Anthropic Releases Mythos-Class Model to the Public - DTNS 5286

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, it's Paige from Giggly Squad, and I want to let you know that if you want the newest and hottest beauty products, you have to go to Sephora.

0:36I'm obsessed with the One Size Liquid Blotting Paper Spray. It's the first of its kind blotting spray that mattifies for up to eight hours. It's unreal. There's also the Summer Friday Shade Drops SPF 50, which basically lives in my bag at this point. Oh, and the Kayali Eden Plush Pear smells so good. It's sweet, warm, and addictive. like people literally stop me and ask what perfume I'm wearing, which you know I love. So if you want the makeup, skincare, hair, and fragrance products everyone's about to be obsessed with, shop only at Sephora. Hey there, it's Wayfair here, where delivery and setup are as easy as a few taps on your phone.

1:13You're relaxing in an old hammock, scrolling Wayfair's app when you spot it, a brand new patio set. Next thing you know, Wayfair delivers it right to your patio and sets it up. Oh, you need a new grill too? All right, Wayfair's got you covered. With Wayfair's room of choice delivery and fast experts set up on qualifying orders, life gets a little easier. Visit Wayfair.com or the Wayfair app. Wayfair, every style, every home. Still waiting in line? Again? That's time you will never get back. Save time and money with Stamps.com. Over 4 million businesses have skipped the line with Stamps.com. Join them to save up to 90 % off carrier rates from your computer or phone right now.

1:54Print postage for certified mail, registered mail, and packages in seconds. Then schedule a pickup right from your home or office. For a limited time, go to stamps.com and use code podcast for a free welcome gift. Taxes and fees apply.

2:13This is the Daily Tech News for Tuesday, June 9th, 2026. We tell you what you need to know, give you the important context on all of it, and help each other understand. Today, Anthropic decided to just go and release Mythos for Everybody with guardrails. Okay? All right. Plus, I'm going to explain how Apple explained how it uses Gemini but still maintains your privacy. Yeah, there's a lot there to get into. There's a lot of Apple, by the way. If you thought you were done with it, you're not. We're going to talk all about it. I'm Jason Howell. We don't apologize, though. I'm Tom Merritt. No. I certainly did not say I'm sorry.

2:49I just said you better believe it's coming. Let's start with what you need to know with the big story. I wouldn't apologize for Apple. This is Apple Week. This is what you do on Apple Week. Yesterday, Apple told the world, as you know, all the things that its new Apple intelligence can do. later in the day it told the press more about how apple's new models work how it implements private cloud computer and or compute and what google and nvidia's role in all of this has been there's a lot to get into and i know you've been wrapping your head around that uh quite a bit tom yeah yesterday uh in our discussion uh you you'll you'll know that we we were speculating like how much of Gemini is in here and what does it mean?

3:38And thankfully, Apple explained that to us later in the day. Apple has four third generation models of its own Apple Foundation models trained with proprietary data and reinforcement learning. Apple was at pains to explain how little of Gemini is in these, but there is some Gemini in these. There are two on-device models that you will use. AFM stands for Apple Foundational Model. AFM3 Core is a Gemini-based model co-developed with Google, but none of these models are the Gemini you get from Google in the public. They are taking Gemini and using it to improve their own model. Apple's trying to imply like there's barely any Gemini in there, and we're not using any Gemini at all, if you mean the Gemini product that you would go get from Google, but they are using it to tune their own models.

4:34However, if you have a device with 12 gigabytes of RAM and an M3 or newer Mac, an M4 or newer iPad, or an A19 Pro iPhone, which is just the iPhone 17 at this point, then you get AFM 3 Core Advanced. Also, Gemini tuned, but capable of a lot more. So you'll be able to do on-device things no matter what iPhone you use. if you use iOS 27, but you'll be able to do more things on device if you have these newer models. Now, when it will decide whether it has enough power to do it on device or whether it should send it to the cloud is decided by the system orchestrator. That is an on-device system that decides based on your query and based on your power and based on how much personal data it's going to need, whether to kick it to private cloud compute or keep it on your device.

