1 00:00:00,000 --> 00:00:02,240 Welcome to the Daily AI News Podcast. 2 00:00:02,240 --> 00:00:05,840 We're gonna take a deep dive into all the AI news today. 3 00:00:05,840 --> 00:00:08,120 We've got news articles, company announcements, 4 00:00:08,120 --> 00:00:10,360 research snippets, the whole nine yards. 5 00:00:10,360 --> 00:00:13,840 It's amazing how much is happening in AI every single day. 6 00:00:13,840 --> 00:00:15,080 Right. 7 00:00:15,080 --> 00:00:16,400 Things are changing so fast. 8 00:00:16,400 --> 00:00:19,560 And speaking of fast moving, let's talk about open AI. 9 00:00:19,560 --> 00:00:20,960 Yeah, open AI. 10 00:00:20,960 --> 00:00:23,160 There's a lot of buzz around their next big model. 11 00:00:23,160 --> 00:00:24,680 Orion. Orion, right. 12 00:00:24,680 --> 00:00:26,200 December seems to be the target. 13 00:00:26,200 --> 00:00:27,680 At least for some companies, 14 00:00:27,680 --> 00:00:29,960 they're letting a few get early access. 15 00:00:29,960 --> 00:00:31,000 A little preview. 16 00:00:31,000 --> 00:00:33,120 Exactly, for product development, things like that. 17 00:00:33,120 --> 00:00:35,520 And get this, Microsoft engineers are getting ready 18 00:00:35,520 --> 00:00:37,080 to host it on Azure. 19 00:00:37,080 --> 00:00:38,680 On Azure, already. 20 00:00:38,680 --> 00:00:40,520 It could be as early as November, even. 21 00:00:40,520 --> 00:00:41,400 Wow. 22 00:00:41,400 --> 00:00:42,440 So it's really happening. 23 00:00:42,440 --> 00:00:44,760 It is, and there's no official name yet, but... 24 00:00:44,760 --> 00:00:46,120 Everyone's saying Orion, right? 25 00:00:46,120 --> 00:00:50,000 Yeah, Orion, it's likely the successor to GPT-4. 26 00:00:50,000 --> 00:00:52,320 And they're saying it could be a hundred times more powerful. 27 00:00:52,320 --> 00:00:54,160 A hundred times, really? 28 00:00:54,160 --> 00:00:55,920 What could you even do with that? 29 00:00:55,920 --> 00:00:58,000 The possibilities are kind of mind blowing 30 00:00:58,000 --> 00:00:59,000 when you think about it. 31 00:00:59,000 --> 00:01:00,400 But wait, there's more. 32 00:01:00,400 --> 00:01:03,480 Open AI wants to combine all their big models. 33 00:01:03,480 --> 00:01:04,320 All of them? 34 00:01:04,320 --> 00:01:06,200 To create an even more capable model, 35 00:01:06,200 --> 00:01:09,400 which might even lead to artificial general intelligence. 36 00:01:09,400 --> 00:01:11,560 AGI, that's like the ultimate goal, right? 37 00:01:11,560 --> 00:01:15,480 Exactly, a machine that thinks and learns like we do. 38 00:01:15,480 --> 00:01:18,360 But it's not all smooth sailing at open AI. 39 00:01:18,360 --> 00:01:19,280 What do you mean? 40 00:01:19,280 --> 00:01:22,120 Well, some important people have left recently. 41 00:01:22,120 --> 00:01:23,920 The CTO, Mir Merati, 42 00:01:23,920 --> 00:01:26,120 and the chief research officer, Bob McGrew. 43 00:01:26,120 --> 00:01:27,720 Ah, that's interesting. 44 00:01:27,720 --> 00:01:29,120 Any idea why they left? 45 00:01:29,120 --> 00:01:31,640 Well, there's talk that things aren't so great 46 00:01:31,640 --> 00:01:33,440 between open AI and Microsoft. 47 00:01:33,440 --> 00:01:35,840 Seems like open AI needs more resources. 48 00:01:35,840 --> 00:01:37,200 Like what kind of resources? 49 00:01:37,200 --> 00:01:40,960 Mostly computing power and server capacity, 50 00:01:40,960 --> 00:01:43,680 all that stuff to run these huge AI models. 51 00:01:43,680 --> 00:01:44,680 Yeah, I can imagine. 52 00:01:44,680 --> 00:01:46,000 Apparently it's putting a strain 53 00:01:46,000 --> 00:01:47,640 on their partnership with Microsoft. 54 00:01:47,640 --> 00:01:49,600 Oh, so they're not too happy about it. 55 00:01:49,600 --> 00:01:50,840 You could say that. 56 00:01:50,840 --> 00:01:53,080 Open AI actually renegotiated their deal 57 00:01:53,080 --> 00:01:54,560 to partner with Oracle. 58 00:01:54,560 --> 00:01:56,080 With Oracle? For what? 59 00:01:56,080 --> 00:01:57,680 To get even more servers. 60 00:01:57,680 --> 00:01:59,720 Seems like they need all the power they can get. 61 00:01:59,720 --> 00:02:02,200 But you have to wonder who's really in control here. 62 00:02:02,200 --> 00:02:04,000 Yeah, open AI or Microsoft. 63 00:02:04,000 --> 00:02:08,280 Right, Microsoft needs open AI's models for their products. 64 00:02:08,280 --> 00:02:10,160 Like for 365 Copilot, right? 65 00:02:10,160 --> 00:02:11,960 That makes them look pretty good in the AI game. 66 00:02:11,960 --> 00:02:15,720 Exactly, but if open AI reaches AGI first, 67 00:02:15,720 --> 00:02:17,400 then Microsoft could be left behind. 68 00:02:17,400 --> 00:02:19,000 It's a possibility. 69 00:02:19,000 --> 00:02:21,600 And it looks like both of them are taking it seriously. 70 00:02:21,600 --> 00:02:22,440 What do you mean? 71 00:02:22,440 --> 00:02:25,160 They've both hired investment banks to advise them. 72 00:02:25,160 --> 00:02:27,560 On what? On equity stakes. 