Is OpenAI a Modern Yahoo? - DTNS 5102

0:00This episode is brought to you by Capital One. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade-in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. Hey guys, it's Drew Afualo from the comment section. Listen, it's time to stop settling for hair products that just don't work.

0:37With all the prestige brands at Sephora, upgrading your routine is easy. If your hair needs repair, they've got the K18 Leave-In Molecular Repair Hair Mask. If you're trying to level up your wash day, the Kerastase Elixir All-Time Hair Oil leaves your hair looking ridiculously shiny without feeling heavy. And if you're into that glossy, healthy hair vibe, the Gisu honey infused hair oil is a must have. Also, Sephora makes it super convenient with free shipping and same day delivery. Shop hair at Sephora and give your hair the upgrade it's been waiting for. This episode is sponsored by Morning Brew Daily.

1:12If you're anything like me, you want to stay informed about what's happening in business and finance, but you also don't have an hour every morning to sift through headlines. Morning Brew Daily is the go-to business news podcast for millions of people. It's a daily business news podcast that's a quick, fun, and easy way to start your day. Whether it's the latest in tech, markets, or the trends shaping the economy, the show breaks it all down in a way that's clear, engaging, and actually enjoyable to listen to. Hosts Neil Freiman and Toby Howell make the show feel like you're catching up with two friends while giving you the context you need to feel smarter and more confident throughout your day.

1:48New episodes drop every weekday wherever you get your podcasts. Tune in to Morning Brew Daily wherever you get your podcasts. Alienware's Back to School event is the perfect time to score top gaming gear with incredible features and Intel Core processors to go beyond performance. Save big on gaming desktops, laptops, and more. Start your Alienware journey with the Alienware 15 gaming laptop featuring Intel Core processors. Smoothly game, live stream, and multitask for hours on end. Plus, play through every game exactly how you want with customizable AlienFX lighting across your Alienware ecosystem, creating your very own gaming profile, and more.

2:26Finally, pair your incredibly smooth gaming experience with immersive visuals and sound by saving on sleek Alienware monitors, headsets, and more. This limited-time sale awaits you now at Alienware.com slash deals. That's Alienware.com slash deals.

2:46Huyen Tue Dao: This right here, my friends, is the Daily Tech News for Thursday, September 11th, 2025. We tell you what you need to know. We follow up on context. We're going to do that a lot today and help each other understand. Today, Andy Beach tells us how Pocket FM will use AI to become the Netflix of audio and Oracle risks its future on OpenAI's successes. I'm Tom Merritt. I'm Wynne Twetnau. Let's start with what you need to know with that big story. It was a bit of a heady Wednesday for Oracle and its CEO, Larry Ellison. Oracle earnings could get pretty wonky, but the top line important bit seems to be that because of a bunch of contracts that have been signed but haven't started bringing in the money yet, Oracle projects its cloud infrastructure business will grow 77 % this next year to$18 billion and then skyrocket to$144 billion by 2030.

3:43Huyen Tue Dao: Oracle said it signed four multi-billion dollar contracts with three different customers. And it didn't name those three, but it did roll off a list of what it called its significant customers. And you could probably guess these folks, OpenAI, X, Meta, NVIDIA, AMD, and more. Wall Street Journal found some sources that tell it that that open AI contract is one of the four multi-billion dollar ones. In fact, it's the majority of Oracle's future revenue, reportedly bringing in$300 billion over five years. But that won't start until 2027. All of this led to an Oracle stock surge. That briefly made CEO Larry Ellison worth more than Elon Musk and therefore the richest person in the world.

4:28Huyen Tue Dao: But then the stock fell back from its peak. So we didn't mention it yesterday and now it's not true anymore. But he's close. Do you feel bad for him, Wyn? Not really. OpenAI, and apparently Oracle, believe that OpenAI will increase its revenue from the$10 billion it's making now. OpenAI makes$10 billion, spends much more than$10 billion, and as we know, is planning to spend much more than$10 billion. CEO Sam Altman has previously said he expects OpenAI to become profitable in 2029. A lot of times I think people go like, ah, this, this business is a scam. They're just losing money. And it's like, that's every business ever, right?

