1 00:00:00,000 --> 00:00:02,680 Welcome to the Daily AI News Podcast. 2 00:00:03,280 --> 00:00:06,040 Looks like you are really into this AI stuff, huh? 3 00:00:07,400 --> 00:00:11,240 This week, we're diving deep, especially with all the recent developments. 4 00:00:11,280 --> 00:00:12,120 And the drama. 5 00:00:12,240 --> 00:00:13,880 Yeah, for sure, some drama too. 6 00:00:14,280 --> 00:00:16,520 And we've got your stack of articles right here. 7 00:00:16,800 --> 00:00:17,520 Awesome. 8 00:00:17,640 --> 00:00:20,920 So let's try to figure out what's really important and what's just hype, you know. 9 00:00:20,960 --> 00:00:23,080 It's been a wild ride lately, that's for sure. 10 00:00:23,160 --> 00:00:23,840 Yeah, no kidding. 11 00:00:24,160 --> 00:00:28,280 Okay, so first up, remember that big chatgy PT outage on the 26th? 12 00:00:28,320 --> 00:00:29,080 Oh yeah. 13 00:00:29,080 --> 00:00:30,760 Everyone was freaking out. 14 00:00:30,800 --> 00:00:32,000 The memes were great. 15 00:00:32,040 --> 00:00:32,960 Yeah, it was crazy. 16 00:00:33,000 --> 00:00:34,240 There's a pretty big outage. 17 00:00:34,280 --> 00:00:35,080 So what happened? 18 00:00:35,120 --> 00:00:37,000 OpenAI was kind of vague about it. 19 00:00:37,120 --> 00:00:37,800 What did they say? 20 00:00:37,840 --> 00:00:41,040 They blamed like an upstream provider or something. 21 00:00:41,080 --> 00:00:42,080 What's that even mean? 22 00:00:42,240 --> 00:00:45,800 I don't know, but it looks like it was probably connected to some power issues 23 00:00:45,840 --> 00:00:47,880 Microsoft was having at one of their data centers. 24 00:00:48,000 --> 00:00:48,600 Oh really? 25 00:00:48,640 --> 00:00:52,600 Yeah, because remember Microsoft is like their exclusive cloud provider. 26 00:00:52,640 --> 00:00:53,440 Oh right, right. 27 00:00:53,480 --> 00:00:57,680 And that outage messed up a bunch of other stuff too, like Xbox cloud gaming. 28 00:00:57,680 --> 00:00:59,000 That's interesting. 29 00:00:59,040 --> 00:00:59,480 Yes. 30 00:00:59,520 --> 00:01:00,720 It is kind of funny though, right? 31 00:01:00,760 --> 00:01:07,320 I mean this super advanced AI that's supposed to be able to write like Shakespeare and all that. 32 00:01:07,360 --> 00:01:07,840 Yeah. 33 00:01:07,880 --> 00:01:09,200 Taken down by a power outage. 34 00:01:09,360 --> 00:01:11,880 I think they'd have some kind of backup system, right? 35 00:01:11,920 --> 00:01:15,000 Yeah, well this isn't even the first time chatgy PT has gone offline. 36 00:01:15,040 --> 00:01:15,600 Really? 37 00:01:15,640 --> 00:01:20,960 Yeah, they had one back in June and then just earlier this month, their new text to video thing. 38 00:01:21,000 --> 00:01:21,840 Oh, Sora. 39 00:01:21,960 --> 00:01:23,360 Yeah, Sora went down too. 40 00:01:23,480 --> 00:01:26,720 Oh, make sure you wonder how reliable all this stuff really is. 41 00:01:26,720 --> 00:01:29,760 Well, especially with all the hype about it being the future of everything, you know? 42 00:01:29,800 --> 00:01:30,760 Yeah, for sure. 43 00:01:30,800 --> 00:01:39,720 And speaking of things that make you wonder, the Guardian found this security flaw in that new chatgy PT search. 44 00:01:39,760 --> 00:01:40,320 Yeah. 45 00:01:40,360 --> 00:01:41,760 You know, the AI search engine. 46 00:01:41,800 --> 00:01:42,200 Right. 47 00:01:42,240 --> 00:01:43,960 Some researchers were able to trick it. 48 00:01:44,000 --> 00:01:44,880 How? 49 00:01:44,920 --> 00:01:52,640 Well, they got it to ignore like bad reviews and even to spit out harmful code just by using hidden text on websites. 50 00:01:52,680 --> 00:01:53,400 Oh wow. 51 00:01:53,440 --> 00:01:54,680 Crazy, right? 52 00:01:54,720 --> 00:01:55,840 Yeah, that's a pretty big deal. 53 00:01:55,880 --> 00:01:56,440 I know. 54 00:01:56,440 --> 00:02:00,120 It's the first demonstrated attack on a live AI search product. 55 00:02:00,160 --> 00:02:01,160 That we know of at least. 56 00:02:01,200 --> 00:02:06,360 Yeah, Google's probably been dealing with this kind of thing for years with their regular search, you know? 57 00:02:06,400 --> 00:02:07,560 Yeah, I guess so. 58 00:02:07,600 --> 00:02:11,400 But AI search is so new, it's kind of uncharted territory. 59 00:02:11,440 --> 00:02:13,000 It really makes you think, right? 