73 00:02:27,560 --> 00:02:29,200 And governance rights. 74 00:02:29,200 --> 00:02:31,480 You know, open AI started as a nonprofit 75 00:02:31,480 --> 00:02:32,880 focused on ethical AI. 76 00:02:32,880 --> 00:02:33,720 What did I remember? 77 00:02:33,720 --> 00:02:36,400 And now they're a multi-billion dollar company 78 00:02:36,400 --> 00:02:38,880 that makes you think about their original mission, you know. 79 00:02:38,880 --> 00:02:39,720 That's a good point. 80 00:02:39,720 --> 00:02:40,880 It seems like things have changed a lot. 81 00:02:40,880 --> 00:02:41,720 They have. 82 00:02:41,720 --> 00:02:42,920 And on top of all that, 83 00:02:42,920 --> 00:02:45,480 they're competing more and more directly with each other. 84 00:02:45,480 --> 00:02:49,320 Open AI's GBC4O, the faster, cheaper model. 85 00:02:49,320 --> 00:02:53,400 That kind of hurts Microsoft's paid AI services on Azure. 86 00:02:53,400 --> 00:02:55,440 Ouch, so what's Microsoft doing about it? 87 00:02:55,440 --> 00:02:57,920 They're working on new APIs for things like search 88 00:02:57,920 --> 00:02:59,280 and speech recognition. 89 00:02:59,280 --> 00:03:00,560 All connected to GPT4. 90 00:03:00,560 --> 00:03:01,400 You got it. 91 00:03:01,400 --> 00:03:02,240 They might even launch them 92 00:03:02,240 --> 00:03:04,200 at their Ignite conference in November. 93 00:03:04,200 --> 00:03:05,840 So the battle continues. 94 00:03:05,840 --> 00:03:06,680 Absolutely. 95 00:03:06,680 --> 00:03:08,360 It's like an AI arms race. 96 00:03:08,360 --> 00:03:11,440 Open AI is really changing the game, definitely. 97 00:03:11,440 --> 00:03:12,760 And they're not the only ones. 98 00:03:12,760 --> 00:03:13,600 Oh, right. 99 00:03:13,600 --> 00:03:14,440 What else is going on? 100 00:03:14,440 --> 00:03:16,840 Clawed.AI, they're a competitor to open AI, 101 00:03:16,840 --> 00:03:18,880 and they just came out with a new analysis tool. 102 00:03:18,880 --> 00:03:19,960 For their Clawed model. 103 00:03:19,960 --> 00:03:20,800 Exactly. 104 00:03:20,800 --> 00:03:21,640 What can it do? 105 00:03:21,640 --> 00:03:25,960 Well, it can write and execute JavaScript code. 106 00:03:25,960 --> 00:03:27,760 Wait, the AI can actually code now? 107 00:03:27,760 --> 00:03:28,760 Yep. 108 00:03:28,760 --> 00:03:30,440 I don't even know what that means, really. 109 00:03:30,440 --> 00:03:31,600 What's that good for? 110 00:03:31,600 --> 00:03:35,080 Well, for one, it lets Clawed do some pretty complex math. 111 00:03:35,080 --> 00:03:35,920 Okay. 112 00:03:35,920 --> 00:03:38,160 It can analyze data in real time, 113 00:03:38,160 --> 00:03:41,440 and find insights that go beyond simple questions 114 00:03:41,440 --> 00:03:42,280 and answers. 115 00:03:42,280 --> 00:03:43,760 So it's not just giving you information, 116 00:03:43,760 --> 00:03:45,080 it's actually understanding it. 117 00:03:45,080 --> 00:03:45,920 Exactly. 118 00:03:45,920 --> 00:03:48,320 It uses its reasoning and its coding skills 119 00:03:48,320 --> 00:03:49,200 at the same time. 120 00:03:49,200 --> 00:03:50,920 Wow, that's pretty impressive. 121 00:03:50,920 --> 00:03:51,840 What else is happening? 122 00:03:51,840 --> 00:03:53,160 Well, Met has been busy, too. 123 00:03:53,160 --> 00:03:55,680 They've released what they call quantized versions 124 00:03:55,680 --> 00:03:56,760 of their Lama models. 125 00:03:56,760 --> 00:03:57,600 The Lama model. 126 00:03:57,600 --> 00:04:02,760 Yeah, their Lama 3.2, the 1B and 3B versions. 127 00:04:02,760 --> 00:04:03,680 I know what that means. 128 00:04:03,680 --> 00:04:06,800 It means they're smaller and faster, 129 00:04:06,800 --> 00:04:08,400 enough to work on mobile devices. 130 00:04:08,400 --> 00:04:11,080 Wait, so my phone could run a powerful AI? 131 00:04:11,080 --> 00:04:12,520 Potentially, yeah. 132 00:04:12,520 --> 00:04:13,520 And they've managed to do it 133 00:04:13,520 --> 00:04:16,080 without sacrificing quality or safety. 134 00:04:16,080 --> 00:04:16,920 That's amazing. 135 00:04:16,920 --> 00:04:18,400 How do they even do that? 136 00:04:18,400 --> 00:04:21,520 They used some pretty advanced techniques, 137 00:04:21,520 --> 00:04:24,320 like something called Quantization Aware Training 138 00:04:25,320 --> 00:04:26,720 and Spin-Hwant. 139 00:04:26,720 --> 00:04:27,760 I have no idea what that means, 140 00:04:27,760 --> 00:04:29,000 but it sounds impressive. 141 00:04:29,000 --> 00:04:29,840 It is. 142 00:04:29,840 --> 00:04:32,160 It means the models need less memory, 143 00:04:32,160 --> 00:04:34,600 they run faster, and they're easier to move around. 144 00:04:34,600 --> 00:04:36,120 And Met is not stopping. 145 00:04:36,120 --> 00:04:39,680 Right, they're working with partners to use NPUs. 146 00:04:39,680 --> 00:04:40,680 NPUs, what are those? 147 00:04:40,680 --> 00:04:44,720 Neural Processing Units, special hardware just for AI. 148 00:04:44,720 --> 00:04:46,880 So even more power and efficiency is coming. 