1 00:00:00,000 --> 00:00:01,080 All right, let's dive in. 2 00:00:01,080 --> 00:00:04,000 Today we're gonna be talking about AI infrastructure. 3 00:00:04,000 --> 00:00:05,160 Ooh, sounds exciting. 4 00:00:05,160 --> 00:00:06,080 It is actually. 5 00:00:06,080 --> 00:00:08,000 It might not sound like the most thrilling topic, but. 6 00:00:08,000 --> 00:00:08,840 I know, right? 7 00:00:08,840 --> 00:00:09,680 Infrastructure. 8 00:00:09,680 --> 00:00:11,200 Yeah, but we're basically talking about like 9 00:00:11,200 --> 00:00:13,080 the keys to the AI kingdom, right? 10 00:00:13,080 --> 00:00:14,520 Who's building this tech? 11 00:00:14,520 --> 00:00:15,560 Who's partnering up? 12 00:00:15,560 --> 00:00:18,720 And what are the stakes in this whole global power play? 13 00:00:18,720 --> 00:00:19,560 High stakes. 14 00:00:19,560 --> 00:00:20,400 Super high. 15 00:00:20,400 --> 00:00:23,480 I actually, I've got some notes here from a conversation 16 00:00:23,480 --> 00:00:26,480 with an AI expert who is just at this huge 17 00:00:26,480 --> 00:00:28,360 international summit in Paris. 18 00:00:28,360 --> 00:00:29,200 Oh, wow. 19 00:00:29,200 --> 00:00:30,920 Yeah, and things are getting really interesting. 20 00:00:30,920 --> 00:00:31,760 I bet. 21 00:00:31,760 --> 00:00:33,440 So one of the things that really struck me was 22 00:00:33,440 --> 00:00:35,320 how AI is being talked about now. 23 00:00:35,320 --> 00:00:38,400 It's not just like, you know, another cool gadget. 24 00:00:38,400 --> 00:00:39,240 Right. 25 00:00:39,240 --> 00:00:42,360 It's being recognized as this like general purpose technology. 26 00:00:42,360 --> 00:00:43,920 Like electricity or the printing press? 27 00:00:43,920 --> 00:00:44,760 Exactly. 28 00:00:44,760 --> 00:00:46,520 Those are like huge game changers, right? 29 00:00:46,520 --> 00:00:47,360 Exactly. 30 00:00:47,360 --> 00:00:48,760 And that's what's so fascinating is like, 31 00:00:48,760 --> 00:00:50,880 it's being recognized as on that level now. 32 00:00:50,880 --> 00:00:52,840 It's not just some, you know, fun little thing. 33 00:00:52,840 --> 00:00:53,680 It's like. 34 00:00:53,680 --> 00:00:54,520 It's fundamental. 35 00:00:54,520 --> 00:00:55,360 Yeah, it really is. 36 00:00:55,360 --> 00:00:56,400 So help me understand this a little bit more. 37 00:00:56,400 --> 00:00:59,440 What's the big deal to have this general purpose idea? 38 00:00:59,440 --> 00:01:01,240 Why is that comparison so important? 39 00:01:01,240 --> 00:01:02,360 Well, you know, think about it. 40 00:01:02,360 --> 00:01:05,240 Electricity completely changed how we live and work. 41 00:01:05,240 --> 00:01:08,440 The printing press transformed how we share information. 42 00:01:08,440 --> 00:01:10,240 And it's not just that they were inventions. 43 00:01:10,240 --> 00:01:12,120 It's that they completely reshaped the world. 44 00:01:12,120 --> 00:01:13,600 Like how we interact with the world. 45 00:01:13,600 --> 00:01:14,640 Now they changed everything. 46 00:01:14,640 --> 00:01:15,480 They changed everything. 47 00:01:15,480 --> 00:01:17,080 And that's what, you know, world leaders 48 00:01:17,080 --> 00:01:18,800 are starting to see with AI. 49 00:01:18,800 --> 00:01:20,000 They're seeing that potential. 50 00:01:20,000 --> 00:01:20,840 Exactly. 51 00:01:20,840 --> 00:01:23,520 It's like this key to unlocking, you know, all this progress. 52 00:01:23,520 --> 00:01:25,520 And so, you know, you had both President Macron 53 00:01:25,520 --> 00:01:28,280 and Prime Minister Modi at this summit in Paris. 54 00:01:28,280 --> 00:01:30,400 And both of them were emphasizing 55 00:01:30,400 --> 00:01:34,280 how absolutely crucial it is for any country that wants to, 56 00:01:34,280 --> 00:01:35,520 you know, stay ahead of the curve. 57 00:01:35,520 --> 00:01:38,080 So it's not just about having a cool AI gadgets. 58 00:01:38,080 --> 00:01:40,560 It's about having the actual infrastructure 59 00:01:40,560 --> 00:01:42,760 to build and control that technology. 60 00:01:42,760 --> 00:01:43,600 100%. 61 00:01:43,600 --> 00:01:46,840 Yeah, there's this increasing sense of infrastructure 62 00:01:46,840 --> 00:01:48,120 independence. 63 00:01:48,120 --> 00:01:50,200 Infrastructure independence. 64 00:01:50,200 --> 00:01:53,480 Like this idea that countries are realizing that they can't just 65 00:01:53,480 --> 00:01:56,280 rely on other countries for their AI needs. 66 00:01:56,280 --> 00:01:58,840 You know, they need to have that foundation themselves. 67 00:01:58,840 --> 00:02:01,360 You know, the chips, the data centers, the models, all of it. 68 00:02:01,360 --> 00:02:02,360 They want to own it. 69 00:02:02,360 --> 00:02:03,200 They want to own it. 70 00:02:03,200 --> 00:02:05,000 Yeah, they don't want to be dependent on someone else 71 00:02:05,000 --> 00:02:06,200 for something this crucial. 