OpenAI's First Hardware is a Codex Accessory by Work Louder - DTNS 5300

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.

0:36this is the daily tech news for tuesday june 30th 2026 we tell you what you need to know give you the important context on all of that stuff and help each other understand today open ai teased some hardware oh they're such teases maybe not the hardware you're looking for but then again oh maybe it is maybe yeah we'll find out i'm jason owl i'm todd merritt and down here is Bronson. You might see him too, my dog. Let's start with what you need to know with the big story. Thank you, Bronson. He's like, you're not feeding me fast enough. I already fed him, but he could always use more. Let's talk OpenAI.

1:10OpenAI is teasing a Codex-branded hardware collaboration, let's say. It's set to launch on July 15th, which is only, you know, half a month away, right around the corner. posted a short teaser video on X saying, quote, Codex shortcuts are getting an upgrade. And then the video, if you watch it, it's a very short, you know, kind of your standard short promotional teaser with like lots of really up close shots and everything. Hardware, keyboard type device, colorful LEDs, backlighting, that sort of thing. A silhouette of, to my count, around 13 buttons and knobs, a little squarish form factor device.

1:49So it's not like a full size keyboard. It's definitely a more compact thing. It looks like OpenAI has partnered with a company called WorkLouter on the hardware. They are a boutique keyboard company that makes a whole portfolio of custom keyboard and controller options for specific software, services, that sort of stuff. If you use something like Framer, for example, they've got a keyboard. And in fact, speaking of that, I was looking through the portfolio to just kind of see, like, get some clues because we don't know a whole lot from what OpenAI had mentioned. I came across the Framer Creator Micro 2, and to my eyes, it looks like a very similar form factor.

2:29It's a Creator Micro designed for Framer, which is why I brought it up. And it gives some clues on, I think, what we can expect from the controller. It's got this radial joystick for selecting multiple tools when you move it in different directions. It's got a bunch of replaceable key sets so you can swap out and assign specific functions to different keys. It's got a rotating knob that I'm not entirely sure what that does, but it looks really satisfying and cool. Anyways, that controller runs$200. It is on sale right now for$179, by the way, but normally$200. And I think it's important to note that, like, is this the first hardware coming with OpenAI's kind of collaborative name on it?

3:12Sure. Is this the collaborative project that we've been waiting for or expecting from Johnny Ives' love from? Certainly not. So it's not quite that yet. No, the framer example is perfect because you can go get a Creator Micro 2 on its own, and it's a dull, not black. But if you get the framer one, it's sleek and black. I see. Right? It's made for Framer. And I think you brought up that one as a perfect parallel. You're going to get one that's made for Codex, and it's going to be glowy with the cool LEDs and stuff. This definitely is not open AI hardware. This is open AI. Just totally from the outside, having to guess, this is open AI talking through its user base, talking through its internal users of codex saying, what makes your job easier?

4:04And they're like, well, I tried Stream Deck, but that's a little fiddly because it's folder control. Like some kind of custom keyboard would be great. I've been using this one. And maybe some people internally were using WorkLouder and they approached WorkLouder or maybe WorkLouder approached them. Who knows? But they said, oh yeah, no, people using this say that it's great. It allows you to trigger some things automatically. The reason it's better than Stream Deck also is that you can use that little joystick nubby thing to like scroll through stuff. So yeah, this is for the Codex Power user, basically.

4:37And, you know, it's a smaller kind of like side of keyboard accessory is what I imagine. Yeah, it's a little squarish looking thing. Yeah, kind of larger than a puck, but, you know, a bunch of keys, you assign them, you know, probably what makes this more Codex than anything else is if Codex, which I should be fully honest, I have personally have not spent much time, if at all, with codex. So I don't know, but I'm assuming that in the codex experience, there probably are certain buttons that have different graphics and everything. And usually with controllers like this, if they've got the swappable keys, you can take those and physically put them onto the controller so that it's easy to look down and go, oh, that's the button that I go to all the time to do this one particular function that has that graphic that lines up with the thing that I'm looking at on the screen.

