 Coming up on DTNS which facial recognition systems can be fooled by a mask delivery drivers in Mexico take some safety into their own hands and surgeons that'll give you Instagram face. This is the Daily Tech News for Monday December 16th 2019 in Los Angeles. I'm Tom Merritt and from Studio Redwood. I'm Sarah Lane and I'm the show's producer Roger Chang and I'm Joe just fine that shows technical director. Oh look at that. We got it. We got Joe in here. Roger's got staff. This is amazing. This is a big step. Thank you, Patrons. Ginger Fed. Woo-woo. We lost our story. It wasn't her dog. If you want to find out how that turned out you got to get the wider conversation a good day. Internet by becoming a member at patreon.com slash DTNS. Let's start with a few tech things you should know. An FBI investigation led to Daryl Polo and Luis Viorino pleading guilty to copyright infringement charges for operating iStreamItAll, a subscription based streaming site and JetFlix, a TV streaming service that had a lot of legal content involved. As part of his plea agreement Polo said that he emailed members of iStreamItAll to cancel their licensed services in favor of this pirated content. He also admitted to earning $1 million from his privacy operations and to downloading the content from torrent websites. ICANN is going to review the pending sale of PIR. That's the management of the .org domain. Remember we talked about how the internet society would like to sell PIR to private equity company Ethos Capital. ICANN has now requested more information on the terms of the sale from PIR. Once it gets that information ICANN has 30 days with which to withhold its consent for the sale or approve it. ICANN approved a change to eliminate price caps on .org domain names a few months ago and that's where a lot of the concern lies. Ernest Biarawanga, an executive at the non-profit entity Africanic, has resigned following accusations that he secretly operated several companies which sold tens of millions of dollars worth of IPv4 addresses to online marketers. The allegation stemmed from a three-year investigation by California-based freelance researcher Ron Gilmet who since 2016 has been tracking several large swaths of IP address blocks set aside for use by African entities that ended up at internet marketing firms based in other continents. Tech Crunch's sources say Uber is in advance talks to sell its Uber Eats India business to Indian rival Zomato. Uber launched Uber Eats in India in mid-2017 and reportedly peaked at 600,000 orders per day compared to the more than one million daily orders from Zomato and another local rival Swiggy. And Derek Ozerashahi has said if you don't get a million, we're going to get out of that market. So I guess they're getting out of that market. Alright, let's talk about a roadblocker with Google Chrome. Bit of one, yeah. Google paused the rollout of Chrome 79 on Android following reports of missing data for apps that use Android's built-in web view. Chrome 79 changes where web data is stored, but it didn't properly mitigate data from local storage and web SQL. So as a result, some apps and sites didn't show local data, not deleted, just inaccessible after the update. Google confirmed the bug and is working on a fix, which it says may take up to a week. I am often one to try to calm everyone down and say, hold on, you know, bugs happen, doesn't mean people are incompetent, it's good to catch them. This one I feel like from the limited amount of knowledge I know about it should have been caught. If the data isn't getting moved, that seems like something you catch in your testing, like you checked, did the data move? So I'm curious what the explanation is of why this happened, and I understand people's frustrations if they feel like their data was deleted. Thankfully it's not, it's just temporarily inaccessible. Also, I mean, I don't know, it depends on how you got to fix these sort of things, but up to a week seems like a long time for a fix. Yeah, well, maybe they don't know what's wrong. Maybe that's possible as well. We tested it and it did work, Tom. We don't know why this is happening. Give us a week. It's Christmas. That's true too. The Verge is Casey Newton has another write up of the experiences of content moderators. Remember, he's written two stories about Facebook, both in Florida and Arizona. This time, he's talking about moderators for Google and YouTube, including Google's contractors at Accenture in Austin. Moderators view five hours of video per day in a lot of cases, even though Susan Wojcicki has said that they were going to limit the YouTube video moderation to four hours. While those working on some of the content enjoy the job, they're just looking for copyright violations, things like that. Those working on disturbing content, as we have heard from Casey Newton before, describe feelings of anxiety, depression, night terrors, and some severe mental health consequences as a result of seeing disturbing images. While full-time Google