72 00:02:06,200 --> 00:02:08,840 OK, so where does the US fit into all of this? 73 00:02:08,840 --> 00:02:10,880 Because, you know, we're always talking about, you know, 74 00:02:10,880 --> 00:02:12,600 competition, especially with China. 75 00:02:12,600 --> 00:02:15,720 Are we talking like another space race, but with AI this time? 76 00:02:15,720 --> 00:02:16,640 You could definitely say that. 77 00:02:16,640 --> 00:02:19,840 Yeah, the US is very, very focused on keeping its lead, 78 00:02:19,840 --> 00:02:21,720 especially with China making huge strides. 79 00:02:21,720 --> 00:02:22,040 Right. 80 00:02:22,040 --> 00:02:24,640 And there's a really big push to ensure that AI models that 81 00:02:24,640 --> 00:02:28,520 are American made or by our allies or open source 82 00:02:28,520 --> 00:02:30,760 can really hold their own against anything else out there. 83 00:02:30,760 --> 00:02:32,160 Compete on a global stage. 84 00:02:32,160 --> 00:02:32,800 Exactly. 85 00:02:32,800 --> 00:02:36,640 Yeah, Judy Vance, who is an expert on the USAI strategy, 86 00:02:36,640 --> 00:02:38,840 she spoke at the summit about this very thing. 87 00:02:38,840 --> 00:02:39,520 That's fascinating. 88 00:02:39,520 --> 00:02:41,200 So the US wants to make sure it can compete. 89 00:02:41,200 --> 00:02:42,240 But what about everyone else? 90 00:02:42,240 --> 00:02:44,160 What about other countries who, you know, 91 00:02:44,160 --> 00:02:46,640 maybe don't have the resources of a superpower? 92 00:02:46,640 --> 00:02:47,960 What options do they have? 93 00:02:47,960 --> 00:02:50,200 That's where it gets really strategic, right? 94 00:02:50,200 --> 00:02:52,440 Because they basically have a choice to make. 95 00:02:52,440 --> 00:02:55,400 Build their own AI infrastructure from scratch, 96 00:02:55,400 --> 00:02:58,200 buy it from someone else, or partner with another nation. 97 00:02:58,200 --> 00:02:59,560 Build, buy, or partner. 98 00:02:59,560 --> 00:03:01,600 So it's like a high stakes game of risk. 99 00:03:01,600 --> 00:03:02,280 It kind of is. 100 00:03:02,280 --> 00:03:04,280 And the stakes are huge because, you know, 101 00:03:04,280 --> 00:03:05,640 this isn't just about technology. 102 00:03:05,640 --> 00:03:07,000 It's about global power. 103 00:03:07,000 --> 00:03:07,640 Absolutely. 104 00:03:07,640 --> 00:03:09,560 So let's unpack these options a little bit. 105 00:03:09,560 --> 00:03:10,720 Let's talk about build. 106 00:03:10,720 --> 00:03:13,720 Building your own infrastructure sounds incredibly expensive. 107 00:03:13,720 --> 00:03:14,840 It's massively expensive. 108 00:03:14,840 --> 00:03:15,600 It's complex. 109 00:03:15,600 --> 00:03:16,560 Yeah, you're not wrong. 110 00:03:16,560 --> 00:03:19,280 It requires a ton of investment, like top talent. 111 00:03:19,280 --> 00:03:21,600 You need access to all the cutting edge resources. 112 00:03:21,600 --> 00:03:22,560 Not for the same of heart. 113 00:03:22,560 --> 00:03:23,680 Definitely not, yeah. 114 00:03:23,680 --> 00:03:25,400 I mean, to be a real player in AI, 115 00:03:25,400 --> 00:03:27,760 you need what are being called hyper centers. 116 00:03:27,760 --> 00:03:28,520 Hyper centers. 117 00:03:28,520 --> 00:03:29,560 What are those? 118 00:03:29,560 --> 00:03:31,040 So it's kind of like, you know, imagine 119 00:03:31,040 --> 00:03:33,080 a super powered Silicon Valley. 120 00:03:33,080 --> 00:03:34,280 Oh, OK. 121 00:03:34,280 --> 00:03:36,440 Right, they're these regions with just massive computing 122 00:03:36,440 --> 00:03:39,480 power, tons of data, the best talent, you know, 123 00:03:39,480 --> 00:03:40,120 the whole package. 124 00:03:40,120 --> 00:03:41,040 The whole shebang. 125 00:03:41,040 --> 00:03:41,560 Yeah. 126 00:03:41,560 --> 00:03:43,520 And right now, honestly, the US and China 127 00:03:43,520 --> 00:03:45,280 are the only real hyper centers. 128 00:03:45,280 --> 00:03:46,800 So if you're not one of those two, 129 00:03:46,800 --> 00:03:48,800 build is a long and costly road. 130 00:03:48,800 --> 00:03:50,200 It is, yeah. 131 00:03:50,200 --> 00:03:51,280 Really tough to pull off. 132 00:03:51,280 --> 00:03:53,720 OK, so maybe buy is a little bit more appealing. 133 00:03:53,720 --> 00:03:56,040 You just, you know, cut a check and get your AI 134 00:03:56,040 --> 00:03:57,560 infrastructure delivered, right? 135 00:03:57,560 --> 00:03:57,960 Right. 136 00:03:57,960 --> 00:03:59,160 It's tempting, for sure. 137 00:03:59,160 --> 00:04:00,600 I mean, you skip the whole development phase, 138 00:04:00,600 --> 00:04:01,600 you can get started faster. 139 00:04:01,600 --> 00:04:02,680 But there's a but. 140 00:04:02,680 --> 00:04:03,400 There's always a but. 141 00:04:03,400 --> 00:04:05,440 You're basically handing over a lot of control. 