OpenAI Pauses Training Again After More Agent Alignment Issues - DTNS 5362

0:00Robb Dunewood: This 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, everyone. It's Adam, host of Ted's podcast, Rethinking with Adam Grant. And this episode is powered by AT &T Business.

0:36When you're building a business, momentum is everything. There's nothing worse than hitting your stride only to have a spinning loading icon bring everything to a halt. When navigating competing priorities, reliability isn't a luxury. It's a foundation. AT &T Business is built to work so you can spend less time troubleshooting and more time focusing on what actually matters. powered by AT &T Business, built to work. Get today at business.att.com.

1:04Robb Dunewood: 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. 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.

1:46Robb Dunewood: 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. Think of a toilet as just an everyday object? Your entire perspective shifts the moment you experience a Kohler smart toilet. Design changes everything, transforming a basic routine into something extraordinary. With standout aesthetics and intuitive controls, these are functional works of art. Stunning design that never sacrifices performance. Enjoy customizable features for elevated comfort and convenience.

2:28Robb Dunewood: It's more than a fixture. It's a cleaner routine and a more refined space. Experience the difference of Kohler Smart Toilets. Find more at Kohler.com.

2:50This is the Daily Tech News. For Monday, September 28th, 2026, we tell you what you need to know, give you the important context, and help each other, hopefully, understand.

2:59Robb Dunewood: Today, OpenAI has a whole slew of containment issues to contend with. Yeah, there's a lot to talk about, and it extends deep into the show. So we got a lot to get to. I'm Jason Howell. And I'm Rob Dunwood. We're going to start with what you need to know with the biggest of big stories.

3:18Yeah, so a lot to get to today on the topic of AI agents misbehaving. I like to call it breaking bad. I don't know. It brings a little levity to it. It reminds me of a really great show. But anyways, we're going to start with OpenAI and its agent alignment issues, if you want to call it that. And then we're actually going to get to more that's related to all of this on the other side of the break because it's just that kind of day. There's been a lot of news over the weekend. So let's start with OpenAI. OpenAI's agents, some of them anyways, may have used proxies and an encoding trick to get around API limits and scan a UN, that is United Nations data website, and that site is UNCTADstat.

4:03There's probably an easier way to say that, UNCTADstat. Anyways, Anyways, around 16 ,500 times between April 13th through June 19th. Now, this is all according to Rowan Howard Jones, who went digging into the Transluse report that actually the DTNS crew discussed on Thursday's episode. UNCTAD stat, there's got to be a better way, is a public data site of UN trade and development. And the requests were actually looking for figures on food trade, industry, productive capacity. So they were looking for that. It's important to note here that all of the data that accessed was public, right? So this is not considered a hacking incident.

4:55And this is something that I think we're going to have to talk about in this top block, kind of the differences between those two. But that's not the only thing. OpenAI also says its agents posted 53 images uploaded to ChatGPT to image hosting websites. Those images were anonymized and they were being used for training. So since this report, OpenAI has announced that it's removed some of those images, is working to remove all of them. But we know about that. Agents bypassed security controls to reach some websites, including the Securities and Exchange Commission, the U.S. Department of Commerce, the U.S.

5:35Census Bureau. But again, here, right, like they used publicly available developer keys to read public information. Some of that information was then reposted elsewhere. But again, we're talking about public info. And an agent escaped DNS sandboxing due to insufficient DNS filtering, reached an external chatbot. OpenAI stopped that training run and, quote, decided to pause all other training, evaluation, and inference with tools use, defined broadly, for our most capable models until we have both validated that the gap is resolved and performed additional red teaming of the system, end quote. So they're kind of putting at least a pause on that particular model.

6:21And apparently this pause has been in place since September 20th. So it's been going for a little bit of time now. So we got containment issues. Containment, of course, being the terminology for when agents do things that maybe the people who are implementing or using those agents didn't intend for. Breaking out a sandbox would be one example. But I don't know, Rob, what do you think of as far as the kind of the determination here or the delineation, if you want to call it that, between agents and the story that we've heard a lot in recent months from various publications that they're hacking into systems or doing this sort of like hackery type action versus just kind of using the system to access and recognize public facing information.

7:11information that's just out there for anyone to get, of course, the agents got a hold of it. Do you see a difference between those two?