moderators get medical care and months off of paid medical leave for things like treating post-traumatic stress disorder, contractors do not. Moderators from both Google and Accenture have been diagnosed with things like PTSD, chronic anxiety, other long-term mental health issues. There's some great case studies in this write up from Casey Newton, and I highly encourage people to go read it. But it does shine the light on the fact, in my opinion, Sarah, that humans are not the solution to this kind of moderation. I'm not saying we shouldn't have human moderators, but this is a dangerous job for your mental health. There needs to be more attention paid to the safety of the moderators, and we shouldn't be thinking that humans are the solution. I know Google and Facebook are all looking at AI ways of doing this, but this is where I start to think that maybe Twitter's got a better idea of just changing the whole system of content so that you don't have to moderate it as much, making it harder for bad content to get eyeballs. Yeah, I mean, the whole AI replacing humans, in many cases, we're not there yet. This is one of those cases where we sure, let's get there, right? I also think that anybody working on the disturbing content team of anything should be getting real good benefits. It doesn't mean that you're not going to get PTSD. It doesn't mean that you're not going to feel anxiety. It doesn't mean that you're not going to have to observe a lot of stuff that is detrimental to your mental health. But I think that would be, well, we still have to rely on humans to do a lot of this content moderation. That would be a really good will first step for companies who can afford it. Yeah, I mean, we really, I keep harping on this, but we have to go back to the idea that when the internet was smaller, when there were fewer users, self moderation could happen. But the scale at which these companies work now, and the fact that there really is no community level moderation happening in a lot of cases, not all cases, but in a lot of cases, means that the things that kept the internet working well in its earliest days don't work for these companies. And that's why I think, yeah, decentralize things, make it smaller, make the problems easier and make the access harder. In fact, there was an article earlier today about the fact that ISIL is having a harder problem or diash or ISIS, whatever you want to call them, having a harder problem getting exposure now because they've been kicked off a lot of platforms. Not saying de-platforming is the solution either, but when you make the community smaller, it makes it easier for moderation to happen. It makes it harder for these kinds of disturbing videos to become the problem that they've become on these large platforms. And I think there needs to be more attention paid to that in the interim between now, and like you say, that someday when AI can do an acceptable job at identifying disturbing content before it gets published. And at the same time, I never want anyone to have a horrible day or worse because of having to view a bunch of disturbing content, and it is pretty disturbing. The article lays out some of the stuff that people have to absorb, and it's pretty awful. At the same time, the people running these companies shouldn't be blind to the fact that this stuff exists, even if there is AI to be able to filter it out. That's important as well. Yeah, for sure. Moving on. Intel has acquired Israel-based Habana Labs, which develops programmable AI and machine learning accelerators for cloud data centers. Its GAUTI AI training processor promises four times the throughput of an equivalent number of GPUs at half the energy per chip. And its Goya AI interface processor promises three times the inference performance measured by throughput and latency. Habana also has an environment called Synapsys AI for deploying AI workloads, and will come on board along with previous acquisitions Movidius, Altera, Nirvana, and Vertex.ai. Intel's AI chip revenues reached $3.5 billion this year, up from $1 billion in 2017, and is predicted to grow annually by 30%. So, it's something that Intel's investing in. We've talked a lot about Intel getting out of businesses, cutting back on certain business lines, both under the current CEO and the previous CEO, and how these are smart moves. But we also continue to see Qualcomm, AMD, etc., taking away traditionally markets that Intel has dominated. And I think it's smart for Intel to say, rather than try to fight a rearguard action, maybe we focus also on moving forward into an area where we can dominate. And acquisitions like this will be a way to get Intel into that space quickly. AI chips are going to be things that are in consumer devices in a lot of cases, so they will sell into a lot of consumer devices, but also will be incredibly important in the data center, in enterprise settings, and that's where you get a lot of money. So I think these are, this is a very smart strategy that's unfolding by Intel. And