5:23So it's going to be very customizable. But I think that's the thing here is that this is a real – like this is a niche keyboard company with a very niche keyboard accessory, which is not to say that Codex doesn't have an insane amount of users. It does. But the amount of those users that are going to buy something like this is probably a lot smaller that would find the value in it. But, you know, hey. Which is why it's a modified creator micro most likely is that they didn't want to invent something. They just wanted to customize something. And yeah, you should play around with codecs some more.

5:57I should. I should. I haven't spent any time. Yeah. Because I think already just off the top of my head, some of the things I could see this being useful for is the constant like, hey, is it OK if I do this? And if you had a keyboard where you can be like, yes or no buttons that only go to that so you don't have to switch your focus out, you could just be like, yeah, yeah, yeah, yeah, do that. And the scrolling that I talked about earlier is not something I run into a lot. And I know a lot of people who use this who do, or like, I just need to review something real fast. So those are just a couple of things off the top of my head.

6:29Anything that allows you to interact with it while it's doing its thing without having to leave the thing you're doing is probably going to be a big part of this. Yeah. What it reminds me of is in the audio world over the years, I have fallen for marketing hype on different controllers, like controller keyboard. Oh, this controller has 16 sliders and 16 faders, and you can assign them, and you can do all this stuff. And I've fallen for it more than a couple of times where I'm like, oh, my goodness, that's just going to make me so much faster. That's going to make the entire process better. And for whatever reason, it never quite sticks.

7:06And I'm not saying it's that way for everyone, but that's been my personal history time and time again with things that are very specific like this. You really do have to commit yourself to using it instead of the things you're used to doing by default. And I'm sure at some point enough of that commitment, you kind of rewire a little bit and you start to depend on these things. So they're useful, but sometimes they promise more than they deliver. And that's not necessarily their fault. It's just you got to commit to it. I got an idea. Okay. solve that. Use Codex to help you figure out how to use the new tool that you're using for Codex.

7:41Oh, man, that's so recursive. I love it. Yeah. Well, actually, folks, if you're like, you know what? I would make good use of this because it would help solve this problem for me. Let us know. Feedback at DailyTechNewsShow.com. Love to hear some of those ideas from folks who are actually in the audience pounding away on this stuff. In fact, the entire show is made possible by the folks who listen. Big thanks to Kirk Stephenson, Miranda Janelle, ThatCharlieDude, and Ruben Gonzalez, and every other patron who helps support the show directly.

8:20This 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.

8:53All right, we got some fun stuff to know today here in the briefs. Let's get to it. Yes. I don't know how fun this is. Well, I don't know how fun. But you can no longer hail a Waymo autonomous taxi in Phoenix through the Uber app. Yeah, that's enjoyable. How fun. I have to use the Waymo app now. This ends a three year long partnership. Apparently it happened in May and it took now till now for everybody to kind of notice. They're like, how come the Waymos aren't coming up in my Uber app anymore? You can still hail a Waymo in Phoenix through the Waymo app. This was one of the few markets where Uber and Waymo were operating side by side.

9:31You could hail a Waymo through the Uber app or you could hail it through the Waymo app. And so they're basically saying, OK, that was a fun experiment. They both have nice things to say about their collaboration and their press releases today. And now you can hail Waymo directly through Waymo. You can also hail Waymo in the Uber app in Austin and Atlanta. What's different about that is there is no Waymo app in Austin and Atlanta. Uber is the way you hail them. In fact, I saw Waymo trundling down the road in Austin when I was there last week. Now, there are nine other markets, including Los Angeles, where you will soon be able to hail a Waymo through the Waymo app, and 20 new markets coming later this year, including London.

10:17So Waymo expanding and Uber also expanding, just not always with Waymo. They've got Atlanta, they've got Austin, but they continue to work with companies like AV Ride in Dallas or We Ride in Abu Dhabi to offer autonomous taxis on its app. Don't forget, at CES, Uber announced a partnership with Lucid, the carmaker, and Nuro, the autonomous driving tech platform, and they plan to launch an actual car on the Uber platform later this year. So this isn't a company making the car and then Uber handles the writing. This is Uber being more involved in the making of the car. And then it's part of Uber's actual fleet, which kind of brings us back full circle to those early days when Uber was trying to develop its own autonomous car system.