60 00:02:13,040 --> 00:02:13,800 Oh. 61 00:02:13,840 --> 00:02:16,720 Like is AI search inherently less secure? 62 00:02:17,960 --> 00:02:18,760 Interesting question. 63 00:02:18,800 --> 00:02:21,320 Or is it just, you know, still early days? 64 00:02:21,920 --> 00:02:22,760 What do you think? 65 00:02:22,800 --> 00:02:25,560 I think it's definitely something to watch closely. 66 00:02:25,560 --> 00:02:27,560 But AI is here to stay. 67 00:02:27,600 --> 00:02:28,120 Right. 68 00:02:28,160 --> 00:02:30,040 And search is just one part of it. 69 00:02:30,080 --> 00:02:30,880 Yeah, true. 70 00:02:31,400 --> 00:02:33,960 Okay, well, let's keep talking about these potential problems with AI. 71 00:02:34,000 --> 00:02:34,280 Okay. 72 00:02:34,320 --> 00:02:35,480 Let's talk about Microsoft. 73 00:02:35,520 --> 00:02:38,280 What are they doing with their AI ambitions, you know? 74 00:02:38,320 --> 00:02:42,280 They're putting a ton of resources into co-pilot, their AI assistant. 75 00:02:42,320 --> 00:02:42,840 Right. 76 00:02:42,880 --> 00:02:45,480 But it doesn't seem like it's going as well as they hoped. 77 00:02:45,520 --> 00:02:47,880 It's interesting what they're doing though, their strategy. 78 00:02:47,920 --> 00:02:48,520 What do you mean? 79 00:02:48,560 --> 00:02:53,720 Well, they're basically bundling co-pilot with all their Microsoft 365 subscriptions. 80 00:02:53,760 --> 00:02:54,360 Oh, really? 81 00:02:54,360 --> 00:02:57,160 And raising prices, even if people don't want it. 82 00:02:57,200 --> 00:02:57,800 Oh, wow. 83 00:02:57,840 --> 00:03:00,560 It's causing some problems, to say the least. 84 00:03:00,600 --> 00:03:01,120 Yeah. 85 00:03:01,920 --> 00:03:04,400 Remember that guy in Australia, Alistair Fleming? 86 00:03:04,440 --> 00:03:06,480 He was getting all these co-pilot pop-ups. 87 00:03:06,520 --> 00:03:07,280 Yeah. 88 00:03:07,320 --> 00:03:09,480 And then they upped his subscription price. 89 00:03:10,440 --> 00:03:12,320 He was comparing to Clippy, you know. 90 00:03:12,360 --> 00:03:13,440 That little paperclip? 91 00:03:13,480 --> 00:03:13,960 Yeah. 92 00:03:14,520 --> 00:03:15,520 From like way back. 93 00:03:15,560 --> 00:03:18,040 But you know, from Microsoft's side, it makes sense. 94 00:03:18,080 --> 00:03:18,680 Right there. 95 00:03:18,720 --> 00:03:20,800 They've invested a ton in AI. 96 00:03:20,840 --> 00:03:21,240 Right. 97 00:03:21,280 --> 00:03:22,680 Especially with open AI. 98 00:03:22,720 --> 00:03:23,080 Yeah. 99 00:03:23,080 --> 00:03:24,640 And they want to see a return. 100 00:03:24,680 --> 00:03:26,840 So they're betting big on co-pilot. 101 00:03:26,880 --> 00:03:30,600 Yeah, but co-pilot is really behind chat GPT in the consumer market. 102 00:03:30,640 --> 00:03:31,480 Oh, yeah. 103 00:03:31,520 --> 00:03:33,240 Yeah, download numbers are way lower. 104 00:03:33,280 --> 00:03:34,680 What do you think that is? 105 00:03:34,720 --> 00:03:37,560 Well, I think there are concerns about accuracy. 106 00:03:37,600 --> 00:03:38,080 Right. 107 00:03:38,120 --> 00:03:38,960 Data privacy. 108 00:03:39,000 --> 00:03:40,120 You're sure. 109 00:03:40,160 --> 00:03:42,280 And the cost is a factor. 110 00:03:42,320 --> 00:03:43,000 Definitely. 111 00:03:43,040 --> 00:03:45,720 And honestly, I think people just prefer chat GPT. 112 00:03:45,760 --> 00:03:46,600 I think you're right. 113 00:03:46,640 --> 00:03:49,280 And then there's also that whole thing with Microsoft and open AI. 114 00:03:49,320 --> 00:03:50,560 Yeah, what's that? 115 00:03:50,600 --> 00:03:52,960 I mean, Microsoft needs open AI's technology. 116 00:03:52,960 --> 00:03:53,440 Yeah. 117 00:03:53,480 --> 00:03:56,080 But at the same time, open AI is a competitor. 118 00:03:56,120 --> 00:03:56,520 Right. 119 00:03:56,560 --> 00:03:58,440 Especially with chat GPT enterprise. 120 00:03:58,480 --> 00:03:59,280 Yeah, exactly. 121 00:03:59,320 --> 00:04:00,680 It's a tricky situation. 122 00:04:00,720 --> 00:04:01,360 It is. 123 00:04:01,400 --> 00:04:03,960 But remember, this is just the beginning for Microsoft. 124 00:04:04,000 --> 00:04:04,560 What do you mean? 125 00:04:04,600 --> 00:04:07,640 They're already talking about developing AI agents. 126 00:04:07,680 --> 00:04:08,440 Oh, wow. 127 00:04:08,480 --> 00:04:11,040 That can do even more complex tasks. 