5:29It also can access world knowledge when it goes to the cloud, so it can go out on the web and stuff. I think it can do that on device as well. So it can still get things from the web, but without sharing your data to the web. System Orchestrator will only send the data needed to answer a query into private cloud compute. But Apple says that's for efficiency reasons, not privacy, because private cloud compute is private enough. That data, however, is not stored. So even if it does send some personal data, it uses it to answer the query and then it's not stored. Queries are anonymized and tokenized, but they're also encrypted.

6:07So Apple is overdoing it. They're like, yes, no one can see what your query is because it's encrypted between you and the data. The data is then not stored and it's anonymized and it's tokenized. In the cloud, there are three models. These are also co-developed by Apple and Google. AFM3 Cloud, which is general use, runs on Apple Silicon. So this is not something that has been tuned in any different way. This can use the private cloud compute that Apple has talked about before, and that has been independently audited. ADM3 Cloud, which is for image generation, that also runs on Apple Silicon in the traditional private cloud compute under Apple's control.

6:47where people are going to start to raise an eyebrow is AFM three cloud pro that's for agentic uses that's for complex reasoning that one runs on nvidia blackwell b200s in google's cloud so there's the idea of using gemini's systems to tune apple foundation models and there's a little bit of hand waving about what that means but i i can understand that you know you're you're basically saying, oh, we're kind of maybe not distilling exactly, Jason. You probably understand this part more than I am, but using the Gemini outputs to train our foundation model to be better. Is that fair before I move on here?

7:28Yeah, I think so. And I'm trying to, it's interesting as we talk about this, because the word that keeps popping into my head, and it's probably just vaguely familiar or similar but not is DeepSeq and how kind of the DeepSeq model had been distilled, but in a very much less official sense to a certain degree. I would say not official. They did not get approval from OpenAI. Not at all official. No, you're absolutely right. Whereas this is official and cooperated. But I think it's more than just distillation. But I think that may be – if you know about DeepSeq, then there's something similar there.

8:04Yeah, it's just interesting to me that that is the word that comes to mind when we're talking about Apple and its relationship with Google and Gemini. The other aspect of this that kind of comes to mind for me is if Apple has its own models, as it does with these models that you're talking about here, what would they have looked like if Apple had just kind of like stuck out on its own the way the other companies have done? They look like what we saw up until now. That's what they look like. Not very good. So this is basically just them saying like, look, we know we need to inject this with a little bit of extra jet fuel.

8:39And that's why the DeepSeek thing is a good comparison. Not because of what Apple's doing with Gemini is distillation exactly, but because DeepSeek was able to take a model and use distillation to supercharge it and make it great really fast. I think there's something similar going on where Apple's like, our foundational models were okay, but then we trained them on Gemini output and that supercharged them and they got really good. But they are still Apple's models. However, AFM3 Cloud Pro is not running on Apple Silicon. It is running on NVIDIA Blackwell B200s in Google's cloud. So they have to do a new version of private cloud compute.

9:17Private cloud compute is verified and runs well on Apple Silicon. It runs well in Apple's clouds, which are actually hosted elsewhere like AWS and stuff, but they're under Apple's control. These are Blackwell B200s in Google's cloud. They use a combination of NVIDIA's confidential computing, something they recently rolled out, Intel's trust domain extensions, which has been around for a while, and Google's Titan security chip to create a third-party version of private cloud compute. Apple says this was more complicated, but it was worth doing. they've done it. Apple also said they keep a cryptographically verifiable append-only ledger of all Google Cloud hardware that is part of the private cloud compute fleet.

9:59You don't have to know what that means, but if you do, that's going to mean something important to you. Like, okay, they're doing the work to keep track of this. And Apple's devices will only trust software on those servers that is signed by Apple. So if Google were to try something, not that they will, but if they were, Apple's devices would refuse those and say, no, we're only going to go with stuff that's signed by Apple. This, however, is not fully implemented yet. So if you're using the developer's version of iOS 27 and you're sending something to this one model, again, this is AFM3 Cloud Pro, the agentic complex model, this entire system isn't entirely there yet.

