Microsoft Brings 1 Trillion Parameters to the Desktop - DTNS 5369

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. Therapy's biggest impact doesn't show up in some dramatic breakthrough. It shows up in the smallest moments.

0:35You bite your tongue instead of escalating. You put your phone down instead of firing off that text. You say what you actually mean instead of what would hurt someone else the most. Nobody applauds those moments, but they quietly change your life. That's what growing feels like. And that's exactly what Grow Therapy is here for. Grow is therapy that works, covered by insurance. Sessions average$21 with insurance, and some pay as little as$0, depending on their plan. There are no subscriptions, no long-term commitments. You just pay per session. Grow connects you with thousands of high-quality, licensed therapists across the U.S., offering both virtual and in-person sessions, nights and weekends.

1:17You can search by what matters, like insurance, specialty, identity, or availability. And you can book a therapy session in minutes and begin in as little as two days, not the national 48-day average. See if you could pay as little as$0 for therapy in seconds at growtherapy.com slash ACAST. That's growtherapy.com slash ACAST. growtherapy.com slash ACAST. Availability and coverage vary by state and insurance plan. Mod.com is changing the way people battle excessive fatigue. Ever hit a wall in the middle of the day and a third coffee just doesn't do it anymore? Here's what's actually happening. Coffee only briefly blocks adenosine, the chemical that tires you out.

2:04When it stops working, your system is flooded and you crash. Modafinil is a prescription treatment that works in a completely different way. Modafino works by stimulating multiple receptors in the brain associated with wakefulness, including dopamine, serotonin, norepinephrine, histamine, glutamate, and orexin. It's a more complete solution that gives you 10 to 12 hours of steady focus and energy. To learn more, go to mod.com. Fill out your intake form, meet with a licensed provider, and if approved, ModAlert ships straight to your door. Go to mod.com to get started. That's M-O-D dot com.

3:13Granger, for the ones who get it done.

3:28Huyen Tue Dao: This is the Daily Tech News for Wednesday, October 7th, 2026. We tell you what you need to know, give you important context, and help each other understand. Today, Microsoft and NVIDIA have revolutionized personal computing. Again. Again. At least that's what they say. I'm Tom Merritt. I'm Windward Al. Let's start with what you need to know with the big story. And of course, because we've got some Microsoft announcements to talk to, we're bringing in our co-host Rob Dunwood to the show. Rob, thanks for joining us today to talk about the Microsoft Surface announcement that didn't look like it was going to be about Microsoft Surface at first.

4:09Huyen Tue Dao: It took him about halfway through the announcement before they got to the microsoft surface laptop ultra uh which it felt like a build developer conference address up until that they were they were doing a lot of really developer heavy stuff but i think i understand why let's start with the hardware though uh microsoft surface laptop ultra we previously did hear teased at build uh available for pre-order today starting at 2 599 but you can configure this thing to be a lot more expensive and shipping on Friday, October 16th. The star of the show is it's an NVIDIA RTX Spark laptop. So it's got an ARM processor in there, but it's got NVIDIA RTX Spark, built for developers, built for creators.

4:55Huyen Tue Dao: Although they did do a section on gaming because developers and creators play games, I guess. They're gonna bring Call of Duty to Spark later this year. 100, or I'm sorry, 18 millimeters thick, four and a half pounds, 15 inch mini LED touch display, 2.5x the thermal capacity of an existing surface laptop. That's probably more of an important spec here when you're talking about the power under this. You can do LPDDR5X memory, 24 gigs, 32 gigs on up to 128 gigs. User serviceable, replaceable, solid state drive. I think when you're talking about people who want to do this kind of thing, that's important.

5:33Huyen Tue Dao: And I liked this the most, I think, out of anything. a magnetic connect charging adapter that still uses USB-C. So it's kind of mag safe, but also standard. So you can plug it in with a regular USB-C power source if you need to. There's also the Surface RTX Spark dev box, which they have previously announced all the specs about, with a price now,$5 ,999. Again, you can order it today, ship in October 16th. And then they announced a DGX station for Windows coming from HP and Dell called the supercomputer, which is overdoing it. But it can run trillion parameter models locally. So that trillion parameter model we talked about earlier this week that we said you can't run locally, hey, you can run it on this.

6:18Huyen Tue Dao: Those are coming later this year, though we don't have price or release dates on any of that. Rob, let's start with you. What do you think of this spark box in a laptop form factor? I like it. I like it so much that I can't wait to see what Dell comes out with and says XPS 15 on it and has all the same parts on the inside. So I think that this is a really nice box. And I was I was surprised by the price. I thought it was going to be I thought this was going to be like twenty eight ninety nine to twenty nine ninety nine. And the basement RAM being thirty two gigs. The base is twenty four, but it's several hundred dollars less expensive than that.

