1 00:00:00,000 --> 00:00:02,360 Welcome to Daily AI Crash News. 2 00:00:02,360 --> 00:00:04,720 Today is May 6th, 2025. 3 00:00:04,720 --> 00:00:05,540 - Hello. 4 00:00:05,540 --> 00:00:08,440 - In this deep dive, we're digging into some 5 00:00:08,440 --> 00:00:11,880 really key headlines making waves in AI right now. 6 00:00:11,880 --> 00:00:13,200 - Yeah, lots happening. 7 00:00:13,200 --> 00:00:15,720 - We'll be talking about Apple teaming up with Anthropic 8 00:00:15,720 --> 00:00:18,160 on something called vibe coding. 9 00:00:18,160 --> 00:00:19,680 Very mysterious. 10 00:00:19,680 --> 00:00:21,000 - Intriguing name. 11 00:00:21,000 --> 00:00:25,240 - Also, Google's plan to let kids under 13 use Gemini, 12 00:00:25,240 --> 00:00:26,240 with some controls of course. 13 00:00:26,240 --> 00:00:27,080 - It's a big one. 14 00:00:27,080 --> 00:00:28,840 - And then there's the whole ongoing thing about Google 15 00:00:28,840 --> 00:00:31,720 using web content for training it's search AI, 16 00:00:31,720 --> 00:00:32,780 even if people opt out. 17 00:00:32,780 --> 00:00:33,820 - Still a hot topic. 18 00:00:33,820 --> 00:00:36,720 - And finally, Visa getting into AI agents for, 19 00:00:36,720 --> 00:00:37,920 well, spending your money. 20 00:00:37,920 --> 00:00:39,440 - Right, AI powered shopping. 21 00:00:39,440 --> 00:00:41,800 So, quite a range today, from coding tools 22 00:00:41,800 --> 00:00:45,720 to kids' safety, data rights, and finance. 23 00:00:45,720 --> 00:00:47,080 - Exactly, a lot to unpack. 24 00:00:47,080 --> 00:00:48,320 Let's dive in. 25 00:00:48,320 --> 00:00:49,160 - Okay. 26 00:00:49,160 --> 00:00:51,240 - So first up, Apple and Anthropic, 27 00:00:51,240 --> 00:00:53,600 this vibe coding platform. 28 00:00:53,600 --> 00:00:54,680 What exactly is it? 29 00:00:54,680 --> 00:00:55,960 - Well, the name's a bit catchy, 30 00:00:55,960 --> 00:00:58,320 but essentially it's about using AI agents 31 00:00:58,320 --> 00:01:00,120 to actually help write software code. 32 00:01:00,120 --> 00:01:01,360 - Like an AI coding assistant. 33 00:01:01,360 --> 00:01:02,200 - Pretty much, yeah. 34 00:01:02,200 --> 00:01:04,200 A very advanced one integrated directly 35 00:01:04,200 --> 00:01:06,840 into Apple's Xcode development environment. 36 00:01:06,840 --> 00:01:11,360 - Okay, so developers building apps for iPhone, Mac, 37 00:01:11,360 --> 00:01:12,200 all that stuff. 38 00:01:12,200 --> 00:01:13,040 - Exactly. 39 00:01:13,040 --> 00:01:15,840 And it's powered by Anthropic's Claude Sonnet model. 40 00:01:15,840 --> 00:01:18,320 - Which is known to be pretty decent, right? 41 00:01:18,320 --> 00:01:20,920 - Oh yeah, it's a capable large language model. 42 00:01:20,920 --> 00:01:23,800 The idea is to automate parts of the coding process. 43 00:01:23,800 --> 00:01:25,000 - Like what specifically? 44 00:01:25,000 --> 00:01:26,280 Generating code snippets? 45 00:01:26,280 --> 00:01:29,120 - Generating code, yeah, helping with editing, 46 00:01:29,120 --> 00:01:33,800 maybe even testing and finding bugs, speeding things up. 47 00:01:33,800 --> 00:01:35,280 - But it's internal for now. 48 00:01:35,280 --> 00:01:36,840 Apple's just using it themselves. 49 00:01:36,840 --> 00:01:38,000 - That seems to be the case. 50 00:01:38,000 --> 00:01:40,160 No word on a public release. 51 00:01:40,160 --> 00:01:43,020 It makes you think, especially after those reported delays 52 00:01:43,020 --> 00:01:45,440 with their Swift Assist AI tool earlier. 