149 00:04:46,880 --> 00:04:47,960 It looks like it. 150 00:04:47,960 --> 00:04:49,560 That's pretty wild. 151 00:04:49,560 --> 00:04:51,120 Okay, what else have we got? 152 00:04:51,120 --> 00:04:52,800 Well, Apple's up to something interesting. 153 00:04:52,800 --> 00:04:55,000 They're offering a pretty big reward. 154 00:04:55,000 --> 00:04:56,520 A reward, what for? 155 00:04:56,520 --> 00:04:58,560 Up to a million dollars, actually. 156 00:04:58,560 --> 00:05:00,920 A million dollars, what do you have to do to get that? 157 00:05:00,920 --> 00:05:05,800 Find vulnerabilities in their new private AI cloud service. 158 00:05:05,800 --> 00:05:06,640 What's it called? 159 00:05:06,640 --> 00:05:07,600 Private cloud compute. 160 00:05:07,600 --> 00:05:10,120 And they'll pay a million dollars if you can break into it. 161 00:05:10,120 --> 00:05:12,600 We'll find a serious vulnerability, yeah. 162 00:05:12,600 --> 00:05:15,080 Wow, they must be really confident about their security. 163 00:05:15,080 --> 00:05:15,920 They are. 164 00:05:15,920 --> 00:05:17,960 Crowd sourcing security testing. 165 00:05:17,960 --> 00:05:19,440 With a big incentive. 166 00:05:19,440 --> 00:05:22,720 So what is this private cloud compute anyway? 167 00:05:22,720 --> 00:05:26,000 Well, Apple says it's designed to extend the capabilities 168 00:05:26,000 --> 00:05:28,000 of their on-device AI. 169 00:05:28,000 --> 00:05:28,840 Which is? 170 00:05:28,840 --> 00:05:29,680 Apple Intelligence. 171 00:05:29,680 --> 00:05:31,960 You know, their on-device AI model. 172 00:05:31,960 --> 00:05:34,800 But they also emphasize that they want to keep 173 00:05:34,800 --> 00:05:37,400 user privacy a priority. 174 00:05:37,400 --> 00:05:41,240 So more power without compromising privacy. 175 00:05:41,240 --> 00:05:43,520 That's the idea, it's a tough balance. 176 00:05:43,520 --> 00:05:45,400 It is, I wonder how they'll pull it off. 177 00:05:45,400 --> 00:05:47,640 OK, so what else is going on in the world of AI? 178 00:05:47,640 --> 00:05:51,000 Well, remember Google and that whole antitrust case? 179 00:05:51,000 --> 00:05:52,200 Oh, yeah, right. 180 00:05:52,200 --> 00:05:54,320 The Justice Department said they had an illegal monopoly. 181 00:05:54,320 --> 00:05:55,200 Exactly. 182 00:05:55,200 --> 00:05:56,480 What's happening with that? 183 00:05:56,480 --> 00:05:58,760 Well, the Justice Department is saying that Google's using 184 00:05:58,760 --> 00:06:01,200 generative AI to stay in control. 185 00:06:01,200 --> 00:06:02,080 In control of what? 186 00:06:02,080 --> 00:06:03,000 Of search. 187 00:06:03,000 --> 00:06:06,320 They think Google is using AI to keep its dominance, 188 00:06:06,320 --> 00:06:08,720 even with the remedies they have to implement. 189 00:06:08,720 --> 00:06:11,840 So AI could actually be used to make a monopoly stronger? 190 00:06:11,840 --> 00:06:13,000 That's what they're arguing. 191 00:06:13,000 --> 00:06:15,240 And the judge seems to be taking it seriously. 192 00:06:15,240 --> 00:06:16,440 What's he going to do about it? 193 00:06:16,440 --> 00:06:19,640 He wants to look at more documents about AI, 194 00:06:19,640 --> 00:06:22,760 specifically about products like Google's overview AI 195 00:06:22,760 --> 00:06:25,360 function and their Gemini chatbot. 196 00:06:25,360 --> 00:06:27,920 So they're trying to show that these AI features are giving 197 00:06:27,920 --> 00:06:29,720 Google an unfair advantage. 198 00:06:29,720 --> 00:06:30,440 That's right. 199 00:06:30,440 --> 00:06:33,120 And they're using Google's own definition of AI. 200 00:06:33,120 --> 00:06:35,000 Their own definition, where did they get that from? 201 00:06:35,000 --> 00:06:37,400 From a subpoena they sent to Microsoft. 202 00:06:37,400 --> 00:06:38,640 Huh, interesting. 203 00:06:38,640 --> 00:06:40,280 So what does that definition say? 204 00:06:40,280 --> 00:06:42,840 Basically, it says any product that combines search engine 205 00:06:42,840 --> 00:06:46,200 technology and AI counts as AI. 206 00:06:46,200 --> 00:06:48,680 So they're really trying to understand how Google is using 207 00:06:48,680 --> 00:06:50,240 AI in search. 208 00:06:50,240 --> 00:06:50,760 Yeah. 209 00:06:50,760 --> 00:06:52,840 And whether it's hurting competition, 210 00:06:52,840 --> 00:06:56,080 this case could be a big deal for AI regulation in the future. 211 00:06:56,080 --> 00:06:57,720 This is definitely something to watch. 212 00:06:57,720 --> 00:07:00,520 But it's not all doom and gloom with AI. 213 00:07:00,520 --> 00:07:01,280 True. 214 00:07:01,280 --> 00:07:03,120 There's some pretty positive things happening too. 215 00:07:03,120 --> 00:07:03,600 Oh yeah. 216 00:07:03,600 --> 00:07:04,520 Like what? 217 00:07:04,520 --> 00:07:07,320 Well, there was a Google survey recently about developers. 