6:56So I think you're getting a heck of a lot of laptop. This thing is beefy. It's got a back end on it compared to what laptops are today at four and a half pounds. But it is it is a powerhouse. You can you can pretty much do all the things that can be done on a laptop on this laptop.

7:12Huyen Tue Dao: I'd almost call it a luggable at that weight these days. Right. Because this is this is a spark box you can carry around in a bag, I guess. And Juan, I know you don't do this kind of development necessarily, but from the developer point of view, what does this look like to you? I like it a lot. I think Luggable is a very apt description, especially for the price point. So I'm an Android developer, and I know this is shocking to a lot of people still. We tend to use Mac devices just because of the Linux space, so it kind of helps us along. And the benefit of having a MacBook Pro, of course, is that it's lighter and it's powerful, but it's also extremely expensive.

7:50and in fact, it's getting almost like crazy expensive now. And so the trade-off there, it being more luggable, which is something I used to do actually back when I did use a Microsoft laptop back in the day, it's very familiar. And I do think that there's a really good trade-off here because I worry that some of the more, you know, the kind of professional models that we do use for certain kinds of development are getting prohibitively expensive with Ramageddon. So I like this a lot too. Decent price point, you know, especially with more, with different configurations. And yes, absolutely. I like the MagSafe adjacent charging, which is very important if you're accident prone, raising my hand.

8:26And also, again, like at least some servicial parts is really, really nice. So, yeah, if I could, technically I can use Windows for development. So if I did invest in it, I think this is pretty cool.

8:36Huyen Tue Dao: Yeah. And Microsoft started with a lot of throat clearing about what Windows can do to run AI models locally as a way to justify why you might want to spend$2 ,600. on the Surface Laptop Ultra. Lots of local AI stuff coming to Windows, which will, of course, be useful on all kinds of machines, like the Dell that Rob was just suspecting might come out soon. Local AI in Windows, they showed perplexity portable computer running locally. They announced a native gateway for OpenClaw. And I think what most people might find interesting is taskbar search improvements. So you can just use the search bar like a chatbot.

9:17Huyen Tue Dao: You can find files, you can execute actions. Starting this fall, you'll be able to do things like take a screenshot, enable dark mode by just going to the taskbar and saying, do that. You could also use it like you use Copilot now and have it answer questions right from within the taskbar. That's a little less surprising. But that seems like a more useful way to integrate these kinds of tools into Windows. They also announced Muse for Windows native app coming. I guess that goes along with the local AI, the idea of being able to have enough power on these kinds of laptops to run things locally.

9:51Huyen Tue Dao: And they talked a lot about hybrid because you can't run even on a machine this powerful. You can't do everything locally necessarily. But they're really pushing hybrid intelligence. GitHub's Hydrofusion will be able to work with Windows local models so that you can go to the cloud when you need to, but keep everything local when you don't. And of course, with agents, a lot of people still talking about security. We're not talking about the security of rogue agents here. We're talking about your local agent doing something that you didn't want it to do. So Microsoft Execution Containers, or MXC, is coming out of preview.

10:25Huyen Tue Dao: The billionaires had a long discussion about that at the very end of this announcement. But MXC, the idea is this is the harness you need to make sure that your agent does the things you want it to do and not anything unexpected that you didn't want it to do. Rob, does this kind of shine a light of why you might want to spend money on this sort of thing? So I am not surprised because I kind of knew some of this stuff was coming. But I think that Microsoft is doing what Microsoft needs to do in this area. So number one, Microsoft, although they have all the stuff they do in the cloud, with Azure and Copilot and SharePoint, all that stuff that can be done in the cloud, They still sell enormous numbers of servers that go in data centers that people run, the companies run on their own and on the on-prem and also into just server closets to where there just might be five or six servers for, you know, a small or medium sized business to where they've got boxes in on, you know, on-prem that do things.

11:20and it's significantly less expensive to do this stuff on that hardware, even though your power bill is probably going to be a little higher for the power for the computers running the stuff and definitely for the cooling to keep that closet cool. But it's way less expensive than just doing everything on frontier models in the cloud. You can do some of this stuff on some of these open weight models that are running locally on the hardware and the fact that they're able to figure out very easily what needs to go to the cloud, what model it needs to run on to keep those costs down. I think that a lot of businesses are going to be looking at that.