142 00:04:05,440 --> 00:04:08,320 You become reliant on the vendor for updates, support, 143 00:04:08,320 --> 00:04:09,640 potentially even security. 144 00:04:09,640 --> 00:04:10,400 Right. 145 00:04:10,400 --> 00:04:12,320 And you might be limited in what you can actually do 146 00:04:12,320 --> 00:04:13,160 with the tech. 147 00:04:13,160 --> 00:04:15,440 So buy comes with strings attached. 148 00:04:15,440 --> 00:04:16,480 Definitely does. 149 00:04:16,480 --> 00:04:18,120 All right, so what about partner? 150 00:04:18,120 --> 00:04:19,400 What's the upside there? 151 00:04:19,400 --> 00:04:20,840 Partnering can be a really smart move, 152 00:04:20,840 --> 00:04:22,360 especially if you're a country that 153 00:04:22,360 --> 00:04:24,560 doesn't have all the resources to go in alone. 154 00:04:24,560 --> 00:04:26,400 By pooling expertise and resources, 155 00:04:26,400 --> 00:04:28,440 you can get into the AI game faster 156 00:04:28,440 --> 00:04:30,800 and carve out your own little niche. 157 00:04:30,800 --> 00:04:34,000 This sounds especially interesting for US allies, 158 00:04:34,000 --> 00:04:35,320 particularly those in Europe. 159 00:04:35,320 --> 00:04:38,920 I mean, could we see some transatlantic AI power 160 00:04:38,920 --> 00:04:40,040 couples emerging? 161 00:04:40,040 --> 00:04:41,040 It's a real possibility. 162 00:04:41,040 --> 00:04:44,320 There's a ton of potential for US-Europe collaboration, 163 00:04:44,320 --> 00:04:47,400 especially if they share values and concerns about AI. 164 00:04:47,400 --> 00:04:49,040 We could see countries working together 165 00:04:49,040 --> 00:04:51,520 on different levels of that whole infrastructure stack, 166 00:04:51,520 --> 00:04:53,280 like some specialize in the chips, 167 00:04:53,280 --> 00:04:55,120 others focus on specific models. 168 00:04:55,120 --> 00:04:57,440 There's this really interesting vibe shift at the summit 169 00:04:57,440 --> 00:04:58,240 from what I heard. 170 00:04:58,240 --> 00:04:58,680 Oh. 171 00:04:58,680 --> 00:04:59,080 Yeah. 172 00:04:59,080 --> 00:05:03,200 Previous summits were like really pessimistic about AI risks. 173 00:05:03,200 --> 00:05:05,200 But in Paris, there was this definite sense of, hey, 174 00:05:05,200 --> 00:05:07,480 maybe we can work together and actually make this thing 175 00:05:07,480 --> 00:05:08,520 benefit everyone. 176 00:05:08,520 --> 00:05:10,920 That's a really interesting change in attitude. 177 00:05:10,920 --> 00:05:12,560 So what about France specifically? 178 00:05:12,560 --> 00:05:13,920 I mean, they hosted the summit. 179 00:05:13,920 --> 00:05:15,000 What are their game plan? 180 00:05:15,000 --> 00:05:17,560 How are they building up their own AI infrastructure? 181 00:05:17,560 --> 00:05:19,360 France is definitely making moves, yeah. 182 00:05:19,360 --> 00:05:20,960 They have this strength in nuclear energy, 183 00:05:20,960 --> 00:05:23,360 which is crucial for those power hungry data centers. 184 00:05:23,360 --> 00:05:24,320 Oh, yeah, of course. 185 00:05:24,320 --> 00:05:26,320 But they really need to catch up on actually building 186 00:05:26,320 --> 00:05:27,440 the data centers themselves. 187 00:05:27,440 --> 00:05:29,400 So it sounds like they might be looking to partner 188 00:05:29,400 --> 00:05:30,320 to fill that gap. 189 00:05:30,320 --> 00:05:31,440 Yeah, they're definitely open to it. 190 00:05:31,440 --> 00:05:33,840 They're already talking with American chip companies, 191 00:05:33,840 --> 00:05:37,240 like NVIDIA, which are major players in the AI hardware 192 00:05:37,240 --> 00:05:37,760 world. 193 00:05:37,760 --> 00:05:38,280 Makes sense. 194 00:05:38,280 --> 00:05:41,280 OK, so we've got this global race for AI dominance, right? 195 00:05:41,280 --> 00:05:43,960 Countries choosing to build, buy, or partner. 196 00:05:43,960 --> 00:05:45,760 But there's another layer to this, isn't there? 197 00:05:45,760 --> 00:05:48,760 This whole open source versus closed source debate. 198 00:05:48,760 --> 00:05:50,120 What's the latest on that? 199 00:05:50,120 --> 00:05:50,800 It's a big one. 200 00:05:50,800 --> 00:05:53,520 And the gap between those two approaches 201 00:05:53,520 --> 00:05:56,440 is shrinking fast, which is really shaking things up. 202 00:05:56,440 --> 00:05:58,600 Remind me, what's the difference between open source 203 00:05:58,600 --> 00:06:00,520 and closed source AI? 204 00:06:00,520 --> 00:06:03,280 So open source AI is like a shared recipe. 205 00:06:03,280 --> 00:06:04,880 Anyone can access the code. 206 00:06:04,880 --> 00:06:07,680 They can tweak it, build their own applications on top of it. 207 00:06:07,680 --> 00:06:09,880 Close source is more like a secret recipe, right? 208 00:06:09,880 --> 00:06:12,480 Developed by a specific company, kept under wraps. 