128 00:04:11,080 --> 00:04:12,640 So they're playing the long game. 129 00:04:12,680 --> 00:04:13,560 Definitely. 130 00:04:13,600 --> 00:04:14,440 Interesting. 131 00:04:14,480 --> 00:04:16,720 Okay, so let's shift gears a bit. 132 00:04:16,760 --> 00:04:17,280 Uh-huh. 133 00:04:17,320 --> 00:04:21,680 Let's talk about the rise of open AI. 134 00:04:21,720 --> 00:04:21,960 Okay. 135 00:04:21,960 --> 00:04:22,960 Definitely with that. 136 00:04:23,000 --> 00:04:25,240 So there's this Chinese lab, DeepSeq. 137 00:04:25,280 --> 00:04:26,200 Okay. 138 00:04:26,240 --> 00:04:30,280 And they just released DeepSeq V3. 139 00:04:30,320 --> 00:04:31,080 V3. 140 00:04:31,120 --> 00:04:31,560 Yeah. 141 00:04:31,600 --> 00:04:33,200 And it's making some noise. 142 00:04:33,240 --> 00:04:33,800 What is it? 143 00:04:33,840 --> 00:04:36,760 It's one of the most powerful open AI models out there. 144 00:04:36,800 --> 00:04:37,800 So what does that mean? 145 00:04:37,840 --> 00:04:39,280 Like open. 146 00:04:39,320 --> 00:04:40,160 Good question. 147 00:04:40,200 --> 00:04:42,440 Open means that developers can actually download it. 148 00:04:42,480 --> 00:04:43,440 Oh, wow. 149 00:04:43,480 --> 00:04:46,680 And they can tinker with it, even use it for commercial stuff. 150 00:04:46,720 --> 00:04:48,200 So it's different from chat GPT. 151 00:04:48,240 --> 00:04:51,000 Yeah, totally different from these closed models like chat GPT. 152 00:04:51,000 --> 00:04:53,160 You can only access those through their API. 153 00:04:53,200 --> 00:04:55,240 So chat GPT is closed. 154 00:04:55,280 --> 00:04:55,680 Right. 155 00:04:55,720 --> 00:04:56,320 And get this. 156 00:04:56,360 --> 00:04:58,640 DeepSeq V3, it's not just open. 157 00:04:58,680 --> 00:05:00,120 It's really good. 158 00:05:00,160 --> 00:05:00,880 Is it? 159 00:05:00,920 --> 00:05:01,720 Yeah. 160 00:05:01,760 --> 00:05:05,960 In some tests, it even beats out some closed models like GPT-4O 161 00:05:06,000 --> 00:05:10,640 and Meta's Lama 3.1 405B. 162 00:05:10,680 --> 00:05:11,600 Impressive. 163 00:05:11,640 --> 00:05:15,640 And the crazy part is they did it all for way less money than open AI 164 00:05:15,680 --> 00:05:17,080 spent on GPT-4. 165 00:05:17,120 --> 00:05:17,960 Really? 166 00:05:18,000 --> 00:05:18,360 How? 167 00:05:18,360 --> 00:05:21,960 They used a whole data center full of those NVIDIA E800 GPUs. 168 00:05:22,000 --> 00:05:22,680 Oh, wow. 169 00:05:22,720 --> 00:05:24,960 The ones that the US is trying to stop them from getting. 170 00:05:25,000 --> 00:05:25,920 Yeah, exactly. 171 00:05:25,960 --> 00:05:26,960 That's pretty amazing. 172 00:05:27,000 --> 00:05:27,560 It is. 173 00:05:27,600 --> 00:05:29,680 I mean, even with those restrictions, 174 00:05:29,720 --> 00:05:32,000 they're still innovating and competing. 175 00:05:32,040 --> 00:05:35,160 Okay, so DeepSeq V3 is open. 176 00:05:35,200 --> 00:05:35,720 Right. 177 00:05:35,760 --> 00:05:37,080 It's powerful. 178 00:05:37,120 --> 00:05:38,160 It's cost effective. 179 00:05:38,200 --> 00:05:39,160 Yeah. 180 00:05:39,200 --> 00:05:40,560 But there's got to be a catch, right? 181 00:05:40,600 --> 00:05:42,440 Well, there is one thing. 182 00:05:42,480 --> 00:05:42,880 Okay. 183 00:05:42,920 --> 00:05:44,160 Being a Chinese company, 184 00:05:44,200 --> 00:05:46,800 DeepSeq has to play by the government's rules, you know? 185 00:05:46,840 --> 00:05:47,000 Right. 186 00:05:47,000 --> 00:05:50,680 So the model, it avoids anything politically sensitive. 187 00:05:50,720 --> 00:05:52,080 Makes sense in that context. 188 00:05:52,120 --> 00:05:53,520 It's a different environment for sure. 189 00:05:53,560 --> 00:05:54,280 Yeah, definitely. 190 00:05:54,320 --> 00:05:56,160 But it's still a pretty exciting development. 191 00:05:56,200 --> 00:05:57,000 Absolutely. 192 00:05:57,040 --> 00:05:58,640 Open AI could really change things up. 193 00:05:58,680 --> 00:06:01,520 Yeah, it has the potential to really change the game. 194 00:06:01,560 --> 00:06:02,360 How so? 195 00:06:02,400 --> 00:06:05,080 Well, more competition, more innovation, 196 00:06:05,120 --> 00:06:08,280 and more accessible AI for everyone. 197 00:06:08,320 --> 00:06:09,360 Yeah, it's a good thing. 198 00:06:09,400 --> 00:06:10,760 It's definitely a space to watch. 