11:57Huyen Tue Dao: Yeah. When, what do you make of the software? Because I think it's a really good point that Microsoft, they're hoping you're still going to want to use Azure for all kinds of things, right? And I think they're right. But it is interesting to see them push the local model of this and that sort of almost Apple-like privacy focus. I've heard a lot of buzz, especially among kind of like my career, like developers on the emphasis on local AI models, just because, again, as I think general public, we are very much aware. And we read the news, we see what's going on. And we, I think even before this AI period, we're very, we become all very sensitive to what goes to the cloud and what doesn't.

12:32And even more so now with like this kind of agentic era, where it's not just the data being exposed, but now something can do something with it beyond just humans. So I do think that was really interesting. And I do think it appeals. And again, I think a lot of developers are interested in running local AI, not just for the security reasons, but also sort of almost in a tinker, kind of like, what can I do with this thing? As well as some of the cost improvements as well, or just being able to kind of efficiently use AI without having to kind of go into some of the more expensive subscriptions.

13:04So I think that is very pressing. And also it does vibe with what I understand a lot of us developers are thinking about, including myself. I literally spent this last weekend playing around with trying to set up local AI models. So it jives.

13:15Huyen Tue Dao: Yeah, I think Microsoft's moving a little bit from the cutting edge to the just bruising edge. They're making it a little bit easier for people who are like, well, I'm not comfortable just trying anything, but you give me the Microsoft IQ that chooses the model for me and you got Agent 365 carrying out the test and MXC keeps me safe. And now there's more local models. NVIDIA's Nematron is going to launch on October 15th. DeepSeek V5 Flash will be able to run locally on these machines. there's a new get started experience to help you bootstrap into running your agents. This is far from making it foolproof, right?

13:54Huyen Tue Dao: But this makes it easier for those people to tinker around if they want to. Yeah, there's all that. And the fact that you can, you know, you can debug code on a local agent while killing bugs, playing Call of Duty all on the same machine simultaneously. So there's that as well. You could also smash a bug with the four and a half pounds of this laptop. So there's, there's all kinds of ways. I know folks are saying that four and a half pounds doesn't sound like it's that heavy. Yeah. So you've had a two and a half pound laptop or even a three pound, you know, Mac book, three and a half pound Mac book, five pounds, four and a half pounds.

14:26That's, that's, that is, that is, that is heavy.

14:29Huyen Tue Dao: Four and a half pounds, uh, was light at a time. And in fact, uh, Jensen Huang and Satya Nadella were talking about those times quite a bit at the end of this announcement when they were musing about direct x and their their days on terminal in the 80s and stuff uh but yeah it feels a lot heavier now well rob thanks man uh for jumping in and uh talking with us about this this morning i really appreciate your perspective on it no problem have a good show all right take care i was really sad though that they didn't find some way to include uh most mxc x most extreme elimination challenge did you ever watch that mx ec or is it mx ec i thought it was mxc because they was it just mxc most extreme elimination yeah i don't know why the e they dropped the e but i'm pretty sure it's mxc okay if y 'all aren't aware they like it was a show on comedy central or something yeah and it was a dubbed over old japanese show but with english dubbing and and just nothing to do it wasn't a translation it was just let's just remix this show into something ridiculous and anyway i would have And MXC is the most extreme elimination of risks to running agents.

15:35See? That needed to be the marketing collab there that we didn't have.

15:39Huyen Tue Dao: It's not too late, Microsoft. Give us a call. Gwen and I will talk to you. There you go. Well, we have the most extreme MXC. I want to do MXG. We have the most extreme gratitude for you, the listener, because DTNest right here is made possible by you. So let's take time to thank Kevin Morgan, Paul Thiessen, Alisa Anjabi, and Ben McKinney. And you, because, wait, I'm sorry. You. And you. Maybe if you become a patron now, and we'll be thanking you tomorrow. Yeah. In the most extreme way. If you do that, do it extremely soon. Episode prep gets messy when research, Slack threads, and script trackers are all in different places.

16:20That's where ChatGPT work comes in. Give it the goal and connect to the tools where work happens, like apps and files, and it can help pull the pieces into a draft you can review without you having to re-explain the project at every step. For example, ChatGPT work can gather team feedback from Slack and use it to suggest updates to a project brief in Google Docs. You stay in control reviewing the draft and deciding what moves forward. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT.com by selecting Work Mode. Available on Plus and Pro plans. I used to be the person who hunched over my laptop mid-meeting, typing furiously to catch every word, and missing half the conversation because I was too busy transcribing it.

17:04That's the problem Granola fixed for me. Granola is an AI-powered notepad that makes your meeting notes useful. It captures what happens in your meetings and turns it into clean, structured notes with decisions and action items pulled out and made easy to find. And the best part? It integrates seamlessly with how you already work. No setup, no configuration, no friction. Before Granola, my post-meeting routine was triage. 20 minutes after every call decoding my own shorthand, figuring out who owed me what. Now the notes are just done. Organized action items pulled out before I've even closed my laptop.