209 00:06:12,480 --> 00:06:15,840 So open is about collaboration, closed is about control. 210 00:06:15,840 --> 00:06:16,440 Exactly. 211 00:06:16,440 --> 00:06:18,360 And for a while, there was this assumption 212 00:06:18,360 --> 00:06:20,880 that closed source models, especially those 213 00:06:20,880 --> 00:06:23,920 by big US tech companies, were just so advanced, 214 00:06:23,920 --> 00:06:25,800 open source couldn't compete. 215 00:06:25,800 --> 00:06:27,200 But that's changing. 216 00:06:27,200 --> 00:06:27,600 It is. 217 00:06:27,600 --> 00:06:29,080 Open source is catching up fast. 218 00:06:29,080 --> 00:06:30,600 Think about chat GPT. 219 00:06:30,600 --> 00:06:32,400 Oh, yeah, the chat bot, everyone's talking about that. 220 00:06:32,400 --> 00:06:35,600 That's the one released December 5th by Open AI. 221 00:06:35,600 --> 00:06:38,760 Less than two months later, DeepSeek, a Chinese company, 222 00:06:38,760 --> 00:06:40,080 released a comparable model. 223 00:06:40,080 --> 00:06:40,960 Wow. 224 00:06:40,960 --> 00:06:42,160 That's a tight turnaround. 225 00:06:42,160 --> 00:06:42,660 It is. 226 00:06:42,660 --> 00:06:43,880 Things are moving so fast. 227 00:06:43,880 --> 00:06:45,960 OK, so I got to ask, what does all of this 228 00:06:45,960 --> 00:06:49,560 mean for me or anyone else who's not an AI expert? 229 00:06:49,560 --> 00:06:52,120 But trying to understand where this is all going, 230 00:06:52,120 --> 00:06:54,000 what should we be paying attention to? 231 00:06:54,000 --> 00:06:54,840 It's a great question. 232 00:06:54,840 --> 00:06:57,160 And it leads us to this really interesting perspective 233 00:06:57,160 --> 00:07:01,000 from a venture capital firm, Andreessen Hortowitz, or A16Z, 234 00:07:01,000 --> 00:07:01,560 as they're called. 235 00:07:01,560 --> 00:07:02,520 A16Z, OK. 236 00:07:02,520 --> 00:07:05,000 They're big in the tech world, heavily invested, 237 00:07:05,000 --> 00:07:06,960 and they're putting a lot of faith, a lot of money, 238 00:07:06,960 --> 00:07:08,320 into open source AI. 239 00:07:08,320 --> 00:07:09,040 I'm intrigued. 240 00:07:09,040 --> 00:07:13,840 Why, what makes a firm like that so bullish on open source? 241 00:07:13,840 --> 00:07:15,560 Well, it's not just idealism, though. 242 00:07:15,560 --> 00:07:17,960 They do have this strong belief in making technology 243 00:07:17,960 --> 00:07:18,800 accessible. 244 00:07:18,800 --> 00:07:19,840 They're savvy investors. 245 00:07:19,840 --> 00:07:22,520 They see open source as this potential gold mine. 246 00:07:22,520 --> 00:07:23,200 A gold mine? 247 00:07:23,200 --> 00:07:24,040 How so? 248 00:07:24,040 --> 00:07:26,560 How do you make money off something that's freely available? 249 00:07:26,560 --> 00:07:27,440 Think about it like this. 250 00:07:27,440 --> 00:07:30,760 Open source software has been the foundation of the internet, 251 00:07:30,760 --> 00:07:31,320 like Linux. 252 00:07:31,320 --> 00:07:33,640 It powers a huge chunk of the web. 253 00:07:33,640 --> 00:07:35,400 Companies build businesses on top of it. 254 00:07:35,400 --> 00:07:36,800 They offer services, support. 255 00:07:36,800 --> 00:07:38,200 They customize it. 256 00:07:38,200 --> 00:07:40,600 A16Z believes the same can happen with AI. 257 00:07:40,600 --> 00:07:43,040 The open source AI provides the base, 258 00:07:43,040 --> 00:07:46,040 then companies specialize, build specific stuff 259 00:07:46,040 --> 00:07:46,640 on top of it. 260 00:07:46,640 --> 00:07:47,120 Exactly. 261 00:07:47,120 --> 00:07:48,800 It's about creating this whole ecosystem 262 00:07:48,800 --> 00:07:51,440 where innovation just flourishes. 263 00:07:51,440 --> 00:07:53,320 And history shows us that open source 264 00:07:53,320 --> 00:07:56,480 can be a seriously powerful driver of progress. 265 00:07:56,480 --> 00:07:57,000 OK. 266 00:07:57,000 --> 00:07:57,560 That makes sense. 267 00:07:57,560 --> 00:07:59,400 But I'm sure there are still people out there who believe 268 00:07:59,400 --> 00:08:01,120 that their real breakthroughs are going 269 00:08:01,120 --> 00:08:04,800 to come from those big, well-funded, closed AI labs. 270 00:08:04,800 --> 00:08:06,000 Oh, absolutely, there are. 271 00:08:06,000 --> 00:08:08,160 And those labs have definitely had some amazing results. 272 00:08:08,160 --> 00:08:08,920 No doubt. 273 00:08:08,920 --> 00:08:12,360 But the open source community is like right on their heels. 274 00:08:12,360 --> 00:08:13,720 And there's this argument to be made 275 00:08:13,720 --> 00:08:15,880 that open source might actually be better for handling 276 00:08:15,880 --> 00:08:17,680 all the ethical challenges we talked about. 277 00:08:17,680 --> 00:08:18,680 Oh, how so? 278 00:08:18,680 --> 00:08:19,440 Transparency. 