318 00:10:56,760 --> 00:10:58,400 and Deep Seek, 319 00:10:58,400 --> 00:11:00,680 they've thrown a real wild card into the mix. 320 00:11:00,680 --> 00:11:03,720 They really have, and the implications, they're huge. 321 00:11:03,720 --> 00:11:06,800 We're seeing this shift in how nations view AI, 322 00:11:06,800 --> 00:11:09,240 a move away from just the economics of it all 323 00:11:09,240 --> 00:11:11,960 to something much more strategic, even ideological. 324 00:11:11,960 --> 00:11:14,600 Like, it's becoming this battle of ideas almost. 325 00:11:14,600 --> 00:11:15,840 And for the listener, 326 00:11:15,840 --> 00:11:20,640 it's this reminder that AI isn't just some tech industry trend. 327 00:11:20,640 --> 00:11:23,520 It's this force that's changing the world. 328 00:11:23,520 --> 00:11:25,320 And the choices we make now, 329 00:11:25,320 --> 00:11:27,320 they're gonna have ripple effects for years to come. 330 00:11:27,320 --> 00:11:28,440 What we're seeing with Deep Seek, 331 00:11:28,440 --> 00:11:31,040 it's like just a grim of what's to come. 332 00:11:31,040 --> 00:11:32,800 This whole thing, it raises questions 333 00:11:32,800 --> 00:11:34,840 that go way beyond AI models, 334 00:11:34,840 --> 00:11:37,800 or how fast a ship can process data. 335 00:11:37,800 --> 00:11:42,640 It's about power, control, the future of our digital world. 336 00:11:42,640 --> 00:11:46,160 It's been a mind-bending deep dive, to say the least. 337 00:11:46,160 --> 00:11:48,560 And I gotta say, I think I've got more questions now 338 00:11:48,560 --> 00:11:53,400 than answers, but honestly, I think that's a good thing, right? 339 00:11:53,400 --> 00:11:54,240 I agree. 340 00:11:54,240 --> 00:11:55,760 It's good to be a little uncomfortable 341 00:11:55,760 --> 00:11:57,840 to acknowledge what we don't know. 342 00:11:57,840 --> 00:11:59,880 This landscape is changing so fast, 343 00:11:59,880 --> 00:12:02,280 the important thing is to stay curious, 344 00:12:02,280 --> 00:12:05,280 keep learning, and keep having these conversations. 345 00:12:05,280 --> 00:12:07,760 So that wraps up our deep dive on Deep Seek 346 00:12:07,760 --> 00:12:09,040 and the future of AI. 347 00:12:09,040 --> 00:12:11,800 We hope this conversation, you know, got you thinking. 348 00:12:11,800 --> 00:12:14,080 Remember, the future isn't set in stone. 349 00:12:14,080 --> 00:12:16,920 The choices we make today, they matter. 350 00:12:16,920 --> 00:12:18,840 So stay informed, stay engaged, 351 00:12:18,840 --> 00:12:28,840 and help shape this future of AI.