17:36You get better notes in one place and you get to actually stay present instead of typing the whole time. Once you try it on a first meeting, it's hard to go without. You can try it totally free. Head to granola.ai slash dailytech. That's granola.ai slash dailytech to get your time back. Go to granola.ai slash dailytech and try it on your next call.

18:22advanced, intuitive, and deployed. That's how they sack. That's technology at Capital One. When demand slowed for HydroJug, Amazon Ads rising star Hayden Wadsworth pivoted. When our bank account hit near zero, my brother Jake and I had one last shot to save HydroJug. So we engineered the Traveler, a leak-proof, cup holder-friendly Tumblr. Amazon Ads enables a business to be able to expand their product shelf because of the visibility and the reach you can get. Each week, our new-to-brand customers make up about 95 % of all of our sales. Watch HydroJug's story at advertising.amazon.com slash rising-stars.

18:59Mod.com is changing the way people battle excessive fatigue. Ever hit a wall in the middle of the day and a third coffee just doesn't do it anymore? Here's what's actually happening. Coffee only briefly blocks adenosine, the chemical that tires you out. When it stops working, your system is flooded and you crash. Modafinil is a prescription treatment that works in a completely different way. Modafinil works by stimulating multiple receptors in the brain associated with wakefulness, including dopamine, serotonin, norepinephrine, histamine, glutamate, and orexin. It's a more complete solution that gives you 10 to 12 hours of steady focus and energy.

19:41To learn more, go to mod.com. Fill out your intake form, meet with a licensed provider, And if approved, ModAlert ships straight to your door. Go to mod.com to get started. That's M-O-D dot com.

20:01Huyen Tue Dao: There's more we need to know today. Let's get to the briefs. All right, well, Google has launched a site at synthid.com that lets you upload an image, audio clip, or video to tell if it was watermarked by SynthID and therefore generated by a machine. Google, OpenAI, NVIDIA, and Kakao all support SynthID. Apple will soon support it as well. You do need to sign in with a Google, Apple, or OpenAI account to use it. Now, it will tell you if a watermark is present, but not whether the whole item was generated or if it was only edited by a model. OpenAI has a page on its website where you can check images and audio, but not video for SynthID watermarks and see to PA metadata from OpenAI tools.

20:46Microsoft and Meta use their own systems for watermarking, and Anthropic uses C2PA for images.

20:51Huyen Tue Dao: Yeah, it's that simple, everybody. Just, yeah, just very easily, straightforward, and consistent. Yeah. The takeaway here, I guess, is that Google is launching a tool that can tell if SynthID is watermarked across platforms. OpenAI's tool only looks for OpenAI marks for whatever reason. I don't know why. And Anthropic really doesn't do image generation, so that's not surprising. They do use SynthID for text, but that's not part of this tool. This tool is for images, videos, and audio. And I think this is good. It is useful. I know that the knee-jerk response is always like, well, what if somebody strips out SynthID?

21:29Huyen Tue Dao: And then the answer is, well, then this tool won't detect that. This is really for checking if SynthID is there, which tells you something. That tells you like, oh, this was generated or a generator was used to edit it. and that can be useful. It's not a way to foolproof tell if everything was generated or not, but I think it's still useful, don't you? Yeah, and I think at the scale at which we're seeing AI-generated content, certainly, as you said, there's plenty of people out there who would probably be willing to take the time to strip out the SynthID, but for the vast tsunami of generated content of all types that we see, having at least some kind of almost automatic system where probably most people will not kind of...

22:12will not bother with it. At least kind of there's a fallback there. I think having fallbacks is important, even if it's not perfect. It's a start for sure.

22:19Huyen Tue Dao: Yeah, I'm blanking on falsifiable versus other terms for this. But basically, it can tell you, yes, this was generated. It can't tell you, no, it wasn't. But that's still a useful thing. Google Labs and Unity launched a game generation engine called Playground on Wednesday. You enter a text prompt and Playground can generate a browser-based game. So it's a way for you to make mobile games. You can pick a genre. They give you some seed prompts for particular genres like puzzles or racing, or you can just start from scratch. You can also decide whether you want it to be 2D or 3D, single or multiplayer, and you can even upload your own visual prompts.

22:59Huyen Tue Dao: If you're like, I want it to look like this, those would then be modified into game assets, modified in the sense of turned into assets that work in the game, not necessarily regenerated. You can also choose from a gallery of human-made art. They say they populated this with human-made art and those humans are gonna be compensated for the art they made. Your game can then be played on mobile or desktop browsers. You can share your game with a link. You just send it to your friends. You can keep it private so nobody but you plays it. Or you can publish it on the Playground Explorer Gallery. This does not go into the Google Play Store if you're worried about that.