5:10Huyen Tue Dao: Like they lose these tech companies. Google lost money until it started making money. Facebook lost money until it started making money. Um, Twitter, I don't know if it ever really got to that point. I think it got break even, but, but anyway, that's, that's normally the plan. So 2029 is pretty reasonable. However, Oracle is dependent on a lot of that money coming from OpenAI. It plans to work with a data center builder called Crusoe, if you're familiar with them, to locate new facilities in Wyoming, Pennsylvania, Texas, Michigan, and New Mexico. Meanwhile, a story that caught my eye related to this, kind of building on top of it, was an opinion piece from Bloomberg's Parmy Olson about the risk of companies like OpenAI relying too much on large language models.

5:55Huyen Tue Dao: Olson cites examples like a company called Covariant in the Bay Area, which is working on computer vision to sense surroundings, something great for robotics. Or the UK's Atman Labs, betting on alternatives like game playing and reinforcement learning, something that DeepMind, Google's DeepMind, has pioneered. Olson quotes Meta's chief AI scientist, Jan LeCun, saying that large language models are a dead end, and he called them token generators because they don't understand their surroundings. It may be a little harsh, but I think what he's saying is, you know, LLMs are only going to go so far at what they do.

6:29Huyen Tue Dao: If you want super intelligence, which everybody always talks about, you're going to have to look at something else. Well, the money is certainly on OpenAI right now, but it's intriguing to think at this stage, could OpenAI be the Yahoo to some future companies, Google? And we just haven't met Google yet? When? Am I talking crazy? No, I think that actually that's a really cool way to tie that together because I think that there is something to that there. Because it's like Yahoo was like almost a template for what a central search engine-y type company could be. And it was a thing for a while and it still exists out in the wild.

7:09Actually, it's very big in Japan. But it's like a template, right? and that there was the special sauce of Google, you know, the way that they did search, like their particular search algorithm that took that template and really made something kind of new and, you know, like kind of, I guess, you know, world impacting with it. And I think especially tying that into, you know, the harsh, and I think actually not too unreasonable, like unreasonable, not unreasonable, But yeah, slightly harsh kind of, you know, point that Lacoon made is that, you know, LLMs are just a small segment of what AI are.

7:50Unfortunately, like right now with the zeitgeist, I think most people when they think of AI are just automatically thinking of this generation of LLMs. But there's a lot of different AIs out there. And I think it is fair to say that, you know, and I've heard many people that know a lot more about LMs than I do that, you know, like they're, you know, if you want to be harsh, they're basically, you know, autocorrect. spell check on steroids. They are pattern generators. And I think it's fair that, you know, there are other types of AI that, as Lacoon said, maybe have better understanding of their surroundings and that the ultimate solution is not just picking the right one.

8:23That's honestly not how tech has worked, right? It's usually like the collation of a lot of different knowledges and technologies to create something truly, you know, just like anything else, any other product you think about, it's like a synthesis of things. And so maybe, you know, OpenAI is making the template of what, you know, an AI company, an AI super giant might be. And that, yeah, I love that. And that there's somebody out there that's going to understand either the right connection, the right synthesis, the right, you know, something, business model, whatever it is, and actually make this, you know, and be the Google.

9:01I think that's actually really apt. Yeah, leap ahead. Absolutely. Absolutely. Yeah.

9:05Huyen Tue Dao: That's kind of my premise here is that because originally I was like, is OpenAI the AltaVista? But OpenAI is better than AltaVista. And I think the Yahoo metaphor works better because AltaVista was using search and Yahoo said search isn't good enough. Let's use the directory with search. And that made Yahoo more valuable and has kept Yahoo alive. Yahoo still exists. We've got some Yahoo finance links in the lineup today. So it's not like, you know, I think OpenAI is going to die like AltaVista. So I think Yahoo is a better metaphor in that sense. And search was a previously existing thing that didn't work as well as the directory until Google came along and figured out how to make it work better.

9:46Huyen Tue Dao: Right. And so that's kind of what I'm wondering is, is OpenAI going to get, you know, subverted like that. I don't expect OpenAI to have a peanut butter spread too thin memo or Marissa Meyer to come in and take over. The metaphor can only go so far. And I honestly think Sam Altman has a good idea of how to avoid that sort of thing from happening. We have already seen OpenAI do a lot of things with its GPT to adjust for the problems that Jan LeCun is talking about with LLMs and provide refinements and pivots. And I don't know this, but I can't not believe that there aren't researchers and scientists within OpenAI who are working on those other things, you know, besides just the LLM.

