 This 10th year of Daily Tech News Show is made possible by you, Brandon Brooks, Hector Bones, Tim Ashman and all of you. Coming up on DTNS, GPT-4 is here, but will you notice? Plus the space race for cell phone coverage is on and why Marvel is picking a fight with Reddit and Google. This is the Daily Tech News for Tuesday, March 14th. Hey, it's Pi Day 2023 in Los Angeles. I'm John Merri. Happy Pi Day from Studio Reddit. I'm Sarah Lane. Deeper than 314, where it's 314 Day and Pi Day simultaneously. I'm Patrick Dorton. Oh, good one. And I'm the show's producer, Roger Chang. Oh, that's extra special to be in the area code that is always, it's always Pi Day in St. Louis. I can imagine everyone in St. Louis is just like eating Pi and high-fiving and like normal, but a little extra. What's not to love? Because that's usually the way things are in St. Louis. Did you know that Sam Altman went to St. Louis, went to high school in suburban St. Louis? He went to Burroughs out in Ladoo. Let's say out in the county. Did not know that. And how do you know that? I was looking up his actual title earlier to make sure I got it and I was like, Hey, look at that. Born in Chicago though. All right, let's start with the quick hits, shall we? Microsoft assigned another 10-year deal this time with cloud gaming company Boosteroid to commit to delivering Call of Duty. Microsoft says Boosteroid has about 4 million users globally. This follows similar deals related to Call of Duty with Nintendo and Nvidia and Microsoft says more deals are coming. The company is trying to show regulators it will not prevent Call of Duty from appearing on other platforms if it's allowed to acquire Call of Duty's maker, Activision Blizzard. Find out if Boosteroid is right for you. Platform reports that when Microsoft laid off approximately 10,000 workers back in January, remember that, it eliminated an ethics and society team focused on artificial intelligence. Now, Microsoft still has its Office of Responsible AI and Microsoft says its investment in responsibility work for AI is increasing. Like, yes, we got rid of that team, but we have increased our efforts overall. Many of the former members of the ethics and society team are directly on product teams now or work inside that Office of Responsible AI. However, there is no central hub devoted to thinking about AI ethics design at Microsoft any longer. That's what platform found out. According to documents seen by Reuters, the Indian IT Ministry is considering new rules that would require smartphone makers to let users remove pre-installed apps. The rules might also require phone makers to submit major OS updates for government screening before being shipped. Meeting records show the IT Ministry met with representatives of Xiaomi, Samsung, Apple and Vivo to discuss these new proposed rules. A California State Appeals Court reversed a lower court ruling that Proposition 22 was unconstitutional. If you don't remember, Prop 22 defined gig workers for ride hailing services specifically, much to my annoyance, as a small business owner in California, as a new class of independent contractor with more rights than most independent contractors, but fewer rights than actual employees. So Prop 22, if this decision holds, would be back to being the law in California, though the court did strike down one clause in Prop 22 that made it more difficult to organize a union. Now, it may not be the end of the road, as the group that sued over Prop 22's constitutionality can still appeal it to the California Supreme Court. They haven't decided if they're going to do that yet. And whatever is decided is going to set a precedent for similar laws being pursued in other states. Meta CEO Mark Zuckerberg announced the company will cut 10,000 jobs over the next two months. You might recall that Meta let go around 11,000 employees in November, leaving an estimated 87,000 employees. Still a lot of employees, but apparently 10,000 will be cut. The company will also close hiring for 5,000 open roles and cancel low priority projects. Related to those projects, Meta's head of commerce, Stephanie Casriel, Stefan Casriel rather, announced the company will wind down support for NFTs and digital collectibles on Instagram and Facebook in order to pursue other projects. The NFTs are going to pursue solo projects. They're breaking up with them. No, the company is, but the NFTs, they're free to do that as well. Yeah, they're going to go off and make their own albums. Yeah, good luck with them. All right, that is a look at quick kits. A couple of space bits of news today, CNBC reporting that by the end of this year, SpaceX and T-Mobile will start testing using Starlink to provide satellite to cell service. So in other words, where you don't have a cell service, Starlink would be available to fill in. The idea is to end mobile dead zones. Now, they're not the only ones doing this. Amazon's Project Kuiper has a similar