7:16Robb Dunewood: I do. I see the difference because these are not hacks. These are this system was open. This, you know, the AI had access to it. What happened is that we didn't think that the AI was going to access it in this way, a way that we gave it at you. We gave it the ability to do this thing. We used to anticipate that the thing that we gave it the ability to do, it was going to be very efficient at doing it and get around this thing that we didn't even know was a way that we could get around something. So these are not hacks, but we have to think it's like, I know, because when you, when you say hack, there's like, you know, you automatically think there's something nefarious going on.

7:54Robb Dunewood: It's a very loaded term, right? These are computers that as long as we have water and electricity, they will just keep doing what they were programmed to do over and over and over again. They don't get tired. They essentially will try every single way to figure something out. So what we are surprised by is that we're giving them instructions on what they can do. We're giving them parameters on what they can do. Oh, we didn't anticipate this. It falls under the parameters that we've given it. We just didn't anticipate that this was the way that it was going to get around this particular problem.

8:29Robb Dunewood: And so, no, that's not a hack. Should we be concerned with, you know, about these systems doing that? Probably a little bit. But I don't think that we jump to the sky is falling, the sky is falling. You know, we're being hacked by our AI. These systems are, this stuff is publicly available. We're getting the stuff that we are actually able to get to. We just didn't know we were going to get to it in this particular way. Yeah, I mean, like you said, and I think I absolutely agree, is that it doesn't, what we're saying here doesn't mean that we're saying this is not concerning. Certainly it's concerning.

9:02There are legitimate failures. There are legitimate warnings that are probably warranted as a result of this. Much of the press, though, and the reaction often jumps right to skies falling with something like this. And like you said, these are just systems that have been tasked with carrying out a certain action. And then in some cases, people are surprised when they do it really, really well, based on the information that they have access to. And, you know, it's just, it's a, it's a, an example of a system just being very effective at carrying out a task. And that, again, that's not concerning in its own right, because there are certainly ways in which that can take you down roads that were unintended, potentially harmful in some way, shape or form.

9:56Define that how you will. But it doesn't always constitute hack because like you said, hack as a word, it carries a lot of loaded kind of armor with it. And it in itself can instill a lot of fear, a lot of uncertainty and doubt. And I'm not entirely certain that's what we're looking at here. At the same time, Anthropic, OpenAI, they are facing an accountability crisis, I think. They need to be better at being accountable for the things that are happening, in my personal opinion, whether it is a hack or whether it is just, hey, it did this thing. And I guess we're seeing some of it. It's just like, are we seeing it as a reaction to something else?

10:36Are we seeing it because it behooves them? They have something to gain from sharing the information. It's hard to know.

10:44Robb Dunewood: Yeah, once again, I don't think this is a nefarious hack. Where my concern comes in is that, okay, AI companies, this has happened enough times to enough of you that you now need to say, you know what, these things are really good at doing things we didn't anticipate. So we need to build an anticipation moat so that they can't do the things that we didn't anticipate, because potentially some of the things that we don't anticipate could be very, very, very bad. Once again, not nefarious, but we don't want to accidentally do something that can be incredibly harmful. And I think that these companies have to do a bit more to think about that as these systems get more and more complex and just get better and better and better at doing what it is that they do.

11:27Yeah, yeah. And more and more capable, right? And related to all of this is this idea of recursive self-improving systems. We're going to, on the other side of the break, actually, like I mentioned at the top of the show, this story kind of goes further in a few different ways. And one of these ways is about recursive self-improving systems. We're going to talk about that after the break. We'll get to what that is, whether it actually poses a real threat, if you want to call it a threat, or if it's just perceived. But that's going to be coming up here in a few.

12:02Robb Dunewood: DTNS is made possible by you, the listener. Thanks to Larry Bailey, Michelle Sergiu, Kirk Stephenson, and Justine MacDonald.

12:14Robb Dunewood: 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. That'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.

12:46Robb Dunewood: 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. You 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 daily tech. That's granola.ai slash daily tech to get your time back. Go to granola.ai slash daily tech and try it on your next call. Episode prep gets messy when research, Slack threads and script trackers are all in different places.

13:21Robb Dunewood: That's where chat GPT 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. This episode is brought to you by Capital One.

14:01Robb Dunewood: 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. Think of a toilet as just an everyday object. Your entire perspective shifts the moment you experience a Kohler Smart Toilet. Design changes everything, transforming a basic routine into something extraordinary.