11:02Well, yeah, I think at the end of the day, Uber probably wants to be the kind of the universal app for hailing all of these things. And sometimes pilots just end, right? Sometimes pilots are pilots. But I mean, this was a long pilot. It does make me wonder about three years is a good amount of time to be testing, you know, in air quotes, how this works. But it does make me wonder, like, if Waymo really wants, you know, its app and its fleet to work together in its strongest markets like Austin, Atlanta, those are, I'm imagining, pretty strong markets for them. Is it only a matter of time for those areas?

11:37Yeah, I wonder if Waymo will continue to just use Uber in those markets or if we'll see a split there. It's a little more difficult because Waymo doesn't have its own system up and running the way it did in Phoenix. Um, but I think the, the, the Uber trend here to coin a phrase, uh, that that's confusing. If I say it like the meta trend, well, no, that could be anyway, a trend you could identify here is that we are in the business plan optimization stage of autonomous taxis. Uh, that, that means that these companies are getting past that. Can this work? How can this work? Let's test the market to, okay, we're ready to expand.

12:22Uber's going to hit 20 new markets. Waymo's going to hit 20 new markets. How do we want to operate that? Do we want to partner with someone? If so, who? Or do we want to run things ourselves? And of course, you've got several autonomous taxi companies operating in Southeast Asia and mainland China. You got a handful of them here in the United States. So I think before long, we're suddenly going to go, oh, things have changed on the road, right? That's been the promise for a long time. For 10 years, it's been like, yeah, one of these days, autonomous cars. It does look like autonomous cars are not a thing necessarily that you are going to run, at least at the beginning.

13:04It's a thing that you are going to hail. And that's going to change things. I don't know what it's going to change, but we are slowly approaching the crest of that hill where this goes from a, oh, interesting, they have Waymo in this neighborhood to, I use autonomous taxis all the time. Right, all the time. Yeah. We still don't have Waymo in my neighborhood. Come on. Yeah, same here. I don't have it here either. And I think that's kind of the canary in the coal mine when you and I get it. Yes, exactly. Slightly, not remote, not rural, but suburban, you might say, locations. Yeah. Yeah. I'll be curious.

13:41Tidal Music Streaming Service is cracking down on fully AI-generated music on its platform on July 15th. So, you know, a couple of weeks from now. tracks that its system identifies as entirely generated will be blocked from earning royalties, direct-to-fan revenue, and other types of monetizations surrounding the platform. Those tracks will also display a prominent badge to basically signify to listeners that this was generated by AI so that they can spot them in their library easier anyways. Tidal will use its automated detection system along with reporting from distributors to also find and delete generated tracks that impersonate artists.

14:27Unsurprisingly, EVP and EIC Tony Gervino says the new rules are meant to protect, quote, organic creativity. I hadn't heard that that those terms before. Organic creativity. But it makes sense. And also they're meant to protect subscribers from being pushed towards fully synthetic music. because they're saying like, you know, some people are cool with it. Some people just don't even want to be led down that track or exposed to them. This gives them potentially the option to kind of steer clear, I guess. Titles, the first streaming company, it seems, to eliminate any payment or royalties to generated music on its platform.

15:04We have had other streaming platforms make similar announcements and changes to their policies. Deezer, Spotify, Apple, Apple Music. Focus more so, I'd say, on labeling and also anti-fraud tools than explicit royalty cutoff. Title seems to be kind of one of the first to just be drawing a line in the sand and saying, nope, it's fully generated. You can't make money off of it on our platform. Man, I love organic creativity because it means both like it was organic, like it came out of me. But also, I'm an organic, not a robot. Exactly. Yeah. Yeah. Yeah, it makes sense. This is an interesting and very Tidal policy.

15:45Tidal markets itself as being artist first. So they are going to always try to do the thing that is the most artist friendly. And this is the most artist friendly. The only thing that would be even more artist friendly is just banning AI music from the platform, which I wouldn't have been shocked if Tidal had done. But even Tidal is willing to let you decide if you want to listen to AI music. They're just not going to reward the makers for putting it there, which will probably dry it up quite a bit on Tidal without them having to go to the extra expense of identifying and deleting things. It avoids the controversy of actually accidentally deleting a non-generated track.

