 This 10th year of Daily Tech News show is made possible by its listeners. Thanks to all of you, including Adler, the puppy engineer, Dale McKayhee, Matt Zaglin, and our new patron, Ryan. Welcome, Ryan! On this episode of DTNS, ARM's IPO has a pretty positive initial day. MGM isn't out of the hacker woods, but it's got company. And Google wants to extend the life of Chromebooks in schools. This is the Daily Tech News for Thursday, September 14th, 2023. From Studio Secret Bunker, I'm Sarah Laney. From Columbus, Ohio, I'm Rob Dunderwood. From a balmy Southern California on the shows, producer, Roger Chang. And joining us... From Scottsdale, Arizona, it's me. I was about to introduce you, Teja. And Teja's like, I can do this myself. Teja Kastodi, YouTuber, host of AI Named The Show, and also talk techie to me, the podcast. We are so glad to have you back on the show. Welcome back. Well, thanks, guys, for having me back, even though I talked so long last time. Please talk long again. That's why we have you. That's why we love you. All right, without further ado, we've got some tech news. So let's start with the quick heads. Now valued at more than $5 billion, ARM's IPO was shaping up to be the largest of 2023. The UK chipmaker priced 95.5 million shares at $51, which climbed to $56 in just a few hours on the NASDAQ market where it was listed. Semi-analysis notes that ARM success is due to its innovative architecture, flexible licensing model, and strong ecosystem of partners. HP announced the HP Spectre Fold, a Windows laptop with a 17-inch OLED screen that folds in half. HP says the Fold is the world's thinnest and smallest 17-inch foldable PC, weighing in at just under three pounds. If you're ready, if you're already sold, consider the price, which is $4,999 and 99 cents. Oh, dear. Yeah. Sony threw us a little Thursday curve ball because this week wasn't weird enough, announcing a surprise slate of play, a state of play rather, a streaming event later today, September 14th at 5 p.m. Eastern Daylight Time. The streamed announcement promises something for everyone, including updates on major releases and PSVR2. The company says the showcase will have a focus on indie and third-party titles and will primarily feature updates to previously announced games. So probably not a huge, huge announcement, but something if you're in the market. And good gaming news. Bethesda Starfield announced it will add deep learning super sampling support, or DLSS, and other PC features that were missing at launch. The developer says an HDR calibration menu, a field of view slider, ultra-wide monitor support, and brightness and contrast controls are also new features on the rollout. Yesterday we talked about Game Engine Unity's planned install fees that did not go over well in the game developer community. Unity says in a blog response to the backlash that, quote, 90% of Unity customers will not be affected by this change, end quote. Reminder that Unity previously said that it will charge 20 cents per install for any game with more than 200,000 installs that sparked a lot of questions about multiple installs, Game Pass, OS reinstalls, a variety of other issues. And a lot of people said, you didn't really answer those questions. The company did say it will only count net new installs on any devices starting January 1st, 2024. It also says developers won't pay fees on reinstallations, fraudulent installs through botnets and other tricks, trial versions, web and streaming games, or charity-related installs. Well, the whole Unity thing is a bit of a blur, but what isn't a... I was about to say what isn't a blur, Rob, but it may be a blur is what's going on in Las Vegas. We have a bit more to discuss from our conversation yesterday about MGM's hack. The hackers who allegedly breached the security at MGM's casinos earlier this month. The new news is that they originally planned to manipulate the software running the slot machines and, quote, recruit mules to gamble and milk the machines. End quote. On the Telegram messaging app, a person who claimed to represent the group the claimed responsibility for this hack described the techniques used to evade detection. Security researchers familiar with the group nicknamed Scattered Spider said that the technical descriptions given to the Financial Times did match attacks on at least 100 other victims over the past two years, which makes you feel like, okay, maybe this group that's claiming responsibility is indeed responsible. But, Rob, there's a little bit more here, isn't there? Yeah, there's definitely more. So Bloomberg sources initially said that Caesars not to be confused with MGM apparently paid tens of millions of dollars to hackers who broke into the company systems prior to the MGM leak and threatened to release the company's data. Thursday after Bloomberg reported that Caesars had been hit by a cyber attack, Caesars disclosed the hack in a regulatory filing. So yeah, this gambling companies in general are getting hit here. And here's the weird thing about this. Caesars didn't say anything until MGM got hit. They kind of waited on this. The laws have changed. They don't think they go into effect. The SEC has changed reporting rules. So the AK that Caesars ended up doing yesterday, I don't know that they would have been able to wait to do that. I think the new rules stipulate you have to do that within four days of being aware of a hack. So it's interesting to see how they're still using the old rules, but how things are going to change also soon. Well, we've talked about this quite a bit in the past of, okay, if a company that has