279 00:08:19,440 --> 00:08:20,280 It's key. 280 00:08:20,280 --> 00:08:22,080 With open source models, anyone can 281 00:08:22,080 --> 00:08:23,600 see how the algorithms work. 282 00:08:23,600 --> 00:08:25,920 They can look for biases, suggest improvements. 283 00:08:25,920 --> 00:08:28,040 It's like this global team of auditors, 284 00:08:28,040 --> 00:08:29,760 constantly scrutinizing the tech. 285 00:08:29,760 --> 00:08:31,120 The more eyes, the better. 286 00:08:31,120 --> 00:08:32,080 Exactly. 287 00:08:32,080 --> 00:08:33,840 Catch those potential problems early. 288 00:08:33,840 --> 00:08:37,280 So A16Z sees this as not just a business strategy, 289 00:08:37,280 --> 00:08:40,280 but a way to steer AI in a more ethical direction. 290 00:08:40,280 --> 00:08:41,000 Exactly, yeah. 291 00:08:41,000 --> 00:08:43,400 Building AI that's powerful, but also responsible. 292 00:08:43,400 --> 00:08:44,720 AI for good, as you said. 293 00:08:44,720 --> 00:08:45,160 That's it. 294 00:08:45,160 --> 00:08:45,760 That's the goal. 295 00:08:45,760 --> 00:08:47,160 OK, so let's zoom out for a second. 296 00:08:47,160 --> 00:08:48,760 We talked about the US and China 297 00:08:48,760 --> 00:08:52,000 being those dominant AI hyper-centers right now. 298 00:08:52,000 --> 00:08:53,280 What are the implications of that? 299 00:08:53,280 --> 00:08:56,600 Are we heading toward a winner-takes-all situation? 300 00:08:56,600 --> 00:08:59,080 It's definitely a huge point of attention, for sure. 301 00:08:59,080 --> 00:09:03,280 Both countries see AI as crucial for the economy, 302 00:09:03,280 --> 00:09:04,120 for their militaries. 303 00:09:04,120 --> 00:09:05,000 Right, makes sense. 304 00:09:05,000 --> 00:09:06,800 But I wouldn't say it's a simple winner-takes-all. 305 00:09:06,800 --> 00:09:10,680 Remember, other countries have choices, build, buy, or partner. 306 00:09:10,680 --> 00:09:13,520 And for many, partnering seems like the most attractive option. 307 00:09:13,520 --> 00:09:15,120 Especially for US allies, right? 308 00:09:15,120 --> 00:09:18,400 It lets them tap into American tech and expertise 309 00:09:18,400 --> 00:09:20,600 while still having some control. 310 00:09:20,600 --> 00:09:21,160 Exactly. 311 00:09:21,160 --> 00:09:22,880 And we're already seeing this happen. 312 00:09:22,880 --> 00:09:25,680 The US is encouraging collaborations, especially with Europe. 313 00:09:25,680 --> 00:09:29,240 There's a lot of talk about a transatlantic AI alliance 314 00:09:29,240 --> 00:09:31,000 to counter China's influence. 315 00:09:31,000 --> 00:09:31,600 Interesting. 316 00:09:31,600 --> 00:09:34,320 So how does all of this play out for France specifically? 317 00:09:34,320 --> 00:09:37,840 We talked about their strengths and weaknesses with AI infrastructure. 318 00:09:37,840 --> 00:09:39,200 What role do you see them playing? 319 00:09:39,200 --> 00:09:40,640 France has a lot to offer, yeah? 320 00:09:40,640 --> 00:09:42,880 Strong research, talented workforce, 321 00:09:42,880 --> 00:09:45,240 and a government that's committed to investing in AI. 322 00:09:45,240 --> 00:09:47,520 But as we said, they've got to catch up on data centers. 323 00:09:47,520 --> 00:09:49,080 And that's where those partnerships come in. 324 00:09:49,080 --> 00:09:51,400 Right, like with the US, with American ship companies. 325 00:09:51,400 --> 00:09:51,960 Exactly. 326 00:09:51,960 --> 00:09:55,080 They could become like a hub for AI research in Europe, 327 00:09:55,080 --> 00:09:56,640 attract talent and investment. 328 00:09:56,640 --> 00:09:58,240 So yeah, I think they have a big role to play. 329 00:09:58,240 --> 00:10:00,400 This is painting a really complex picture. 330 00:10:00,400 --> 00:10:02,960 Nations competing open source versus closed source, 331 00:10:02,960 --> 00:10:04,840 the race to build hyper centers. 332 00:10:04,840 --> 00:10:08,080 It feels like the future of AI is still very much up for grabs. 333 00:10:08,080 --> 00:10:08,920 It absolutely is. 334 00:10:08,920 --> 00:10:11,040 And that's what makes it so fascinating to watch. 335 00:10:11,040 --> 00:10:12,960 The decisions made in the next few years 336 00:10:12,960 --> 00:10:15,560 will shape how AI impacts all of us. 337 00:10:15,560 --> 00:10:17,120 All right, so let's talk about those decisions. 338 00:10:17,120 --> 00:10:20,960 We keep coming back to this build, buy, or partner dilemma. 339 00:10:20,960 --> 00:10:23,400 What advice would you give to a country that's 340 00:10:23,400 --> 00:10:25,120 trying to figure out their AI strategy? 341 00:10:25,120 --> 00:10:26,960 What are the things they need to consider? 342 00:10:26,960 --> 00:10:29,880 That's a tough one, because there's no one-size-fits-all answer. 343 00:10:29,880 --> 00:10:33,400 All depends on the resources, their goals, their risk tolerance. 