23:33Huyen Tue Dao: But if you put it in the gallery, you can activate some things like leaderboards and stuff like that. Google is using Unity for this, and they plan to add Unity Spark to Playground later this year. That would allow for more professional-level 3D graphics. You can try it for free if you're 18 years old or older and inside the United States. They're going to expand it later, but right now that's where it is. And there is a token limit, so you can't go hog wild and just make endless amounts of games. At a certain point, it's going to say, you're done for the week, come back next week, unless you're a Google Ultra subscriber, then you can go farther.

24:07Huyen Tue Dao: It's still a token limit, but you can go farther. I took literally like one minute while watching the Microsoft announcement to see just how silly this could be to create a space racing game. And I'm not good at it, but it worked. It's a space racing game. It's not particularly amazing or anything, but it created one and it's in space and it's racing. I don't know when. What do you think of this tool? Of your game? It's amazing, dude. You should send me the link and we can race asteroids? Are those asteroids? There's three tracks. This is an asteroid belt game. I don't know why there's a boost there.

24:54Huyen Tue Dao: I should probably stop playing this. It'll just become a let's play. DTNS Today is now let's play with james uh with um your name's not james tom's generated i'll be james for the purpose of this game um so i have many thoughts on this i i suppose that i will briefly state my reticence with more artistic generated uh the the the the what's the um the providence of games and artistic assets when generated this way that i i'm very very leery about where did these assets come from it but i will also recognize that again i feel like a lot of us especially gamers i've always wanted to make our own games and of course games being one of the probably most high bar in terms of skill resourcing and time entry like kind of um like hobbies or industries that you can get into so i recognize that and also i recognize it that and you made a good comparison to this before we started rolling is that this is more like a sora where people can kind of you know just play with ideas it is a playground it is not actually releasing games for commercial So this can be something you can play around with if you have an idea.

26:00Like I actually had a manager generate a racing game for us as part of like a team building activity. And I think, you know, taking apart the provenance and also the ethics of like how generative AI art is being sourced, which is a big deal, but we can put that aside for now. I do think this is really interesting. And I do understand that we like games. We like sharing things and that in the framing of Sora being a place where you have an idea and generate it and share with people kind of in small groups and just for fun. I recognize that this is really interesting and I do appreciate that it is basically sandbox.

26:34It's not, again, meant for release on the Google Play Store and it allows you to explore ideas or just do something really fun and cute and small for your friends or your family or your small group. So I do think it's interesting.

26:44Huyen Tue Dao: Yeah, it is a playground. I think that Google, you know, for once picked a really good name because it's something where like, I'm just going to mess around. I'm going to try stuff. And you can make things that would be very narrowly inside jokes for your family or your friends, put those in there and play around. And it might actually bootstrap someone into becoming a game developer. They start playing around with this. They see how the mechanics work. And they're like, well, wait a minute. I want to make something more sophisticated. And then they start moving into Unity. That's probably why Unity is doing this.

27:16Huyen Tue Dao: They want to bring people into their world. So I'm not expecting millions of people to become game developers overnight for this, but it could be that kind of inspiration, I could see. I will say that when watching the sizzle reel that they had for it, I noticed that one of the games they made was some kind of first-person cowboy shooter, and there was like a Lego cowboy, and I was like, man, that really looks like Woody from Toy Story, which is to say, again, to my earlier point, but also because of the player, and I doubt anyone's going to get some kind of nasty notice from a big company with like a large IP.

27:50But I just think it's interesting and funny. But as long as it stays a playground and it is something that you can just kind of like imagine, share, I don't mind. Yeah.

27:59Huyen Tue Dao: When I created my space racing game, I said, make the spaceships representative of many different genres of non-infringing sci-fi. I used the term non-infringing because I was like, I don't want to be on the hook for the enterprise being in there when i didn't intend it to be uh but i did want you know lots of different kinds of ships and stuff so yeah um well i i'd be interested if anybody plays around with this uh what they find so give us a give us an email feedback at dailytechnewsshow.com all right well bloomberg's mark german says regulatory filings and his sources indicate lg will launch a line of smart home devices including a doorbell a thermostat door lock and security cameras in partnership with Apple's new smart home platform.

28:46The line will be announced along with a new HomePod mini and TV set top box on October 13th. So the LG products won't be released for a few more months. An account called PDFU posted images on X of what they say are the new LG smart home devices. Yeah.

