 Daily Tech News show is made possible by you listening right now. You great person. You maybe your Paul Boyer or Brad or Kevin Morgan or Colin, whoever you are. Thank you. On this episode of DTNS, a social network that doesn't let you type. YouTube increases its fight against ad blockers and a closer look at the disappointment of Humane AI. This is the Daily Tech News for Tuesday, April 16th, 2024 in Los Angeles. I'm Tom Merritt. And from Studio Animal House, I'm Sarah Lane. And I'm the show's producer, Roger Chang. Oh dear, my friends. I almost said you can't spell disappointment without AI because it's true. You need that A and that I. But we're going to get to, we're going to get to some more of the humanity of an all in a little bit. Shall we begin? We shall. Let's start with the quick hits. Motorola's Edge 50 series of Android smartphones come with an option for a back cover made of wood. Motorola also claims that the phones are the first to get Pantone validation for the camera. That means Pantone's color scientists check that the camera can simulate the full range of Pantone colors and skin tones as well. And the Motorola taste like wood. The independent oversight board that hears appeals on Meta's moderation decisions has agreed to hear two cases regarding images of public figures that were modified with generative models to be explicit, a.k.a. deepfakes. Facebook and Instagram prohibit nudity, but the board wants to examine the generative model side of the issue. So they're hearing the appeals anyway. In one case, an Instagram account that only posts pictures of Indian women posted a nude picture that was reported but not reviewed even after appeal. It was removed after the appeal to the oversight board. The second case is a post in a Facebook group about AI art. One post showed a man groping a nude woman who resembled a public figure from the United States. That one was taken down automatically and the poster appealed the decision to the oversight board saying it shouldn't have been taken down. The board will take public comments for two weeks and publish its decisions shortly after. The new Insta 360 video camera can shoot 360 degree video in 8k at 30 frames per second. That'll let you crop out a perspective and maintain a high resolution pan around 360 degree footage and more. It can also do 5.7k at 60 frames per second and 4k at 100 frames per second. The camera can also do non 360 degree video and 72 megapixel stills. This is a nice cam. It also comes with a removable lens guard for protecting the lens in harsh environments like snow. The Insta 360 is available now for $500. Ness. Intel will launch two special versions of its Gaudi 3 AI chip that it can sell in China without having to get that special license from the US that the US doesn't give anyone. The HL328 and the HL388 will come in June and September respectively and they'll max out at 450 watts of thermal power. That is well below the 600 and 900 watts of the versions available outside of China and supposedly that's the thing that's going to keep the performance down. The Intel chips are similar to Nvidia's H20 which is also specially made for China. Gaudi 3 has more memory capacity but slightly less bandwidth than the H20 and of course Nvidia's software is another selling point for Nvidia. Positive technologies discovered that a malicious group is spreading images that use steganography, knew I was going to get that wrong, to hide malicious code inside. It's a great word though. The researchers call it stegano armor and have identified 320 attacks using it mostly focused in Latin American countries. The attacks are conducted by email with attachments that exploit a flaw in Microsoft Office's equation editor that was patched in 2017. If the system is unpatched, the exploit downloads a visual basic script that then loads a JPEG with a base 64 encoded payload which then compromises the machine. So patch that, patch that machine. Yeah, you do. YouTube is extending its efforts to stop users with ad blockers from viewing YouTube videos. YouTube already disables videos when it detects a viewer is using an ad blocker themselves like in their web browser. This new extension of the policy targets apps that use YouTube's API to display YouTube videos. A lot of apps do this. If you ever watched a YouTube video in another app, they're probably using that API. And some apps use the API to let users view YouTube videos without ads. YouTube will now block those apps access to the API when it detects that behavior and display the message. The following content is not available on this app or it might just buffer forever. YouTube said that like you'll either get this message or it just won't work. It'll just buffer and you'll never know. One popular app called Adguard did say that it does not use YouTube's API. It has another way to play YouTube videos in its app. So it is unaffected. I want you all in chat to get mad. Tell me why YouTube is bad for doing this because people are going to get mad about this. And I'm curious why because Sarah, do you have a right to be mad about YouTube stopping you from not from seeing their stuff without ads? I don't not really. I mean, okay. So YouTube advertising business. That's how not only YouTube makes money, but how creators make money. Now you could argue that creators should make more money than they do if they're on YouTube's platform, but that's a different conversation. That is how the money machine works. It's kind of like saying, well, you know, I can't share passwords on my Netflix account with my friend anymore. And I'm mad about that. It's like, well, yeah, you had a good run. Maybe there'll be another solution for stripping ads out of YouTube videos that doesn't require you to pay for YouTube premium, which is an option, by