199 00:06:10,800 --> 00:06:11,400 For sure. 200 00:06:11,440 --> 00:06:14,000 So speaking of things that could change things up. 201 00:06:14,040 --> 00:06:14,520 Yeah. 202 00:06:14,560 --> 00:06:16,520 Let's talk about AI and art. 203 00:06:16,520 --> 00:06:17,000 Okay. 204 00:06:17,040 --> 00:06:19,440 You hear about that new AI art magazine. 205 00:06:19,480 --> 00:06:20,200 I have. 206 00:06:20,240 --> 00:06:21,120 It's pretty cool. 207 00:06:21,160 --> 00:06:21,640 I have. 208 00:06:21,680 --> 00:06:23,520 A whole magazine about AI art. 209 00:06:23,560 --> 00:06:24,560 Yeah. 210 00:06:24,600 --> 00:06:25,280 It's amazing. 211 00:06:25,320 --> 00:06:28,920 And it's not just like showcasing cool artwork. 212 00:06:28,960 --> 00:06:32,760 They're really trying to chronicle how AI art is changing. 213 00:06:32,800 --> 00:06:33,280 Okay. 214 00:06:33,320 --> 00:06:37,280 And exploring like all the big questions about creativity. 215 00:06:37,320 --> 00:06:37,960 That makes sense. 216 00:06:38,000 --> 00:06:40,160 And this new digital age, you know? 217 00:06:40,200 --> 00:06:41,320 Yeah, for sure. 218 00:06:41,360 --> 00:06:43,320 I did see that they are independently funded. 219 00:06:43,360 --> 00:06:44,360 Right. 220 00:06:44,400 --> 00:06:44,960 That's great. 221 00:06:45,000 --> 00:06:45,720 Which is awesome. 222 00:06:45,720 --> 00:06:49,160 They can really dive into all those like complicated issues 223 00:06:49,200 --> 00:06:51,080 without any commercial pressure. 224 00:06:51,120 --> 00:06:55,320 Plus, their first issue has a jury member created by AI. 225 00:06:55,360 --> 00:06:56,800 No way, really. 226 00:06:56,840 --> 00:06:58,120 Talk about blurring the lines, right? 227 00:06:58,160 --> 00:07:00,480 Between human and machine creativity, right? 228 00:07:00,520 --> 00:07:00,880 Yeah. 229 00:07:00,920 --> 00:07:02,960 And there's an essay in that first issue. 230 00:07:03,000 --> 00:07:03,360 Oh, yeah. 231 00:07:03,400 --> 00:07:06,120 By that famous designer, David Carson. 232 00:07:06,160 --> 00:07:06,920 Okay. 233 00:07:06,960 --> 00:07:09,600 And he asks this really thought-provoking question. 234 00:07:09,640 --> 00:07:10,080 What's that? 235 00:07:10,120 --> 00:07:12,320 If you just give a command to a machine, 236 00:07:12,360 --> 00:07:13,880 are you really an artist? 237 00:07:13,920 --> 00:07:15,400 Wow, that's a good one. 238 00:07:15,400 --> 00:07:16,080 Right. 239 00:07:16,120 --> 00:07:18,560 That's a question that's going to be debated for a long time. 240 00:07:18,600 --> 00:07:19,560 I think so. 241 00:07:19,600 --> 00:07:24,880 So it's cool that AI art is pushing the boundaries like that. 242 00:07:24,920 --> 00:07:25,480 Yeah. 243 00:07:25,520 --> 00:07:27,880 And starting all these philosophical conversations. 244 00:07:27,920 --> 00:07:28,720 Definitely. 245 00:07:28,760 --> 00:07:32,040 But it's also having some unintended consequences. 246 00:07:32,080 --> 00:07:32,960 Like what? 247 00:07:33,000 --> 00:07:36,600 Specifically in the world of like traditional crafts. 248 00:07:36,640 --> 00:07:39,680 You talk about those AI-generated lace making books. 249 00:07:39,720 --> 00:07:40,720 Yeah, exactly. 250 00:07:40,760 --> 00:07:42,040 Oh, yeah, those are weird. 251 00:07:42,080 --> 00:07:42,800 It's crazy. 252 00:07:42,800 --> 00:07:47,000 So scammers are using AI to create these fake lace making books. 253 00:07:47,040 --> 00:07:47,560 Right. 254 00:07:47,600 --> 00:07:50,000 And they're full of either stolen designs. 255 00:07:50,040 --> 00:07:50,640 Oh, wow. 256 00:07:50,680 --> 00:07:53,080 Or instructions that just don't even make sense. 257 00:07:53,120 --> 00:07:53,800 Seriously. 258 00:07:53,840 --> 00:07:55,560 And then they sell them online to beginners. 259 00:07:55,600 --> 00:07:56,160 Oh, no. 260 00:07:56,200 --> 00:07:57,200 Who don't know any better. 261 00:07:57,240 --> 00:07:58,520 It's a real problem. 262 00:07:58,560 --> 00:07:59,000 It is. 263 00:07:59,040 --> 00:08:00,560 And it's not just lace making. 264 00:08:00,600 --> 00:08:04,560 It could happen anywhere that AI can generate content. 265 00:08:04,600 --> 00:08:05,400 Oh, that's true. 266 00:08:05,440 --> 00:08:06,160 Think about it. 267 00:08:06,200 --> 00:08:06,600 Yeah. 268 00:08:06,640 --> 00:08:09,480 Music, writing, even scientific research. 269 00:08:09,520 --> 00:08:10,360 Wow, you're right. 