53 00:01:45,440 --> 00:01:47,720 - Right, so maybe this partnership with Anthropic 54 00:01:47,720 --> 00:01:50,000 is a way to kind of jumpstart their efforts. 55 00:01:50,000 --> 00:01:51,880 - It certainly looks like a strategic move. 56 00:01:51,880 --> 00:01:54,440 You know, we're seeing more of this big tech giants 57 00:01:54,440 --> 00:01:57,040 partnering with these specialized AI labs. 58 00:01:57,040 --> 00:01:58,720 - Because building this stuff is hard. 59 00:01:58,720 --> 00:02:00,440 - Incredibly complex, yes. 60 00:02:00,440 --> 00:02:02,700 And Anthropic has that focused expertise. 61 00:02:02,700 --> 00:02:04,960 It says a lot about where software development 62 00:02:04,960 --> 00:02:05,800 might be headed. 63 00:02:05,800 --> 00:02:06,920 - Faster development maybe. 64 00:02:06,920 --> 00:02:09,180 - Faster cycles potentially, yeah. 65 00:02:09,180 --> 00:02:11,600 But also maybe it changes the developer's role, 66 00:02:11,600 --> 00:02:16,140 more focus on like high level design overseeing the AI. 67 00:02:16,140 --> 00:02:17,040 - Interesting. 68 00:02:17,040 --> 00:02:20,600 And separately, Anthropic's been in the news too, right? 69 00:02:20,600 --> 00:02:22,120 Something about a share buyback? 70 00:02:22,120 --> 00:02:23,720 - That's right, offering employees a chance 71 00:02:23,720 --> 00:02:24,880 to sell some shares. 72 00:02:24,880 --> 00:02:29,520 The valuation mentioned was around $61.5 billion. 73 00:02:29,520 --> 00:02:30,680 So they're doing okay. 74 00:02:30,680 --> 00:02:31,520 - Clearly. 75 00:02:31,520 --> 00:02:33,580 And they upgraded their cloud research mode too. 76 00:02:33,580 --> 00:02:36,200 - Yeah, longer runtime, up to 45 minutes 77 00:02:36,200 --> 00:02:38,460 and a new integrations feature connecting cloud 78 00:02:38,460 --> 00:02:41,440 to other apps like Jira or PayPal. 79 00:02:41,440 --> 00:02:43,240 So they're pushing forward on multiple fronts. 80 00:02:43,240 --> 00:02:44,940 - Okay, so a lot going on there. 81 00:02:44,940 --> 00:02:48,040 Let's shift gears now to Google and Gemini for children, 82 00:02:48,040 --> 00:02:49,300 under 13s. 83 00:02:49,300 --> 00:02:50,720 - Yes, quite a development. 84 00:02:50,720 --> 00:02:53,420 The plan is to make Gemini accessible on Android devices 85 00:02:53,420 --> 00:02:55,480 that are managed using Google's family link 86 00:02:55,480 --> 00:02:56,840 parental controls. 87 00:02:56,840 --> 00:02:58,400 - So parents have to okay it. 88 00:02:58,400 --> 00:03:00,300 - Right, parents get an email notification 89 00:03:00,300 --> 00:03:02,260 and presumably have to enable it. 90 00:03:02,260 --> 00:03:06,580 The idea, Google says, is for things like homework help 91 00:03:06,580 --> 00:03:08,400 or maybe interactive storytelling. 92 00:03:08,400 --> 00:03:10,660 - Okay, sounds potentially useful, 93 00:03:10,660 --> 00:03:11,940 but there are warnings too. 94 00:03:11,940 --> 00:03:13,440 - Oh, definitely. 95 00:03:13,440 --> 00:03:16,720 Google's being upfront that Gemini can make mistakes, 96 00:03:16,720 --> 00:03:20,280 hallucinations, inaccuracies, and importantly, 97 00:03:20,280 --> 00:03:23,600 that kids might stumble upon inappropriate content. 98 00:03:23,600 --> 00:03:25,080 - Yeah, that's the worry, isn't it? 99 00:03:25,080 --> 00:03:25,920 - It is. 100 00:03:25,920 --> 00:03:28,100 So Google's advising parents to talk to their kids 101 00:03:28,100 --> 00:03:31,080 about AI's limitations, you know, critical thinking 102 00:03:31,080 --> 00:03:34,440 and warning them not to share sensitive personal info. 103 00:03:34,440 --> 00:03:35,960 - And family link has controls, 104 00:03:35,960 --> 00:03:40,200 but how effective can they really be with a chat bot 105 00:03:40,200 --> 00:03:41,720 that can generate almost anything? 