218 00:07:07,320 --> 00:07:08,200 What kind of developers? 219 00:07:08,200 --> 00:07:11,320 Bofor developers and IT professionals. 220 00:07:11,320 --> 00:07:12,280 What did they find? 221 00:07:12,280 --> 00:07:16,360 More than 75% of them are using AI for their work. 222 00:07:16,360 --> 00:07:16,920 Really? 223 00:07:16,920 --> 00:07:17,560 That many? 224 00:07:17,560 --> 00:07:19,040 At least one task every day. 225 00:07:19,040 --> 00:07:20,160 What kind of tasks? 226 00:07:20,160 --> 00:07:21,360 All kinds of things. 227 00:07:21,360 --> 00:07:24,920 Writing code, fixing bugs, analyzing data, 228 00:07:24,920 --> 00:07:26,640 even automating some of their work. 229 00:07:26,640 --> 00:07:29,320 So they're using AI to help them get their work done. 230 00:07:29,320 --> 00:07:29,720 Makes sense. 231 00:07:29,720 --> 00:07:30,600 It does. 232 00:07:30,600 --> 00:07:32,880 And most of them said it was actually a good thing. 233 00:07:32,880 --> 00:07:33,760 Good in what way? 234 00:07:33,760 --> 00:07:37,040 It made them more productive, improved the quality of their work, 235 00:07:37,040 --> 00:07:39,640 and even made them feel better about their jobs. 236 00:07:39,640 --> 00:07:42,440 So AI is actually making work better for a lot of people. 237 00:07:42,440 --> 00:07:43,800 It seems that way. 238 00:07:43,800 --> 00:07:44,760 But there's a catch. 239 00:07:44,760 --> 00:07:45,440 That's that. 240 00:07:45,440 --> 00:07:49,440 Almost 40% of them don't really trust AI-generated code. 241 00:07:49,440 --> 00:07:52,000 So even though it helps them, they don't fully trust it. 242 00:07:52,000 --> 00:07:53,080 Not yet, anyway. 243 00:07:53,080 --> 00:07:53,720 Interesting. 244 00:07:53,720 --> 00:07:54,960 So what's the takeaway? 245 00:07:54,960 --> 00:07:58,800 Well, we have to be careful about how we use AI. 246 00:07:58,800 --> 00:08:01,720 We need strategies to make sure it's implemented properly. 247 00:08:01,720 --> 00:08:02,720 What kind of strategies? 248 00:08:02,720 --> 00:08:04,720 Focus on empowering employees. 249 00:08:04,720 --> 00:08:07,400 Get rid of those boring, repetitive tasks. 250 00:08:07,400 --> 00:08:10,040 Have clear guidelines for how to use AI, 251 00:08:10,040 --> 00:08:12,280 and be open about how it's affecting everyone. 252 00:08:12,280 --> 00:08:14,920 It's all about finding a balance between humans and AI. 253 00:08:14,920 --> 00:08:16,160 Exactly. 254 00:08:16,160 --> 00:08:18,120 So what else is going on in the world of AI? 255 00:08:18,120 --> 00:08:23,120 Well, the UK's Competition and Markets Authority, CMA, 256 00:08:23,120 --> 00:08:25,120 they're investigating Google again. 257 00:08:25,120 --> 00:08:27,760 What, the same watchdog that was looking at Microsoft 258 00:08:27,760 --> 00:08:28,400 and Activision? 259 00:08:28,400 --> 00:08:29,640 That's the one. 260 00:08:29,640 --> 00:08:33,360 This time, it's about Google's partnership with Anthropic. 261 00:08:33,360 --> 00:08:34,240 Anthropic. 262 00:08:34,240 --> 00:08:35,640 Remind me what they do again. 263 00:08:35,640 --> 00:08:36,840 They're an AI startup. 264 00:08:36,840 --> 00:08:39,920 They focus on AI safety and making sure it's reliable. 265 00:08:39,920 --> 00:08:41,080 OK, right. 266 00:08:41,080 --> 00:08:42,960 And what's the problem with their partnership? 267 00:08:42,960 --> 00:08:46,080 Well, Google invested a ton of money in them, billions, 268 00:08:46,080 --> 00:08:46,960 actually. 269 00:08:46,960 --> 00:08:50,360 The CMA is worried that it could be bad for competition. 270 00:08:50,360 --> 00:08:53,840 Other AI companies might not be able to keep up. 271 00:08:53,840 --> 00:08:55,720 So Google's getting too big, maybe. 272 00:08:55,720 --> 00:08:57,720 That's what they're trying to figure out. 273 00:08:57,720 --> 00:08:59,360 What does Google have to say about all this? 274 00:08:59,360 --> 00:09:01,600 They're saying that Anthropic is free to work 275 00:09:01,600 --> 00:09:04,520 with other companies, too, and that they don't demand 276 00:09:04,520 --> 00:09:07,120 any special rights to Anthropic's technology. 277 00:09:07,120 --> 00:09:09,400 It sounds like the CMA is really trying to keep things 278 00:09:09,400 --> 00:09:11,120 fair in the AI world. 279 00:09:11,120 --> 00:09:11,880 They are. 280 00:09:11,880 --> 00:09:12,720 It's a big deal. 281 00:09:12,720 --> 00:09:14,520 Governments are getting more and more involved 282 00:09:14,520 --> 00:09:16,480 with regulating AI. 283 00:09:16,480 --> 00:09:17,680 Makes sense. 284 00:09:17,680 --> 00:09:19,280 It's a powerful technology. 285 00:09:19,280 --> 00:09:21,880 It's going to affect everyone, so it's important to make sure 286 00:09:21,880 --> 00:09:23,400 it's used responsibly. 287 00:09:23,400 --> 00:09:24,520 Absolutely. 288 00:09:24,520 --> 00:09:26,520 And speaking of responsibility, we 289 00:09:26,520 --> 00:09:29,600 can't forget about the legal battles brewing between AI 290 00:09:29,600 --> 00:09:31,680 companies and traditional media. 