10:41Apple says it's ramping towards the complete set of protections throughout the summer preview period. So what I think that means is it's mostly there. It's probably safe to use. However, they won't be bulletproof enough for Apple to roll out to customers on the public beta until later in the summer. They want to get it done there. A couple other things to keep in mind here. The contract, Apple says, the contract they have with Google bans them from training using any Apple data. They shouldn't be able to access any Apple data, but even if they could, they wouldn't be allowed to use it to train e so any metadata or anything like that they won't be able to use now granted contracts do get renegotiated at some point uh but right now that's the contract and the independent audit of the google cloud portion as i said has not been published so we don't yet have the verification of the private cloud compute in that one pro model because it is a new instance of it.

11:39I assume we will get that before this goes public in October. I don't know if we'll get it before it goes public for the preview later this summer. Yeah. Well, you know, Apple continues to be the company that users and folks out there kind of watching how it does business compared to others in the smartphone world and how they handle all their data and everything. really portrays itself as a, you know, we are the company that is paying close attention to your privacy. We want you to feel safe. We want you to feel comfortable. I think considering the fact that they're working with Google on something like this, and it sounds like it's a pretty complex kind of arrangement, I think Apple is doing a good job at convincing people, hey, we are doing this, but we're doing it right.

12:25And we're keeping ourselves, you know, close to that mission. It seems like reaction to it has been pretty positive. Yeah. I think there's a lot of people, no matter what Apple says or even what we say on this episode, are going to be like, so Apple uses Gemini now. Okay. And they say it's private, but how could it be private if Google's involved? And Apple's gone to great lengths to explain how that can be possible. And I buy their explanations that it can be possible. I would like that third-party verification. I assume we'll get it, but I would like to have that just to be a hundred percent sure.

13:00Um, I also would love to have more details on how they used Gemini to fine tune the output of their own foundation models, which is basically what they said they did. Um, and, and maybe it is just distillation that we were talking about where you, you basically, you know, do you take the output of another model and you train your model on that output and that improves that model. Um, so, so it could be as simple as that. I expect there's probably something more to it since Apple was able to do it with Google. But it does mean that what you're getting is the Apple foundation model improved by Google, but not Gemini.

13:39It's an Apple foundation model that has been improved and integrated, but isn't the same thing as Gemini. That's one of those things that people are going to be like, that's too many words. That's a distinction that's going to be hard for some people to understand the difference between. I totally agree. Yeah. And there's a whole separate question of how well this works, right? And so far, the people who've had a day to play around with Siri AI say, yeah, it seems better, but it's a little too early to really have an opinion on that yet. Also worth noting, yesterday they said that Apple's Siri AI and Apple intelligence improvements would not come to the EU or China.

14:22There's regulatory hoops they have to jump through in China. In the EU, Apple says the DMA, the Digital Markets Act rules there, delayed it from offering Siri AI. EU's digital affairs spokesman, Thomas Regnier, said Apple was simply unable to develop interoperability solutions that meet essential EU privacy and security standards. Apple apparently sought an exemption to that requirement for Siri. And I think what it means is Apple doesn't want to allow you to swap out Siri AI with Alexa or ChatGPT or Claude. Which is kind of interesting because, I mean, there was a part of me that kind of assumed like that was going to be it.

15:08That was going to be part of the real sales pitch on something like this. And I think the reason they don't want to do that is because they've done all this work for private cloud compute for on-device models. And they don't want someone to just undermine that, be like, great, I'm just going to use actual Gemini. Yeah, totally. I'm going to make that my default. That whole Gemini description that you gave, that's too confusing. I'm used to using it, so I'm just going to stop. I'll just use Gemini. I don't care. Exactly. Yeah, I can either have the distilled version or the real version. I'll take the real version.