270 00:08:10,400 --> 00:08:11,040 Yeah. 271 00:08:11,040 --> 00:08:15,080 So experienced lace makers are worried, understandably. 272 00:08:15,120 --> 00:08:18,120 They're saying that this is going to devalue the skill. 273 00:08:18,160 --> 00:08:18,760 Right. 274 00:08:18,800 --> 00:08:22,960 The artistry of lace making and make it harder for people to learn the craft. 275 00:08:23,000 --> 00:08:23,520 Yeah. 276 00:08:23,560 --> 00:08:26,800 And in the end, it could hurt the people who actually make a living from it. 277 00:08:26,840 --> 00:08:28,240 That's a good point. 278 00:08:28,280 --> 00:08:32,160 It's a reminder that as AI gets more and more sophisticated. 279 00:08:32,200 --> 00:08:33,040 Right. 280 00:08:33,080 --> 00:08:36,680 We have to watch out for how it can be misused. 281 00:08:36,720 --> 00:08:37,200 Definitely. 282 00:08:37,240 --> 00:08:39,400 We need to like harness its power. 283 00:08:39,440 --> 00:08:40,200 For good? 284 00:08:40,200 --> 00:08:41,000 Yeah, for good. 285 00:08:41,040 --> 00:08:41,440 Yeah. 286 00:08:41,480 --> 00:08:44,800 But also like protect against the downsides, you know. 287 00:08:44,840 --> 00:08:45,640 Absolutely. 288 00:08:45,680 --> 00:08:46,480 It's a balancing act. 289 00:08:46,520 --> 00:08:47,040 It is. 290 00:08:47,080 --> 00:08:48,520 For sure. 291 00:08:48,560 --> 00:08:49,440 But on a lighter note. 292 00:08:49,480 --> 00:08:50,360 Okay. 293 00:08:50,400 --> 00:08:51,640 This is seriously cool. 294 00:08:51,680 --> 00:08:52,480 What? 295 00:08:52,520 --> 00:08:55,120 Researchers at Osaka University are working on something. 296 00:08:55,160 --> 00:08:55,760 Uh-huh. 297 00:08:55,800 --> 00:08:58,040 They could help us get over that uncanny valley thing. 298 00:08:58,080 --> 00:08:59,200 Oh, in androids? 299 00:08:59,240 --> 00:09:00,880 Yeah. 300 00:09:00,920 --> 00:09:02,520 You know that creepy feeling you get? 301 00:09:02,560 --> 00:09:05,880 Yeah, when something's like almost human, but not quite. 302 00:09:05,920 --> 00:09:06,960 Exactly. 303 00:09:07,000 --> 00:09:09,800 The uncanny valley, that's a big hurdle in making androids. 304 00:09:09,800 --> 00:09:11,560 That people actually want to connect with. 305 00:09:11,600 --> 00:09:12,440 I know, it's weird. 306 00:09:12,480 --> 00:09:14,000 It's like our brains can't handle it. 307 00:09:14,040 --> 00:09:15,720 So what are they doing? 308 00:09:15,760 --> 00:09:17,280 They're developing this new technology. 309 00:09:17,320 --> 00:09:18,120 What is it? 310 00:09:18,160 --> 00:09:20,360 It uses waveform movements. 311 00:09:20,400 --> 00:09:21,040 Okay. 312 00:09:21,080 --> 00:09:23,720 To make Android faces more natural. 313 00:09:23,760 --> 00:09:24,960 Oh, cool. 314 00:09:25,000 --> 00:09:25,760 And expressive. 315 00:09:25,800 --> 00:09:27,080 How does it work? 316 00:09:27,120 --> 00:09:29,800 So instead of having those pre-programmed expressions. 317 00:09:29,840 --> 00:09:30,400 Right. 318 00:09:30,440 --> 00:09:33,160 They're representing different gestures, you know. 319 00:09:33,200 --> 00:09:33,960 Like what? 320 00:09:34,000 --> 00:09:34,920 Like breathing. 321 00:09:34,960 --> 00:09:35,360 Okay. 322 00:09:35,400 --> 00:09:35,960 Blinking. 323 00:09:36,000 --> 00:09:36,440 Uh-huh. 324 00:09:36,480 --> 00:09:37,440 Even yawning. 325 00:09:37,480 --> 00:09:38,240 Wow. 326 00:09:38,280 --> 00:09:39,400 As individual waves. 327 00:09:39,400 --> 00:09:40,400 That's interesting. 328 00:09:40,440 --> 00:09:44,000 And those can be combined to create these more complex expressions. 329 00:09:44,040 --> 00:09:44,480 Oh, wow. 330 00:09:44,520 --> 00:09:45,840 So they're not just robots anymore. 331 00:09:45,880 --> 00:09:46,840 Yeah. 332 00:09:46,880 --> 00:09:48,240 And they can do it in real time too. 333 00:09:48,280 --> 00:09:48,840 Really? 334 00:09:48,880 --> 00:09:50,840 So the androids can react to stuff. 335 00:09:50,880 --> 00:09:51,840 Oh, wow. 336 00:09:51,880 --> 00:09:52,640 It's pretty mind blowing. 337 00:09:52,680 --> 00:09:53,120 Yeah. 338 00:09:53,160 --> 00:09:56,120 The researchers think this is a game changer for making androids. 339 00:09:56,160 --> 00:09:57,240 I bet. 340 00:09:57,280 --> 00:10:02,040 They're hoping that it's going to lead to androids that are seen as having like a heart. 341 00:10:02,080 --> 00:10:02,640 Interesting. 