344 00:10:33,400 --> 00:10:35,320 But if I had to offer some general guidance, 345 00:10:35,320 --> 00:10:39,320 I'd say first, be realistic about your strengths and weaknesses. 346 00:10:39,320 --> 00:10:41,080 Don't try to be everything to everyone. 347 00:10:41,080 --> 00:10:41,960 Play to your strengths. 348 00:10:41,960 --> 00:10:42,480 Exactly. 349 00:10:42,480 --> 00:10:43,600 Find your niche. 350 00:10:43,600 --> 00:10:45,080 Second, think long term. 351 00:10:45,080 --> 00:10:45,920 This isn't a sprint. 352 00:10:45,920 --> 00:10:47,200 It's a marathon. 353 00:10:47,200 --> 00:10:50,200 Building a solid AI ecosystem takes time. 354 00:10:50,200 --> 00:10:52,040 Yeah, I've got to have that sustained investment. 355 00:10:52,040 --> 00:10:52,640 You do. 356 00:10:52,640 --> 00:10:54,280 No overnight miracles here. 357 00:10:54,280 --> 00:10:55,920 And finally, don't be afraid to collaborate. 358 00:10:55,920 --> 00:10:57,880 Partnering can really speed things up. 359 00:10:57,880 --> 00:10:59,640 Access to resources and expertise, 360 00:10:59,640 --> 00:11:00,840 you might not have yourself. 361 00:11:00,840 --> 00:11:01,360 Exactly. 362 00:11:01,360 --> 00:11:02,600 It's not just about competing. 363 00:11:02,600 --> 00:11:04,480 It's about finding the right partners. 364 00:11:04,480 --> 00:11:04,920 Makes sense. 365 00:11:04,920 --> 00:11:07,720 So in the AI world, it sounds like collaboration is key. 366 00:11:07,720 --> 00:11:08,400 It really is. 367 00:11:08,400 --> 00:11:09,640 Yeah, more than ever. 368 00:11:09,640 --> 00:11:11,200 OK, I want to shift gears a bit and talk 369 00:11:11,200 --> 00:11:13,640 about the bigger picture, the geopolitical entity. 370 00:11:13,640 --> 00:11:15,600 We mentioned before the US and China, 371 00:11:15,600 --> 00:11:18,800 those two dominant AI hyper-centers. 372 00:11:18,800 --> 00:11:20,160 What are the implications of that? 373 00:11:20,160 --> 00:11:22,880 Is it going to be like a winner takes all kind of thing? 374 00:11:22,880 --> 00:11:25,440 It's definitely a major point of tension. 375 00:11:25,440 --> 00:11:26,200 That's for sure. 376 00:11:26,200 --> 00:11:30,240 All countries see AI as absolutely essential 377 00:11:30,240 --> 00:11:32,040 for their economies, their militaries. 378 00:11:32,040 --> 00:11:32,960 Yeah, high stakes. 379 00:11:32,960 --> 00:11:33,760 Really high. 380 00:11:33,760 --> 00:11:35,880 But I wouldn't say it's a simple winner takes all. 381 00:11:35,880 --> 00:11:38,040 Remember, other countries have choices. 382 00:11:38,040 --> 00:11:40,600 They can build, buy, or partner. 383 00:11:40,600 --> 00:11:43,240 And for many, partnering seems like the best option. 384 00:11:43,240 --> 00:11:44,720 Especially for US allies. 385 00:11:44,720 --> 00:11:47,280 They get to use American technology and expertise, 386 00:11:47,280 --> 00:11:50,240 but still maintain a degree of autonomy. 387 00:11:50,240 --> 00:11:51,040 100%. 388 00:11:51,040 --> 00:11:52,480 And that's already playing out. 389 00:11:52,480 --> 00:11:55,160 The US is actively encouraging these collaborations, 390 00:11:55,160 --> 00:11:56,280 particularly with Europe. 391 00:11:56,280 --> 00:11:58,640 That transatlantic AI alliance we were talking about? 392 00:11:58,640 --> 00:11:59,760 Yeah, exactly. 393 00:11:59,760 --> 00:12:01,720 Trying to counter China's influence. 394 00:12:01,720 --> 00:12:02,320 Makes sense. 395 00:12:02,320 --> 00:12:04,200 So how does this affect France? 396 00:12:04,200 --> 00:12:06,200 We've talked about their strengths, their weaknesses, 397 00:12:06,200 --> 00:12:07,640 and AI infrastructure. 398 00:12:07,640 --> 00:12:08,600 Where do they fit in? 399 00:12:08,600 --> 00:12:09,800 France has a lot to offer. 400 00:12:09,800 --> 00:12:12,440 Strong research base, a talented workforce. 401 00:12:12,440 --> 00:12:15,440 And the government is really committed to investing in AI. 402 00:12:15,440 --> 00:12:17,760 But like we said, they need to catch up on data centers. 403 00:12:17,760 --> 00:12:19,040 That's where partners come in. 404 00:12:19,040 --> 00:12:19,560 Makes sense. 405 00:12:19,560 --> 00:12:22,680 We talked about collaborations with American chip companies. 406 00:12:22,680 --> 00:12:26,440 Do you see them becoming like a key player in this transatlantic 407 00:12:26,440 --> 00:12:27,480 AI alliance? 408 00:12:27,480 --> 00:12:28,680 I think it's very likely, yeah. 409 00:12:28,680 --> 00:12:30,920 France has a lot to gain from working with the US. 410 00:12:30,920 --> 00:12:32,360 And they bring a lot to the table. 411 00:12:32,360 --> 00:12:35,920 It could be like a central hub for research and development 412 00:12:35,920 --> 00:12:36,400 in Europe. 413 00:12:36,400 --> 00:12:38,480 Attract talent, investment, the whole thing. 