the way. If you hate ads, a lot of people do that. In fact, YouTube was just touting its YouTube premium subscriber numbers recently. A lot of people say, yeah, I want the ad free experience. Great. You have an option. But you know, to get the ad free experience for free, it's like, you know, you knew that ship was going to sail eventually. Yeah. Especially because they let you pay for the ad free experience. Now, maybe you don't want to pay. Maybe you can't afford to pay. Those are fair. In which case, they have the free version that has ads. In fact, Friday, we're going to talk with Nate Langston from Bloomberg about the European approach to this where Facebook has said, all right, we won't track you for ads if you pay to use the service. Or you can use it for free and be tracked for ads. Otherwise, we can't give you the service because we got to make money on it. That's a little finer point than this. YouTube isn't even talking specifically about tracking, although they do track you when they give you the ads. They're just saying, look, you don't want to watch the ad, then we don't want you to see the video. I get being frustrated. I get being disappointed. I get being annoyed. But I don't get thinking that YouTube is somehow wrong or evil for doing this. Granted, they are overly dominant in this space. But even if there was more competition for online video, which you could argue TikTok and Instagram or online competition, so Snapchat. But even if there was, what would the competition do? Let you use their product for free without making any money on you in any possible way. I guess it could be freemium where you get some videos for free and you pay if you watch more than a couple hours. There's other business models, but advertising is pretty normal. Going after people who are essentially trying to get the service without paying for it, it seems reasonable for YouTube. I'm not trying to say they're wonderful or knights in shining armor, but this is what I would expect them to do. This is what I would expect any business to do. I know that this is focused at least in this certain way of a lot of mobile apps. I can't think of any mobile app I'm using that would allow me to watch YouTube videos that strip them of ads, probably just because I haven't gone looking for one specifically. I don't know what AdGuard is doing to get around YouTube's API and maybe YouTube doesn't care. I think probably what happened is the company didn't care until it had to care because too many people were using these apps. A few people here and there getting around the system and being like, yeah, free stuff. YouTube doesn't care about that. It's when too many people start to do the same thing that we had in browsers, ad blockers and browsers. It's like the companies get hip to the game and they figure out a way around it so that that extension doesn't work anymore to ad block stuff. It's a little bit different on mobile, of course, because you're talking about apps where YouTube videos live inside. I find it hard to say, wow, YouTube, what a greedy machine already was. That's what it's here for. This isn't why it's greedy. I guess I'm with you because Ken Otafian makes a good point. He says part of the issue is YouTube has recently cranked up the number of ads it runs, and I could see that being a friction point where the response should be, well, there's too many ads I'm going to stop watching YouTube, but maybe there's nothing else for you to watch in that category. Maybe you really want to watch that particular creator and you want to support that creator. Maybe you have other ways of supporting them through their Patreon or giving them chats, super chats, things like that, and you just can't handle the ads because there's too many. That's the best I've seen, well, there isn't a lot of choice. Nick with a C says there's not a lot of places to go, and they've put so many ads, it's become hard to watch. If you really can't support, can't afford paying for the service to get the ad free, then yeah, you are in a pickle there, but honestly, I don't think this is the thing, like you said, Sarah, that makes YouTube to be evil. And frankly, if the economics were better for YouTube right now, economics are really bad for media and entertainment right now. If they were better, they wouldn't be bothering spending the money to go fight this battle because they don't want to spend money if they don't have to. Right. Well, software engineer Benjamin Sandowski published a post on Monday about that humane AI pin and it's how should we say less than stellar reception from early reviewers. I know you and Nika talked about this a little bit on GDI yesterday, Tom, but let's take a closer look at some of the variables that went into making this product. Now humane raised $230 million and spent six years to ship what is now an AI pin that reviewers have, which you wear on your body, has a projector to beam content onto your hand, does a lot of stuff that's AI focused and it runs you $700. Now Sandowski goes into a lot of detail about the venture capital part of how humane got to this point. Excellent read, but let's focus more on what we know about the product itself and where we are now. So in May of 2023, and we talked about this on the show at the time, humane co-founder and former Apple employee Imran Chowdhury delivered a TED talk with a parent. Some people say this was a rigged demo that focused on AI and how the AI pin was going to replace the smartphone. It was very, it was bold. It looked cool. People were like, all right, can't wait to see this. Well, users now complain now that they have it. And again, early reviewers have had it for a little bit of time. Complain has way fewer features than that demo promised. Too heavy, runs too hot, terrible battery