15:42So yeah, I'm curious how Apple threads that needle because while normally I think Apple protests too much about giving people choice over default browsers and stuff, I do think, I agree they should be allowed, they should offer you a default assistant change, but I don't think they should be forced to offer you a default underlying operating system Apple intelligence change because I think that's too complicated. And I think Apple has done a lot of good work to make sure that the operating system is private. And the EU on the one hand is saying you need to make sure that you respect privacy. And on the other hand saying, but also could you undermine it in the operating system?

16:24Yeah, that's a good point. Well, folks, these kinds of explanations and all the good work we do on Daily Tech News Show is made possible by you. Big thanks to every single one of you who supports us directly, including Matt Schultz, Johnny Hernandez, High Tech Oki, and Chris Zaragoza.

16:47This 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.

17:45every home. Hey, it's Paige from Giggly Squad, and I want to let you know that if you want the newest and hottest beauty products, you have to go to Sephora. I'm obsessed with the One Size Liquid Blotting Paper Spray. It's the first of its kind blotting spray that mattifies for up to eight hours. It's unreal. There's also the Summer Friday Shade Drops SPF 50, which basically lives in my bag at this point. Oh, and the Kayali Eden Plush Pear smells so good. It's sweet, warm, and addictive. Like people literally stop me and ask what perfume I'm wearing, which you know I love. So if you want the makeup, skincare, hair, and fragrance products everyone's about to be obsessed with, shop only at Sephora.

18:25Your 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. Get started at Vanta.com. A BetterHelp ad. Hold on. One second. I just need to...

19:03What if you had a room where no one interrupts? No notifications. No expectations. Just space to talk. With BetterHelp, therapy happens in a space that's yours. Visit betterhelp.com slash randompodcast for 10 % off your first month of online therapy.

19:26Okay. You thought we were done with Apple. I'm sorry to tell you. Now I'm not. Now I am apologizing. I'm not going to. I'm not sorry. I'm happy to bring you more Apple right now. We got lots. Dang, I almost did it. I mean, honestly, there's just a lot of interesting stuff coming out of WWDC. There's always interesting stuff coming out of Google I.O. and Microsoft Build and all of them. Give us a day. Give us an extra day, y 'all. Yeah, yeah, yeah. Apple did something really interesting yesterday, all kidding aside, that led us to have more things in this show than I think we normally would in the day after, which is they did a very short keynote.

20:04And we did the show immediately after the keynote. and there's a lot of details that they left for other presentations later in the day. For example, here are some things for iOS 27 that were not in the keynote. Apple is adding full screen home screen widgets. So not just a little piece, you can have the whole home screen. Separate volume control for your alarms and alerts. A button in the keyboard for easy copy and paste. A refreshed weather app with a highlight summary. a drawing mode in messages. So now you can annoy people with sketches, not just voice memos. Cleaner camera control interface, which they cleaned up in the last iOS and now they're tidying it up even more.

20:47And lots of other little refinements and quality of life improvements, of course. Bloomberg reported that iOS 27 beta, as well as developer framework, have code that hint at a foldable. This is not a shock, but we found the evidence. something everybody calls the iPhone Ultra. The iOS 27 frameworks have reference to fold state and angle degrees, along with a check of the total count of included displays. So just iOS 27. We've got some interesting additions there, some nice features, and confirmation like anybody's surprised. But yes, it looks like they are building the operating system to work with the foldable, which I think you can credibly think means they expect to release the foldable during the lifetime of iOS 27.

21:35Yeah, I would imagine so. If it's already kind of proliferating its way through there, possibly everybody's pointing to the September hardware, the iPhone 18 Pro family. We'll see what's coming, but I am intensely curious to see what Apple does in the foldable space, if and when that happens. Apple also pushed out tvOS 27, now including a handful of actually really good accessibility features or enhancements. On-device subtitle generation, larger on-screen text, tighter pairing for made-for-iPhone hearing aids on Apple TV. So those are going to be really handy to have. The update also includes a redesigned podcast app, more responsive control center behavior, AppleCare details inside of settings, and smart downloads.