414 00:12:38,480 --> 00:12:40,640 This all paints a really complex picture. 415 00:12:40,640 --> 00:12:43,160 Nations competing, open source versus closed source, 416 00:12:43,160 --> 00:12:44,960 this race to build hypercenter. 417 00:12:44,960 --> 00:12:49,440 It feels like the future of AI is still very much undecided. 418 00:12:49,440 --> 00:12:49,880 It is. 419 00:12:49,880 --> 00:12:50,760 It really is, yeah. 420 00:12:50,760 --> 00:12:52,160 And that's what makes it so fascinating. 421 00:12:52,160 --> 00:12:54,120 And the decisions made in the next few years, 422 00:12:54,120 --> 00:12:57,080 they're going to have a huge impact on how AI shapes our world. 423 00:12:57,080 --> 00:12:57,600 Absolutely. 424 00:12:57,600 --> 00:13:01,400 So let's circle back to that build by or partner dilemma. 425 00:13:01,400 --> 00:13:04,640 What advice would you give to a country trying to figure out 426 00:13:04,640 --> 00:13:05,840 its AI strategy? 427 00:13:05,840 --> 00:13:08,960 What are the key factors they should be thinking about? 428 00:13:08,960 --> 00:13:10,040 It's a tough question. 429 00:13:10,040 --> 00:13:11,160 No easy answers. 430 00:13:11,160 --> 00:13:13,520 Because there's no one size fits all solution. 431 00:13:13,520 --> 00:13:15,120 It depends on their resources, their goals, 432 00:13:15,120 --> 00:13:16,880 how much risk they're willing to take. 433 00:13:16,880 --> 00:13:18,880 But if I had to give some general guidance, 434 00:13:18,880 --> 00:13:19,280 Yeah. 435 00:13:19,280 --> 00:13:20,200 What would you say? 436 00:13:20,200 --> 00:13:23,080 First, be honest with yourself about what you're good at, 437 00:13:23,080 --> 00:13:23,880 what you're not. 438 00:13:23,880 --> 00:13:25,600 Don't spread yourself too thin. 439 00:13:25,600 --> 00:13:27,720 Find those areas where you have a real edge. 440 00:13:27,720 --> 00:13:29,080 Play to your strengths, like we said. 441 00:13:29,080 --> 00:13:29,720 Yeah, exactly. 442 00:13:29,720 --> 00:13:30,800 Find your niche. 443 00:13:30,800 --> 00:13:32,360 Second, this is a long game. 444 00:13:32,360 --> 00:13:34,240 Don't expect results overnight. 445 00:13:34,240 --> 00:13:36,880 Building a strong AI ecosystem takes time 446 00:13:36,880 --> 00:13:38,240 and consistent investment. 447 00:13:38,240 --> 00:13:39,440 Got to be in it for the long haul. 448 00:13:39,440 --> 00:13:40,080 You do. 449 00:13:40,080 --> 00:13:42,200 And finally, don't be afraid to team up. 450 00:13:42,200 --> 00:13:44,760 Find the right partners, people who share your vision. 451 00:13:44,760 --> 00:13:47,600 Working together can really accelerate your progress. 452 00:13:47,600 --> 00:13:49,160 So it's not just about competition. 453 00:13:49,160 --> 00:13:50,360 It's about collaboration. 454 00:13:50,360 --> 00:13:51,200 More than ever, yeah. 455 00:13:51,200 --> 00:13:53,320 In the AI world, that's how you succeed. 456 00:13:53,320 --> 00:13:55,120 This has been an incredible deep dive. 457 00:13:55,120 --> 00:13:57,320 We've covered so much from the technical stuff 458 00:13:57,320 --> 00:14:01,080 to the geopolitical landscape to the ethical considerations. 459 00:14:01,080 --> 00:14:02,840 Is there anything else you think our listeners should 460 00:14:02,840 --> 00:14:07,160 keep in mind as they try to make sense of this whole AI world? 461 00:14:07,160 --> 00:14:10,880 I think the most important thing is to just stay curious, 462 00:14:10,880 --> 00:14:11,960 stay engaged. 463 00:14:11,960 --> 00:14:13,320 AI is constantly changing. 464 00:14:13,320 --> 00:14:15,520 So keep learning, keep asking questions, 465 00:14:15,520 --> 00:14:16,800 be part of the conversation. 466 00:14:16,800 --> 00:14:18,160 Don't be a passive observer. 467 00:14:18,160 --> 00:14:18,840 Get involved. 468 00:14:18,840 --> 00:14:19,720 Exactly. 469 00:14:19,720 --> 00:14:22,600 This technology has the power to completely transform 470 00:14:22,600 --> 00:14:24,960 our world in good ways and bad. 471 00:14:24,960 --> 00:14:26,480 We need to make sure it's used for good. 472 00:14:26,480 --> 00:14:27,320 AI for good. 473 00:14:27,320 --> 00:14:27,880 I love that. 474 00:14:27,880 --> 00:14:28,760 That's the goal, right? 475 00:14:28,760 --> 00:14:29,320 It is. 476 00:14:29,320 --> 00:14:30,280 Well said. 477 00:14:30,280 --> 00:14:33,480 And on that note, I think it's time to wrap up this deep dive. 478 00:14:33,480 --> 00:14:36,640 But before we go, we want to leave you with one final thought. 479 00:14:36,640 --> 00:14:39,360 AI is this incredibly powerful tool. 480 00:14:39,360 --> 00:14:42,360 Like any tool, it can be used for good or for bad. 481 00:14:42,360 --> 00:14:43,640 The choices we make today, they're 482 00:14:43,640 --> 00:14:46,480 going to determine how AI shapes our world for generations 483 00:14:46,480 --> 00:14:47,160 to come. 484 00:14:47,160 --> 00:14:49,120 So we encourage you to really think about the things 485 00:14:49,120 --> 00:14:52,920 we've discussed, stay informed, and be part of the conversation. 