 Coming up on DTNS, the saga of YouTube DL is way more autonomous driving safe and why you should not fear of the skinless Disney robot. This is the Daily Tech News for Monday, November 2nd, 2020 in Los Angeles. I'm Tom Merritt and from Studio Redwood. I'm Sarah Lane and I'm producing the show today. Amos, Amos is stepping in for Roger Chang, who has the week off because he's barnstorming the United States. I don't know something like that. We were just talking about watching TV, not only shows we like, but like the services we use. If you want that wider conversation, get that expanded show, Good Day Internet. Become a member of Patreon.com slash DTNS to get that. Let's start with a few tech things you should know. Twitter announced that it will consider U.S. election results official when they're confirmed by state election authorities or confirmed by two of the following outlets, ABC News, Associated Press, CBS News, CNN, Decision Desk HQ, Fox News and NBC News. Twitter will add labels to tweets that claim victory before this confirmation. The Raspberry Pi Foundation announced the Raspberry Pi 400, which is basically a keyboard that has a Raspberry Pi 4 system on a chip inside of it. The Raspberry Pi 400 cost you 70 bucks for just the keyboard, with the bundle of course, the chip inside, or $100 if you want to add in a mouse, power supply, micro SD card, HDMI cable and a beginner's guide. It's available in the U.K., U.S., France and Germany as we speak, coming to Italy and Spain next week with releases planned in India, Australia and New Zealand by the end of the year. India's open architecture with payment infrastructure, unified payments interface, or UPI, surpassed 2 billion transactions in October, doubling on the year. Walmart's phone PE, payment app had the most volume in the month, beating Google Pay's year-long lead with 835 million transactions versus 820 million. Paytm and Amazon Pay rounded out the top four with 245 million and 125 million respectively. The Financial Times sources say Huawei plans to build a dedicated chip factory in Shanghai to make components for its telecom infrastructure business. The factory will launch using a 45 nanometer process, that's right, 45. The factory will eventually scale down over the years to 20 nanometers by late 2022. Combined with existing stockpiles of chips, this should let Huawei build back its domestic telecom business with minimal disruption. Well, Tim Cook, he teased it and Apple has officially announced it will hold its first fall third, rather, fall event. I know how many are there going to be on November 10th at 10 a.m. Pacific Standard Time. The event announcement is titled, One More Thing. Apple had previously said it would release Apple Silicon based Macs by the end of 2020. So maybe this is about that. That's probably what it's going to be about. And maybe we'll finally get a date that Fitness Plus launches, something like that. All right, let's talk a little more about coughing. Coughing sends a lot of people into fleeing these days. MIT scientists have published a paper in the IEEE Journal of Engineering in Medicine and Biology, describing an algorithm they developed that can identify whether you have COVID-19 by the sound of your cough. That's even true if you're asymptomatic. In other words, if I mean a cough is a symptom, but you don't have the classic symptoms of COVID-19 because a cough could be a symptom of anything. COVID changes the sound you produce even when you're asymptomatic. In testing, the algorithm was 98.5% accurate on patients with a positive COVID-19 test. So they were able to use coughing to detect 98.5% of people who were definitely positive with COVID-19 and 100% accurate for those with no other symptoms. The algorithm was trained on a data set of 70,000 audio samples with multiple coughs, 2,500 of which were from confirmed COVID-19 cases. So that's how the algorithm was able to go, OK, that's somebody who doesn't have it. That's somebody who does and figure out the patterns that it would listen for. The scientists hoped to get regulatory approval to use it as a way to take quick, non-invasive daily screenings and for pool testing to quickly detect outbreaks in groups, pool meaning like a group of people, a test of a bunch of people at once. Cambridge University, Carnegie Mellon University, and a UK health startup called Novoic are all working on similar projects. So it's not just MIT, but they're the most recent to publish a paper on it. I mean, this is great. Anybody who's gone through getting a COVID test, well, it depends on where you live, of course. But it's not just sort of like, hey, I want to test. And you go, OK, meep. And then you are like, oh, you're fine, whatever. Over the weekend on Saturday, I actually had to go for an optometry test. And so it was just nothing major. But when I walked in, they said, you know, can we take your temperature? And they kind of point something at my head and I said, yeah, of course. And they go, OK, your temperature is fine. And, you know, at the time, you know, I always think, well, I'm glad that I don't have a fever. You know, I'd be nice if you told me that I did. But I'd probably feel crappy otherwise. But if I had, if there was any easy way to be able to tell me, hey, you don't feel like you're sick, you're asymptomatic, you don't think you're going to be infecting anybody. But if for any reason you were going to, you know, let's keep you out of the hospital, I would do that. There are so many things that need this kind of rapid testing, too. And granted, this isn't the be all end all. If if this thing says you've you're probably you probably have COVID, you probably do is the way that I'm reading the paper. And