partnership with Verizon and Project Kuiper just showed off its customer satellite internet terminals. The 100 megabit per second one looks like a little squirrel picnic table. They all look like tables. In fact, the larger one looks like a table for a raccoon. All of the terminals use Amazon's own Prometheus chip. And of course, Apple and Qualcomm have partnerships with GlobalStar and Iridium, respectively. That's for emergency data service, though. Then there are other satellite companies out there like OneWeb, Link, and AST SpaceMobile just wants to do full-on space telephone service. Patrick, are you excited for no more dead zones? This is absolutely, utterly and totally what I would call a mixed blessing. The idea of never ever being out of range of someone calling me is abjectly, utterly and totally horrifying at some levels. On the other hand, having been in a situation where I had to drive eight miles to get to a road, to drive 10 miles to get to a town where I had to get changed and then punch change into the machine, there's part of me that's like, hey, satellite everywhere. It's cool. I'm a lot more excited about the idea of Starlink having competition just so we can see prices start to come down. I'm also not really surprised that everyone's partnering with mobile companies because mobile companies make a lot of money and satellites are really expensive to put up in space. A lot of this makes sense to me. Really, really curious to see performance, what kind of line of sight it needs, because one of the big challenges with Starlink because satellites in space is that to get optimal performance, you need to be able to see a large chunk of sky. I'm assuming it's going to be the exact same thing for him. I'm already picturing the PC MAG article that explains why T-Mobile says this isn't a dead zone, but because of the tree coverage, you would have to climb the tree to get your signal. I can see stuff like that coming. If you're on the top floor, you can get the Starlink, no problem. It's interesting. One of the things about the mobile product they put together was that it does not try to scan all of the sky. I am exaggerating. It scans less of the sky than the fixed version does, but it's really interesting to look at this because people are like, I spent a lot of time, we're full-time RVing, having this kind of internet access would have been a game changer in a way that it's really hard to explain if you haven't tried to balance having mobile coverage as your primary way of making money. The flip side is when you're in RV park, or for example, when you are camping in places that are not the desert, you often want to be under trees. Trees will not work with the satellites because of the space and the leaves and the needles interrupting the view of the satellites. He said sounding like a really competitive chat GPT-1 children's story. This is going to become more common and we're going to start complaining that we can't get service under trees when we couldn't get service in the entire forest before. That's going to happen. I think it's also interesting to point out that these companies that are putting up satellites probably don't see consumers as their main revenue source. Starlink, yeah, they've got more than a million consumers and that helps them publicly test it and get the word out. But selling back hold to T-Mobile, that's where they're going to make their money. I'm actually really curious how that's going to roll out because there's a finite number of people they can put in a cell. I'm curious. I'm very, very curious to see how that rolls out and what the financials look like in a few years. The Prometheus chip that Amazon made, apparently Amazon says, can handle thousands of customers at once and has two point-to-point microwave back hauls in it. Obviously, the squirrel size one is not for that, but the big commercial size ones, the larger than raccoon size ones, yeah. Sorry, I've been visioning the outer sunset raccoons that were the size of my dog. Yeah. Oh, you should have seen the raccoon I saw the other night. Oh, boy. They are commercial satellite internet sized raccoons out there. You know, I've got some friends who live, not super far from me, but they live in an area that's even more of a cell dead zone than where I live. My Wi-Fi is fine. When the Wi-Fi goes down, I have to drive 10 miles to the east just to be able to call my mother and be like, I'm coming to your house. Right. And they got Starlink probably about six months ago. And I was like, how is it? And they were like, well, we have internet. So it's amazing and we'll pay anything because that was the only option, at least, you know, according to them at the time. So, you know, anything like this that starts to become more norm helps people to be able to live wherever you want to live and get the, you know, all the internet. Cannot make a compromise, yeah. Exactly. And, you know, maybe prices come