291 00:09:31,680 --> 00:09:31,920 Right. 292 00:09:31,920 --> 00:09:33,160 There's a lot of tension there. 293 00:09:33,160 --> 00:09:36,440 Who owns the information that's used to train these AI systems? 294 00:09:36,440 --> 00:09:39,440 That's the million dollar question, literally. 295 00:09:39,440 --> 00:09:40,880 Because perplexity. 296 00:09:40,880 --> 00:09:43,400 Perplexity, the AI search startup. 297 00:09:43,400 --> 00:09:44,320 Yeah, that's the one. 298 00:09:44,320 --> 00:09:45,400 They're being sued. 299 00:09:45,400 --> 00:09:46,000 By who? 300 00:09:46,000 --> 00:09:48,040 Dow Jones and the New York Post. 301 00:09:48,040 --> 00:09:49,480 And they're both owned by? 302 00:09:49,480 --> 00:09:50,280 Newscore. 303 00:09:50,280 --> 00:09:52,760 Big players in the media world, so what's the problem? 304 00:09:52,760 --> 00:09:54,760 Copyright infringement. 305 00:09:54,760 --> 00:09:57,280 The publishers say perplexity used their content 306 00:09:57,280 --> 00:09:58,480 without permission. 307 00:09:58,480 --> 00:09:59,040 To do what? 308 00:09:59,040 --> 00:10:00,880 To build their search engine. 309 00:10:00,880 --> 00:10:04,200 So it's a classic case of old media versus new technology. 310 00:10:04,200 --> 00:10:05,240 You got it. 311 00:10:05,240 --> 00:10:07,480 The news organizations are protecting their content. 312 00:10:07,480 --> 00:10:10,280 And the AI companies say they need access to tons of data. 313 00:10:10,280 --> 00:10:11,400 Exactly. 314 00:10:11,400 --> 00:10:12,640 It's a tough situation. 315 00:10:12,640 --> 00:10:13,040 It is. 316 00:10:13,040 --> 00:10:16,360 It seems like we're going to see a lot more of these kinds of cases 317 00:10:16,360 --> 00:10:18,120 as AI gets more popular. 318 00:10:18,120 --> 00:10:19,080 Probably. 319 00:10:19,080 --> 00:10:20,240 But you know what else is interesting? 320 00:10:20,240 --> 00:10:25,000 How AI is being used to understand other species. 321 00:10:25,000 --> 00:10:27,000 Other species, like animals. 322 00:10:27,000 --> 00:10:27,920 Yeah. 323 00:10:27,920 --> 00:10:31,760 Researchers used AI to have a conversation with a whale. 324 00:10:31,760 --> 00:10:32,640 A whale? 325 00:10:32,640 --> 00:10:34,040 How is that even possible? 326 00:10:34,040 --> 00:10:36,000 It was a humpback whale named Twain. 327 00:10:36,000 --> 00:10:36,680 Twain? 328 00:10:36,680 --> 00:10:37,360 Like Mark Twain? 329 00:10:37,360 --> 00:10:38,480 Yeah, apparently. 330 00:10:38,480 --> 00:10:40,160 They were off the coast of Alaska. 331 00:10:40,160 --> 00:10:41,240 So what happened? 332 00:10:41,240 --> 00:10:43,800 This team, they call themselves Whale SETI. 333 00:10:43,800 --> 00:10:46,160 They played recordings of whale calls. 334 00:10:46,160 --> 00:10:47,320 Twain responded. 335 00:10:47,320 --> 00:10:49,520 They talked to a whale using AI. 336 00:10:49,520 --> 00:10:51,720 Well, they think it was intentional communication, 337 00:10:51,720 --> 00:10:52,960 like a conversation. 338 00:10:52,960 --> 00:10:53,720 That's amazing. 339 00:10:53,720 --> 00:10:55,840 So the AI is helping us to understand whales. 340 00:10:55,840 --> 00:10:56,960 Exactly. 341 00:10:56,960 --> 00:11:00,200 And maybe even other intelligent life out there, who knows. 342 00:11:00,200 --> 00:11:01,000 Wow. 343 00:11:01,000 --> 00:11:02,800 This AI stuff is really incredible. 344 00:11:02,800 --> 00:11:03,400 It is. 345 00:11:03,400 --> 00:11:05,360 And it reminds us that we need to protect 346 00:11:05,360 --> 00:11:07,880 these amazing creatures and their habitats. 347 00:11:07,880 --> 00:11:08,720 I agree. 348 00:11:08,720 --> 00:11:12,160 It's amazing how AI can help us connect with other species. 349 00:11:12,160 --> 00:11:13,760 So what else is going on? 350 00:11:13,760 --> 00:11:14,720 Well, how about this? 351 00:11:14,720 --> 00:11:16,600 Scientists have created an AI that 352 00:11:16,600 --> 00:11:18,720 can understand pig emotions. 353 00:11:18,720 --> 00:11:19,880 Wait, seriously? 354 00:11:19,880 --> 00:11:20,240 Yeah. 355 00:11:20,240 --> 00:11:22,280 Pig sounds like they're grunts and oinks. 356 00:11:22,280 --> 00:11:24,360 So it can tell if a pig is happy or sad. 357 00:11:24,360 --> 00:11:26,800 It can tell if they're feeling positive or negative. 358 00:11:26,800 --> 00:11:27,680 That's incredible. 359 00:11:27,680 --> 00:11:28,760 What's that good for? 360 00:11:28,760 --> 00:11:30,040 Well, think about it. 361 00:11:30,040 --> 00:11:32,080 Farmers could use it to understand 362 00:11:32,080 --> 00:11:33,560 how their pigs are feeling. 363 00:11:33,560 --> 00:11:36,320 That's a great idea to make sure they're being treated well. 364 00:11:36,320 --> 00:11:37,120 Exactly. 365 00:11:37,120 --> 00:11:39,720 They could create a better environment for the pigs, 366 00:11:39,720 --> 00:11:41,040 make them more comfortable. 