29:02Huyen Tue Dao: So this is really interesting because with Matter and Threads, you can use Apple's smart home platform right now for anything made by anyone, right? And what we're expecting them to announce is something that's a little more Siri AI integrated so that it's easier to manage, easier to set up. I'm guessing Apple didn't want the risk of creating and managing a bunch of smart home devices themselves, like smart cameras and that. LG wants to get into this business more. So there's a match made in heaven. And Apple has done this before with Logitech and others, where they're basically like, here's the design, you go build it, you go manage it.

29:41Huyen Tue Dao: And I think that's what they're doing with LG. So yeah, you can use these LG devices on Google or Amazon because they're Matter, but they're designed to work really well and really easily with the HomePod mini and the TV set-top box. At least that's what I'm guessing this is going to be. No, it makes sense because I think, as you said, for Apple, it's like a lower risk and liability on them to not create hardware, and hardware is hard. But then they get the benefit, of course, of ecosystem where… Right. Their, their particular like smart home platform works with all of these things. And so it's, it's one of those like, well, it's, I wouldn't say a win-win for them, but it makes a lot of sense because ecosystem and locking and people who are locked in can be like, oh, well, I'm locked into Apple, but also look at all these things that I can have.

30:23It does. It kind of, it kind of offsets that because it feels like you do have, and you, you do have a lot of choice.

30:28Huyen Tue Dao: It's that shelf at the Apple store that has all the Belkin things and Logitech things, right? I love that shelf so much. Like, hey, look, here are our friends. Yeah. Right over there, just on that wall. Yeah. And yeah, I don't think this should be misinterpreted as like, oh, so if I buy a matter compatible device and I use it with Apple, it won't work as well. No, it'll work fine. I'll be curious what the special Siri AI integrations that these LG devices take advantage of that other devices might not be able to are. I think any device should be able to be set up and controlled. But is there some special feature that these are really good at doing or easier to take advantage of?

31:08Huyen Tue Dao: That's what I'll be looking for when they make this announcement. Always interesting when there's a common protocol or common interface. And when corporations collaborate on it, then it's like, how do they individually take the most advantage of it for themselves? Yeah, yeah, exactly. Exactly. And that's the great thing about a standard like Matter is that it raises that playing floor so that it's less harmful to the consumer when they do this kind of integration. But it actually might give you some cool perks because of it. Yay, standards. And also, hey, yay, Dr. Nikki and I's podcast, Technical Tea, just dropped our new episode, the third one ever this month.

31:43It is a Halloween episode about tech vampires and science zombies. but it ends up being a lot. And also if you're squeamish, this might be less fun of a listen to you.

31:54Huyen Tue Dao: There's a lot of talk about humors. There's humors, but there's also talk about blood sucking creatures and also actual real life diseases that could theoretically make you into a zombie. That is Nikki's Nikki, Nikki's speciality, but we found a zombie spider. She did it in real life too. And she had pictures of it. It was amazing. There were a lot of funny moments. There was a lot of both of us, both cringing and laughing. and we really had a great time with it. I really, we're really proud of it. We dressed up for it as well. You can find this, you can find us Technical Tea wherever podcasts are found.

32:29We also have a YouTube channel and I want to point that out because I spend a lot of time on visuals even though some of them are kind of gross and scary. If that doesn't bother you or if you just want to get in the mood for Halloween, check out Technical Tea on YouTube and on your podcatcher of choice.

32:46This 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. Therapy's biggest impact doesn't show up in some dramatic breakthrough. It shows up in the smallest moments.

33:22You bite your tongue instead of escalating. You put your phone down instead of firing off that text. You say what you actually mean instead of what would hurt someone else the most. Nobody applauds those moments, but they quietly change your life. That's what growing feels like. And that's exactly what Grow Therapy is here for. Grow is therapy that works, covered by insurance. Sessions average$21 with insurance and some pay as little as$0 depending on their plan. There are no subscriptions, no long-term commitments. You just pay per session. Grow connects you with thousands of high-quality, licensed therapists across the U.S., offering both virtual and in-person sessions, nights and weekends.

34:04You can search by what matters, like insurance, specialty, identity, or availability. And you can book a therapy session in minutes and begin in as little as two days, not the national 48-day average. See if you could pay as little as$0 for therapy in seconds at growtherapy.com slash ACAST. That's growtherapy.com slash ACAST. growtherapy.com slash ACAST. Availability and coverage vary by state and insurance plan.