I think that there may come a day in the not too distant future. I wouldn't be too shocked when you no longer hear about Intel inside your product, but that's fine. Just the way we don't have IBM personal computers on our desktop anymore. Yeah, with the projected growth, 30%, okay, that's just prediction. If you're Intel, you're like, all right, let's shed some business, some legacy businesses of ours that aren't working as well, and pick up some stuff that we believe is the future, and that's kind of what the company has to do. Yeah, this is a good move for Intel. Fortune reports on a test of facial recognition systems by AI company Neron, K-N-E-R-O-N. Neron creates high quality, or did create, they don't normally do this, but they had created some high quality 3D masks to fool payment systems from Alipay and WeChat in China. The Neron team also was able to access a self boarding terminal in the Netherlands, as well as pay fare to board a train in China, both just using the picture on a phone screen. Now the masks were high quality items. They were made by a specialty mask maker in Japan. It's not something that you or I could do easily. It's not something that's going to be a mass vulnerability, but you know, if you wanted to go after a high quality target, that might be something you do. Neron also said it could not fool Apple's face ID on the iPhone X, so it tried some facial recognition that did not work. Neron CEO Albert Liu said, quote, the technology is available to fix these issues, but firms have not upgraded it. And of course, it's worth taking into account that Neron makes its own facial recognition technology. So this was a bit of a publicity stunt for them, but it's grounded in reality where they really were able to fool these systems that are out in the public right now. Yeah, it's, you know, I hopefully no one's ever going to make a mask of me and, you know, log me into some stuff. You know, you don't have any money anyway. But just the fact that this is, it's highly unlikely this would happen to you, but it's possible. That's something to think about. Yeah, and I think even even beyond that, just the idea that these are in the public, these are in deployment systems that are at least in the case of the phone easily fooled. And in the case of the masks, foolable, they shouldn't be deployed yet. I hear a lot of people upset about facial recognition and there's certainly a privacy aspect to it. I personally believe there is a responsible way to use facial recognition to let me board a plane and go through security and all of that sort of thing. I don't labor under the misconception that people's faces are somehow unknowable. But I think we need responsible legislation on how that data is managed, just like we need responsible standards on when facial recognition can be deployed. A bigger thing than the privacy right now is the fact that facial recognition systems are being put in place that don't work. And that's where the privacy gets amplified because you have false positives, people getting in trouble or being harassed for something they didn't do because the facial recognition system was fooled. And I guess if you take it a step further, it's like, OK, let's fix this issue. What do you do to facial recognition systems so that there is some deeper level where the system can say, yeah, no, that's a mask. It looks a lot like that person, but it's not that person. Well, I mean, Apple does it. Apple said you can't fool our system with a mask. Nearon says their system can't be fooled either. And I go back to that quote from Albert Lute, the CEO of Nearon. He says, the technology exists. This is not a problem that needs to be solved technologically. It's just the companies are putting systems out in the world without wanting to implement the fixes, whether that's just lagging behind the times, whether it's not knowing buying the wrong products or just not wanting to budget the proper amount to upgrade systems. It shouldn't be happening. Well, in other news that shouldn't be happening, Gia Tolentino at the New Yorker wrote a detailed account of what she dubs the Instagram face, which is a real life version of a filtered person based on cosmetic surgery. Sometimes not invasive, sometimes invasive. Tolentino describes it as a young face without pores, with high cheekbones, with cat like eyes, with long lashes. We all know this face. You see it all the time. The nose is small. The lips are full. She says it looks like a direct descendant of Kim Kardashian West, Bella Hadid, Emily Radikowski, and Kendall Jenner. Great. Tolentino met with multiple plastic surgeons in LA to actually get price quotes and suggestions from these surgeons if she were to want to achieve the Instagram face. Some of these quotes were quite high in the 20s of thousands of dollars and walked away somewhat despondent that this is our new normal. Yeah, the surgeons, all three of the surgeons that she talked to in this article said, yes, we definitely get people coming in, holding