106 00:03:41,720 --> 00:03:43,240 - That's the million dollar question. 107 00:03:43,240 --> 00:03:45,000 What are the ethical implications here? 108 00:03:45,000 --> 00:03:47,800 The potential developmental impacts on young kids. 109 00:03:47,800 --> 00:03:50,000 Parental controls are a tool, 110 00:03:50,000 --> 00:03:53,760 but they might not be foolproof against a powerful, 111 00:03:53,760 --> 00:03:55,440 sometimes unpredictable AI. 112 00:03:55,440 --> 00:03:57,960 - It feels like a space where we need a lot more discussion, 113 00:03:57,960 --> 00:03:59,200 maybe more guard rails. 114 00:03:59,200 --> 00:04:01,320 - Absolutely, responsible deployment is critical, 115 00:04:01,320 --> 00:04:02,800 especially with younger users. 116 00:04:02,800 --> 00:04:04,920 - Okay, let's move to the next Google story. 117 00:04:04,920 --> 00:04:05,880 This one's about data, 118 00:04:05,880 --> 00:04:09,100 specifically how Google trained its search AI. 119 00:04:09,100 --> 00:04:10,480 - Right, this came out 120 00:04:10,480 --> 00:04:12,940 during the Department of Justice antitrust trial, 121 00:04:12,940 --> 00:04:15,780 confirmation that Google can use web content 122 00:04:15,780 --> 00:04:18,600 to train its search AI models, 123 00:04:18,600 --> 00:04:22,660 even if the website publisher has tried to opt out 124 00:04:22,660 --> 00:04:23,720 using certain methods. 125 00:04:23,720 --> 00:04:26,900 - Wait, even if they opted out, how does that work? 126 00:04:26,900 --> 00:04:30,080 - Well, it seems the specific opt out protocols, 127 00:04:30,080 --> 00:04:32,480 like the one for the common crawl.org bot, 128 00:04:32,480 --> 00:04:35,080 mainly restrict Google's deep mind division. 129 00:04:35,080 --> 00:04:36,880 - Ah, not the main search division. 130 00:04:36,880 --> 00:04:38,760 - Apparently not to the same extent. 131 00:04:38,760 --> 00:04:40,760 The search team seems to have, 132 00:04:40,760 --> 00:04:42,480 let's say broader permissions 133 00:04:42,480 --> 00:04:44,720 to use publicly available web data 134 00:04:44,720 --> 00:04:46,320 for training its own systems. 135 00:04:46,320 --> 00:04:47,600 - So if you run a website, 136 00:04:47,600 --> 00:04:49,920 the only real way to stop Google search 137 00:04:49,920 --> 00:04:52,220 using your content for AI training is, 138 00:04:52,220 --> 00:04:54,120 what, blocking Google entirely? 139 00:04:54,120 --> 00:04:56,960 - Pretty much, using the robots.txt file 140 00:04:56,960 --> 00:04:58,880 to block Google bot from crawling your site. 141 00:04:58,880 --> 00:04:59,720 - But that's huge. 142 00:04:59,720 --> 00:05:01,800 That means you disappear from Google search results. 143 00:05:01,800 --> 00:05:02,920 You lose all that traffic. 144 00:05:02,920 --> 00:05:05,980 - Exactly, it puts content creators in a really tough spot. 145 00:05:05,980 --> 00:05:07,620 You potentially lose your audience, 146 00:05:07,620 --> 00:05:09,920 but your content might still be ingested and used 147 00:05:09,920 --> 00:05:12,760 to make Google's AI better, which competes with you. 148 00:05:12,760 --> 00:05:15,400 - That seems fundamentally unfair to many creators. 149 00:05:15,400 --> 00:05:17,380 - It highlights a huge tension, doesn't it? 150 00:05:17,380 --> 00:05:19,540 The AI needs vast amounts of data, 151 00:05:19,540 --> 00:05:21,640 often publicly accessible data, 152 00:05:21,640 --> 00:05:22,960 but the creators of that data 153 00:05:22,960 --> 00:05:24,480 feel like they're losing control 154 00:05:24,480 --> 00:05:26,600 and maybe not being compensated. 