14:41Robb Dunewood: With standout aesthetics and intuitive controls, these are functional works of art. Stunning design that never sacrifices performance. Enjoy customizable features for elevated comfort and convenience. It's more than a fixture. It's a cleaner routine and a more refined space. Experience the difference of Kohler Smart Toilets. Find more at Kohler.com. There are moments that reveal who we are and who we're meant to become. For those born to lead, such moments call for a vehicle of equal distinction. Dynamic by design, uncompromising in execution, the Range Rover Sport was engineered for those rare individuals who demand the world and possess the conviction to claim it.

15:24Robb Dunewood: When your time to lead arrives, you need a vehicle that rises to meet it. The Range Rover Sport. Exclusive offers available now. Explore further at RangeRover.com.

15:36All right, there's plenty more we need to know today, so let's get right to the briefs.

15:42Robb Dunewood: 22 AI researchers publishing personally, not on behalf of their companies, published a paper today called What If Automating AI R &D Triggers an Intelligence Explosion? Among the authors are OpenAI Chief Scientist Jacob Pachaki, Anthropic Co-Founder Jack Clark, Microsoft Chief Scientific Officer Eric Horvitz, and Meta's Vice President of AI Research Don Song. First, what is AI intelligence explosion? In this context, it has to do with Anthropic and OpenAI who are beginning to perform more internal research and development tasks and automate a sizable portion of their engineering tasks. As they do this, their operations get closer to recursive self-improvement, or RSI.

16:20Robb Dunewood: They say an intelligent explosion is not a certainty because of computing constraints, automation challenges, diminishing returns from expanding capabilities, or time-consuming training runs. But they do believe that there is a chance of that happening and suggest mandatory reporting of incidents, independent evaluators, safety requirements for testing powerful systems, and creating ways to pause certain AI work. The concern is that once an intelligent explosion begins, the window for action may close. Hmm. So the real challenge becomes if the systems can move, they can improve themselves faster than, say, a human evaluator can actually keep up with it, then that ultimately potentially limits the visibility, the human visibility, into why they work the way they do.

17:09And if there's maybe a limit that needs to be implemented or a safeguard that needs to be integrated, if the human is behind because these systems are so good at updating and revising themselves, right, that they just move at this rapid rate, which is largely what we've felt, I think, in the tech industry the last couple of years is how quickly these systems have gone from writing a dumb little poem to doing all the things that they do now in a matter of a few short days. So it's not impossible to see that these systems might be able to, you know, update at a rapid pace. And basically, these, you know, these folks who wrote this are, you know, saying outside of their company, basically saying like, hey, this is something to pay attention to, because there might be a point at which they are so capable and so advanced at updating themselves that the humans can't even like you can't keep up with the rate of improvement.

18:07And that could potentially lead to, you know, whatever bad things they think it could lead to.

18:12Robb Dunewood: Yeah. If you if you just think about it, two years ago, it was it was difficult for robots to to basically just walk in a straight line efficiently. You know, humanoid looking robots to just just just walk straight in, you know, you know, you know, and and be good with doing that. And then we've seen like, you know, global dynamics and, you know, these companies that that, OK, these things are getting really, really good. Like you can you can hit them with a baton and they keep their balance and they're just getting better and better and better. And then we just recently had a robot that can run faster than Usain Bolt, the fastest human that we know of.

18:48Robb Dunewood: And it's not like these things are going to get any worse at this. They're going to continue to get better and better and better. When you move it into the realm of the A.I., it's that we right now, at least with the models that we are privy to, the models that we are aware of in the general public, humans are writing these things. and humans are making these things better. But as these things code themselves, and they get better and better at doing that, you're going to get to where we have an AI that wrote an AI. And if it's good, and that works, that work becomes ridiculously fast. It's not going to be every couple of months or three, you know, every, you know, two, three months that we'll see a new engine come out.

19:27Robb Dunewood: It'll be every few hours, maybe every few minutes that we see a new one. Because once again, And I said, this is the top of the show. All these things need is power and water, and they can just go and do what they do. So if they're if they're making themselves better at scale and at speed, that absolutely can become a problem somewhere down the road. Yeah, this really reminds me last. I guess it was last Thanksgiving. So at this point, you know, closing in on a year ago, Google had invited me out to this like Friendsgiving thing for their team. And it was right around the release of a major Gemini milestone.

20:04I think Nano Banana Pro was happening that week. And then there was Gemini. I can't remember what the major release it was. 3.5 possibly, although don't quote me on that. And that was one of the questions that I asked the team at the time because it seemed really far away. But I was very curious. I was like, are you using your model to develop your models that come forward? And even then they were saying, yeah, definitely, you know, a large part of the development process of this current model was using the models that we have access to to do that. Now, they weren't, you know, it wasn't like top to bottom, like, all right, make the new model.