life. You're supposed to be expected to swap out batteries every few hours. Looks weird on a lot of different types of clothing. Projector can't handle sunlight or even really daylight at all. Input system sluggish. Uses the cloud too much. The list goes on. One reviewer who got a lot of attention, Marquis Brownlee, a very, very popular YouTuber, 18 million YouTube followers posted a review, which if you watch it about 25 minutes long, pretty fair and balanced, but it was, it was, it was critical. The title was the worst product I've ever reviewed for now. Now he does in his review. And I think this, I think this is a good way to look at this, say this product right now is not good. But if the company can fix the problems that the product has right now, it could be pretty cool. They're just nowhere near that yet. And that's where you, you get into this weird gray area of, well, how do you convince somebody who thinks the product is terrible now that they just have to, you know, let you take another couple of years to make it really cool. Yeah, I liked Sandowski's note in his column that the founders being former Apple employees may have made it harder for them to do this properly. So Sandowski's argument is, normally what a startup does is they make a version of the product, they get it out there, it fails in some ways, but it proves that there's a market and then they iterate and improve the product and add features as they go along. And you build interest and then you build investors, etc, etc, etc. Apple doesn't do that. Apple is very famous for polishing and polishing and polishing and polishing and being behind the curve and putting out smart watches, smartphones, tablets after other companies have put them out because they've been working hard to present something that they feel is super functional. The problem in humane's situation that Sandowski pointed out is there's no established product to imitate, right? There's no previous smart pin for them to look at and go, okay, let's learn from their mistakes. There's no evidence that people wanted this because there is no product like that and humane didn't seem to understand the user mindset. And so Sandowski says they should have launched a less polished product earlier rather than launching what is a fully developed product with lots of possible capabilities that doesn't work very well, right? It's not fully developed in the fact that it works well, it's fully developed in that it's got a lot of things in it that it's supposed to be able to do and it can't do most of them. To Sandowski also makes another note that I thought was interesting that humane launched in 2018. 2018 was a time where some people were already pretty worried about it, but you started to get a lot of chatter of people saying, we're on our phones too much, is this routing our brains, especially the kids, and that between then and now, Apple's screen time is a great way to monitor your internet habits and be like, hmm, maybe I'm overdoing it, let's dial back next week. You have a variety of tools, social networks trying to help you not be on Instagram too much type thing. And at the same time over the last couple years, everything has become AI focused. And the company may have been in sort of a tight spot where this whole like ditch the phone, you'll never need it again. That's a hard sell. That's a really hard sell because everyone goes, but it's great. My phone is great. It does everything. So why would I ditch it? You know, for something that doesn't even tether to the phone, and you could say the same thing about, I mean, my Apple watch, it does a lot of things really well, but it's not a replacement for my phone, but talks to my phone. So they work in tandem really well for me. And it doesn't, they don't even both have to be Apple products. I mean, they're, you know, I could have a variety of different watches. So the whole wearable thing I think is, I think it's bold and I think it's cool. And I wanted to think it was really cool. When I saw the, the TED Talk demo, I thought, well, this seems weird, but you know, reach for the stars, everybody. So, you know, it's, it's innovation often does fail. I think the, I think in theory it's cool, but yeah, this is not a shippable product. I, I think my theory is most of the problems seem to be hardware or I'm sorry, most of the problems seem to be software. I have seen very few critiques of the actual hardware, the interface, the ability to tap and all of that. It's been in the response time and what features actually work and software is developable, developable. So I agree with Marquez so far in that what he's saying is, Hey, there's a lot of fixable problems here. They should try it again. Do you think they can, do you think the world will give them a second chance? I tend to think we overestimate how much people punish products for failing. Witness Samsung. People said they'll never sell a phone again after the exploding notes and Samsung's fine. Samsung's leading the world in, in market share. So, you know, humane's not Samsung though. This is a lot more negative press for a company that has built up no good will. Can they survive? Yeah. I mean, it's, if I were a betting man, I'd say, Nope, I don't think so. And again, you know, to all the points that you were making, it's, it's to, yeah, to, to, to have a lot of people expecting something and it feels so half baked. And yes, the, they raise so much money is a part of this conversation because you kind of go like, What was everybody giving them money for, you know, with no product, you know, to speak of just sort of a vision, but you know, they wouldn't be the first company to, to have a lot of investment and then have a product that doesn't land. Yeah. Well, from former Apple employees to current