155 00:05:26,600 --> 00:05:27,840 - You'd expect more pressure now 156 00:05:27,840 --> 00:05:29,620 to revisit those opt-out standards, right? 157 00:05:29,620 --> 00:05:31,020 - You would think so. 158 00:05:31,020 --> 00:05:33,480 This whole situation really brings the complexity 159 00:05:33,480 --> 00:05:38,360 of AI data usage and creator rights right to the forefront. 160 00:05:38,360 --> 00:05:39,700 What's fair use? 161 00:05:39,700 --> 00:05:41,020 What are the economic impacts? 162 00:05:41,020 --> 00:05:42,840 Big questions for the future of the web. 163 00:05:42,840 --> 00:05:44,540 - Definitely ones we'll be watching. 164 00:05:44,540 --> 00:05:46,240 Okay, finally, let's talk about Visa 165 00:05:46,240 --> 00:05:49,260 and AI stepping into our shopping carts, essentially. 166 00:05:49,260 --> 00:05:50,620 - Yeah, Visa's looking ahead 167 00:05:50,620 --> 00:05:54,080 with something they're calling Visa Intelligent Commerce. 168 00:05:54,080 --> 00:05:55,300 - Sounds futuristic. 169 00:05:55,300 --> 00:05:56,180 - It does. 170 00:05:56,180 --> 00:05:58,860 The idea is to enable AI agents think, 171 00:05:58,860 --> 00:06:00,500 like your personal AI assistant, 172 00:06:00,500 --> 00:06:03,360 to securely make purchases for you online. 173 00:06:03,360 --> 00:06:04,260 - How secure? 174 00:06:04,260 --> 00:06:05,660 - Well, they're partnering with big names 175 00:06:05,660 --> 00:06:07,380 like OpenAI and Microsoft, 176 00:06:07,380 --> 00:06:10,060 and the plan involves using tokenized credentials. 177 00:06:10,060 --> 00:06:12,660 So instead of sharing your actual credit card number, 178 00:06:12,660 --> 00:06:15,500 the AI uses a secure one-time use token. 179 00:06:15,500 --> 00:06:18,760 - Ah, like Apple Pay or Google Pay uses tokens. 180 00:06:18,760 --> 00:06:20,120 - Exactly, similar principle, 181 00:06:20,120 --> 00:06:23,920 but applied to AI agents making the purchase autonomously. 182 00:06:23,920 --> 00:06:25,680 - So what could this be used for? 183 00:06:25,680 --> 00:06:28,000 Like automatically reordering groceries? 184 00:06:28,000 --> 00:06:29,260 - That's the kind of thing, yeah. 185 00:06:29,260 --> 00:06:32,120 Automating routine, everyday purchases, 186 00:06:32,120 --> 00:06:34,120 paying recurring bills, perhaps. 187 00:06:34,120 --> 00:06:36,540 It could really change how we handle, you know, 188 00:06:36,540 --> 00:06:39,220 those mundane shopping tasks. 189 00:06:39,220 --> 00:06:41,100 - Convenience is the big selling point, I guess. 190 00:06:41,100 --> 00:06:42,840 - Convenience, efficiency, sure. 191 00:06:42,840 --> 00:06:46,800 But it immediately brings up some pretty serious questions. 192 00:06:46,800 --> 00:06:47,640 - Like what? 193 00:06:47,640 --> 00:06:48,460 Trust. 194 00:06:48,460 --> 00:06:49,300 - Trust, definitely. 195 00:06:49,300 --> 00:06:52,400 Privacy, what data is the AI accessing? 196 00:06:52,400 --> 00:06:54,360 Control, how much oversight do you have? 197 00:06:54,360 --> 00:06:56,760 And reliability, can you really trust an AI 198 00:06:56,760 --> 00:06:58,240 not to mess up your finances? 199 00:06:58,240 --> 00:06:59,840 - Yeah, letting an AI spend your money 200 00:06:59,840 --> 00:07:01,600 requires a huge leap of faith. 