342 00:10:02,680 --> 00:10:04,640 And can have more meaningful interactions with us. 343 00:10:04,680 --> 00:10:09,240 It's a glimpse into a future where androids could be more than just machines. 344 00:10:09,240 --> 00:10:09,840 Right? 345 00:10:09,880 --> 00:10:10,640 Yeah, exactly. 346 00:10:10,680 --> 00:10:14,520 Imagine them as companions or caregivers or even collaborators. 347 00:10:14,560 --> 00:10:15,000 Yeah. 348 00:10:15,040 --> 00:10:16,160 And it is exciting. 349 00:10:16,200 --> 00:10:18,120 But also a little unsettling. 350 00:10:18,160 --> 00:10:19,120 Yeah, for sure. 351 00:10:19,160 --> 00:10:21,800 It makes you think about what it even means to be human. 352 00:10:21,840 --> 00:10:22,800 It does. 353 00:10:22,840 --> 00:10:25,800 Okay, so we've talked about some potential problems with AI. 354 00:10:25,840 --> 00:10:26,240 Yeah. 355 00:10:26,280 --> 00:10:27,840 But let's get back to the positive stuff. 356 00:10:27,880 --> 00:10:28,800 Okay. 357 00:10:28,840 --> 00:10:32,640 Did you see that DeepSeq actually open sourced DeepSeq v3? 358 00:10:32,680 --> 00:10:33,600 Wait, really? 359 00:10:33,640 --> 00:10:33,960 Yeah. 360 00:10:34,000 --> 00:10:37,280 I thought we were just saying that it was open for developers to use. 361 00:10:37,320 --> 00:10:37,560 Right. 362 00:10:37,560 --> 00:10:39,280 So open sourcing goes even further. 363 00:10:39,320 --> 00:10:39,760 How? 364 00:10:39,800 --> 00:10:41,320 They're making the whole code-based public. 365 00:10:41,360 --> 00:10:42,760 Oh, wow. 366 00:10:42,800 --> 00:10:45,280 Anyone can download it, study it, even change it. 367 00:10:45,320 --> 00:10:46,400 Wow. 368 00:10:46,440 --> 00:10:47,920 They can even redistribute it. 369 00:10:47,960 --> 00:10:52,360 So they're like giving away the recipe for their secret sauce? 370 00:10:52,400 --> 00:10:52,960 Exactly. 371 00:10:53,000 --> 00:10:54,840 Why would they do that? 372 00:10:54,880 --> 00:10:56,240 Well, maybe they're just being nice. 373 00:10:56,280 --> 00:10:56,720 Nice. 374 00:10:56,760 --> 00:10:59,040 You know, sharing the AI love. 375 00:10:59,080 --> 00:11:01,040 But it's also a good business move, right? 376 00:11:01,080 --> 00:11:01,400 Yeah. 377 00:11:01,440 --> 00:11:06,400 So by open sourcing it, they're getting the whole global developer community involved. 378 00:11:06,440 --> 00:11:06,720 Right. 379 00:11:06,720 --> 00:11:10,400 And more people using it and improving it means faster innovation. 380 00:11:10,440 --> 00:11:12,080 So it benefits DeepSeq, too. 381 00:11:12,120 --> 00:11:12,880 Exactly. 382 00:11:12,920 --> 00:11:13,640 It's a win-win. 383 00:11:13,680 --> 00:11:13,960 Yeah. 384 00:11:14,000 --> 00:11:20,520 Developers get this powerful AI and DeepSeq gets tons of people helping them make it better. 385 00:11:20,560 --> 00:11:25,920 It's a great example of how open source can really speed up progress in AI. 386 00:11:25,960 --> 00:11:26,520 For sure. 387 00:11:26,560 --> 00:11:32,040 And it might even put pressure on other companies to be more transparent with their AI development. 388 00:11:32,080 --> 00:11:33,400 It's definitely a trend to watch. 389 00:11:33,440 --> 00:11:33,760 Yeah. 390 00:11:33,760 --> 00:11:37,200 Who knows, maybe one day all the big AI models will be open source. 391 00:11:37,240 --> 00:11:38,040 That would be something else. 392 00:11:38,080 --> 00:11:39,200 It would be an all new world. 393 00:11:39,240 --> 00:11:39,800 It would. 394 00:11:39,840 --> 00:11:46,680 But for now, let's switch gears again and talk about something a little more tangible. 395 00:11:46,720 --> 00:11:47,320 Like what? 396 00:11:47,360 --> 00:11:48,680 Goyai generated lace. 397 00:11:48,720 --> 00:11:51,520 Wait, are we talking about those fake lace making books again? 398 00:11:51,560 --> 00:11:53,080 No, no, not the books. 399 00:11:53,120 --> 00:11:56,960 This is about real lace being made by AI. 400 00:11:57,000 --> 00:11:57,400 Seriously. 401 00:11:57,440 --> 00:11:58,640 Yeah, there's this company. 402 00:11:58,680 --> 00:11:59,280 What are they doing? 403 00:11:59,280 --> 00:12:03,920 They're using AI to design and make these really intricate lace patterns. 404 00:12:03,960 --> 00:12:05,160 I didn't even know that was possible. 405 00:12:05,200 --> 00:12:09,760 So they've trained an AI on a huge data set of traditional lace designs. 