34:36Huyen Tue Dao: To the quick headlines. They will make you smart. Well, thanks to DJ Fission for noting this one on our subreddit. Best DJ name ever, by the way. Belgian-American physicist Frances Halsen has been awarded the 2026 Nobel Prize in Physics for creating an observatory to detect high-energy neutrinos that rarely interact with matter. Hausen presented the idea for using the South Pole to create a detector in 1988, and the IceCube observatory was completed in 2011. Yeah. Side note, DJ Fission worked in Antarctica for a while. He's Craig, the guy I interviewed who worked in Antarctica, now works in data centers.

35:17Huyen Tue Dao: So this was relevant to his interest because I think he worked near where the ice cube detector is. Xreal's Aura Android XR smart glasses are available for pre-order now starting at$1 ,279. The glasses are tethered to a smartphone-shaped screenless pack. It's smaller than a smartphone, but it runs Android XR on the little puck. And then that puck works as a trackpad and a fingerprint sensor. and then you wear the glasses and see all the cool XR things. Yeah, these are not XR glasses. I'm just covering up an IA a month that I have, by the way. If you're on video wondering why Win Again is wearing sunglasses during the day.

35:58Or actually, you do wear sunglasses during the day. It's at night. That's weird. Okay, anyway.

36:06Amazon announced the Ring smart lock that comes with a hand crank to recharge the battery if it runs low on power and you're locked out. Three or four turns generate enough power to last 30 seconds, which is good. You won't be there for like, you know, an hour. Like the battery, by the way, lasts one year on a full charge. You can register interest now, but it won't ship until early 2027 for$249.

36:34Huyen Tue Dao: You know, or you could just have a key as a backup. In a fake rock on the side. Yeah, I mean, a lot of these smart locks do have key backups, but I admire the ingenuity of the three or four turns to get you 30 seconds to get you in the door. That's pretty great, actually. I like the sustainability of that. Could you imagine just driving down the street and seeing your poor neighbor just at their door just cranking? Just like, yeah, it's a bad day. It's just been a bad morning, dude. Yeah, locked out again, Dave? Starting this Friday, October 23rd. Well, that's not this Friday. Starting on a Friday, October 23rd Friday to be specific, the U.S.

37:14Huyen Tue Dao: Grand Prix stream on Apple TV will offer Dolby Atmos sound. First time they've been doing Dolby Atmos on one of their Formula One race streams. All right. Well, Garmin fans unite because Garmin has announced the Enduro 4 smartwatch with up to 320 hours of battery along with the fitness and health track you get from the Fenix 9 Pro. the Enduro 4 does cost$900. It does. That is correct. Yeah. I think it has to be, I think it must have some solar charging to extend that battery because it says you have to be around light to get the most out of it. But I couldn't find that for sure. Common Sense Media published a report saying it found that chat GPT for teens took too long, in its opinion, to alert parents of concerning behavior and in some cases failed to send an alert.

38:09Huyen Tue Dao: Basically, Common Sense Media said it often needed hundreds of prompts in a concerning conversation before it generated an alert. Common Sense Media thinks it should happen the first time one of those concerning prompts happens. And therefore, they're recommending that OpenAI pause marketing the product and prevent teens from using ChatGPT until those safety concerns are addressed. Well, OpenAI released 722 manuscripts on GitHub with solutions or progress towards solutions or progress towards solutions, rather, for 372 major math problems. The results will now be need to will now need to be obsessed by mathematicians.

38:48Huyen Tue Dao: This is just making the mathematicians madder. Madder mathematicians with all their manuscripts. Mad math. uh barcelona's in-brain neuroelectronic uh led clinical work on using a magnet to steer ultra thin graphene electrodes into the brain usually they're metal aluminum or something these are made of graphene uh that makes them extra light it's called the mini graph project and the idea is to be able to target better treatment they're going to start with parkinson's disease which responds to deep brain stimulation. This will allow them to get more precise and get feedback on what a particular brain looks like and how it works so they can tailor the treatment to that particular brain.

39:35That is so cool. That's some sci-fi stuff right there. All right. Well, the Dutch Tax and Customs Administration has stopped its migration to Microsoft 365 and will instead use its own servers and open source tools. The move was made possible because the agency added its own data center capacity.

39:53Huyen Tue Dao: Ah, see? He who controls the data center. Some headlines are floating about saying retailers are pulling back from self-checkout, causing many to assume that that means it might end, and therefore they're rejoicing because they hate self-checkout. The stories are pulling from the 2026 Voice of the Retail Industry Survey from Toast. You may have run into Toast. They do some delivery orders and stuff, but they make retail point of sale software. That's why they're doing this survey. The line that is getting the attention is that among the 340 operators surveyed, 36 % said they plan to invest in self-checkout technology.