down because, you know, as you mentioned, Tom, there are competitors as well. I think this is the one for everybody. Yeah. If you were on T-Mobile and this was in existence and you suddenly got cell service without having to use your Wi-Fi when the Wi-Fi was down, it's huge. I could totally see that. Right. Or if you're one of the millions of places that have DSL only, it was kind of fascinating to watch certain parts of the country that were just dense enough to have a lot of people fill up their cells. You know, there are a lot of space and hexagons for Starlink. And you were like, there's not that, oh, I guess there were a lot of people there that didn't want internet that could get DSL. Or wanted something at least better than DSL in some cases. Well, I might ask you, what would the Marvel superheroes do in this situation? Fly there. That was a terrible segue. Before the release of Ant-Man and the Waff's Quantumania, some of the movie's dialogue was posted in a Google Doc and linked to from a post on Reddit. This falls under copyright violation, so under the rules of the US Digital Millennium Copyright Act, aka the DMCA, Disney issued a takedown notice and the doc was deleted by Google and the link was deleted on Reddit the same day that it was posted. That's done, right? No. Disney said, we'd like to go further with this. Marvel has requested the district court for the Northern District of California to issue a subpoena to Reddit and Google for the identities of users involved in posting or editing the content, as well as for the identities of the users of the account MSSmods, which is shared by moderators of the Marvel subreddit. Now, I'm going to start on the side of the studio here and then end up where I actually am. On the side of the studio, you want to find the person who got a hold of this. The suspicion is that they got a hold of a foreign language subtitle script and were able to get the stuff out of that. You want to find out who that is, you want to track down the source of the link or the leak, and you want to stop it. I get that. And to be able to get that person, you want to do everything in your power, which means you want to know who that person was responding to, who else helped that person. How was the communication done? Because if you find out who they were communicating with, you have a better chance of figuring out who the person was. So they're asking for a wide net. That's why they're asking for a wide net. This isn't really punitive for the mods on Reddit. It's that we want to see what the mods were saying because that'll help us figure out who the person responsible was. That is not what the DMCA is for. The DMCA is for stopping you from subverting copyright. And the DMCA was executed as soon as Google took this down. I don't think they even really should have been able to get the link removed from Reddit because the link was dead at that point. But if I'm Reddit, yeah, I'll remove the dead link if that makes you happy. Why not? Going after the identity of the Google Doc holder, I don't love but make sense. It's beyond the scope of the DMCA, but Disney has a right to go after that person. However, beyond that, I'm not sure if a subpoena should be issued. I think you're starting to say like, well, where does it end? What else would you like to just have all the server logs of everyone who was on Google that day? Because you have a chance of maybe catching the person who did it. It starts to go beyond a reasonable request at that point, at least where I stand. What do you think, Patrick? We know whenever one of these things happen, on one hand, I'm incredibly sympathetic. You have staggering amounts of money invested in the creation of this property. You need to earn your money back. Their shareholders just may have, but invariably every time I read one of these, I hear Joe Pesci playing Anthony Spilotro in Casino going, maybe it was Robert De Niro, but I want them dead. I want their families dead. I want their goats, their chickens dead. I want you to burn the house down. It's salty. I think it was Joe Pesci. Yeah, I think you're right. Sounds like him. Yeah, but you know what I mean? It's always this salted earth. We want everyone. We want to sue the cable manufacturers who provided the cable in the centers where the packets were shipped. Right. Exactly. There's never any end to it. Part of me sympathetic and part of me is like, you know, just yeah, there's so much of this that's like, I 100% understand where Disney is like, oh man, you know, like this is like such a buzzkill, right? Like, let's go after the people who have ruined the movie for everybody. The people who want to watch the movie don't care about that. I didn't see this before I saw Ant-Man, Quantum Media did not affect my enjoyment at all. Even if you could watch it, you know, it's like, you know, let's, you know, let's get real here. I'm more sympathetic. I'm less sympathetic to the copyright