Apple employees, what are they doing these days, Sarah? Well, hopefully they're listening to Apple vision show that myself and Eileen Rivera host every week. In fact, we just recorded our 11th episode yesterday and we're having a really good time. We like to think of it as an Apple show with a lifestyle twist. You might be a power user. You might just love Apple products or you might be somebody who wants to feel more productive and talk with like-minded folks. If so, we hope you give our show a try, applevisionshow.com and smash that subscribe or like. On Monday this week, my friend Patrick Beja sent me an invite to something called air chat. It's essentially Twitter, but instead of typing your post, you record your post with your voice and they do video too. An algorithm does a decent job of taking what you say and turning it into text. So you see a typical timeline of written posts, but as you browse, some of them play in the original voice, the original recording. If you don't actively scroll, in fact, it'll just play through them unless you press the play and pause button. There's a video button too, although few seem to be using that. And a lot of folks on air chat are comparing it to podcasting, which I think is a bit of a stretch, but it is a little bit like clubhouse, which is kind of like podcasting. So it's somewhere on that spectrum somewhere. Air chat co-founder, Naval Ravikant had to cut off new signups temporarily over the weekend because Silicon Valley noticed air chat and finally got excited about it and overloaded it. Ravikant joined the project a year ago, which was started under a different name with a slightly different focus by Tinder's former chief product officer, Brian Norgaard last spring. Ravikant told Wired that air chat will not sell the voice data. It will not try to use your data to make a clone of you. They did not say whether they would use it to improve their own transcription model, but I think it would be reasonable to let them do that. Just tell us in the terms. Sarah, you got the buzz from this too over the weekend. What do you think of air chat so far? It's weird. It's not humane AI pin weird, but it's different kind of weird. Yeah. Okay. So I was sort of expecting a clubhouse. Let's resurrect clubhouse and call it something different with fancy wrapping. Very much not. So clubhouse, what I didn't like about clubhouse, I know a lot of people did. It was trendy and buzzy, but it was a place that a lot of people really did hang out was you've got these curated conversations where maybe you have one person talking, maybe you have five people talking, and there's bunch of people that are essentially in a hall listening to people on a stage talk about whatever. You can maybe take audience questions, but it was all around conversations. It was like a conference panel discussion in clubhouse. The audio was often kind of crappy depending on what mic, I guess somebody was feeding into it. I don't know what kind of encoding process was going on there, but anyway, clubhouse was not for me. It just felt to kind of like, I don't have time for this meeting of sorts right now. What I like about air chat is that it really, when I first, because you invited me, Tom, when I first loaded it up, you know, I followed a few people and I get it just people I know and I was kind of like, why are all these other people in my feed that I haven't followed? And I realized, and this is, you know, maybe they'll change this later as they have more people that are just on the platform. But if I follow someone that follows someone else, that person is in my feed because air chat is still relatively small right now. I don't have a problem with that. In fact, some of the people that I wasn't following, I know. So it wasn't like I wasn't following them on purpose. I just hadn't. And to hear the voices, some of those voices are familiar to me because I've heard their voices before. I go, this is interesting, especially because I'm reading along because they all have transcriptions. And somehow that makes me pay more attention. You know, it's kind of like having subtitles on when you're watching a movie. You don't necessarily need the subtitles if it's in the language that you understand. But sometimes it like gives more of an imprint on the information. I couldn't figure out how to get video on for myself. It's grayed out and I just couldn't figure it out. It's, I don't know, it's noisy. And I feel like this is an app that maybe if I really got a lot of joy out of hearing audio stuff, rather than reading, I would open up occasionally. But it's too noisy for me, I think, to, to, to replace something like Twitter or even threads. Yeah. It reminds me a little bit of B Real in that it's got a unique take that leans into authenticity. So I like that when I see a post from Sarah or I see a post from Patrick Beja, I hear their voice. That's, that's personal. That is one of the appeals of podcasting is people like to hear me or Sarah or whoever they're listening to talk to them, right? It's a very personal relationship. So it brings some of that to social media. And I think that that has the chance to change how we react to social media because instead of the cold text and the problems in tone that that brings, you can actually hear the tone. You can hear, were they saying this with a smile? Were they saying this sadly? Were they saying this jokingly? Were they saying it angrily? You can actually hear that. So it has the ability to change the way people react to posts, which I think is very interesting. I think that it is fun to try. And I think that's why Silicon Valley is all excited about it right now. I do think that it's a little inconvenient for old people like me. I know that younger users don't tend to have this problem, but