so that then they can do like a confirmatory test to make sure. Or in the case of going into the salon, just be like, yeah, sorry, come back when you've got a negative test. It would let some people who are positive slip through. So you wouldn't want it to be the only thing you rely on. But like you say that the COVID test that we're doing right now, they aren't perfect either, as far as positivity. And they they take a while. I look at all the sports leagues who are relying on tests that can take a day to get results back. And in the meantime, you've got somebody wandering around infecting their teammates. If you could do this cough test and be like, you know what, hold on, we're going to keep you off the field today while we confirm a test that could be really useful. And when you mentioned the fever thing, I know they're using that as a proxy to these days. But this week in science was talking last week about the fact that the temperature, the average temperature of humans is lower than the normal 98.6 that we think. So somebody could technically have a little bit of a fever and still appear 98.6 normal. Yeah, in fact, when they took my temperature, I said, just out of curiosity, like what was it? And they were like, it's like 97.8. And I was like, oh, gosh, am I a reptile? You know, they're like, hey, you know, it's a lot of people are lower than that. Don't worry about it. You don't have a fever. But yeah, it's I think that there's so much lost in right right now because there's, you know, everybody is very you know, waiting with with bated breath for better testing, vaccinations, you know, wondering, you know, are we all you know, when is life going to go back to normal? Right. That's what everybody wants life to do. And something like this can help in so many cases where you really feel fine. But yes, they're your cough can can can give it away. And, you know, as a responsible adult, which I know everybody is here, you want to know. So you keep the people that are vulnerable safe. You want to get into the bar, turn your head and cough. That's right. The Internet Archive announced that web pages archived in the way back machine that had been found to be part of misinformation campaigns will have banners added noting that. So saying, you know, you can read it if you want, but we've we've we've noted this. The Internet Archive will rely on fact check.org. Check your fact. Lead stories. Politifact. Washington Post fact checker. AP News fact check. USA Today fact check. Graphica and Stanford Internet Observatory plus our dot news to identify those pages. The content of the archived post will not be changed. So they they're not going to change that what you're seeing, they're just going to say you might want to think twice about what you're reading. A thin yellow header will be added between the normal web app machine header and then beginning of the archive page. And that will link to an article for additional context. Yeah, this is interesting because it points out one of the arguments that Twitter and Facebook and others have been making with moderation is historical record. And a lot of people have said you're just trying to get out of having to be responsible for what's on your platform. With archive.org, I think it becomes clear that, no, their only purpose is historical record. So there is a reason to say, like, yes, this misinformation was posted at this place. We want to have a record of that so that you can go back and look and say, OK, what, who posted it? Why was it posted? How long was it up? Those are important for historical research. But archive.org isn't trying to say that that they shouldn't point out if it is generally accepted that, like, hey, this wasn't actually a factual article. And here's some resources if you want to check check that out. I I can see people objecting and saying it's a historical record. It's a historical record. You should be objective and you shouldn't, you know, weigh in on fact checking or not. But in this day and age, I think a lot of people might try to use the way back machine to spread some misinformation in the archive. Doc.org is just trying to protect itself. Yeah, yeah, I think it's, you know, this is this is the reality that we live in right now. Also, there are a lot of historical record scenarios I can think of that people kind of went back and said, actually, hmm, new light has been shed on the story. And it's not necessarily what we thought it was for the last 100 years or so. So, you know, it's it's this doesn't seem like it doesn't seem like a bad thing to me, especially because no actual content is being doctored. Eventually, I would like us to get to a point out as a society where we could just say, oh, well, it's common knowledge that when you're looking at archive.org, that that stuff may or may not be right, that it's just preserved. It's it's not necessary. You have to find other signals about whether it's reliable, but maybe we're not there yet. Oh, this one feels like a throwback. I feel like I'm I'm 35 again. It's the DMCA. It's the RIAA in a tech story. YouTube DL, you probably have heard of it if you know what the story is about. It's a tool that lets you download video from YouTube for offline viewing and use. It's a streamer. There are all kinds of reasons you would use a streamer. We used to use a streamer to get our own video off of YouTube faster than YouTube would let us download it sometimes. And until recently, YouTube DL was hosted on GitHub. It's an open source platform, and it was hosting its code on GitHub. Now, it's also important to understand that the US Digital Millennium Copyright Act says that, and I'm just going to quote you the act, no person shall