10:37Huyen Tue Dao: The LLM is just the most public facing part of this. Yeah. I mean, we're already seeing like agentic AI kind of come out and be more like integrated into a lot more companies. And so, yeah, I mean, I think it's just, it's still kind of remarkable to me how LLMs, I guess, quote, like took off in a way. And it is because they are so good at being user-facing, yeah, because they're good at sounding, they're good at the things that human beings like, which is - They play on our psychology. Exactly, exactly. Although they're really good masks, I think. And I think if someone can find like the actual intelligence, all of the weird sci-fi or other otherwise ethical, putting all those issues aside for just 20 seconds.

11:23I think if someone could actually find like the connections that put a real brain or how a real thinking process behind it, that would be the Google, the Google to open AI is Yahoo.

11:35Huyen Tue Dao: Yeah, for sure. Yeah. It's something that I think you would use LLM, right? Yes. Maybe the metaphor is LLM is like replicating the speech center of our brain, but not the rest of it. Yeah. You know, and somebody is going to be able to replicate more of the brain, you know, the vision center and the reasoning center. And that is one thing that OpenAI is working on is reasoning, you know, reasoning models to inform the LLM. But the LLM is just the speaking part, you know, it's not the whole thing. Which makes it great for like user interfaces, but not for anything behind it. So yeah, I like that a lot.

12:10It's not the only thing. It's like we need like Voltron. We need like the other arms and the head and stuff.

12:15Huyen Tue Dao: They need to all assemble. Exactly. It's Voltron. We finally figured it out. Well, Voltron is made of multiple, you know, small robots, but DTNES is made up of us and is made possible by you, the listener. Thank you to Joni Hernandez. Hi, Tech Oki, Chris Zaragoza, and we want to welcome new patron TM. Welcome, TM. And we want to thank you, Ukit, for a raise. Yes, we have hundreds and thousands of Voltrons. This episode is sponsored by Morning Brew Daily. If you're anything like me, you want to stay informed about what's happening in business and finance, but you also don't have an hour every morning to sift through headlines.

13:00Morning Brew Daily is the go-to business news podcast for millions of people. It's a daily business news podcast that's a quick, fun, and easy way to start your day. Whether it's the latest in tech, markets, or the trends shaping the economy, the show breaks it all down in a way that's clear, engaging, and actually enjoyable to listen to. Hosts Neil Freiman and Toby Howell make the show feel like you're catching up with two friends while giving you the context you need to feel smarter and more confident throughout your day. New episodes drop every weekday wherever you get your podcasts. Tune into Morning Brew Daily wherever you get your podcasts.

13:36This episode is brought to you by Capital One. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade-in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. This message is brought to you by Coligard. If you recorded songs off the radio for a mixtape, it might be time to scream for colon cancer.

14:13The Coligard test is non-invasive, requires no special prep or time off, and you start screening right from your home. It's just three simple steps. Set up, sample, and ship. If you're 45 or older and at average risk, ask your doctor about the Cologuard test. Available by prescription only. Learn more or request a prescription at Cologuard.com slash podcast. Your team just added its 67th AI tool and also your 67th security blind spot. The good news? The Vanta agent works like a GRC engineer in the background, finding every app your team uses, scoring the risk, and drafting fixes for you. Vanta is the platform used by over 16 ,000 fast-moving companies like Ramp, Cursor, and Harvey, who are shaping the future with AI and staying ahead of AI risk.

15:00Get started at Vanta.com. There's a reason millions of people start their day with Morning Brew Daily. Each morning, we break down the biggest news in business from the latest tech headlines to why nobody can afford a house right now. All in a format that fits seamlessly into your day. You'll leave each episode of Morning Brew Daily smarter and ready to take on the world around you. I'm Toby Howell. And I'm Neil Freiman. And we're ready to show you that business news doesn't have to be boring. So join us on Morning Brew Daily, available wherever you get your podcasts.

15:35Huyen Tue Dao: There's more we need to know today. Let's get to the briefs. All right, a few Microsoft announcements of note today. Developers may now submit apps to the Microsoft Store without having to pay a registration fee. Microsoft will also handle hosting and signing for the apps. And if it's not a game, you can still use your own in-app payment system. Developers will need to verify their identity with a government ID and selfie. Microsoft also released the first preview of Visual Studio 2026, the first major version update since 2021. The register reports that it has more hooks to co-pilot and a refreshed design, but still runs on the.NET framework.