201 00:07:01,600 --> 00:07:02,920 - It really does. 202 00:07:02,920 --> 00:07:05,800 What are the safeguards against errors or fraud 203 00:07:05,800 --> 00:07:08,640 or the AI just misunderstanding your instructions? 204 00:07:08,640 --> 00:07:10,160 The potential benefits are clear, 205 00:07:10,160 --> 00:07:12,260 but the risks around security, privacy, 206 00:07:12,260 --> 00:07:14,920 and user control need very careful handling. 207 00:07:14,920 --> 00:07:17,720 - So a lot of potential, but also a lot to figure out 208 00:07:17,720 --> 00:07:19,640 before AI becomes our personal shopper. 209 00:07:19,640 --> 00:07:20,480 - Precisely. 210 00:07:20,480 --> 00:07:22,480 Robust security and clear user controls 211 00:07:22,480 --> 00:07:25,000 will be absolutely essential if this is gonna take off. 212 00:07:25,000 --> 00:07:27,120 - Okay, so let's just quickly recap. 213 00:07:27,120 --> 00:07:28,640 A really busy landscape today. 214 00:07:28,640 --> 00:07:29,480 - For sure. 215 00:07:29,480 --> 00:07:32,300 - We saw Apple partnering with Anthropic, 216 00:07:32,300 --> 00:07:34,720 pushing into AI assisted coding, 217 00:07:34,720 --> 00:07:36,520 potentially changing software development. 218 00:07:36,520 --> 00:07:37,840 - We like that vibe coding. 219 00:07:37,840 --> 00:07:40,360 - Google planning to bring Gemini to kids, 220 00:07:40,360 --> 00:07:42,560 which comes with a whole host of opportunities 221 00:07:42,560 --> 00:07:45,600 and serious concerns. 222 00:07:45,600 --> 00:07:47,180 - Big ethical questions there. 223 00:07:47,180 --> 00:07:49,840 - The confirmation about Google's broad use of web data 224 00:07:49,840 --> 00:07:51,660 for training search AI, 225 00:07:51,660 --> 00:07:53,620 putting content creators in a bind. 226 00:07:53,620 --> 00:07:55,660 - The data rights debate continues. 227 00:07:55,660 --> 00:07:57,840 - And Visa exploring AI agents 228 00:07:57,840 --> 00:08:01,340 for automated secure commerce, hinting at a future 229 00:08:01,340 --> 00:08:04,120 where AI manages more of our daily transactions. 230 00:08:04,120 --> 00:08:06,880 - Yeah, AI moving into finance in a new way. 231 00:08:06,880 --> 00:08:09,080 - It really shows how AI is weaving itself 232 00:08:09,080 --> 00:08:10,840 into, well, almost everything. 233 00:08:10,840 --> 00:08:13,720 From how software gets made, to how kids learn, 234 00:08:13,720 --> 00:08:16,760 how information is used, and even how we might shop. 235 00:08:16,760 --> 00:08:19,400 - It's becoming deeply integrated, often behind the scenes. 236 00:08:19,400 --> 00:08:22,520 - So as you, our listener, hear about all this, 237 00:08:22,520 --> 00:08:23,680 it's worth thinking. 238 00:08:23,680 --> 00:08:25,520 How are these trends gonna shape your interactions 239 00:08:25,520 --> 00:08:26,960 with technology down the line? 240 00:08:26,960 --> 00:08:28,480 - Yeah, and maybe the bigger question 241 00:08:28,480 --> 00:08:30,520 that emerges from all this is, 242 00:08:30,520 --> 00:08:32,940 as AI gets embedded more deeply, 243 00:08:32,940 --> 00:08:35,580 what are those fundamental issues of trust, 244 00:08:35,580 --> 00:08:38,320 control, and maybe even responsibility 245 00:08:38,320 --> 00:08:39,720 that we really need to grapple with? 246 00:08:39,720 --> 00:08:41,320 Where is this all pointing? 247 00:08:41,320 --> 00:08:42,600 - Definitely something to mull over. 248 00:08:42,600 --> 00:08:44,200 Thanks for joining us for this deep dive 249 00:08:44,200 --> 00:08:45,940 into today's AI news. 250 00:08:45,940 --> 00:08:47,680 - Always fascinating, thanks for listening. 251 00:08:47,680 --> 00:08:48,520 - Until next time.