manufacture, import, offer to the public, provide or otherwise traffic in any technology, product, service, device, component or part thereof that is primarily designed or produced for the purpose of circumventing a technological measure that effectively controls access to a work protected under this title. In other words, it is illegal to make tech primarily meant to circumvent copyright protection. That's what that means. So on October 23rd, the RIAA filed a copyright complaint with GitHub requesting it take down the YouTube DL project. The RIAA claimed that YouTube DL is primarily designed and marketed for circumventing YouTube's technological measures. They pointed to an example in the code where they say, like, here's some things you can do with YouTube DL and it used a copyrighted song title in the example. And that's what the RIAA is so concerned about. They're like, people are using this to rip music and download it themselves without having to pay a streaming service for it. Now, the DMCA requires GitHub to respond to that take down request. They can't decide whether they think it's true or not without exposing themselves to liability. So it is normal for a platform to say, like, hey, man, we got the request. We have to take it down. The that action immediately caused an uptick in searches for YouTube DL on Google. Torrent Freak has a great graph of like the hockey stick of YouTube DL getting taken down. And suddenly everyone decided to go, what is this thing that got taken down? Many users began to mirror and clone the source code. So distribution rose as well. The developers also continue to host the source code from their own website. The RIAA contacted the website and said, you need to take this down. But at the time they contacted them, the source code wasn't hosted there yet. So it hasn't been taken down from the YouTube DL website. A lot of people forked the code and uploaded it back to GitHub to saying, this isn't YouTube DL. This is my own fork of YouTube DL. In fact, a Twitter user got really interested and created images that have the source code encoded on it so that you can just share it on Twitter and Instagram and wherever you want, Galactic Furball is the Twitter user who did that. Now the CEO of GitHub, Nat Friedman, has reached out to YouTube DL's developers to try to help them defeat the DMCA claim. His hands are tied about what GitHub can do in response, but he's going to try to help them file counterclaims, et cetera. Friedman suggested to Torrent Freak that YouTube DL could be restored easily if it removed the examples from its code that showed a copyrighted song, and also potentially removing a particular piece of code that circumvets something called the YouTube Rolling Cipher that was cited in the complaint by the RIAA. However, Torrent Freak has shown that the Rolling Cipher can be circumvented using your browser. Just inspect the code in Firefox or Chrome, and you don't need any stream ripping tools. You can get the audio out of any YouTube stream yourself. So are we going to go after browsers now, because it's capable of doing that? GitHub did have to warn its users that reposting the exact same code is a violation of GitHub's DMCA policy. They're like, look, we're trying to help these guys, but we are going to have to enforce this if you just repost the same code on GitHub. Yeah, this does feel like a throwback story, doesn't it? The game of whack-a-mole, I mean, the RIAA isn't going to win this one. You know, if there's a way to, unless, I don't know, YouTube did a big platform redesign where somehow this type of thing wasn't possible anymore, like you said, sure, there are reasons that people are doing this for, you know, and it would be technically illegal. There are also reasons that content creators use tools like this. I used to use a tool like this at a previous job to be able to use what, you know, we were very, very convinced was kind of news fair use purposes. But again, there wasn't a great way in the YouTube platform itself to be able to get videos downloaded and be able to, you know, put them into Premiere Pro and that sort of thing. So I don't, I mean, sure, yeah, I mean, go after browsers next, I suppose. But I don't, I just, I don't see why this is anything more than kind of grandstanding. Before, when I respond with my thoughts, let me pull out my here. Here he goes. What's Tom got this? What? Oh, these are my arguments from twenty twenty ten against the RAA doing similar things here. OK, yeah. Turns out that stream ripping in this case probably isn't a big threat to your bottom line. Let me see what else I've got in here. Oh, going after people. Well, back then it was, you know, suing individuals just hurts your cause and makes people more likely to try to infringe and you you've Streisand affected YouTube DL and now way more people know about it than if you had just left it alone. So yeah, I think I don't I don't think the I think the RIA has a legal case here. I think they can say like YouTube DL had some stuff in its in its text that implied it should be used for piracy. And that's in the that's in the law that if you're primarily marketing this thing for that, that it is a violation. It's the law doesn't say that the existence of this is a violation. The law says if it's primarily made for circumvention, that's what YouTube DL got in trouble is they left the door open for the RIA to make that argument. But that all that said, I don't think the RIA made a good move in going after this tool because I doubt it was harming them really that much. And it's certainly now way more people's radar than it used to be. Yeah, it seems like it was like a marketing thing that, you know, ruffled some