16:15And lastly, Microsoft is integrating previously standalone sales service and finance copilot tools into Microsoft 365 copilot commercial.

16:24Huyen Tue Dao: Yeah, all very developer focused developments here, but good ones. I think Visual Studio might be the more disappointing one, given what the register said about it, but I don't really know. Microsoft 365, that's just going to make it easier for folks to get more tools and it's going to make it easier for Microsoft to sell more tools. What do you think about the Microsoft store saying free? Great, dropping the registration fee. No, that's awesome. I mean, I don't know actually how much it is, but it always feels a little bit weird. And this might be kind of like more of the open, not open, like, I mean, obviously Microsoft and open source are kind of two weird words to put together in the same sentence.

17:00But I always feel like you want to promote, like you want to like promote developments or promote participation in your store and not having a registration fee feels like kind of an easy way to do that. And also, yeah, hosting and signing for the apps, like having Microsoft handle that. Maybe there's some downside to that that I'm not well aware of. But these are the kind of things that feel like barriers to entry to getting your first app anywhere. So I like it, Microsoft. Someone please let me know if I'm being a little naive. But I don't know. I like it. I'm easy.

17:29Huyen Tue Dao: Yeah, yeah. Counter arguments respectfully submitted are very welcome. Feedback at dailytechnewshow.com. A few more trailing pieces of news from the Tuesday Apple announcement. R.W. Nash noted on our subreddit that the iOS feature compatibility page says that that live translation feature coming to AirPods will not be available, and I quote, if you are in the EU and your Apple account country or region is also in the EU. So I guess your translation feature doesn't stop working if you travel to the European Union, but your region is outside of it, is the reason it's phrased that way. Anyway, Apple did not make clear why this is, but the EU has stricter compliance rules for AI than other regions.

18:12Huyen Tue Dao: So AI features often are delayed. For example, Apple Intelligence did not launch in the EU until this past March. Also, the new MagSafe charger from Apple is now fully Qi 2.2 compliant. That means it will not only charge iPhones at 25 watts, which it would before in the previous version, but it's also now compatible with non-iPhone. So your Android phone, if it's Qi 2.2 compliant, will charge fast. A couple of think pieces that are worth looking at regarding Apple. and Gadget's Ian Carlos Campbell argues that the addition of features to AirPods like conversation ducking and hearing assistance and now live translation and heart monitoring is moving AirPods from being headphones to a just always-on wearable that you'll use whether you're listening to headphones or not.

18:58Huyen Tue Dao: And Bloomberg's Mark Gurman notes that while the iPhone starting prices haven't risen much in eight years, about$100 total for the starting price, the top end prices have. And notes particularly that the iPhone 17 Pro Max with two terabytes of storage sells for$2 ,000. First time an iPhone has reached that price. So we are not raising the entry level price, but definitely stretching what the top level folks who want to max out the specs will pay. I don't know if this is a bad way of doing it so that there's always kind of like a benchmark, but then obviously people who want the premium, the premium as premium will get the premium as premium.

19:38Huyen Tue Dao: Yeah. The small price brings you in and then you want the two terabytes and you rationalize spending$2 ,000. Well, after two years in testing, YouTube has fully launched its multi-language audio feature. Over the coming weeks, all YouTube creators will have the option of adding a generated translation in multiple languages that uses Gemini to replicate voice and tone. YouTube says those who have tested the feature saw 25 % of their views come from people watching using the generated language. YouTube is also testing thumbnails that can display text in multiple languages based on the user's default.

20:17Huyen Tue Dao: I have only run into one video that was using this. It was a Japanese TV host. It was fine for her, but she was interviewing other people and it was kind of the same voice for all of them. So it's a little hard to follow. That was a while ago though. That was about a year ago. This has only been in testing for two years. So I'm curious if this works well on multiple voices or if it's kind of meant for, you know, single or, you know, what we're doing right now. One voice, one voice, one voice, one voice. It's crazy because I think that, you know, this kind of translation and even like the voice replication and tone and have come an insanely long way and are really, really good.