part of this case where I would 100% and agree with you, Sarah. It's like, it's really isn't that damaging. I'm more sympathetic with, but we want to find out who was responsible for leaking it because we want to make sure this doesn't happen again. Maybe in the future, you know, something like that would, you know, be more harmful to the studio. Yeah. You know, and Reddit's fighting this whole separate court battle with studios over who want identities of users who posted comments in a thread about piracy because that might lead them to people who are pirating, which I think is absolutely ridiculous. So yeah, studios have the wherewithal to overdo it. So they're going to overdo it because what do they have to lose? They're going to spend the same money on the case anyway. And if nothing else, they sow some fear, uncertainty, and doubt out there that might dissuade some people from trying stuff in the future. It's also a spectacular example of what you can do to get people to click on something, to download something, to hopefully infest their computer with malware. I mean, none of you should have clicked on this to trap. Yeah, right. Exactly. Even though it wasn't in this case, another time it will be. Yeah, absolutely. Folks, if you have a thought about this, are you on the side of Ant-Man in all of this? You're like, I'm not for Disney or for the piracy. I'm for Ant-Man. Send us an email feedback at dailytechnewshow.com. Well, y'all, Google seems to have a little bit of an issue with timing. It's AI news because it has a lot of its own AI news. Tuesday, Google announced new features coming to Google Workspaces. Trusted testers would get writing tools and docs and Gmail this month that can do things like compose emails for you from bullet points, summarize blocks of texts. Many of us say pretty cool. Developers are getting something as well. Google launched API access for its large language model, also known as Palm, alongside an app called Maker Suite that can help fine tune your use of this model. The company is also launching a platform called Generative AI App Builder to speed up shipping new apps like Chatbots and Digital Assistance. Now, this all comes a couple of days before Microsoft's future of work with AI, which is an event that is taking place on March 16th. Sounds like a scoop, doesn't it, Tom? Oh, yeah. Open AI released GPT4 today. Google can't catch a break. They had the whole thing with Bing and AI last month. Today, they're like, this time we've got it. Microsoft's things later this week. We're going to put out all our AI news and open AI is like, hey, we've got the next model of GPT, GPT4. And we've even got API access just like you do, Google. Developers can get on a wait list for GPT4 API access. It's going to already power Chat GPT if you subscribe. So if you're paying $20 a month for the Chat GPT plus service, your Chat GPT will be powered by GPT4. Microsoft says that Bing has been powered by GPT4 for the past five weeks. And if you're like, OK, GPT4, it's one better. What's better about it? It can accept text and images as inputs. So, for example, it could write a caption if you give it an image or you could give it an image and say, please describe what's in the image. It doesn't generate them, but it can describe them. It can react to them. Open AI says it's spent six months aligning the program to give good results on factuality, steerability, staying within guardrails. Though, again, Microsoft says it's been using it and Bing for the past five weeks. So it's not as good as staying in the guardrails there as maybe it is in this release. I don't know. In a blog post, Open AI wrote, in a casual conversation, the distinction between 3.5 and 4 can be subtle. The difference comes out when the complexity of the task reaches a sufficient threshold. GPT4 is more reliable, creative, and able to handle much more nuanced instructions than GPT3.5. But also, Open AI CEO Sam Altman wrote, it is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it. Way to sell a product. Sorry, all right. Sam. I mean, I've heard from lots of people that wait till GPT4 comes out. This is going to be the game changer. But, Patrick, it sounds like you might not notice. Okay. First of all, I have to make the obligatory Skynet joke. Wouldn't it be sad and wondrous if Skynet came out of Bing of all places? So part of what's crazy about what GPT4 can do, now this is coming straight off of the Open AI website, openai.com, when you go to the GP4 stuff. So one, much larger amounts of text, like 25,000 words of text, which is a lot. If you've ever typed 25,000 words, Tom, I know you have. That is a rather large chunk of information to process. But the thing that kind of was kind of the part that freaks me out, right? As he keeps rolling down, it's like chat GPT outperforms chat GPT by scoring at higher approximate percentiles amongst test takers. It's like, oh, that's interesting. And so they