a lot of money and reserves is hit by a hack such as this, and the hackers say, give us millions of dollars and we'll go away. We won't just close all of the data that we have gotten on your customers, et cetera, et cetera. Does the company do that? Do they refuse? And what's to make of this as a just regular person who might go to Vegas and stay at Caesars Palace type thing? I'm not totally sure. Teja, where do you fall on the idea of hackers being paid to not disclose sensitive data on users from which they have gleaned? You mean extortion? That would be another way to put it, yeah. Because the backbone for this whole thing is how the people that claim responsibility, I love when hackers come out and say things like, they're doing this for ethical reasons. You're not some blue hat hacker that was hired to be a penetration expert from MGM. So let's stop there. Not that I'll ever defend casinos, I'm not saying that, but it's just like, come on, let's call it for what it is. But it does become ethical on the end of the actual owner operators because not only they're sensitive data, but all of our sensitive data. So the thing with this latest MGM breach was that where the regular people like us would have really noticed it is like their key cards wouldn't even work. And so stuff like that started happening, which then makes you think, well, what else, what other information do they get from the system? So I'm assuming personal information, addresses, birthdays, maybe more information. But I don't know because some companies have paid these hackers off. I just feel like to stick it to them, I wouldn't want to pay them off. Not that I want my information or anyone else being doxxed or like having our information out there used for malicious purposes. But just the whole thing of it is like, you're not, you're not some good people hacking this for some good cause. It just bugs me so much. Like, no, you're not going to get your money too. Isn't it? You're not going to get your cake and eat it too. Yeah, the whole pay us and we'll go away thing is like, that's a tough, like, is that really the pill we're all swallowing? Nobody wants to do that. No company wants to do that. I feel like nobody wants to, but I can, I can see why Caesars did. They paid tens of millions. This is going to cost MGM so much more than tens of millions of dollars. So I can see why working under the rules as they are, Caesars did it the way that they did it. When these new SEC rules kick into effect the first of the year, it's a different ballgame. Would Caesars even be able to, you know, to just pay and kind of not say anything until someone else has an issue? Oh, yeah, we had one too. That's the big crux of it for me. It's like, to them, there's a lot of trust going on here that you're like, I'm just going to pay to make this go away and we're just going to pretend the whole thing never happened and I'm going to trust the hackers that they're not going to do this again when now I already know they have access to my system. And whatever else I do, they're obviously going to get access again to my system. So it's not actually solving a problem. It's like putting a band-aid on there right now and they're not solving for the future or helping anybody else that's going to be in the same boat, which as we see in this case. MGM and Caesars, by the way, combine about 60,000 hotel rooms in Las Vegas alone. There are others, but that is where the most of people who would be perhaps going to Las Vegas and have been inconvenienced would be a nice way to put it by this entire thing. People who have been to Las Vegas, I don't know, in any past years, also inconvenienced by this, depending on the data dump. So yeah, it's bad news. Like we were talking about Rob yesterday, it's more of the same of, okay, if you are a company that is this big, you're making less much money based on this many consumers. How do you have that second plan when something like this happens? All right, well, let's talk about schools. Let's get a little lighthearted here. Google is lengthening the life of Chromebooks, hoping to aid school administrators who have said the platform doesn't have enough support to be cost-effective because the support isn't that long. Specifically, Google said Thursday it plans to offer software updates for Chromebooks for up to a decade. The new policy starts next year. Chromebooks are the go-to in classrooms in the U.S., but some education software doesn't work after Google's so-called auto-update expiration date, after which if your Chromebook is unsupported, it can be used for mandatory state testing even if the hardware is still working. So schools recycle and replace them at cost. Here's what's changing. Google currently sets expiration dates based on the release date of specific models. Newer models have eight years of support and older Chromebooks have five. Starting next year, Google will support a given laptop platform which it considers having a certain combination of hardware components for 10 years after the first device the platform is available to buy. Yeah, and the change retroactively applies to any devices released since 2021. If you have a pre-2021 Chromebook, you have the option to extend the support, which is nice. So to recap, any Chromebook model that was set to expire within the next two years can receive an extension of two to three years depending on the model. If you have a Pixelbook, for example, that would be slated to expire in June of 2024. Now, it would be June of 2027. So Teja, I know you fall a lot of Google news. What do you think? Is this the right call for Google