21:01But it's kind of like there are certain gaps that I think are common use cases, but that are understandably hard. Like, for example, a truly bilingual video also seems like it would be hard. And I think as I understand it, yeah, that a lot of times the translation model would just be, is set to listen to a particular language. Like, hey, and that's why it kind of works in this set, right? Because it's the user's default language. So it's like, okay, you speak Spanish. So I'm going to be listening for Spanish. But in a situation where there's like probably multiple, like not even like coincidental words, but true multilingual like conversations, it wouldn't be like as good.

21:37But I mean, at the same time, I am so excited for this. And like, obviously I think also as an accessibility and just letting people like have access to more wonderful content or information, all for it, it'll work.

21:51Huyen Tue Dao: Agreed. Google will now add a separate view to Gmail to show you any tracking information of packages that it is aware of based on shipment emails you get. You may say, like, I already get this, right? Yes, you will continue to see shipment notifications at the top of Gmail when a package is arriving within 24 hours. But they're adding a new purchases tab that you can go to in the navigation list that will show all your recent orders. So you can page through them, search for particular ones. And they're making a change to the promotions tab that's sort of related. Promotions will now be sorted by most relevant rather than just chronological.

22:30Huyen Tue Dao: You can change it back if you want. But the idea is to show the ones you engage with more often at the top. And I don't know how I feel about this one. It will also include nudges, similar to what you get in the main inbox. Like when I send an email to someone and five days later, it's like, hey, do you want to follow up on this? They haven't responded. it'll do that in promotions when you haven't responded. Okay. Yeah. Not, um, I, I do love the previous feature actually. I, okay, cool. Yeah. I like the shipment stuff. I like the purchases tab. I think that's great. I've noticed it getting things wrong from time to time.

23:04Huyen Tue Dao: So I don't entirely trust it, but it's still nice, especially with the holidays coming up and a lot of shipments. Uh, yeah, I don't need to be nudged, but I guess it's only nudging you in the promotions tab. So If you never go into the promotions tab, you won't be bugged. Just don't open that door. It'll be fine. Well, Qualcomm announced updates to its Quick Charge feature for mobile chips. Chips with Quick Charge 5 Plus aim to reduce the risk of overheating from the 100-watt fast charging it supports. Qualcomm says it adapts to use lower voltages at higher currents to minimize heat generation.

23:37and it can come to existing Qualcomm powered devices with a firmware update for any device newer than Quick Charge 1.0.

23:46Huyen Tue Dao: Yeah, there's a lot of cooling related technology being put into phones these days. So it was interesting to see this. It's not really Quick Charge 6. It's just a kind of a point upgrade to Quick Charge 5, but also nice that it's a backwards compatible firmware upgrade for most people, you know, only the oldest phones. Opera GX is a gaming-focused version of the Opera browser, but it just added some smart home features. It's a little bit geeky. If you have a smart home system that supports something called MQTT, so if you have Home Assistant or Node-RED or HomeBridge, if you know what those are, you might know what MQTT is, you can enable support in the early bird version of Opera GX and then connect it to an MQTT broker and then install Opera's smart home extension from GitHub.

24:33Huyen Tue Dao: So yeah, this is a little DIY. Not everybody's going to do this, but if you do it, you can play around with features like when you close your browser or have your browser automatically close if your door sensor detects it opening, you know, like you want to hide that you're shopping for that gift or whatever, or something else. It can sync your room lights to your Opera GX theme colors. Another example is automatically dimming the lights when you press play on a movie in your browser. But you get the idea. You can whip up your own formulas. Okay. I like that a lot. It's got a, it's got an if this, then that zap you, but, but like, but for plugged into your smart home.

25:12Huyen Tue Dao: Yeah. That's kind of cool. I like that. I could also see shenanigans resulting from this, but more in-house shenanigans. Uh, well, we're going to try something new today, uh, and do a quick look at a bunch of stories that I feel like you should know exist, uh, because they'll help you understand other stories in the future. So let's, let's get into some things you should know. Well, Sweden's buy now pay later company Klarna debuted on the New York Stock Exchange on Wednesday. Klarna's CEO said the company, quote, over indexed on using AI and is, quote, trying to course correct. That seems like a trend.