486 00:14:52,920 --> 00:14:55,200 How do we want to shape the future of AI? 487 00:14:55,200 --> 00:14:57,200 Thanks for joining us on this deep dive. 488 00:14:57,200 --> 00:15:00,280 I'd say start by being honest with yourself about what 489 00:15:00,280 --> 00:15:01,720 you're good at and what you're not. 490 00:15:01,720 --> 00:15:02,080 Right. 491 00:15:02,080 --> 00:15:03,880 Like we were saying, play to your strengths. 492 00:15:03,880 --> 00:15:05,400 Yeah, find your niche. 493 00:15:05,400 --> 00:15:06,280 Don't try to do everything. 494 00:15:06,280 --> 00:15:07,000 Focus. 495 00:15:07,000 --> 00:15:08,080 Exactly. 496 00:15:08,080 --> 00:15:10,160 And then think long term. 497 00:15:10,160 --> 00:15:10,920 This isn't a sprint. 498 00:15:10,920 --> 00:15:11,680 It's a marathon. 499 00:15:11,680 --> 00:15:12,560 It really is, yeah. 500 00:15:12,560 --> 00:15:15,160 Building a strong AI ecosystem, it takes time, 501 00:15:15,160 --> 00:15:16,680 consistent investment. 502 00:15:16,680 --> 00:15:18,440 Don't expect to see results overnight. 503 00:15:18,440 --> 00:15:19,360 Yeah, to be patient. 504 00:15:19,360 --> 00:15:19,720 You do. 505 00:15:19,720 --> 00:15:20,280 You do. 506 00:15:20,280 --> 00:15:23,480 And then finally, don't be afraid to collaborate. 507 00:15:23,480 --> 00:15:24,360 Find those partners. 508 00:15:24,360 --> 00:15:25,480 Yeah, find the right partners. 509 00:15:25,480 --> 00:15:28,200 People who stare your vision, working together, 510 00:15:28,200 --> 00:15:29,480 it can really make a difference, 511 00:15:29,480 --> 00:15:30,880 really speed up your progress. 512 00:15:30,880 --> 00:15:32,920 So in the AI world, it sounds like it's not just 513 00:15:32,920 --> 00:15:33,840 about competition. 514 00:15:33,840 --> 00:15:34,760 It's about collaboration. 515 00:15:34,760 --> 00:15:35,160 Yeah. 516 00:15:35,160 --> 00:15:37,880 More than ever, that's the key to success. 517 00:15:37,880 --> 00:15:40,040 This has been such a fascinating deep dive. 518 00:15:40,040 --> 00:15:41,600 I feel like we've covered so much 519 00:15:41,600 --> 00:15:44,520 from the technical details of infrastructure, 520 00:15:44,520 --> 00:15:47,800 to the big geopolitical picture, to the ethical questions. 521 00:15:47,800 --> 00:15:48,200 It's a lot. 522 00:15:48,200 --> 00:15:49,160 Yeah, I feel like to think about. 523 00:15:49,160 --> 00:15:49,600 It is. 524 00:15:49,600 --> 00:15:50,120 It is. 525 00:15:50,120 --> 00:15:51,080 Is there anything else you think 526 00:15:51,080 --> 00:15:52,440 our listeners should keep in mind 527 00:15:52,440 --> 00:15:56,320 as they're trying to navigate this whole AI world? 528 00:15:56,320 --> 00:16:00,760 I think the most important thing is to just stay curious. 529 00:16:00,760 --> 00:16:01,600 Stay engaged. 530 00:16:01,600 --> 00:16:02,480 Yeah, exactly. 531 00:16:02,480 --> 00:16:04,120 AI is constantly changing. 532 00:16:04,120 --> 00:16:05,480 So keep learning. 533 00:16:05,480 --> 00:16:06,680 Keep asking questions. 534 00:16:06,680 --> 00:16:08,680 Don't just be a bystander. 535 00:16:08,680 --> 00:16:09,880 Be a part of the conversation. 536 00:16:09,880 --> 00:16:11,360 Be an active participant. 537 00:16:11,360 --> 00:16:12,280 Yes. 538 00:16:12,280 --> 00:16:16,160 This technology has the potential to completely transform 539 00:16:16,160 --> 00:16:18,800 our world in good ways and bad. 540 00:16:18,800 --> 00:16:21,200 It's up to us to make sure it's used for good, right? 541 00:16:21,200 --> 00:16:22,800 Like you said, AI for good. 542 00:16:22,800 --> 00:16:23,240 That's it. 543 00:16:23,240 --> 00:16:23,960 That's the goal. 544 00:16:23,960 --> 00:16:24,560 Well said. 545 00:16:24,560 --> 00:16:26,960 Well, I think it's about time to wrap up this deep dive. 546 00:16:26,960 --> 00:16:30,400 But before we go, we want to leave you with one final thought. 547 00:16:30,400 --> 00:16:31,880 AI is a powerful tool. 548 00:16:31,880 --> 00:16:36,200 And just like any tool, it can be used for good or for bad. 549 00:16:36,200 --> 00:16:37,840 The choices we make today, they're 550 00:16:37,840 --> 00:16:41,760 going to determine how AI shapes our world for generations 551 00:16:41,760 --> 00:16:42,240 to come. 552 00:16:42,240 --> 00:16:44,120 So we encourage you to really think 553 00:16:44,120 --> 00:16:46,360 about the things we've discussed, stay informed, 554 00:16:46,360 --> 00:16:48,840 and be a part of the conversation. 555 00:16:48,840 --> 00:16:52,120 How do we want to shape the future of AI? 556 00:16:52,120 --> 00:17:12,760 Thanks for joining us on this deep dive.