406 00:12:09,800 --> 00:12:10,320 Wow. 407 00:12:10,360 --> 00:12:13,120 And now it can come up with totally new patterns. 408 00:12:13,160 --> 00:12:13,480 Really? 409 00:12:13,520 --> 00:12:17,360 That are both like beautiful to look at and structurally sound. 410 00:12:17,400 --> 00:12:19,480 So it's not just copying old designs. 411 00:12:19,520 --> 00:12:21,160 No, it's creating something new. 412 00:12:21,200 --> 00:12:21,920 That's amazing. 413 00:12:21,960 --> 00:12:25,440 It's like having a master lace maker right there in your computer. 414 00:12:25,480 --> 00:12:28,760 But isn't that going to put traditional lace makers out of business? 415 00:12:28,760 --> 00:12:30,280 That's a good question. 416 00:12:30,320 --> 00:12:33,320 But this company sees it more like a collaboration. 417 00:12:33,360 --> 00:12:33,880 Really? 418 00:12:33,920 --> 00:12:39,960 They're working with lace making artisans to integrate the AI-generated designs into their work. 419 00:12:40,000 --> 00:12:40,560 Interesting. 420 00:12:40,600 --> 00:12:43,360 So it's about helping them, not replacing them. 421 00:12:43,400 --> 00:12:45,640 So it's augmenting their skills with AI. 422 00:12:45,680 --> 00:12:46,560 Exactly. 423 00:12:46,600 --> 00:12:52,440 It's a good example of how AI can be used to preserve and even elevate traditional crafts. 424 00:12:52,480 --> 00:12:53,240 I like that. 425 00:12:53,280 --> 00:12:53,560 Yeah. 426 00:12:53,600 --> 00:12:56,200 That's way more optimistic than those fake lace making books. 427 00:12:56,240 --> 00:12:56,720 Definitely. 428 00:12:56,720 --> 00:12:59,400 It's cool to see how AI can be used for good. 429 00:12:59,440 --> 00:13:04,120 And speaking of inspiration, did you see the AI Art Magazine's latest project? 430 00:13:04,160 --> 00:13:06,520 You mean that AI-generated fashion show? 431 00:13:06,560 --> 00:13:07,360 Yeah. 432 00:13:07,400 --> 00:13:08,640 Yeah, I saw the photos. 433 00:13:08,680 --> 00:13:09,480 Well, actually, it's... 434 00:13:09,520 --> 00:13:10,240 The clothes was so wild. 435 00:13:10,280 --> 00:13:11,200 Yeah, I know, right. 436 00:13:11,240 --> 00:13:12,520 It was like something out of a movie. 437 00:13:12,560 --> 00:13:15,760 It was a powerful statement about what AI can do. 438 00:13:15,800 --> 00:13:16,920 For creativity. 439 00:13:16,960 --> 00:13:17,880 Exactly. 440 00:13:17,920 --> 00:13:21,200 And it got people talking about the future of fashion. 441 00:13:21,240 --> 00:13:22,720 And the role of human designers. 442 00:13:22,760 --> 00:13:23,240 Right. 443 00:13:23,280 --> 00:13:23,520 Or... 444 00:13:23,560 --> 00:13:24,440 It really makes you think. 445 00:13:24,440 --> 00:13:28,480 Where is fashion headed in this age of AI? 446 00:13:28,520 --> 00:13:29,000 Yeah. 447 00:13:29,040 --> 00:13:33,320 It's just one example of how AI is everywhere now. 448 00:13:33,360 --> 00:13:34,120 I know, it's great. 449 00:13:34,160 --> 00:13:36,840 From the clothes we wear to the art we create. 450 00:13:36,880 --> 00:13:38,760 To the way we search for information. 451 00:13:38,800 --> 00:13:39,400 Right. 452 00:13:39,440 --> 00:13:40,880 It's a wild ride, for sure. 453 00:13:40,920 --> 00:13:41,480 It is. 454 00:13:41,520 --> 00:13:42,640 And we're just getting started. 455 00:13:42,680 --> 00:13:45,680 Yeah, it really is crazy how fast everything's moving. 456 00:13:45,720 --> 00:13:46,360 I know. 457 00:13:46,400 --> 00:13:50,160 And we've only, like, scratched the surface of what's happening with AI right now. 458 00:13:50,200 --> 00:13:51,120 I think you're right. 459 00:13:51,160 --> 00:13:52,960 Speaking of, like, keeping up, I read this article. 460 00:13:53,000 --> 00:13:54,240 Oh, yeah. 461 00:13:54,240 --> 00:13:54,880 About what? 462 00:13:54,920 --> 00:13:59,440 It was about how AI is being used to personalize education plans. 463 00:13:59,480 --> 00:14:00,360 Interesting. 464 00:14:00,400 --> 00:14:01,920 It's a pretty fascinating idea. 465 00:14:01,960 --> 00:14:02,480 Yeah. 466 00:14:02,520 --> 00:14:03,920 That's a big area right now. 467 00:14:03,960 --> 00:14:05,480 Right, so using AI. 468 00:14:05,520 --> 00:14:06,040 Yeah. 469 00:14:06,080 --> 00:14:07,400 To, like, tailor learning. 470 00:14:07,440 --> 00:14:08,880 To each student. 471 00:14:08,920 --> 00:14:09,520 Yeah, exactly. 