in schools? Absolutely. Not just from, you know, kind of how they've pitched it here in terms of expenses on the school side. It costs a lot to recycle devices. Also, just we don't actually know how much of the device is getting recycled. And to me, I hate waste. So even outside of the cost excess of this all, I love it from a less waste, a better usability. Like we're all just expected to update our devices all the time. And nobody really gives any second thought about how these get recycled, where the devices go, who could better use a device if a school in this case decides to upgrade. Cool. Now all of these other models anyway can be reused for people in need. Or, you know, like now it just becomes less infrastructure, less waste. I am here for this all day, every day. I'm tired of having FOMO. I'm being told to update all the time. So I definitely in agreement with Google deciding to do this because, and I've said this many times before, a slow Chromebook is better than no Chromebook. And we know that schools are scrapped for cash. But I really feel for the kid that has the 10-year-old Chromebook that's on that last update. I'm just thinking back to when my youngest was in, you know, first started high school. She had a Chromebook from the school. And it was so, it was just so slow. It was still supported, but it was just so freaking slow. It literally took minutes upon minutes for it to boot up and for her to move around in it. And we ended up just buying her a Surface Go that she could run all her Chrome apps on. And she ended up using that. So it was nowhere near 10 years old. It might have been a three or four-year-old Chromebook. When you think about these devices, they really don't have that much hardware in them. Hence keeps, you know, it keeps the prices down. So I can only imagine what a 10-year-old Chromebook runs like with, you know, that has been updated with new software. I kind of wonder, as we're talking about this, you know, what is, you Rob talking about your daughter and that extremely slow Chromebook that she had, you know, school provided, but wasn't realistic. You know, like what does a Chromebook have to do and be in order to be able to keep up with other kids at the school who might have other computers? I don't necessarily expect a Chromebook to act like a MacBook or to act like a $5,000 17-inch HP or Spectra. I don't expect that. But when I say it was slow, I mean, we would turn my, you know... And then like walk away and make dinner and come back. She'd turn it on and she'd come on from tri-practice, sit the Chromebook on the desk, turn it on, then go take a shower and then come back and it's up and running. I mean, it took minutes for it, not as long as her showers, but it took minutes for it to actually just be able to log into it and moving back and forth. So she ended up, she wouldn't turn it off. She would just leave it, you know, running as long as possible, hoping that, oh, I'm not going to have to reboot it. I'm not going to have to restart it because it just took so long for it to come up. And just hearing her complain about it, that's kind of what pushed us into getting her a new device. There's nothing as frustrating as a slow device. This is how accustomed we've come to things right now is just like imagine her workflow too. Like all you're trying to do is your assignment and you don't want to do homework all night and now you're just waiting for like Chrome or Google Drive to load. Yeah. Yes. I agree. That's all it's built for, Sarah. Like, you know, to answer your earlier question is like, that's really all a Chromebook needs to do is open Chrome and run the Google suite of apps. And do the things. That's what you're asking me to do. And let me, you know, finish my math assignment. I mean, how many times have we all have technical issues, you know, here and there, because most of us are working out of home studios in some capacity. But, you know, wherever you work, it's like if you have, if you have a delay of a couple of minutes, you know, we are now conditioned to say, this should not stand. We do not have to deal with that sort of thing. I feel like the sort of, well, all the kids in school, the Chromebooks, you know, they're just, you know, happy to have, happy to have computers in general. And I don't think that's the case. Like you were talking about Rob, it was, you know, difficult for your daughter to use the computer that she was supposedly gifted. It kind of has to work as advertised or else they're going to have to figure out something else. It does sound like Google's at least trying to add support for a few more years for some of these monitors. Well, if you have any feedback about anything that gets brought up on this show, do get in touch with us. We have lots of social accounts that you can use DTNS Show on X, formerly Twitter, Mastodon at mastodonmstdn.social, Daily Tech News Show on TikTok, and DTNS pics, PIX on Instagram and threads. All right, let's talk about some science. Scientists from Columbia University have revealed that current AI models have limitations when it comes to distinguishing nonsense from natural language inputs. So what's junk? What's real? The study tested nine AI models by submitting pairs of sentences and then asking which one would be likely to be heard in regular speech. The control group was 100 people, so small group, who were asked the same questions such as, a buyer can own a genuine product, also one versed in circumference of high school I rambled. Probably nonsense, right? Do the AI get that? Although more sophisticated models like GPT-2 did better than