20:40And it just did it. They were integrating pieces of it into the process. And so you could technically say that they were using the model to create a new model. I think this is kind of talking about taking that to the next level and saying, do these models have the ability to wholesale create the next one and at such a rapid pace that humans can't keep in the loop? And I think that visibility into the development process is personally, I think that's very important.

21:07Robb Dunewood: Jason, think of it this way. We aren't even four years into the large language model takeoffs, sendoff, wherever we are. We're not even four years into it. There are college freshmen that when they were high school freshmen, we hadn't heard about ChatGPT yet. Not even four years ago. So if you think about how fast we've iterated in just four years where humans are still behind the wheel, that we're still controlling what's happening. Just think of how fast it goes if the AIs themselves can actually iterate themselves. And that's where the concern is. I'm just fascinated. As a technology fan, it is a fascinating, interesting, sometimes unnerving time to be living in the technology world.

21:51We are, I don't know, I do a podcast on AI, so of course I'm fascinated by this stuff, but I feel like it's unprecedented times. Now, speaking of keeping agents contained, of this containment thing, theme of today's show, NVIDIA thinks that it has a solution. NVIDIA has announced the open agent safety platform. What NVIDIA is calling a reference design for how agents can be confined, contained, monitored, let's say. NVIDIA has released some of it as open source, along with a slew of partners announced. Cisco, Microsoft, Oracle, CoreWeave, Dell, HPE, Lenovo, ARM, Intel. So a lot of pretty major partners there.

22:35NVIDIA CEO Jensen Wong described it as a, quote, browser for agents, which essentially means that it allows access only to things an agent actually needs, in other words. The system does this in two layers, like a two-layer approach. There's OpenShell, which sets limits on agentic capabilities. Anthropic, by the way, is implementing this part. This is the containment layer. When an agent tries to access something, it's not allowed to access it. OpenShell would essentially attempt to block it. And then Sentry is the other part of it. It monitors and enforces the agents. It runs on network chips, not CPUs, not the GPUs.

23:20And basically, that separation is important because it means the system can effectively remove the watchdog from the thing that it's watching. So it's a separate entity. And NVIDIA claims that this system, by the way, could have prevented the hugging face incident, which has really been the moment that has kind of started this current wave of containment concerns, let's say.

23:44Robb Dunewood: This is good work by NVIDIA, in my opinion. I think other companies are going to do similar things. The reason for this, and we talked about a moat at the top of the show. So we these tools are getting so powerful that we need to create a set of rules for here is how you access this stuff. These are the only ways that you can access this stuff. And here are the parameters for it. That way you do tend to eliminate not eliminate. You tend to mitigate the possible accidents that can happen. If you're saying that you can only access this data in this way that we've approved, then it would be a hack.

24:17Robb Dunewood: It would be it would be getting around that system where we actually said that we there are some things that we don't anticipate. So we're only going to allow you access in this way. I think that is good work because we do not want, especially with agents. And let's take the nefarious stuff out. You know, we talked about Muse here on the show over the last month since it's come out. But it's an agent that basically you can give it access to your bank account. You can give it access to your cell phone. You can give it access to your email, your Gmail. You can give it access to these systems. you know, Walmart, Best Buy, they're working with Meta in order for you to be able to go and buy stuff that's hooked up to your credit card.

24:54Robb Dunewood: Some really bad things can happen if a rogue agent just went and started buying stuff because something happened, it wasn't intended. So I like what NVIDIA is doing here. I want them to continue to do this work. And I want the other players to also do this work as well. So we can just try to mitigate some of the risks that we could, you know, you know, befall ourselves if we just, you know, I'd rather have a system of how you access data as compared to just go access the data however you can get to it. I think the former is definitely better than the latter. Yeah. Frameworks are important. We'll see how developers choose to kind of configure their limits on this, how Sentry enforces them, how reliably all that's done.

25:36But yeah, I like I like what this what NVIDIA is laying out here. I think it's an important timing for something like that.

25:43Robb Dunewood: Bose is getting in on the weird audio comeback with the$100 Bose noise-canceling wired earbuds, a USB-C pair with active noise-canceling, a feature that is rarely found in wired earbuds. The buds don't rely on batteries, but they pull the power needed for the noise cancellation through the USB-C connection. That means that battery life on the connected device will suffer slightly, and the feature won't work when connected to systems that don't deliver power. However, reviews on CNET and Bloomberg do share that the quality of this noise cancellation falls below Bose's more premium products. The buds are now on sale with an expected shipping date of October 15th.