pictures of Kim Kardashian West or others like that wanting that type of face, that Instagram type of face. From a certain point of view, this isn't new. People have always, for a long time anyway, airbrushed photos to match more to an ideal of beauty. And people have put on cosmetics and later engaged in plastic surgery in order to make themselves use technology to reach an ideal of beauty. I feel like Instagram is accelerating that like we see with so many other issues that we're just talking about with moderation and disturbing images and all of that. It's accelerating the access to this ideal such that it's now merging these people's right instead of just having Twiggy be the ideal and you want to be thin. It's like, well, I want her lashes and her cheekbones and the, you know, cobble it together to the perfect face. Well, and you have people who can give that to you saying, yes, you should do this and here's how much it will cost you. It's interesting too that, I don't know, I mean, there are lots of Snapchat filters, for example, where I go like, oh, I look way better. But to actually go ahead and try to get that in real life all the time would be quite an ordeal. And not everybody's doing this, but the fact that it is on the rise in a way that has become much more of a socially acceptable norm than in the past. Even not even something that you kind of go like, you know, maybe I got some work done. Maybe I didn't. Almost like a like a bragging right. Like I care enough to do this is it's disturbing to say the least to me. Well, it's it's familiar in its disturbingness, right? Right. People parroting political lines. People parroting angry opinions about movies. People parroting looks on Instagram. Right. Nice. Oh, sorry. Go ahead. No, I was going to add. It just occurred to me like this is a composite of people. This isn't like Twiggy where there was an actual tweet earlier. Exactly. This is literally a composite of various attributes of someone who does not exist synthesized into an ideal form, which, you know, may or may not be achievable for for a lot of people. And I think the disturbing part is unlike other forms of media that used to be rather, you know, separated from you, whether it was a billboard or magazine, it was something somewhat detached. But in social media, it's like such a pervasive part of a lot of people's lives that you just can't help but be kind of sucked into it sort of. It goes from ripping a page out of a magazine and bringing it into a plastic surgeon saying, can you make me look like this? To not only being able to know from over exposure exactly what the elements you want from various people are, but to create a picture of yourself with a Snapchat or an Instagram filter that shows you looking like how you want to be. And it's not just that it makes it easier to describe that plastic surgeons could do things like that for decades and decades. It ingrains the desire for it in you because you're seeing it regularly. That's that's the point I was making about you hear a political statement. You you repeat a movie thing because you get a reward for doing those you feel good from the community response. And that's what people are chasing with Instagram. They want those likes when they show up with that face. And now they want to they're like, Hey, could I get that same feeling in real life when people see me on the street? That would be amazing if I looked in real life the way I look on Instagram. And at the same time, all this really is fashion. You know, it's like back in the day, if I had those boots, I wish I could be photographed in those boots. They're the coolest boots. Now it's sort of like, Oh, are you going to permanently alter your actual body? Okay, well, that's fashion too. But what happens when the tide changes and all of a sudden that's not what people want to look at it anymore? Because that's what happened to. Yeah. All right. Freelance journalist Martha Piskowski posted at mediums one zero about delivery drivers in Mexico for companies like Didi, Uber, Cine de l'Intel and Rappi who use WhatsApp groups and Google Maps to help each other. The delivery drivers use these to help each other stay safe in high crime neighborhoods. It's not like they're always delivering in high crime neighborhoods, but often their deliveries take them there. And they're helping each other out. While companies do offer their own support hotlines and sometimes even insurance, workers say they don't always get the response from those hotlines that they would like. Sometimes the hotline operators aren't even in their same country. Say they're talking to somebody in Columbia in the case of Cine de l'Intel. And those people don't understand what a certain neighborhood in Mexico City might be and what they're talking about. So they've started to rely on each other. The workers comment on crashes, injuries, harassment, robberies, among other things. WhatsApp helps them monitor each other so they can keep in touch while they're making these