472 00:14:09,560 --> 00:14:10,840 It's a cool concept. 473 00:14:10,880 --> 00:14:13,800 So instead of everyone learning the same stuff at the same pace. 474 00:14:13,840 --> 00:14:14,680 Yeah. 475 00:14:14,720 --> 00:14:16,360 You could have this AI system. 476 00:14:16,400 --> 00:14:19,080 That figures out the best way for each kid to learn. 477 00:14:19,120 --> 00:14:19,720 Yeah. 478 00:14:19,760 --> 00:14:21,000 Based on their strengths and weaknesses. 479 00:14:21,040 --> 00:14:22,040 Even their learning style. 480 00:14:22,080 --> 00:14:23,600 It could be huge for education. 481 00:14:23,600 --> 00:14:27,640 Yeah, especially for kids who don't do well in normal classrooms. 482 00:14:27,680 --> 00:14:29,040 But what about teachers? 483 00:14:29,080 --> 00:14:30,040 What about them? 484 00:14:30,080 --> 00:14:32,120 Will they be replaced by robots? 485 00:14:32,160 --> 00:14:33,480 I don't think so. 486 00:14:33,520 --> 00:14:34,120 You don't. 487 00:14:34,160 --> 00:14:37,840 I think it's more likely that AI will just be a tool for teachers. 488 00:14:37,880 --> 00:14:38,400 Okay. 489 00:14:38,440 --> 00:14:40,040 You know, something they can use. 490 00:14:40,080 --> 00:14:42,680 So to help them personalize instruction. 491 00:14:42,720 --> 00:14:44,960 Right, and track student progress. 492 00:14:45,000 --> 00:14:47,160 And even figure out which students are struggling. 493 00:14:47,200 --> 00:14:48,440 Yeah, stuff like that. 494 00:14:48,480 --> 00:14:49,040 That makes sense. 495 00:14:49,080 --> 00:14:51,200 So it's like AI can handle the boring stuff. 496 00:14:51,240 --> 00:14:51,680 Right. 497 00:14:51,680 --> 00:14:54,000 And free up teachers to focus on the important stuff. 498 00:14:54,040 --> 00:14:58,480 Like mentoring and inspiring the human side of things. 499 00:14:58,520 --> 00:14:58,960 Yeah. 500 00:14:59,000 --> 00:15:01,840 It's about collaboration, not replacement. 501 00:15:01,880 --> 00:15:04,240 That seems to be a recurring theme with AI, doesn't it? 502 00:15:04,280 --> 00:15:05,480 Yeah, for sure. 503 00:15:05,520 --> 00:15:07,240 It's not about getting rid of humans. 504 00:15:07,280 --> 00:15:07,520 Right. 505 00:15:07,560 --> 00:15:08,840 It's about helping us do things better. 506 00:15:08,880 --> 00:15:10,400 It's all about finding that balance. 507 00:15:10,440 --> 00:15:11,200 I think you're right. 508 00:15:11,240 --> 00:15:11,920 Yeah. 509 00:15:11,960 --> 00:15:13,800 Okay, well, I think we've covered a lot today. 510 00:15:13,840 --> 00:15:14,400 We have. 511 00:15:14,440 --> 00:15:16,440 From the glitches to the ethical stuff. 512 00:15:16,480 --> 00:15:16,760 Yeah. 513 00:15:16,800 --> 00:15:18,120 The breakthroughs. 514 00:15:18,160 --> 00:15:20,240 The potential impact on society. 515 00:15:20,280 --> 00:15:20,960 Definitely. 516 00:15:20,960 --> 00:15:24,640 It's clear that AI is a force to be reckoned with. 517 00:15:24,680 --> 00:15:25,240 No doubt. 518 00:15:25,280 --> 00:15:27,160 It's a topic that people are talking a lot about. 519 00:15:27,200 --> 00:15:29,040 It definitely sparks a lot of debate. 520 00:15:29,080 --> 00:15:30,880 And that's good, isn't it? 521 00:15:30,920 --> 00:15:32,640 I mean, we need to be having these conversations. 522 00:15:32,680 --> 00:15:33,440 Oh, absolutely. 523 00:15:33,480 --> 00:15:36,080 We need to think critically about what AI means. 524 00:15:36,120 --> 00:15:37,880 And how we want to shape its future. 525 00:15:37,920 --> 00:15:38,960 Exactly. 526 00:15:39,000 --> 00:15:40,720 It's not just about the technology. 527 00:15:40,760 --> 00:15:41,720 What else is it about? 528 00:15:41,760 --> 00:15:43,880 It's about what kind of future we want. 529 00:15:43,920 --> 00:15:45,160 I like that. 530 00:15:45,200 --> 00:15:48,200 Well, thanks for joining me on this deep dive into AI. 531 00:15:48,240 --> 00:15:49,160 It's just been fun. 532 00:15:49,160 --> 00:15:51,360 It has, it's been a really interesting conversation. 533 00:15:51,400 --> 00:15:52,840 Yeah, I agree. 534 00:15:52,880 --> 00:15:54,120 All right, well, that's it for today. 535 00:15:54,160 --> 00:15:54,800 Okay. 536 00:15:54,840 --> 00:15:57,880 Thanks for listening to the Daily AI News Podcast. 537 00:15:57,920 --> 00:15:59,640 See you next time. 538 00:15:59,640 --> 00:16:19,120 And stay tuned for more.