26:23Yeah, CNET's David Carnaway, by the way, gave these an editor's choice, said they're among the best wired earbuds he has tested. And yeah, it is a little strange to have noise cancellation, active noise cancellation on like wired earbuds because usually those have the battery embedded in them getting that through usbc is great you're going to be limited though like if you get on an airplane and you want to plug into the uh airplane system um if it has usbc at all for audio which usually they have right like they usually have like a 3.5 millimeter although actually i was just on southwest and i didn't see a 3.5 but they did have a usbc charging port i wonder if that would have I wonder if that would have done it anyways that's neither here nor there you know does it deliver power is the other thing you know you are you know you might be able to plug in usbc to get the audio through but it might not deliver enough power to drive the automatic the active

27:19Robb Dunewood: noise cancellation so that's another factor yeah my gut tells me that most people that are going to be using these they're plugging this stuff into into tablets into laptops into their phones these things all do deliver or most of them do many of them do deliver power so that's that's where i think i don't think this is going to be the audio file who's going to be plugging usbc into their system with tubes in it they're probably going to still go with a you know with at least a trs connection if not a full you know quarter inch jack to plug in if they if they want that true uh if they if they want that true analog sound but this is kind of cool because it has been something that if you do get wired headphones even if they're not the super premium ones if you if If you get them wired, we were just talking about earlier, mine, I have open back or semi open back so I can hear around me.

28:06Robb Dunewood: You know, so so people do like to have, you know, you know, sometimes ambient noise coming through. Sometimes you want to you want to you want to kill all that. So just having a wired pair of headphones that allows that is pretty cool. Not to mention it's a it's a wired headphone resurgence happening, you know, and I don't know how much of that is nostalgia for an older time or how much of that is like these things are just good for certain things. where you don't want to have to deal with Bluetooth connectivity issues or latency issues. That's another really great thing about these is that if you're gaming and you plug these in, you're going to have really low latency audio on the other side because it's not converting through codecs and all that kind of stuff.

28:44Oh, yeah.

28:44Robb Dunewood: I didn't even think about plugging them into your system. That's actually pretty cool as well. Yeah. Yeah. Cool stuff. Well, folks, we live stream the show every Friday where we answer the big tech question of the week. Catch them by becoming a subscriber at youtube.com forward slash Daily Tech News Show.

29:25Robb Dunewood: 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. Some days, getting the work done isn't the problem. Staying locked in is. AlphaBrain is Onnit's daily nootropic formula for cognitive support, designed to support memory, focus, attention, and mental clarity. Its formula with science-driven ingredients include theanine and Hooperzine A to support focus and memory, and it's caffeine-free. So when you're deep in the work, AlphaBrain helps support enhanced productivity and tune-out distractions.

30:02Get focused on what matters. Shop AlphaBrain now at onnit.com.

30:10All right, we have a slew of quick headlines that'll make you look really smart. They're good to know as well, so let's get into it.

30:18Robb Dunewood: An eventual phase-out of Chrome OS is looking likely after a Google Book support page listed that Chrome OS updates will continue through mid 2034. At that point, Google plans to phase out the operating system entirely. Many of the newest Chromebook models will be eligible to transition to Google Book OS. So at least there's that. At least you got that. Yeah, I kind of saw this coming, right? Like, yeah, this has been the question. And it kind of seems like the support page is saying, yeah, it's probably going to get phased out eventually, but you got time. So let's just say that. SpaceX's Starship reached orbit for the first time on its 14th test flight.

30:53The planned 10-hour mission, though, was cut to about three hours after losing one of six Raptor engines shortly after stage separation. The 26 Starlink satellites on board deployed successfully.

31:07Robb Dunewood: So the cool thing is that there was a massive failure, but it was still able to do its job and get back. So that was pretty cool. Yeah, indeed. Bloomberg's Mark German reported over the weekend that Apple's Vision team is exploring four concepts for a new headset, codenamed N224. Not expected until at least 2028. Apple is reportedly prioritizing smart glasses for late 2027. The Apple Vision Pro continues to get pushed away. Smart glasses, the new thing. Yeah, be curious what that means. A Santa Fe jury found Meta misled Facebook users about data sharing in the Cambridge Analytica scandal. You remember that?

31:46and identifying more than 43 million violations of New Mexico's consumer protection law. Now, the state is seeking up to$5 ,000 per violation. That could total more than$200 billion.