22:25So your download's getting a little bit smarter. And then on the app front, Apple Music getting a broader refresh, leading with a redesigned artist page layout. It includes a new prominent shuffle button for all you shuffle fans, I suppose. Refinements to the auto mix algorithm that include new transition types so that your blends between tracks during playback are just better. And I'm super curious about that. Also, faster, more reliable playback, lyrics, translation, expanded to a bunch of new languages, seven of them, language pairings. And then high resolution lossless audio coming to the Apple TV 4K through the app.

23:06Yeah, I don't know if I'm going to use these Apple TV captions, but I love that there's some really good accessibility stuff and being able to pair your hearing aid and everything. It seems like accessibility are the best features coming to tvOS. And with Apple Music, we didn't get a whole lot of screenshots of this stuff. So to me, that makes all the differences. What does it look like? Because I didn't love some of the things they did with Apple Music last year, and I've gotten used to them. And now they're going to change it again. So don't change too much. But the better algorithms for the mixing, I will turn it back on.

23:41I had to turn it off because it was doing some weird stuff. So maybe it's going to stop doing the weird stuff. Yeah, I always wonder on features like that. Way back in the long time ago, I was a DJ and I would mix between two records. And I've always thought that that's a thing that an algorithm should be able to nail pretty closely. And yet anytime I see these features or hear them, it's just not quite right. And I always wonder, when are they finally going to get it? They just needed Gemini, and now they can do it. Right. On the App Store, Apple is rolling out personalized app recommendations based on your downloads, your interests, and your engagement.

24:16Personalized Collections will also show app notes explaining why this app was recommended to you. This is something that is Apple intelligence-driven. Developers also get a bigger toolkit with more page media options, an asset library for marketing materials to make your marketing go faster, app bundles so you can put all your subscription apps together and bundle them, as well as multi-user purchasing options for organizations. And finally, Apple refreshed its app review guidelines, expanding on its policy of rejecting copycat apps or overcrowded app types. Apple now says that if you submit an app in a well-established category, they can remove it if it's not being actively updated, improved, or just not getting enough users, essentially targeting low-effort apps.

25:05Yeah, there's just so many apps out there. At a certain point, these markets are just so overfilled. So I don't know. I support moves like this if apps aren't being updated. It cleans up search. Why are we continuing? You know, if there's a bunch of stuff. You always have to wade through a bunch of stuff. You're like, that hasn't been updated in 10 years. Why is that in there? So, yeah. Absolutely. So, good moves. All right. We got to get to the other big news today. Because literally right before we were about to wander over to the studio to record, Anthropic released its Mythos class model. One of its Mythos class models, I guess.

25:39Called Claude Fable 5. So, it's not called Mythos. It's called Fable 5. To the general public, like you said at the top, Tom, with guardrails. It will reject questions about sensitive issues, things like bioweapons, exploiting software bugs, that sort of stuff. And when it recognizes those, it will kick those answers back to Opus 4.8, its previous top-tier model that's been around for a couple weeks now. However, when you can use it, Fable 5 has much better memory. It can handle larger, more complex tasks. Super curious to play around with it. Fable 5 will cost twice as much as Opus 4.8 because it's going to use fewer tokens for certain types of tasks.

26:23So that'll be interesting to kind of learn what you're working within when you use the model. Early access to the Mythos model goes from preview to Mythos 5. And just to confirm real quick, when this news hit, I refreshed my Cloud app, desktop app, and I can confirm I have access to the new model. It shows in my app that I have access until June 22nd. So it kind of seems like this is the preview period, and then things are going to change a little bit. And that's Fable 5, not Mythos 5. Mythos 5 is for organizations. Fable 5 is for us. Right, exactly. Yeah, to be clear, Fable 5 is in my pulldown and sits above Opus 4.8.

27:05This feels like Anthropik really starting to stretch definitions. I'm sure that it's not a lie that parts of the mythos model are in fable five. And it makes you feel really excited if you're like, yeah, I have just think how good I'm going to be at whatever I'm using it for now that it's yeah. And these guardrails would be in there anyway. Like they don't they don't ever want you to make a bio weapon with Claude. They they have never wanted you to try to find exploits for Claude. So, yeah, will they be subject to a jailbreak by somebody at some point? Everybody's going to try. Anthropic seems pretty convinced that they've put in good guardrails.