the simpler models, all of them exhibited blind spots, labeling some sentences as meaningful, when they weren't, that human participants thought were gibberish. Tal Golan, one of the authors of the research paper, said it was premature to allow these models to replace human decision-making in areas like law, medicine and student evaluations. Among the issues highlighted is the ability for people to intentionally manipulate models. So Teja, I don't know how much you've been playing around with LLMs, but I think this is one of the main things where I feel like people kind of do like that, like, I got you AI. You thought that was a real sentence. It was not. It's probably a good thing that we're figuring out a way for robots to know how to weed those out. Oh, 100%. I was just talking to Tristan about this, who I host AI named the show with, and we were talking about it in terms of medicine and health and vaccine development and all this kind of stuff. Like, these large language models are really good at sometimes recognizing forms of language, like, can pull out big, like, mass amounts of text quicker than you or I ever could, but it's not so great at saying whether or not that information is factual or whether or not it's misinformation. And then this is where we get into a lot of these problems in terms of how quickly misinformation can spread. So I can't say I'm surprised by this. At the same time, it's like, are we out here to just say, I got you AI? Like, everybody just needs to be aware of the limitations and then we need to start building systems and balances in place that can also get a little bit better and a little bit smarter and train these systems with what is, in this case, gibberish or misinformation, not factual stuff. And that's going to take a bit of time. But the good news is that all of this advances so quickly. So it's the bad news, but it's also the good news because hopefully we can train these models a little bit better to start recognizing nuance and forming those language structures a little bit more accurately and in ways that just make sense that aren't these sentences. You know what's interesting is that we're talking about all this and we're really talking about these language models and they've been around for a while but we've only really been using them for not even a year yet. ChatGPT hasn't even been out for a year in the way that it is out now. So it's amazing how fast we've moved and okay, yeah, I'm going to figure out how to trick this thing. We're actually trying to fool something that hasn't even been out for a year yet. That being said, they're going to get significantly better. At some point, I would imagine that all the AIs if they're going to be around would be able to figure out what is a gibberish sentence as compared to one that is not. This is what it would remind me of. I don't know if you guys have ever seen this. There's a video. I believe it is an Italian pop star singer that is singing a song in English but the English makes no sense. He's just using English words and the video went super viral and but it sounds like gibberish like this. So clearly an American listening or someone who speaks English would know. I think it is like, yeah, the Romania or something. Yeah, but I think AI is going to get to the point where it'll know as well. It'll figure it out. You know what? That's a great point. We sometimes get emails, feedback at dillitechnueshow.com, by the way, if you want to send your own where people are like, AI, like are we ever going to stop talking about this? Well, probably not because it keeps... AI is part of what many people are going to use in their lives going forward. We are in the early steps to where we... I think we get stories like this where it's like, well, the AI didn't know that I said a nonsensical thing. It just tried to do its best. That's because it's AI. It's not supposed to know that I was trying to trick it. It's supposed to help the person who's not trying to trick it. You can trick things. You can trick all sorts of things that are not human. Now, where do we go? I think what we can get out of a story like this is, okay, you might need to still think about how you phrase a query. Think about how the AI knows certain keywords but doesn't necessarily know the cadence of your speech. And that's all kind of exciting because it's exciting about what it's going to be in probably not the two distant future. But yeah, that's where we are now. So yeah, I'm really excited about AI. I kind of was on a roller coaster with it. Initially, when it first... When chat GPT first came out, I thought it was awesome. And then it's like, is this all we're going to talk about? But when you really think about what these large language models can do and not just you talking to it trying to trick it, but when it can scan your data and actually come back and help you with your thoughts. I mean, I'm using things that are AI related to actually help me with podcast show notes. It's like, that's pretty impressive. This is technology that just... It didn't exist four or five years ago for us to be able to use it the way we are today. So yeah, there's problems with it because it's still a baby. It's still... Like we haven't been talking about these things for a year yet at the way that we are now. Eventually, they're going to get better and they'll become more useful and they'll just be... They'll just be like, I'm not going to say the names, but the thing for the iPhone, the thing for the Android phones, as you say, and they answer questions for you. It's just going to be... They're going to get better