threw it against the uniform bar exam. So chat GPT was in the 10th percentile on the bar exam. This is not good. This is not a lawyer you want, although we could argue that, you know, don't use it as a lawyer. Yeah. Don't use it as a lawyer. Chat GPT4 scored 90th percentile. In the biology Olympiad, chat GPT was 31st. Chat GPT4 with vision was in the 99th percentile. Wow. So the idea is that, you know, and again, I'm going to quote this because I just like looking at the pictures and the gigantic text on the page, following the research path from GPT, GPT2, and GPT3, our deep learning approach leverages more data and more computation to create increasingly sophisticated and capable language models. So if you've been experimenting with chat GPT, you know, there's it is fascinating because I've, you know, I've mentioned this before, I've been dealing with some human powered services versus chat GPT. And in terms of things like grammar and actually having a lead paragraph, and you don't know whether it is just type in lead paragraph on Google, and you'll learn more than you ever wanted to know about how news articles and everything else is put together to grab your attention. But what's fascinating is that, you know, chat GPT can also write incredibly authoritatively and accurately and be dead wrong. So, you know, but I look at this and I'm like this, you know, it takes a significant amount of, you know, mental human processing to do the bar exam. Or I'm sure to go through a biology Olympiad, but I'm not super familiar with that. And to see this kind of performance jump, again, this is not the real world. This is a measure of accuracy, though, to your point, right? Like this means the GPT forward doesn't get things wrong as often because what I'm hearing from from you and Altman's post and what they're saying is this won't visually appear that different because it's going to sound just as authoritative behind the scenes. It actually might be right more often. Your mileage still may vary. You shouldn't rely it. I think that's why Altman's coming out thing. Don't think this thing is perfect. Don't rely on it. You still need to fact check it, but it should be better. I mean, something somebody said to me as an aside, they worked on the Watson team. This is several, several years ago and they were basically like, look, white collar jobs, they're going to get destroyed by AI just the way automation destroyed jobs, you know, in the working sector. If you want to debate me about whether or not this enhances jobs or create jobs or what it means, that's fine. You know, tweet at Patrick Norton. We'll get into a spirited debate about this. But what is happening is when you think about how many times have there have you seen processing errors with a service? How many times have you read about a court case that went off because somebody missed a technicality? And, you know, the thing about computers is when they work, they do a pretty good job of never failing. Right? You know what I mean? We can argue about that too and feel free to roast me on Twitter or anywhere else. Because we've all had cases where the computer was an exception to every rule. But the computers work more reliably than humans when they're good at something. That's the, you know, that's a big caveat. But it's an interesting idea. You know, one of the things that people are saying is, look, there's certain jobs that aren't going away. This will make you more efficient. It will make you faster. It will make you more accurate. And, you know, I got to be honest, I'm pretty sure I would get my ass kicked by the uniform bar exam right now, even if I studied for it for a few weeks. And the fact that they can throw this at it and it could be, you know, fairly successful, I'd say 90th percentile would be a score I would be very happy with. That's impressive. Yeah. And I think you're right. Overall, jobs will be newly created because of this. It's always the short term that where the problem is, right? Are you going to be like accountants and make it through just fine and even have more jobs? So you're going to be like factory workers where your job just doesn't exist. Go train for another one because they exist. And that's what you have to look out for. AI whispering. Your future career. AI whispering. I have a friend who's a prompt engineer and that's going to be a job that is already a job and it's just going to be a bigger job. Yeah. Well, Silicon Labs, a chip maker you may or may not have heard of has a new chipset, the XG27, which is small, very small in fact, energy efficient enough to go where chips have never gone before. For example, into a human mouth. Now you might say, why would you ever want to do that? Well, the socks include the BG27 and the MG27. They're both built around the ARM Cortex M33 processor with the BG27 focusing on Bluetooth, the MG27 supporting ZigBee and other protocols. Now the chipsets range from about two