40:36Huyen Tue Dao: So more than a third are still investing in this, but that's down seven points. And so because it's down, people are saying, aha, everybody's pulling back from this. However, I'd like to point out in the same toast survey, investments in accounting software are also down five points and investments in barcode scanners are down three points. I don't think this means they're pulling back from accounting software or barcode scanners. It just means maybe they're mature and they don't need to add more self-checkout technology. Anyway, in the words of Inigo Montoya, I do not think it means And finally, Biohub, the U.S.

41:18Department of Energy, and the National Institutes of Health are putting$1.8 billion into data to train models that can predict how cells behave. Meta, Google DeepMind, and Isomorphic Labs are also putting a combined$300 million into the project. Resulting data sets will become a public resource.

41:38Huyen Tue Dao: Listen, shout out to the DOE, NIH, and Biohub for putting in the real money. All the headlines I'm seeing are about Meta and Google. They're giving a paltry$300 million. So give Biohub, the Department of Energy, and the NIH a little credit on this one. Yes, please. Yeah. All right. Well, we end every episode of DTNest with some shared perspectives. Today, Doug has a question about full disk access for agents. Yeah, Rob and I were talking about this on Monday's show, how Apple has said they're going to put some more security safeguards around full disk access, probably in the sense of warnings and stuff before you grant it.

42:20Huyen Tue Dao: And Doug wrote, I agreed with the discussion that Tom and Rob had about full disk access for agents, but it left me wondering, does Apple provide an interface for some kind of restricted disk access? I'm especially thinking of iOS, which I have understood to be even more locked down than macOS. This also follows the agents discussion that you had on Friday. I think at least for a while until these technologies mature, we're going to need a more partitioned view of what they can and cannot access and what they are allowed and not allowed to do. The options can be all or nothing, either from the agents, the users, or the operating systems.

42:55Huyen Tue Dao: And yes, Doug is correct. iOS apps run in an isolated container directory. So if an app needs to access another app's data, it must ask the user to grant that access. And you can approve or deny these categories inside settings in the privacy and security area of iOS. And if you want a specific file or folder, you grant temporary or scoped access rather than giving access to all the files on the disk. Yeah. And if you're kind of curious on the Android side, very similar, it has a concept called scope storage where each app has its own sandbox. and there is precedent. I know Messages is a big example.

43:34And there are specific APIs, permissions, including a whole user prompting permission system and services for different kinds of data, whether that's media, document file pickers, and even notification SMS. These are things that they've all thought about and there are very rigorous APIs around. And on Android, if an app wants to share specific data with just open it up to other apps, they do have a provider service where you can make a specific interface to get that safely. And full disk access is a special permission and requires manual app review and approval. And it's generally for things like file managers.

44:10So yes, they've thought about this and that there are ways to get some of that information like messages on these different platforms in very specific ways.

44:18Huyen Tue Dao: Yeah, I think it's really interesting that mobile platforms are way ahead of the game on this stuff. And I guess it's because as mobile was developed, people were more concerned about how many people were using them, whereas computers still come from that age of like, well, you kind of know what you're doing if you get a computer. So we don't need to protect you as much. And phones have a constant Internet connection built into them in a way that most laptops and PCs don't. So, yeah, it's just a different tradition. And laptop operating systems come from an era of, you know, trusting you to be in charge of your security more than phones do.

44:59Huyen Tue Dao: So I think Doug's bringing up a good point of like, well, maybe that era is over. So my read is mobile. We don't trust any of you. But yeah, no, absolutely. It's a really great question to bring up. And if you have anything that you're thinking about, you have questions or things you want to tell us, we would love to hear it. Feedback at DailyTechNewsShow.com. Big thanks to Doug for contributing to today's show. Thank you for being along for Daily Tech News Show. You're the folks who keep us in business. Again, I'm going to give a big virtual hug to the next person who becomes a patron at patreon.com slash DTNS.

45:32Huyen Tue Dao: Please, now. The DTNS family of podcasts. Helping each other understand. Diamond Club hopes you have enjoyed this program.

45:48Huyen Tue Dao: with the spark cash plus card from capital one you earn unlimited two percent cash back on every purchase and you get big purchasing power so your business can spend more and earn more steven brandon and bruno the business owners of sandcloud reinvested their two percent cash back to help build the company's retail presence capital one what's in your wallet find out more at CapitalOne.com slash SparkCashPlus. Terms apply. Hey, I'm running out to drop off a stack of letters and packages. Need anything shipped while I'm gone? Save the gas. Stamps.com lets you print postage and schedule pickups right from your desk.

46:24Sync your contacts into the address book, get detailed accounting insights, and get notified the moment your certified mail is delivered. All from my desk? All from your desk. Join the 4 million businesses using Stamps.com right now with code podcast and get your free welcome gift. Stamps.com code podcast. Taxes and Please apply.