16:26It's a much easier recovery to say, oh, you aren't generated. we can go back and restore your monetization versus like, oh, right, we totally removed you from the platform. So yeah, I sort of get what I want. At first, I was like, why not just get rid of it all? But as I thought it through, I'm like, no, this is going to get rid of it and avoid some pitfalls and costs for Tidal in doing it. And I mean, there are people who exist who are open to listening to this kind of music or curious to see how it develops and what it develops into. So, you know, not, I would totally not be surprised either if title had just said, absolutely not.

17:09You know, we're just going to go on the arms race and start battling these and, you know, hopefully our detection systems stay up. Yeah, I guess, you know, if you're, if you're open to music as an, as an art form, and some people would say generated music is not art, but it is kind of art in the same way that you listen to it and you either enjoy it or you don't. Same with all music, right? Some people are going to find value in that. Some people aren't. I think removing the monetization from it entirely is the move that I would prefer over banning it outright, because there are some people who are just music lovers as a whole, and they don't want to just eliminate the possibility that there might be something that they like that was generated.

17:47Yeah. And it's a more elegant way to do it. Frankly, if you're the kind of person who wants that, you're probably going to go to Spotify or Apple or YouTube, not to Tidal to listen to your AI music. But yeah, I think all the arguments you make are good ones for why these platforms aren't totally deleting it, but labeling it or disincentivizing it in other ways, like Deezer, where it won't put it on playlists and all of that. And I'll be honest, when this sort of stuff started showing up on YouTube at first, I thought it was nice background music. I'm like, oh yeah, no, this is nice. But my ears have tuned to it over time to the point that Eileen and I both are like, yes, it's kind of the same thing.

18:26It's kind of repetitive. It's not that interesting. I'd rather hear a playlist of actual jazz vinyl or actual songs. And so we've moved away from it, even though it seemed like kind of a harmless thing at the beginning. Yeah, yeah, interesting. Well, Gemini is now gonna offer you personalized image generation for free. It used to be only the folks on the paid subscription got that, but now everybody gets it. personalized image generation uses whatever you have let Gemini know about you. So if you haven't approved access, you're not going to get much out of this, but if you let Gemini access your Gmail or your photos or YouTube, and then you tell it, make something I like, it will use what it knows about you, your likes, your preferences, and you don't have to include those likes and preferences in the prompt.

19:16You can say, you know what I like. Make a picture of a coffee cup that you know that I would like. Google has been rolling out a beta, a voice search for Gmail as well. This is called Gmail Live. Feature was announced at Google I.O. earlier this year. And if you get the beta, you'll see it in your Android or iOS Gmail app. Look in the search box. Instead of just seeing the little Gemini star, you'll see some like lines in the Gemini star. That's the new live icon. If you tap that goes full screen and then you just talk to it to say like, find all those photos of coffee cups that I mailed myself.

19:54Some users have also seen keep live show up. That brings voice search to the keep app and docs live is apparently in the works as well. Yeah. And so it sounds like it's not just search. It's like, you know, as with live, it's conversational, right? Yeah. Yeah, yeah, yeah. Let's talk about what's in my Gmail inbox. What this, I don't know if you've had a chance to play around with Spark at all, but what Spark from a handful of weeks ago and then this and everything really reminds me of is like I've played around a lot with, and I've talked about it plenty, with Claude and other services to kind of improve my Gmail experience or make it so that I don't miss things or make it faster and easier, more comprehensive, all that kind of stuff.

20:37And it seems like to me that Google is doing what Google often does, which is they've got all of these different surfaces and just a wealth of information interconnected throughout it. Personal intelligence is a prime example of that. And they're just basically taking what you can do with other things if you piece it all together and just integrating it into the service so that it's there. And like Spark does that really well for a few key things with my email for giving me like a daily summary of things, which I was building something around before. And so it's just the value of what Google has given how many people use all of their different services.