25:48Huyen Tue Dao: Jack Ma's ant group showed off its first humanoid robot, the R1, at a conference in Shanghai built entirely with parts from Chinese suppliers. R1 can act as a tour guide. It can sort medicine at pharmacies. It can perform basic medical kitchen tasks. It's another sign of the return of Jack Ma and also the fact that Ant is focusing on developing the intelligence part of robotic systems. Well, India's bike taxi company Rapido, Rapido, Rapido, now has 50 million monthly active users in the market versus Uber's 30 million. And it's making inroads in four wheel rides with 30 percent of the market there compared to Uber's 50 percent with the rest going to Ola.

26:30Huyen Tue Dao: The UK's Information Commissioner's Office, or ICO, has issued a warning about what it calls the worrying trend of students hacking their own school and college IT systems for fun. The ICO did not add kids these days, but it did warn that it doesn't think teachers understand the insider threat. Kids these days. Yesterday, we mentioned Google's rise in a worldwide premium phone market. share today. Further numbers from CounterPoint Research show Samsung raised its U.S. smartphone market share from 23 to 32 percent, while Apple's market share declined from 56 to 49 percent. Below 50 percent. China's cyberspace authority has penalized social media company Xiaohengshu, which markets its app as Red Note in English.

27:20Huyen Tue Dao: The regulator said the company was, quote, damaging the internet ecology. But we didn't get any more details. And the U.S. Federal Trade Commission ordered Google OpenAI, Meta, Snap, X, and Character Technologies to submit information on the impact of chatbots on children. The FTC wants to evaluate how the companies test and monitor their chatbots and what steps they take to limit use by children and teens. And I can't help it. I'm getting sucked into the nothing hype cycle. They released a teaser for the earbuds, nothing year three, that show a talk button on the case. Nothing calls it a super mic.

27:56Huyen Tue Dao: We don't know what it does and we won't find out until September 18th. But it is intriguing that there's a talk mic thing on the case, not on the buds themselves. I've bought every single nothing audio ear thing or audio thing. So I'm sympathetic. Yes. Thank you. Thank you for being sympathetic. Those are the essentials for today. Let's dive a little deeper into a topic. Well, Pocket FM wants to be the Netflix of audio. But to scale up its production, it's turning to AI. Andy Beach has the details. Hey, Andy, thanks for coming back, man. Tom, good to see you again. Yeah, good to see you too. So Pocket FM is doing some interesting things that are probably going to make a lot of people mad.

28:40Huyen Tue Dao: But it's the kind of thing that if you don't use these tools, I don't think you can do what they want to do. Explain what the topic is here. Yeah, absolutely. This was a TechCrunch article that I found. And I know that you covered it a little bit on DTNS, but I thought it was worth going into a little bit of detail. So PocketFM is an India-based audio platform, and they've come out with their own version of an AI tool set that they're referring to as Copilot. And it's really geared around... Clever name. No one's ever thought of that before. Well, you know, it's funny. Microsoft explicitly didn't patent that because Satya fully believed that people should be able to use that to call.

29:19Huyen Tue Dao: They didn't trademark it because they were like, yep. They didn't because they believe that it should be a co-pilot, not an autopilot. Okay, sure. To reinforce that human collaboration piece. But what PocketFM is doing here with these writer tools is they want to be able to transform narrative text into either dialogues or structured episodes, the ability to add more drama. They're effectively creating simple tools that allow the writer to more quickly adapt to sort of specific timings or specific sort of episodes, styles, as part of what they're doing, including the ability to localize content into other languages, but doing it with sort of a cultural adaptation.

30:09So you're not just like straight translating something, you're moving it into both the names and the local references that make more sense for that region.

30:19Huyen Tue Dao: I'm trying to think of a good example because I think I know what they mean by that. It's not suddenly this is set in Mumbai instead of Chicago. It's more like the way Indians talk about restaurants is different than the way Chicagoans talk about restaurants. That is certainly one, but it also might be adapting where you might go get in your car in the U.S. You might go get on your moped somewhere else or you might make a train. So it really does change the story in some cases. So it could absolutely adapt it. Or even like if you're staying in your car, you're having the localized version of the Toyota or the...

31:00Huyen Tue Dao: Yeah, no, that makes sense. I'm driving a Hyundai in Seoul, but I'm driving a Honda in Tokyo. So, yeah. Yeah, exactly. So, so I think they're, they're looking at how they, how they do that for, for a lot of it. And like I said, they're, they're starting with these writing tools, but they of course are also partnering with 11 labs and bringing out voice models as well so that you have the ability to immediately take some of this written narrative and turn it into generative AI dialogue. Yeah. This is going to make everyone angry if it gets any kind of popularity at all. because just using models to voice stuff already makes people very upset because they assume it puts a lot of voice actors out of work.