367 00:11:41,040 --> 00:11:42,280 That's fantastic. 368 00:11:42,280 --> 00:11:44,480 I love that AI is being used to help animals. 369 00:11:44,480 --> 00:11:45,320 Me too. 370 00:11:45,320 --> 00:11:48,840 And it really shows how AI can be used for good. 371 00:11:48,840 --> 00:11:49,440 Absolutely. 372 00:11:49,440 --> 00:11:51,960 It really shows how AI can be used for good. 373 00:11:51,960 --> 00:11:53,480 And the research actually suggests 374 00:11:53,480 --> 00:11:55,840 that pigs are happier when they have more freedom. 375 00:11:55,840 --> 00:11:56,400 Really? 376 00:11:56,400 --> 00:11:58,760 Yeah, they looked at pigs in different environments. 377 00:11:58,760 --> 00:12:00,680 The ones in more natural settings, 378 00:12:00,680 --> 00:12:03,880 with room to move around, made fewer stress calls. 379 00:12:03,880 --> 00:12:06,560 So the AI could help farmers figure out the best way 380 00:12:06,560 --> 00:12:07,760 to raise pigs. 381 00:12:07,760 --> 00:12:08,800 Exactly. 382 00:12:08,800 --> 00:12:10,800 It's like giving the pigs a voice, 383 00:12:10,800 --> 00:12:14,760 maybe even an app that translates pig sounds into emotions. 384 00:12:14,760 --> 00:12:16,280 That would be amazing. 385 00:12:16,280 --> 00:12:18,600 So are we done talking about animals, or is there more? 386 00:12:18,600 --> 00:12:20,480 That's it for the animal news for now. 387 00:12:20,480 --> 00:12:23,200 But we were talking about governments and AI earlier. 388 00:12:23,200 --> 00:12:24,760 Right, about how they're trying to regulate it. 389 00:12:24,760 --> 00:12:25,640 Exactly. 390 00:12:25,640 --> 00:12:29,160 And it seems like the UK is really taking it seriously. 391 00:12:29,160 --> 00:12:30,280 The CMA, right? 392 00:12:30,280 --> 00:12:32,000 The Competition and Markets Authority. 393 00:12:32,000 --> 00:12:32,720 That's the one. 394 00:12:32,720 --> 00:12:35,200 They're all over this Microsoft Activision deal. 395 00:12:35,200 --> 00:12:38,120 Yeah, and now they're looking into Google's partnership 396 00:12:38,120 --> 00:12:39,240 with Anthropic. 397 00:12:39,240 --> 00:12:39,760 Right. 398 00:12:39,760 --> 00:12:42,200 They want to make sure that Google's not using its power 399 00:12:42,200 --> 00:12:43,760 to squash the competition. 400 00:12:43,760 --> 00:12:46,400 So they're trying to protect the smaller AI companies. 401 00:12:46,400 --> 00:12:47,160 In a way, yeah. 402 00:12:47,160 --> 00:12:48,880 They want a fair market for everyone. 403 00:12:48,880 --> 00:12:49,440 It's interesting. 404 00:12:49,440 --> 00:12:52,040 This whole balance between letting 405 00:12:52,040 --> 00:12:53,640 these big companies innovate. 406 00:12:53,640 --> 00:12:56,200 But also making sure they don't become too powerful. 407 00:12:56,200 --> 00:12:57,240 Exactly. 408 00:12:57,240 --> 00:12:58,800 It's a tough job for these regulators. 409 00:12:58,800 --> 00:12:59,640 It is. 410 00:12:59,640 --> 00:13:01,600 And speaking of tough situations. 411 00:13:01,600 --> 00:13:02,680 Oh, right. 412 00:13:02,680 --> 00:13:04,160 That lawsuit against perplexity. 413 00:13:04,160 --> 00:13:04,880 Yeah. 414 00:13:04,880 --> 00:13:06,840 It really highlights this conflict 415 00:13:06,840 --> 00:13:10,080 between AI and traditional media. 416 00:13:10,080 --> 00:13:11,840 Who gets to use the information? 417 00:13:11,840 --> 00:13:12,840 Who owns it? 418 00:13:12,840 --> 00:13:13,600 It's a mess. 419 00:13:13,600 --> 00:13:14,160 It is. 420 00:13:14,160 --> 00:13:16,000 With publishers want to protect their content, 421 00:13:16,000 --> 00:13:17,400 and understandably so. 422 00:13:17,400 --> 00:13:20,600 But the AI companies need data, lots of it, 423 00:13:20,600 --> 00:13:21,680 to train their system. 424 00:13:21,680 --> 00:13:22,240 Exactly. 425 00:13:22,240 --> 00:13:24,520 And it doesn't seem like there's an easy answer. 426 00:13:24,520 --> 00:13:26,440 It'll be interesting to see how this plays out. 427 00:13:26,440 --> 00:13:29,560 Not just for perplexity, but for the whole AI industry. 428 00:13:29,560 --> 00:13:30,960 Absolutely. 429 00:13:30,960 --> 00:13:33,440 Now remember that million-dollar reward 430 00:13:33,440 --> 00:13:34,760 Apple is offering. 431 00:13:34,760 --> 00:13:35,800 Yeah. 432 00:13:35,800 --> 00:13:37,320 I was wondering what that was all about. 433 00:13:37,320 --> 00:13:39,000 It sounds kind of mysterious. 434 00:13:39,000 --> 00:13:42,080 Well, it's not for solving some AI puzzle 435 00:13:42,080 --> 00:13:42,920 or anything like that. 436 00:13:42,920 --> 00:13:44,200 Oh. 437 00:13:44,200 --> 00:13:45,440 What is it for then? 438 00:13:45,440 --> 00:13:47,960 It's for finding security flaws. 439 00:13:47,960 --> 00:13:49,280 Security flaws? 440 00:13:49,280 --> 00:13:49,960 In what? 441 00:13:49,960 --> 00:13:54,040 Their new private AI cloud service. 442 00:13:54,040 --> 00:13:56,280 Private cloud compute, it's called. 