21:14Interconnecting them like this is powerful if you opt in. Yeah, yeah. If you choose to do it. I ran into it the other day. I was using Codex to create some travel documents. and I was like, you know what, take some of my photos that are appropriate to this trip and throw them in there to make it look nice. And Codex was like, yeah, I don't have access to that. I can't do that. And if I had been in Google at the time, it would have just said, go to Google Photos and find them and it wouldn't have had a problem. So you do have those little things. And again, if you're a power user, I'm sure you're like, well, all you needed to do was create this subroutine and prompt it this way.

21:53At the time, I was like, yeah, I don't need to go down that path. I'll find another way to make this work. And that's the thing with these tools is it really does lower the barrier so that more just regular users are tempted to explore it. Yeah, yeah. Maybe you needed open claw before something, you know, to do something like this. Right. Now it's just like, oh, no, it's just there in the search box in your Gmail app. Yeah, I feel like that is often one of the ways I use these tools is kind of dumb. like I'm like, I know I could do this, but most people won't. So I'm not because I want to see how easy it would be for someone who's just sitting down and like, you know, in a rush or, or not wanting to put that much thought into it can make it work.

22:35Well, on the other side of the scale, I put a lot of thought into new Patreon perks coming July 1st. If it's July 1st, you can see them right now. We have new benefits, including exclusive ways to find out what's going on behind the scenes, even have a voice in the show. And the more pertinent part for you folks who are not patrons yet, from July 1st through July 31st, if you use the code EXPERIMENT, you get 26 % off your first pledge. That's for new patrons. Use the code EXPERIMENT when you check out the new tiers at patreon.com slash DTNS.

23:17This 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.

23:50All right, we got a whole gaggle of quick headlines for you to just fill you with smartness. Yeah, like geese. Yeah. Fill you with smartness. sense. Well, a gaggle of phone makers are announcing cuts in shipment targets. Xiaomi, Oppo, and Vivo, all revising shipment targets by up to 30 % lower because of difficulties getting components. And it's not just memory. It's almost every smartphone component facing either higher prices, tighter supply, or both. Oh, man. Things are going to get really interesting in the smartphone world in the next couple of years, I think. And we'll be following it. Chinese food delivery giant Mai Tuan.

24:33Is it Mai Tuan? Mai Tuan. Mai Tuan. Thank you. Open sourced their LLM Longcat 2.0. It's a 1.6 trillion parameter AI model that it says is the first of its size fully trained on domestic Chinese chips. It uses a cluster of 50 ,000 processors that were reportedly tied to Huawei hardware on Chinese open source models just kind of continuing to close the gap. It's getting real interesting real fast. I got two reactions to this. One is the longer you deny the use of other processors, the more independent from needing your processors companies will get. And two, the food delivery service in China is cutthroat.

25:16There is no margin and no profit. So I'm not shocked that Meituan is like, maybe we make more money if we push an open source model out there. Yeah. Here, here's our model. Look how awesome it is. So they're kind of doing an all birds. They're doing a some birds because they're not dropping the food delivery. But yeah, they're doing a partial all birds, perhaps. That's just a supposition on my part. Thanks to RWNash and the subreddit for noting this one. researchers found three airdrop vulnerabilities that would let a nearby attacker crash airdrop, airplay, handoff, and other Apple services with a single request and be able to keep them down.

25:59That is an annoyance more than anything, but Apple has patched one of them and is working on patching the other two. Okay. All right. Look for that. UK's Competition and Markets Authority Liberty has proposed letting app developers point users to payment options outside of the App Store and the Play Store. Also, they are proposing the requirement of Apple to open its NFC chip to rivals of Apple Pay. Yeah, this is going to be one of those stories where we are going to have to come back to you in a few months with, and now they're doing it. And this is the thing that they're actually doing. Yeah, yeah.

26:37According to Wired, hundreds of meta contractors posed as miners, not like diggers in the ground, but like people who are younger than 17 years old, say, and then went to competitive chatbots and put in prompts about suicide, sex, drugs to see what ChatGPT, Gemini and character AI would respond and then log the answers. okay uh thanks to fuzzy mistborn in the subreddit thank you fuzzy uh d brand has canceled and is refunding a hundred thousand pre-orders of its portal themed companion cube case for the steam machine valve had reached out and said hey you never actually licensed the portal ip you can't do this it looked cool though i'm sorry you can't do that it's probably what it's close to what they said.