31:59Robb Dunewood: And that's just one state. That's just one state. All the states are usually going to have these lawsuits. A federal appeals court upheld the Pentagon's blacklisting of Anthropic in a two-to-one ruling. The case is far from settled, however, as this goes against Anthropic winning a separate but related court challenge in August. Yeah, this is not over. Sony says it's skipping the Consumer Electronics Show entirely next year to focus more on entertainment. By the way, if you had noticed, or if you hadn't noticed, Sony did not have a booth at CES last year, or rather this year. I guess it's the year that we're currently in.

32:36So this isn't entirely unsurprising, I suppose, but CES seems like the place for a company like Sony to be. It's surprising to me. It is.

32:46Robb Dunewood: I missed the Sony booth earlier this year because they had the they had the mass. It's like a city within the convention center. So, yeah, this is this is different here. The Dutch government is building a Nix OS based Linux desktop for government agencies after U.S. sanctions against the International Criminal Court cut off the ICC prosecutors access to Microsoft services. A stable release is planned for 2027. There you go, Linux fans. That's for you. Raspberry Pi's strategy of stockpiling memory before prices spiked, it paid off. They had record first half results, revenue up 90%. Not bad, Raspberry Pi.

33:25Yeah, good on you, Raspberry Pi.

33:27Robb Dunewood: If you have an Xbox Series X, you can now play certain PlayStation 5 games on the console with an unofficial port of the open source KTi PS5 PlayStation 5 emulator, a project called PS5X. It requires developer mode and only a small handful of titles work so far, but development is underway. Wow. Yeah, pigs flying there. That's very interesting. And again, just to reiterate, this is not like an official thing. This is an unofficial mod port, so to speak. And there are not many games that qualify yet, but it might be the beginning. Yeah, it might be the beginning of something. We end every episode of DTNS with some shared perspectives.

34:07today more excellent cable management tips this one courtesy of jay jay writes some time ago my

34:15Robb Dunewood: wife asked me to clean out my junk boxes they were mostly filled with old cables they have come in handy so i don't want to throw them out after fighting with them to get them untangled i had an idea to put all my cables in ziploc bags one cable per bag no more tangles my wife gave me a sheet of stick-on label so i could put what type of cable was in each bag i used labels as my hands turn black from the Sharpie wiping of the Ziplocs. Three boxes condensed into one small box and finding the cable I need is super simple. Work, friends and family all know who to call if they need an obscure cable.

34:46Robb Dunewood: Pro tip, use good freezer style thick Ziplocs. Love the show. Long time listener from BOL Days. I swear, Jay, you're reading my mind. Back in my little cabinet down there in my set is my pile of cables. And probably about a year ago, this exact thing dawned on me. I have one bag, freezer style Ziploc bag. That is my HDMI. You know, one is for USB-C and at least it keeps like, they can be messy inside of there, but at least it keeps them contained. And so when I need a thing, I just pull out the Ziploc bag, reach in, grab it out, and it stays all organized. It's really cool. That's a good idea because I actually have a Tupperware tote full of cables and the cables are nicely wrapped up, but they're just loose.

35:32Robb Dunewood: So over time, you just keep throwing more and more. They're going to get tangled. So yeah, the Ziploc is pretty cool. You know what is also pretty cool? That we have listeners to go all the way back to Buzz Out Loud. How awesome is that? Love it. I always love to see the letters B-O-L appear in an email. Makes me happy. So folks, we love to know what you're thinking about. So if you've got insights into a story, please, please, please share it with us at feedback at dailytechnewsshow.com. Yeah. Yeah, big thank you again to Jay for contributing to today's show and cleaning up your cable mess. And thank you for being along for Daily Tech News Show.

36:06You can keep us in business. All you got to do is become a patron, and you should do that. Patreon.com slash DTNS. See you tomorrow.

36:25Robb Dunewood: The DTNS family of podcasts. helping each other understand. Diamond Club hopes you have enjoyed this program.

36:38Robb Dunewood: This 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. Hi, this is Brooke DeVard from Naked Beauty. I am always looking for ways to elevate our daily rituals.

37:14Robb Dunewood: And the Kohler Smart Toilet is proof that design changes everything. It's this stunning sculptural piece that feels like quiet luxury for your bathroom. It totally transforms your daily ritual into something elevated. with customizable cleansing and touchscreen controls. It's functional art that makes your space feel cleaner and more intentional. A modern home deserves a modern toilet. Experience the difference of Kohler smart toilets. Find more at Kohler.com.