443 00:13:56,280 --> 00:13:59,200 So if you can hack it, you get a million bucks. 444 00:13:59,200 --> 00:14:00,600 Well, not hack it exactly. 445 00:14:00,600 --> 00:14:03,280 More like find a serious vulnerability. 446 00:14:03,280 --> 00:14:03,720 OK. 447 00:14:03,720 --> 00:14:04,360 Same difference. 448 00:14:04,360 --> 00:14:05,880 But why are they doing that? 449 00:14:05,880 --> 00:14:08,880 Well, they're really emphasizing security with this service. 450 00:14:08,880 --> 00:14:11,000 I guess they want to make sure people feel safe using it. 451 00:14:11,000 --> 00:14:11,960 Exactly. 452 00:14:11,960 --> 00:14:14,000 Especially since it's all about expanding 453 00:14:14,000 --> 00:14:16,400 their on-device AI capabilities. 454 00:14:16,400 --> 00:14:16,720 Right. 455 00:14:16,720 --> 00:14:17,680 Apple intelligence. 456 00:14:17,680 --> 00:14:19,840 They want to make sure that user data stays private. 457 00:14:19,840 --> 00:14:20,680 That's good. 458 00:14:20,680 --> 00:14:22,000 Privacy is important. 459 00:14:22,000 --> 00:14:24,640 So what happens if someone actually finds a flaw? 460 00:14:24,640 --> 00:14:28,880 Well, they get paid anywhere from $150,000 to $1 million, 461 00:14:28,880 --> 00:14:30,320 depending on how serious it is. 462 00:14:30,320 --> 00:14:31,680 Wow. 463 00:14:31,680 --> 00:14:33,480 That's a pretty good incentive to find those bugs. 464 00:14:33,480 --> 00:14:33,880 It is. 465 00:14:33,880 --> 00:14:35,080 It's a smart move by Apple. 466 00:14:35,080 --> 00:14:36,680 It shows they're serious about security. 467 00:14:36,680 --> 00:14:38,680 And it gets people talking about their new service. 468 00:14:38,680 --> 00:14:39,200 Exactly. 469 00:14:39,200 --> 00:14:39,760 It's a win-win. 470 00:14:39,760 --> 00:14:43,640 But you know, all this talk about AI and companies and security. 471 00:14:43,640 --> 00:14:46,040 It's easy to forget about the people involved. 472 00:14:46,040 --> 00:14:48,600 The ones whose jobs might be affected by all this. 473 00:14:48,600 --> 00:14:49,880 That's a good point. 474 00:14:49,880 --> 00:14:52,720 There's a lot of fear about AI taking over jobs. 475 00:14:52,720 --> 00:14:53,960 And it's already happening, right? 476 00:14:53,960 --> 00:14:56,840 Factories, customer service, even writing. 477 00:14:56,840 --> 00:14:57,680 Yeah. 478 00:14:57,680 --> 00:15:00,680 But it's important to remember that AI is a tool. 479 00:15:00,680 --> 00:15:04,760 And like any tool, it depends how you use it. 480 00:15:04,760 --> 00:15:06,480 So you're saying it's not all doom and gloom. 481 00:15:06,480 --> 00:15:07,200 Not at all. 482 00:15:07,200 --> 00:15:10,520 AI can also create new opportunities. 483 00:15:10,520 --> 00:15:11,840 New jobs. 484 00:15:11,840 --> 00:15:12,440 In what? 485 00:15:12,440 --> 00:15:15,480 Well, we'll need people to design AI systems, 486 00:15:15,480 --> 00:15:17,680 to build them, to maintain them. 487 00:15:17,680 --> 00:15:19,600 And to make sure they're being used ethically. 488 00:15:19,600 --> 00:15:20,520 Absolutely. 489 00:15:20,520 --> 00:15:22,800 And there are also jobs that AI can't do. 490 00:15:22,800 --> 00:15:23,480 Like what? 491 00:15:23,480 --> 00:15:25,840 Jobs that require creativity, critical thinking, 492 00:15:25,840 --> 00:15:26,560 things like that. 493 00:15:26,560 --> 00:15:27,360 The human touch. 494 00:15:27,360 --> 00:15:27,880 Exactly. 495 00:15:27,880 --> 00:15:29,560 So it's not about AI replacing us. 496 00:15:29,560 --> 00:15:31,760 It's about AI working with us. 497 00:15:31,760 --> 00:15:33,360 But people still need to adapt, right? 498 00:15:33,360 --> 00:15:34,040 Absolutely. 499 00:15:34,040 --> 00:15:37,160 We have to learn new skills, keep up with the technology. 500 00:15:37,160 --> 00:15:39,400 It's like a whole new world of work is opening up. 501 00:15:39,400 --> 00:15:40,000 It is. 502 00:15:40,000 --> 00:15:42,120 And it's going to require us to be flexible, 503 00:15:42,120 --> 00:15:44,960 to be adaptable, to be lifelong learners. 504 00:15:44,960 --> 00:15:46,760 It's a challenge, but it's also exciting. 505 00:15:46,760 --> 00:15:48,240 I think so too. 506 00:15:48,240 --> 00:15:49,920 It's all about how we approach it. 507 00:15:49,920 --> 00:15:52,640 We can choose to see AI as a threat. 508 00:15:52,640 --> 00:15:53,680 Or as an opportunity. 509 00:15:53,680 --> 00:15:54,320 Exactly. 510 00:15:54,320 --> 00:15:57,560 We have the power to shape how AI is used. 511 00:15:57,560 --> 00:15:59,040 And to make sure it's used for good. 512 00:15:59,040 --> 00:16:00,320 That's the key. 513 00:16:00,320 --> 00:16:02,440 Now let's shift gears a bit and talk about, 514 00:16:02,440 --> 00:16:04,400 it's all about how we approach it. 515 00:16:04,400 --> 00:16:08,600 We can choose to see AI as a threat or as an opportunity. 516 00:16:08,600 --> 00:16:10,000 It's really up to us. 517 00:16:10,000 --> 00:16:10,720 That's a good point. 