31:45Huyen Tue Dao: Having writers changing their stories on the fly so that they can not just translate. We're not talking about making it accessible for people who speak different languages, but having different stories, that's just going to offend people's sensibilities, right? They're like, no, the story takes place where it is and the characters are who they are and you should write what you know. And I see a lot of backlash against this. I'm sure there is going to be pushback. That I don't question. But I do think we are sort of beginning to tiptoe into this era where we think of the story as almost a programmable object, the way that we think of a game, where it can be different for me in a different region or me on a different device or me in a different culture as I interact with it.

32:34But the storyteller, and I use air quotes around that, has set up a scenario and then they allow the AI to alter the scenario based on how I am interacting with it. All I'm doing is describing game mechanics, except there's no game. It's a narrative story I go through.

32:54Huyen Tue Dao: Although even games don't change who the characters are, what cars they drive, You know, like, like there is something different to this, but I think it's something that is the kind of thing tools make possible that anyone who grows up with it always being like this is not going to have a problem with it. But it is a big change where we think of stories as immutable. We think of them as a book on a shelf. And this is just not that. Maybe that still exists, but this is a different kind of story where you're like, I just want to be entertained by something. It actually reminds me more of Webtoons, which are semi-disposable, right?

33:29Huyen Tue Dao: Or Wattpad, where you're just getting a lot of really like, you know, throw away disposable stories that entertain versus something that's meant to endure. I don't think we're anywhere near sort of settling in on on where technology like this goes. I think it ends up in multiple areas, like the ones you just described, like the sort of the open ended narrative I was talking about. But I do think in the short term, it's also going to be used for smaller teams or individuals doing what a production team used to do in order to prep something. And maybe that's for a customer. But in many cases, it might be an individual creator who's just looking to not go the studio route with their content and wants a fleet of tools that give them basically a production in a web browser.

34:19Yeah.

34:20Huyen Tue Dao: So what is the advantage to PocketFM doing it this way? I mean, ultimately, it comes down to it being a very cost efficient way of handling it. They believe that it's going to cut time to market for a lot of the content that's out there at, again, a relatively modest production value. They're rolling it out slow. It's not fully in production yet. They're doing a thousand pilots per month at the moment. And a lot of the focus is around comics and micro drama video formats at the moment. Yeah. Yeah. That makes sense. Does this just end up being a bunch of slop though? I mean, I guess it wouldn't succeed if it was, but how do they ensure quality with that level of production?

35:09That's the big fear, right? And I think more content always means more. Just like when YouTube came around, we suddenly had a a lot more short form content that was out there. That didn't mean it was always good, however, and that was the ultimate difference in what happened. So I think what will happen here is we will have some level of AI slop that's part of the audio content that we see out of a platform like this, but we will see good content. And if the platform takes off, we will see more and more good content over time and less and less AI slop.

35:43Huyen Tue Dao: Gotcha. Andy, thank you so much for helping us and dig into this a little more. Appreciate it. You do this on a regular basis. Where can people find what you do? Absolutely. I've been writing about these kinds of topics, really the sort of the intersection of AI and media technology over at enginesofchange.ai. Excellent. Go check it out. Thanks, Andy. Thanks, Tom. I thought about later after we talked, the idea that you would get a different story if you traveled. Like you start a story and then you fly overseas and suddenly all the cars are different because it's like, oh, you're at a different IP address.

36:19Huyen Tue Dao: Let me give you a different story. I kind of like that aspect of it just because I think part of storytelling, it's so interesting is that it's so, like so many stories or maybe most stories, I don't know, are so contextual to the society and like the perspective of the storyteller. So I think there's something very interesting on that aspect of it that has something that in a sense can sort of learn or into it like all like any any look like any location big big big bunny ears on that um i guess i'm just still not i'm still a little uncomfortable with the rest of it because i feel like um and i i really i always appreciate everything that andy brings um and i always try to keep an open mind especially since i tend to be a little bit of a luddite or a skeptic about a lot of these things i guess i just worry even like on the slop side of it that i mean even youtube i think i don't i I don't think the slop short form content problem has gone away.