and better and better over time. This is something that we've become. Yeah, I was about to ask you if there's anything that you've used lately that is definitely AI enhanced. I did that just yesterday when somebody asked me about scope of work, if you're a producer, you know all about that. And I was like, oh gosh, I don't really have anything at the ready. And the person who asked me for it was like, just do a chat, GBT, sir. And I was like, huh, yeah, okay. It's become like an unpaid assistant now. Yeah, right. Truly, like that's what I mean. I just... When I say things like we humans, we just expect so much all the time right away. And it's like as Rob is saying, like we didn't have commercial access to this at our fingertips, like he said, not even a year ago, and we're already so demanding of it. But it has become a big part of my almost daily workflow. I use a few different ones, but like chat, GBT in particular, I use to write a bunch of social posts for me because that stuff to me is just so time consuming and I don't love doing that part of the job. So it can do a pretty good job at writing posts for specific networks. I've used it, you guys, to help me write a contract, a business contract. Of course, check your work. I double-checked it all. Of course, yeah. I changed the language that fits into like video production and stuff like that. But I was kind of struggling with what this one client was asking for. And I got to say chat, GBT knocked it out of the park and I just had to make a few tweaks. And so in something that took me probably 20 to 30 minutes, as opposed to like writing something from scratch or Googling like I need a template. And then, okay, I got to build off this template or I got to figure it out. It was just, no. Yeah. There was no formatting. So it's just like, let's not be so demanding. Like I think where I want us to be demanding is that in the accuracy of it and to understand a bit better what the technology is and the limitations of it and the power of it and the risks of it and the amazingness of it and all the things. But of it itself, it's just pretty cool. I mean, could we just give it a rest? Yeah. I mean, for everybody who has an assistant in the workspace, I do not. You know, where I kind of go like, well, wouldn't that be nice? If this can be my assistant and I'm still the person who makes the final call, amazing. You know, just like you were talking about Teja, you know, putting together some sort of a contract. It's more your friend than not, sounds like. It's also good for summarizing articles. Just as a good little tip. That too. That too. Well, if you feel like, you know what? I'm out of here, guys. I need to take a vacay. You wouldn't be alone. If you're ready to hear about the best way to fly to your next destination from where you are currently, Chris Christensen has an answer for you. This is Chris Christensen from Amateur Traveler with another Tech in Travel Minute. I've got a website that I like to share with you and that's flightconnections.com. It's a simple site. All it does is help you map where you can fly from to. And so, for instance, if you just put in a single airport, you can find out where you can get direct from that airport. Or if you say two airports, you can see what routes will connect you between that airport. So, for instance, I'm interested in a trip to Malta and I was trying to decide could I go from San Jose or from San Francisco. And it was quite clear by typing in San Francisco versus San Jose that there are a lot more different possible connections. There's no direct flight between the two, but there is more than a dozen different connections from San Francisco and only one from San Jose. You can pretty much guess that San Jose is going to be more expensive and it is. The site again is flightconnections.com. It's a fun little interface. And this is Chris Christensen from Amateur Traveler. Ah, Chris. He always comes with the knowledge. I'm actually trying to figure out how to most cheaply get from where I am to New York City because it's very expensive. So this is going to be a real help for me. I don't know if all y'all have been looking at flight connections recently, but I think in the post-COVID world that we're in, if we can even say that, post a really bad COVID world, you know, the airlines are still catching up. They're making their money. Indeed. They are making their money as they continue to try to cash up. Teja Custody, thank you so much for being with us on the show today. Such a pleasure. Let folks know we're going to keep up with the rest of your work. Yeah, you can find me at TejaCustody.com. I'm at Teja Custody on all the things. You can head to YouTube and find me there as well. You can listen to Talk Techie to me or our other great awesome new podcasts that I do with my friend and fellow tech expert, Tristan Jutra, called AI Named This Show. And if you can't get enough of AI, you're going to want to listen to it. New episode drops tomorrow. And patrons, stick around for the extended show Good Day Internet. We'll be talking about Amazon's new search options that might remind you of Pinterest's model. Yeah, but just a reminder, we do the show live. You can catch it after the fact, but if you can join us live, we'd love to have you Monday through Friday at 4 p.m. Eastern, 200 UTC. Find out more at dailytechnewshow.com slash live. We'll be back doing it all again tomorrow with Will Smith joining us. And guess who else will be back tomorrow? Dot, dot, dot. This show is part of the Frog Pants Network. Get more at frogpants.com. I hope you have enjoyed this program.