feathers and they were like, you know what, we're going to take down this tool. And now a lot of more people are know about it and use it. So well, last week in our discord, there was a great conversation about YouTube DL from a bunch of people who were either fans of it or interested in it. And I actually watched that and I was like, y'all are making all the arguments I would make in your conversation. It was really, really fun to watch. So you get those kinds of conversations in our discord. If you become a member at patreon.com slash D T N S. All right, everybody, let's talk autonomous vehicles, shall we? Waymo released two reports detailing the data that it gathered in its autonomous car business, as well as outlining safety methodologies. The reports look at data from all of 2019 and the first nine months of this year 2020. In 2019, Waymo's 300 vehicle fleet logged 6.1 million miles with on board human safety drivers and a 100 square mile service area around Chandler, Gilbert, Mesa and Tempe, Arizona, as well as 65,000 fully autonomous miles. In that time, Waymo vehicles were involved in 18 crashes and 29 events. Waymo counts events where safety drivers took over and avoided accidents. It simulates what would have happened otherwise and uses the result of the simulation to improve the system. Of the actual crashes, 14 were from other drivers rear ending Waymo cars. So, you know, if it was two people, person in the back is in the is in the wrong and simulated crashes, the most common events were lane changes with another car and train a Waymo vehicles lane. Other drivers contributed to eight severe or potentially potentially severe incidents. Three of those occurred in real life. One event involved a Waymo vehicle T boned by another vehicle going 40 miles per hour and only three actual crashes was an airbag deployed in any vehicle involved with Waymo saying there was no serious injuries from incidents. Only one crash occurred with a passenger aboard the car and a four mile per hour rear ending with a human safety driver aboard the car as well. Yeah, the ones where where the autonomous car were at fault were very minor. They were low speed crashes and they were mostly simulated. They were mostly avoided by the human drivers. I think that's a big take away that I had from reading this. And the other thing is that that the ones that were not avoided but were the fault of other drivers weren't serious. They were able to avoid serious accidents that T bone that you know, sounds edgy. But you know, I guess they lucked out on that one and were able to get away. But again, that could happen to you if you're not in an autonomous car, you're just going through an intersection. Somebody runs a red light. Happens all the time. Right. So I think that's why Waymo decided to be so transparent here is this report makes them look really good. I don't mean to minimize the value of them releasing this information without having to they released a very comprehensive report. It's not like they're hiding things, but it's easy to release this kind of report when it looks so good. But it legitimately does look good. Yeah, I of all the folks kind of, you know, doing the hand wringing of, you know, the autonomous vehicles, are we really ready for this? It's going to be bad. It's going to be mayhem on the roads. When you say, all right, over a couple of years, 18 crashes, it's like, oh, 18 crashes. Oh, 14 of those were from humans who ran into the back of a Waymo car. Well, you know, it's I don't think, yeah, it's it's putting into perspective how pretty safe this is seeming. And again, limited test in Arizona and only a small you know, out of Phoenix area. That's that's fine. It's not going to be replicated just anywhere easily. But this is pretty good. It's it's yeah, it's pretty good data, you know, for the last couple of years, because Waymo has taken this seriously. Yeah, I think this is this is worth looking at. And it's a reminder that our accountant for the company here lives in Mesa and or works in Mesa, actually don't know where she lives, but she she works in Mesa has agreed to try to get into one of these fully autonomous Waymo cars without a human driver and report back. So first of all, hats off to her her bravery to be willing to try that out. And I hope she's able to do it by the end of November. We'll we'll get back to you on Good Day Internet with a report. Let's talk about robots. OK, OK, a couple interesting developments in robot news might be nightmarish to you or you might think it's kind of cool. The first comes from Disney Research Team, which published a paper with researchers from the University of Illinois and the California Institute of Technology about a robot designed to hold eye gaze more realistically. Yeah, they're going to stare you down. This robot can turn face to face to people. They can also blink also shift from direct eye contact to the rapid eye movements to better reflect how people actually hold eye contact because you're not always just staring and on blinking, right? So it's a little bit more human like in a demonstration video, the robot has no skin. So a little unnerving if you look at the video, but it does dress in a stylish Oxford, which we appreciated. Not to be outdone, however, researchers from the Chinese University of Hong Kong presented developments on a new kind of modular robots called Freebots at the International Conference on Intelligent Robots. These include small robotic vehicles inside iron spheres with two motorized wheels and a permanent magnet inside. So you can think of it kind of like a sphere robot that can magnetically connect to other units. Right now, these all need to be remote controlled and