millimeter squared to five millimeter squared. They're very small with the BG27 currently being developed by Lura Health. That's a medical device maker. As a tooth-mounted wearable salivary diagnostic sensor that would let clinicians potentially test for more than 1000 health conditions, you could either mount that onto a tooth itself, maybe wear it inside a night guard or other retainer, something that's mouth related. But it's also very, very small, so it wouldn't be too much in your way. Lura Health says it's just finished clinical trials for the sensor with Yukon Orthodontics and will seek FDA approval. This is really cool. The ZigBee thing just stuck out to me. I'm like, what? Is it not matter compliance? Like, I want to be able to set it up easily from my tooth. But yeah, this is wild. I mean, two millimeters. That's like the sharpened tip of a pencil. Oh, teeny tiny. I mean, listen, I have braces three times. This is nothing. Put it on any one of my teeth. You'll never see it. But I was also pretty fascinated with not only is it small, so you go, okay, sure. A mouth is just one example of a place where you would want a chip to be quite small. But to be able to identify or at least test for so many health conditions, over a thousand health conditions, I knew that saliva could be helpful for some things. But I feel like this is a huge medical advance, and we're all going to have these in a few years' time. I mean, it's a matter of cost, right? Cost and insurance coverage, right? But this is just the initial use, too. The article was saying you could use it in a medical patch or as a glucose monitor or even like a wearable EKG, which would be really tiny. So it wouldn't be in your way. It doesn't use a lot of power, 0.8 volts. This is a big advance. This is a good one. I like it. Well, especially because I feel like a lot of the medical devices that aren't invasive tend to be like, let's put it under the skin and you're either weirded out by that or it's just not feasible because it's too big and will be strange. But put it in my mouth. I'm fine with it. Well, but it's small. It's small. It's like a fun little diamond, you know? You could inject this under your skin, and you wouldn't even notice. That, too. It's really tiny. It could go in a little RFID container. Well, Patrick Norton, we're not going to inject you under our skin, but we sure are going to have you back on the show early and often because you're our best friend. See what I did there? That was nice. But yeah, let folks know where they can keep up with everything else that you're doing in between appearances here. Oh, as always at Patrick Norton on the Twitter until Twitter collapses and burns, which I understand is imminent, depending on who you're following and whether or not they're trying to guilt you. It's been imminent for months now. Today, somebody actually, they were blackmailing me. I'm never going to follow you again unless you go to mastodon. And I'm like, man. Is that blackmail? Or is that just like, bye? A threat? I give you a threat. Blackmail is a strong word. You know, their heart was in the right place. That's like the person who complained that we could we cover this chat, GPT nonsense too often. I'm like, well, sorry. We do tend to cover important news. The news does itself. I have somebody who writes for a living, chat GPT and its future and the future of whether or not I will be replaced by a very small shell script is very important. But that's just me. In any case, AVXL, search forward on your favorite podcatcher, audio and video, home theater, personal audio, we cover all that stuff and or search forward on your favorite podcatcher, AVXL. Excellent. We also have two new brand new bosses to thank Joshua and Charlie just started back at us on Patreon. Thank you, Joshua. And thank you, Charlie. Oh my gosh, Joshua and Charlie, everybody. Welcome in. It's good to have you. It's good to have new bosses. You are the light of our world. You make you power this podcast. You are the 0.8 volts that power the two millimeter chip that is Daily Tech News show. Exactly. That's another way of saying you complete me. Joshua and Charlie, stick around. You're getting good day internet now. And we're going to be talking about Apple's new live over video shopping assistant. Are you tired of AI? Well, this isn't AI. This is a live person available from Apple to help you buy Apple stuff, spend more money on Apple. Do we need or want this? I don't know. We're going to talk about that. Stick around to find out. Just a reminder, you can catch DTNS live Monday through Friday at 4 p.m. Eastern. That is 200 UTC. And you can find out more at Daily Tech News show dot com slash live. Love to have you join us live. If you can, we'll be back tomorrow with Scott Johnson joining us. Don't mess it. Talk to you. This show is part of the frog pants network. Get more at frogpants.com. Diamond Club hopes you have enjoyed this program.