27:48I don't think this is different than any other Claude release in that respect. All that's different is that Anthropic has said that this is more capable at those kinds of things. Yep, yep. And we will certainly find out. But, yeah, you better believe once people start getting their hand. They're already working on it. They're already trying to jailbreak this thing. And we're in IPO season. SpaceX IPO has been submitted. Anthropics IPO has been submitted. And now OpenAI has submitted a confidential S1 letting regulators begin review of its financials for public disclosure. They plan to go public as soon as Q4 of 2026, though they have not given an exact date.

28:27OpenAI did note that it has some objectives that will be easier to tackle as a private company before going public that they want to get done before they issue stock. OpenAI is working with Goldman Sachs, Morgan Stanley on the filing. And those are the same banks that do everybody. So, oh, what a surprise. They're also doing the SpaceX IPO. Yep. And yeah, Aldman also shared this is the third phase of OpenAI where he says that this is about the economy shaping itself around AI. If you're curious what those other phases were, what was it? The first phase was research. The second phase is becoming a product company and understanding how these tools are used.

29:09And now they're just saying, oh, my goodness, everything is transforming around us. And we're the leader. Do you find any value in any of that that he said? I don't know. You know, he sounds like a bunch of marketing speak. I know it totally is. I mean, he puts out which is fine. I'm not judging him. I'm just saying I don't I don't necessarily feel that's useful to understanding the IPO. So, yeah, no, I mean, you know, I often run into this because he puts out a lot of blog posts where he's philosophizing around AI. And because he is who he is, it's hard to read that and not think, oh, well, there's there's got to be something to it because there aren't many people in the same position that he is kind of, you know, seemingly driving.

29:51That's called. But it is marketing. At the end of the day, it's marketing. What is that called? There's a there's a name for that where somebody. He's a think fluencer, Tom. The cloaking of authority makes you believe things from them that you wouldn't otherwise. I'm not saying you shouldn't believe them. I'm just saying those are a lot of words that mean open AI is great. Yeah, that's true. At the end of the day, that's all he's saying. And he has it attached to something that's really important to him, which is this upcoming IPO. I know it's good journalists to practice to always give a quote from the source.

30:23But if that quote doesn't say much, I tend to ignore it and just be like, you know what? Open AI is doing an IPO. That's all you really need to do. Yeah, yeah, yeah. Well, on the other side of the fence, maybe it's the other side of the fence. I don't know. Maybe it's the other side of the yard. Perplexity CEO Erevan Srinivas confirmed to CNBC that the company, that is Perplexity, plans to go public in 2028. All right. So we don't have like a fourth major IPO happening right now. And he says this is going to happen regardless of how Anthropic or OpenAI's IPOs perform. Srinivas says that both companies deserve their high valuations, he says, since they operate at the frontier level and can succeed on their own merits.

31:03He views the performance of SpaceX's IPO to be an indicator of whether investors actually have an appetite for other major scale IPOs like those coming from Anthropic and OpenAI. No matter what, though, he says, next year, we're doing our IPO. That's settled. That's the upshot. that you won't get a perplexity IPO this IPO season or next year either. He's targeted in 2028. And this season is graduation season as well. Lots of dads and grads stuff happening out there. And we'd like to congratulate our very own graduate, Live With It, the show that reviews products we actually live with and we use them.

31:41And it's becoming its own show. It already got its own YouTube channel and now it has its own Patreon. If you want to get Live With It ad-free and help it spread its wings, become one of the very first patrons of Live With It by heading to patreon.com slash livewithit.

32:04This 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. Parles-tu français? Hablas espanol? Parli italiano? If you've used Babbel, you would. Babbel's conversation-based technique teaches you useful words and phrases to get you speaking quickly about the things you actually talk about in the real world.