37:16If anything, it almost feels like they're trying to, they're, they're like, they're kind of at a stalemate with it in my perspective where they're trying to, yeah, they're trying to tune the algorithm. They're trying to adjust like whatever mods they're doing, whatever kind of like automated AI probably, you know, powered tools to fight it. But it just feels like it's just always at best 50, 50 in terms of like the content I see that is like genuine and um like of quality let's not say genuine of quality versus slop slop slop slop slop I mean it's getting better but I just kind of worry that you know um I don't know if we're becoming like numb to the slop numb to the slop it's a t-shirt in the making not a good I'm not happy t-shirt yeah it's something you give as a

38:03Huyen Tue Dao: joke gift that no one would ever wear uh if you have feedback about anything that gets brought up on the show, if you would like that t-shirt to be made, get in touch with us on the social networks at DTNS show on X, Instagram threads, Blue Sky and Mastodon. Or if you're on TikTok or YouTube, you can find us at Daily Tech News Show. This episode is sponsored by Morning Brew Daily. If you're anything like me, you want to stay informed about what's happening in business and finance, but you also don't have an hour every morning to sift through headlines. Morning Brew Daily is the go-to business news podcast for millions of people.

38:38It's a daily business news podcast that's a quick, fun, and easy way to start your day. Whether it's the latest in tech, markets, or the trends shaping the economy, the show breaks it all down in a way that's clear, engaging, and actually enjoyable to listen to. Hosts Neil Freiman and Toby Howell make the show feel like you're catching up with two friends while giving you the context you need to feel smarter and more confident throughout your day. New episodes drop every weekday wherever you get your podcasts. Tune into Morning Brew Daily wherever you get your podcasts. This episode is brought to you by Capital One.

39:13Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade-in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One.

39:42We end every episode of DTS with some shared wisdom. We continue to get a lot of great examples of how people say they would use the AirPods live translation feature. Today, we want to highlight Dean's unique use case.

39:57Huyen Tue Dao: Yeah, I'm sorry we can't read all of these. There have been some great examples, but I've tried to pick the ones that stand out as being unexpected or maybe ones you might not have thought of yourself. Dean commented on Patreon, currently on vacation in France. There's a couple of times that I've been eavesdropping a tour group that isn't in English. Live translation would be good to learn a little more. i actually so um um i totally get what you're what you're getting at dean and especially maybe if you go to a country where i don't know you go to for some reason that the first thing that comes to mind is like a brewery tour or some kind of tour like that right um i think we went to like the when we were in hokkaida we went to this like sapro beer they don't have english tours so i mean so and so we're there we're there you know um we're there earnestly but we would love to know more.

40:45And I wonder if there's just something like I would, yeah, thank you. That would be great. I would love to just be able to go to more places where, you know, they don't, you know, they don't need to cater to English speakers as much. And then I could like learn more. I don't know.

40:58Huyen Tue Dao: I like that. It's a time honored tradition to just eavesdrop on the guided tour next to you in a museum, right? I love that Dean's like, I want to add in all the languages, please. Thank you. Yes. Well, what are you thinking about? Do you have some insight? Share it with us, feedback at dailytechnewsshow.com. Big thanks to Dean and Andy Beach for contributing to today's show. Thank you for being along for Daily Tech News Show. You can keep us in business, by the way. Just become a patron. It's easy. Patreon.com slash DTNS. Talk to you soon.

41:35The DTNS family of podcasts. Helping each other understand.

41:41Huyen Tue Dao: Diamond Club hopes you have enjoyed this program.

41:46Huyen Tue Dao: Any pickups, Hammond? Okay.

42:01Huyen Tue Dao: I said there were four multibillion-dollar contracts with three different customers.

42:10Huyen Tue Dao: and then that's why I changed it to four because I was like oh it's one of the three customers or four contracts I changed it to the four no that's a good question yeah thank you I was gonna ask you about that too okay yeah no worries all right ending the recording

42:31this episode is brought to you by Capital One Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. Your team just added its 67th AI tool and also your 67th security blind spot. The good news? The Vanta agent works like a GRC engineer in the background, finding every app your team uses, scoring the risk, and drafting fixes for you.

43:14Vanta is the platform used by over 16 ,000 fast-moving companies like Ramp, Cursor, and Harvey, who are shaping the future with AI

43:21Huyen Tue Dao: and staying ahead of AI risk. Get started at Vanta.com.

43:39Thank you.