don't have any sensors on board. But it solves a major problem of modular robots by making connecting much simpler. Yeah, it's not a big leap to think that they could put some autonomy in these, even at the small size they're at. And then pretty soon you have the replicators from Stargate and stuff's just assembling itself as it goes. But all joking aside, this is this is going to make some problems that were once huge problems, not be problems anymore. If you can have self organizing vehicles of any kind, of any size, there's all kinds of stuff you can do with that. To the Disney robot, I mean, this is obviously being developed by them for the theme parks to make theme park interactions more natural with the animatronics. But I was really impressed by how even without a face, right, even without skin, this thing looked natural. And to me, it was all down to to those those eye movements called saccades. That's the the involuntary shifting of your eye that just kind of goes back and forth. Now, when you describe it like that, it sounds shifty. But when you see it in action, you realize that's why a lot of animatronic stuff and a lot of robots look odd because they their eyes just look where they're looking and never move. Whereas if you're having a conversation with an actual person next time you do that, if you do that, pay attention to how your eyes just kind of involuntarily move a little bit while you're looking at someone. But you also want people to look you in the eye and this thing does both. It's able to look you in the eye, but it has that natural saccade movement that suddenly I was like, this thing looks real. Like you put some realistic skin on that thing and hair and it will be very natural. Like it's very smooth and and the eye movements felt felt really normal to me. Yeah. Yeah. It's a it definitely when you see it, you're sort of like, oh, you know, it is somewhat nightmarish looking, right? Because it's almost looks like, you know, like a little bit of a zombie situation. But but yeah, you're right. There are there are so many very subtle nuances that we as humans do and we don't think about it. But we know when something isn't quite right. So what you're talking about is like, you know, the stare that's a little too direct kind of thing. That is that's definitely what what they're getting at here. It's cool. All right. Let's check out the mailbag. Let's do it. First one comes from Stephen who says about a week behind on DTNS, but I was just finishing listening to episode 3891. Scott Tom and Scott Johnson were discussing gaming subscriptions. This was in the aftershow. Stephen says, I just want to point out some of us play a single game for months before moving on. Currently, I'm on month 10 of playing Ark Survival. It was six months of satisfactory before that. So for one game at a time, people like me, fifteen dollars a month for a subscription, not economical. I agree that subscription works for most gamers like my kids, but there are still some of us that are happy with just the one game. Man, every time you say anything, there's going to be somebody raise their hand and say, not me. I'm not like that. And Stephen's that guy this time. But Stephen, that's me, too. I don't want to subscribe to these things for anything other than research purposes because I don't play that many games either. So I don't think there's that many of us, to be honest. I think that's pretty rare. And that's why these things will probably survive. And Stephen admits he's like, yeah, my kids are going to love this sort of thing. But you're right. There are people who who still just buy one game and just play it until you're done. Yeah, I'm kind of I'm kind of that person as well. If I like it, I'm like, I don't need another game. I like this one. Got another email from Russell. Tom, what do you say? Just quick thought on the discussion about getting a free Apple watch in return for recording a decent level of exercise. Russell says, I guess it's possible that unfit people would ratchet up their exercise enough to reduce the risk of death for a free watch. But I wonder if the insurance company has a slightly different name. I would imagine the offer is especially appealing if you already put in the sweaty calories that might bias the insured population towards fitter people and safe payouts. Russell, it's a good point. I think both are the aim. The insurance company is like anybody that we can get on this life insurance plan to use the watch is going to potentially be healthier because they're using the watch. But if it also attracts fitter people who are more interested in the watch, that means fewer payouts too. So yeah, it works either way. Thanks everybody who emails us feedback all the time. If you'd like to join the crew feedback at DailyTechnewshow.com is where to send them. We'd like to shout out our patrons at our master and grand master levels. And today they include Bjorn Andre, Scott Hepburn and Dale McKayhee. Thanks to everybody who supports our show directly as well. If you're listening to this on the public feed, you may hear some ads from time to time, depending on where you are. But you can get a feed that is all your own that doesn't have those Patreon.com slash DTNS. And another way to support us is you're like, well, I can't do that on an ongoing basis. Get a DTNS hat or a hoodie or a mask or a mouse pad. We have all that and more at the Daily Tech News Show store DailyTechnewshow.com slash store. We're live Monday through Friday for 30 p.m. Eastern 21 30 UTC. Yes, standard time change for a lot of us over the weekend. Find out more at DailyTechnewshow.com slash live and we'll be back tomorrow with Engadget's Nicole Lee. See you then.