27:32Joby Aviation and Toyota announced a joint venture to scale up production of Joby's electric air taxi, Toyota taking a 51 % stake in the partnership. So Joby's been kicking around for a long time. It's got real technology, but it needed something to scale. And Toyota ain't a bad partner. Not bad at all, if that's what you need. Proton announced a new version of its privacy-focused LLM. LUMO 2.0 promises better reasoning, image generation, memory, and more accurate search and citations. LUMO routes queries to the appropriate model. It uses open source or open-weight models like Quen and Zipu. It's available for free with limits, unlimited for subscription, and also for business and government use cases.

28:17Yeah, I have not found too much success in the few times I've tried the free model, but the fact that they're touting it as being so much improved, I'm going to give it a few more tries and see what it's good. Because I use Proton for other things like password management and stuff. So I would like to be able to use them for that as well. Apple told Reuters it is pushing out more bug fixes earlier than it used to. Apple used to say like, OK, we know the bugs out there. It's been documented. It's probably not going to be exploited. So we're going to push them all out in a yearly release. More critical vulnerabilities they would patch on an accelerated schedule.

28:53But by and large, there wasn't a big risk in holding them all till one patch. And that just made it easier to push everything out at once and test it and all of that. However, LLM-generated exploits have shortened that timeline between disclosure and deployment. And so to maintain its good record of patching vulnerabilities before they are exploited, Apple is going to start pushing out bug fixes more often or earlier. Interesting. NASA will launch a robot spacecraft called Link into orbit on June. Well, today, that would be today to attempt to raise the orbit of the Swift Observatory and prevent it from burning up in the atmosphere later this year.

29:34Link will be launched on a Pegasus XL rocket, which will be delivered into orbit by Northrop Grumman L-1011 jet called Stargazer. Yeah, this Swift Observatory does like magnetosphere stuff, I think, gamma ray observations, and it would just burn up and die. You wouldn't be able to use it at all anymore later this year. So it will be interesting to see if they can make this work. apptronic announced a robot training facility called robot park in austin texas this is meant to accelerate the adoption of commercial robots uh developed with google deep mind and also announced uh or at least introduced its newest humanoid robot apollo 2 and finally the new york times has i thought an interesting report about tech workers in san francisco who are finding that their lowly$180 ,000 salaries aren't enough with all of the AI-driven wealth proliferating through the area.

30:33And I just find it an interesting read. It definitely kind of takes me back because I used to live in San Francisco. I still live in the Bay Area. You used to live in San Francisco. I feel like it's a story that's retelling itself because I feel like we've been here before. Yeah. San Francisco tech salaries make it expensive to live in San Francisco is a recurring story, isn't it? Yes. Indeed it is. but interesting article nonetheless definitely read it we end every episode of dtns with some shared perspectives and today we're highlighting chris yes every so often i like to highlight someone in the community who's doing some cool stuff today it's chris chinchilla uh i was going to be able to see chris when we have our european meetups next week but i told him the wrong date in discord so to make up for that and his canceled train ticket which i feel very bad about i want to direct you to his excellent Tech Lounge podcast, full of great content for developers.

31:26Most recent episode when I looked was about the OSS resistance, a call to contribute to open source, trying to get more people to join up on projects that need your help. There's also an episode about an open source project called Dockling for converting docs, especially messy docs, into structured data, all kinds of good stuff over there. So please help me make it up to Chris and get some fine technology info in your ears by checking out Tech Lounge and all of Chris's other fine works at ChrisChinchilla.com. Folks, if you got some insights into a story, if you're like, hey, you know what? I do know what that little keyboard thing from OpenAI is going to be good for or anything else, email us feedback at DailyTechNewsShow.com.

32:10Yes, please. Thank you once again to Chris Chinchilla for being awesome and 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. And then you're awesome, too. I mean, you're all awesome. But Patreon.com slash DTNS. We'll see you tomorrow.

32:31The DTNS family of podcasts, helping each other understand. Diamond Club hopes you have enjoyed this program.

32:43This 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.