playing audio or speaking out loud or something that I don't like to do in places where there's other people around, right? So it's going to be harder for me to post on an airplane, you know, which is one of the things I do on on X and threads and stuff when I'm sitting on an airplane waiting for it to take off, right? Before you have to turn the phone off, I'll just read and post. Probably not going to do that with air chat. And then I'm not going to want to listen in those cases. But maybe that's just me. I don't know. It feels like it's got something interesting. I guess the big question then, Sarah, is where do they go to make this succeed, right? Will enough people use it is one question. But if they do, how do they monetize it without using my data? Yeah, I don't know. That's a good question. Because, because really it's if you, if you want to be around a bunch of people who are either like minded or you respect or, you know, whatever, then there are already social networks for that. You might have a specific reason that you don't like one where you prefer one over another, but this is not reinventing the social network at all. It is adding a layer that you don't really, you've not seen before, at least in this specific kind of user experience. And that's cool. But I think it is a pretty hard sell to say this would replace your favorite social network. You know, maybe images get added, but then it starts to feel like Instagram. It's like, I find, I find apps like this to be in a tough spot because it's like, Hey, we're a social network, but instead of text, we've added the audio layer, but it's still text. You know, you're still able to read stuff and even, you know, turn audio off if you want, although the transcriptions, like you said, Tom, might lack something, you know, if somebody was laughing or being jokey, the context might be lost. But if you start packing it with other features, I think that's where sometimes these apps lose their way. And is it just a, is it just a feature, not a service? And does threads, acts, you know, do they steal it and say like, Hey, let's make that an option, you know, and then, then, and then they get mirror-catted. Yeah. Or like, you know, when Clubhouse was big, Twitter Spaces launched. Yeah. And Twitter Spaces, I think it still exists, but you don't hear a lot of people using it anymore. Right. Yeah. I don't know. Folks, if you're interested in Air Chat and you get an invite, you have to have somebody's phone number or I'd be like, Hey, let me give out a couple of invites to listeners. And, but I don't feel like it's, it's fair to ask everyone to send me their phone number in order to give them an invite. So I will, I will try to spread some of the love as far as possible. But if you do get an invite, let us know what you think. Feedback at DailyTechNewShow.com. Speaking of which, let's check out the mailbag. Gabby writes in and says, Amazon Fresh Stores. Now they made people believe that AI was miraculously handling all of the checkouts when it was really a bunch of people overseas studying cameras and making sure all of the items you took were accounted for. Often it took forever to get receipts because it was really people, not computers during the work. I'm just a bit disillusioned with the tech world at the moment. Gabby had a lot of other reasons she's disillusioned in the email, but we're focusing on this particular one for the show. I think this story gets misunderstood a lot. They were not lying about using AI and computer vision. They were using humans to review what the AI determined. So the AI would flag things of like, hey, we think you picked this one up, did he? Can you review it? So it's like moderation, right? The AI, the algorithm goes so far, and then humans review it to make sure it got it right. And I think a lot of people took this as like, oh, they were never using AI. It was just humans going through and doing it all. That would have taken even longer than it took for you to get your receipt. What was happening was the AI was tallying everything up, flagging a couple of things, and then people were looking over that and signing off on it. That is to me not as bad as they weren't using AI at all, but to Gabby's point, what turned out may not be as bad as Amazon lying about it, but it turned out that that technology just never got good enough for Amazon to feel comfortable saying, yeah, it's getting better. We don't have to review as many of them. We can speed up the receipts. And that seems to be the reason that they have decided not to continue to roll it out, because they are not seeing the improvements with it that they would like. Yeah. Yeah. I mean, I'm sure there was some sort of a timeline like, okay, we've got these many reviewers, after which point, the system will feel confident, and maybe the company just didn't feel confident. Yeah. Yeah. There was just too high of an error correction rate. And so it didn't work. I feel like that's less bad than we said it was AI, but it never was, because it was. It just didn't work as well as they wanted it to work, and they thought they could get it to improve. On the one hand, we say humans should always use AI to assist and not to replace. So I'm not going to get mad at Amazon doing the right thing and having humans review the AI decisions in this particular case. Patrons, stick around for the extended show Good Day Internet. TCL made a soap opera using AI that they intends to stream on its TCL TVs. Oh my gosh, you have to see this trailer. We're going to discuss. Stick around. You can catch our show live Monday through Friday at 4 p.m. Eastern. That's 2100 UTC. And you can find out more at dailytechnewshow.com slash live. We're back doing it all again tomorrow with Scott Johnson joining us. Talk to you then.