32:48With lessons handcrafted by over 200 language experts and voiced by real native speakers, Babbel is like having a private tutor in your pocket. Start speaking with Babbel today. Get up to 55 % off your Babbel subscription right now at babbel.com slash acast. Spelled b-a-b-b-e-l dot com slash acast. Rules and restrictions may apply.

33:15all right we have a bunch of quick headlines that are good to know might make you look smarter in the future google announced it has cut the ai plus subscription price lowered a price from 799 to 499 a month and doubled your storage to 400 gigabytes that will take effect at the next renewal so be on the lookout dang that's a that's a deal google also announced an update to notebook LM that shifts to Gemini 3.5 as one piece of news and now automatically builds a knowledge base of primary sources for users by using its anti-gravity powered research skills at Google Search. So you don't always have to come in with your sources loaded.

33:52You can just have it build it automatically for you. Proton storage product Proton Drive now has a CLI, a command line interface for Windows, Mac and Linux that'll help you with backups and deployment pipelines and stuff like that. Okay. At least 70 of Microsoft's GitHub projects have been disabled as the company investigates how attackers breached the projects and injected password-stealing malware into them. Again. Again. Instagram launched the ability to rearrange the grid of pictures on your profile, regardless of when they were posted. Oh, okay. A lot of people want to do that. Yeah, apparently.

34:26That world's first wind-powered underwater data center off the coast of Shanghai that we told you about has begun operations of its 24-megawatt facility. And Europe has its own SpaceX. Germany's ISAR Aerospace has raised 870 million euros towards scaling production of its Spectrum rocket, starting with a factory outside Munich designed to make 40 rockets a year and the development of a second launch site. They have one in Norway. They're going to make another one in Nova Scotia, Canada. Okay. Remember, we talked about META and the face recognition kind of hiding in the code. Well, Meta has removed its unactivated facial recognition system from the Meta AI app for smart glasses.

35:07This is after Wired had discovered that it was embedded in there that we talked about, though it was unused. Meta told Wired, quote, no final decision has been made on what to do here, if anything. Nintendo confirmed it will release a remake of The Legend of Zelda Ocarina of Time for the Nintendo Switch 2. that'll come sometime this year with a new art style. Yeah, and the art looks like real. It looks more like real. It's a little more photorealistic. Yeah, there you go. And we end every episode of DTNS with some shared perspectives. Today's CCO Mario has some thoughts on the difference between how Android and Apple users approach their assistants.

35:46Yes, Mario wrote, when thinking about Apple versus Google, you have to ask yourself how many Samsung owners, I am one, use Bixby, I do not, but a lot of Apple users use the default. A lot of Android users like to tinker. Yeah. I mean, I can speak for the Android side and say it's always been that way. If you got the Android device, it's kind of at its core. It's a tinkerer's platform. Maybe less so in recent years, but it's definitely always had that at its core. But when we were talking yesterday, Rob Dunwood, who's also a Samsung user, we were speculating how many people will use Siri AI and notice that it's better than Siri before.

36:24So good perspective on that, CCO Mario. Thank you for that. What are you thinking about? You got a good perspective. I know you do. Send it to us, feedback at dailytechnewsshow.com. Yep. Thanks again to CCO Mario for contributing to today's show. Thank you for being along for Daily Tech News Show. You can keep us in business by becoming a patron at patreon.com slash DTNS. We'll see you next time. Bye.

36:54Podcasts. Helping each other understand. Diamond Club hopes you have enjoyed this program.

37:04This 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 there, it's Wayfair here, where delivery and setup are as easy as a few taps on your phone. You're relaxing in an old hammock, scrolling Wayfair's app, when you spot it, a brand new patio set.

37:44Next thing you know, Wayfair delivers it right to your patio and sets it up. Oh, you need a new grill too? All right, Wayfair's got you covered. With Wayfair's room of choice delivery and fast experts set up on qualifying orders, life gets a little easier. Visit Wayfair.com or the Wayfair app. Wayfair, every style, every home.