518 00:16:10,720 --> 00:16:13,640 We have the power to decide how AI is used. 519 00:16:13,640 --> 00:16:15,840 To make sure it's used ethically and responsibly. 520 00:16:15,840 --> 00:16:16,560 Exactly. 521 00:16:16,560 --> 00:16:19,480 And that brings us to one of the biggest challenges with AI. 522 00:16:19,480 --> 00:16:20,280 What's that? 523 00:16:20,280 --> 00:16:21,880 Bias. 524 00:16:21,880 --> 00:16:22,640 Right. 525 00:16:22,640 --> 00:16:24,120 We've seen examples of that already. 526 00:16:24,120 --> 00:16:26,000 Like facial recognition systems that 527 00:16:26,000 --> 00:16:28,520 are less accurate for certain groups of people. 528 00:16:28,520 --> 00:16:29,000 Yeah. 529 00:16:29,000 --> 00:16:30,800 And hiring algorithms that discriminate 530 00:16:30,800 --> 00:16:32,000 against certain candidates. 531 00:16:32,000 --> 00:16:34,360 Even loan approval systems that are biased. 532 00:16:34,360 --> 00:16:35,320 That's a serious problem. 533 00:16:35,320 --> 00:16:35,840 It is. 534 00:16:35,840 --> 00:16:39,480 We need to make sure that AI systems are fair and unbiased. 535 00:16:39,480 --> 00:16:40,520 So how do we do that? 536 00:16:40,520 --> 00:16:42,960 Well, for starters, we need more diversity 537 00:16:42,960 --> 00:16:44,880 in the people who are developing AI. 538 00:16:44,880 --> 00:16:45,680 More diversity. 539 00:16:45,680 --> 00:16:46,680 Yeah. 540 00:16:46,680 --> 00:16:48,200 People from different backgrounds, different cultures, 541 00:16:48,200 --> 00:16:49,320 different perspectives. 542 00:16:49,320 --> 00:16:52,400 So they can spot potential biases that others might miss. 543 00:16:52,400 --> 00:16:53,440 Exactly. 544 00:16:53,440 --> 00:16:56,400 And we also need to be careful about the data that's 545 00:16:56,400 --> 00:16:58,120 used to train AI systems. 546 00:16:58,120 --> 00:17:00,000 Because the data is biased. 547 00:17:00,000 --> 00:17:01,680 The AI will be biased too. 548 00:17:01,680 --> 00:17:03,800 It's like teaching a kid bad habits. 549 00:17:03,800 --> 00:17:05,560 So it's a multifaceted problem. 550 00:17:05,560 --> 00:17:06,440 It is. 551 00:17:06,440 --> 00:17:08,120 We need to be thinking about bias 552 00:17:08,120 --> 00:17:09,840 at every stage of AI development. 553 00:17:09,840 --> 00:17:12,960 From the data to the algorithms to the people 554 00:17:12,960 --> 00:17:13,800 who are building it. 555 00:17:13,800 --> 00:17:14,600 It's a big challenge. 556 00:17:14,600 --> 00:17:15,080 Yeah. 557 00:17:15,080 --> 00:17:18,160 But it's one that we have to address if we want AI 558 00:17:18,160 --> 00:17:19,440 to benefit everyone. 559 00:17:19,440 --> 00:17:20,120 Absolutely. 560 00:17:20,120 --> 00:17:22,720 We need open discussion and collaboration 561 00:17:22,720 --> 00:17:24,320 between everyone involved. 562 00:17:24,320 --> 00:17:27,440 Researchers, developers, policymakers, the public. 563 00:17:27,440 --> 00:17:28,440 It's a team effort. 564 00:17:28,440 --> 00:17:28,940 It is. 565 00:17:28,940 --> 00:17:30,160 We're all in this together. 566 00:17:30,160 --> 00:17:33,280 And we need to figure out how to use AI in a way that's 567 00:17:33,280 --> 00:17:36,560 fair, ethical, and beneficial to all of humanity. 568 00:17:36,560 --> 00:17:37,760 Well said. 569 00:17:37,760 --> 00:17:41,040 This has been an incredible deep dive into the world of AI. 570 00:17:41,040 --> 00:17:42,880 We've covered a lot of ground today. 571 00:17:42,880 --> 00:17:45,600 From whale conversations to pig emotions 572 00:17:45,600 --> 00:17:49,160 to million dollar bounties and antitrust lawsuits. 573 00:17:49,160 --> 00:17:51,600 It's amazing how much is happening in this field. 574 00:17:51,600 --> 00:17:54,640 Every day brings new breakthroughs, new challenges, 575 00:17:54,640 --> 00:17:56,080 new possibilities. 576 00:17:56,080 --> 00:17:59,200 It's an exciting time to be following AI. 577 00:17:59,200 --> 00:18:01,040 But it's also a time to be cautious, 578 00:18:01,040 --> 00:18:03,280 to be thoughtful, to ask the tough questions. 579 00:18:03,280 --> 00:18:03,960 Absolutely. 580 00:18:03,960 --> 00:18:05,640 AI is a powerful tool. 581 00:18:05,640 --> 00:18:08,520 And like any powerful tool, it can be used for good or for bad. 582 00:18:08,520 --> 00:18:11,120 It's up to us to make sure it's used for good. 583 00:18:11,120 --> 00:18:14,240 To shape the future of AI in a way that benefits everyone. 584 00:18:14,240 --> 00:18:15,080 Exactly. 585 00:18:15,080 --> 00:18:18,280 So stay informed, stay engaged, and keep 586 00:18:18,280 --> 00:18:19,600 those conversations going. 587 00:18:19,600 --> 00:18:21,480 The AI revolution is just beginning. 588 00:18:21,480 --> 00:18:23,720 And we'll be here to cover it every step of the way. 589 00:18:23,720 --> 00:18:26,000 Thanks for listening to the Daily AI News Podcast. 590 00:18:26,000 --> 00:18:42,040 And stay tuned for more.