deliveries. And several of them have made some Google Maps marking locations of past robberies and some addresses to avoid. There's even a group called Ni Un Repartidor Menos that has almost a thousand members spread across four different WhatsApp groups to all help each other out. Stay in communication, share tips, et cetera. And they use a Google Doc to share their own contact info. And that Google Doc gives them a registry number that they use to help each other. And that Google Doc gives them a registry number that they write on their backpacks for identification. So if one of them, say, gets knocked out and is in trouble and another driver finds them, they can look them up in the registry and find out their information really fast. You know, I never want to say, ah, you know, the company shouldn't have this infrastructure in place. And it sounds like in many cases they do. But, you know, got to hand it to people taking safety into their own hands in this situation. I read the story and I was like, that's cool. You know, I mean, you should never feel unsafe. You should always be safe. And I hope that you are safe always. But the fact that this community has sort of banded together and saying, all right, well, we all have something in common. And there are some safety issues. Let's help each other out is kind of heartwarming. Yeah, Peskovsky's article is really well, again, all of these articles we're talking about today, we have links in our show notes. You should go read them. And she points out that unlike in, say, the United States where we see gig economy jobs as going from a safe and secure job with insurance and labor protections to something less safe, a lot of times the people filling these jobs in Mexico are coming from in four or sometimes called informal jobs where they had zero protection. They don't even have the minimal protections that the gig economy provides. And so they see this as a better job. They see it as a step up. And it's interesting to see with that perspective, the kinds of systems that they start to create as a community with each other. They're still pushing for the right to unionize and a lot of the same things you would you would see workers doing in the United States. But just coming at it from a different perspective, I thought it was a fascinating read. Indeed, gig economy stories and all the stories that you care about always accepted in our subreddit. You can submit stories and vote on them at DailyTechNewShow.Reddit.com. You can also join in our conversation in Discord. The Discord community is growing and great. Join by linking to a Patreon account at patreon.com. To get all the tech headlines each day in about five minutes, be sure to subscribe to DailyTechHeadlines.com. Chris Christensen, our amateur traveler, now has a handy translation tip for travelers going to China. This is Chris Christensen from amateur traveler with another tech in travel minute. If you are traveling to China, here's an app you might be interested in, which is the PLECO app, P-L-E-C-O. I was interviewing someone recently about their trip to China and they find that this app works much better than Google Translate in China. It has OCR, it can detect handwriting, it has dictionaries, it has 34,000 Mandarin words being pronounced in audio by native speakers. It has flashcards, so it can be a useful app if you're going to China. I'm Chris Christensen from amateur traveler. Well, I'm not going to China, but that's good to know, Chris Christensen. Yeah, that actually sounds extremely helpful. That is super handy. My dad's a native speaker and he still brings the dictionary with him. Yeah, well, because there's so many different accents and stuff, too. Is that why? It's because you need a standard 1000 set of characters just to get by. It goes up a lot more than that. Alright, let's check out the mailbag, Sarah. Let's do it. Brandon wrote in and said, during episode 3671, can you believe we've gotten to that point, the pre-show discussion about the IPv4 heist, I started thinking about the eventual IPv6 future and was wondering if you foresee a future where the rules regarding an IP address not being a person get reevaluated. That's an interesting question. I think where Brandon's going with this is since IP addresses get dynamically assigned and IPv6 addresses don't have to be dynamically assigned because there's so many of them that can be assigned to the device without network address translation, then would it be more reliable to say, well, no, that IPv6 number is assigned to that device and that device is that person. So the IP number and the person match a little more. Yes and no, it's an interesting thought and it probably will make it happen more often. You're still going to get dynamically assigned addresses for a long time. As we can see, people don't want to give up IPv4. We're certainly going to continue to have network address translation on people's routers for a long time. I don't know though, maybe 20 years down the road, everybody has their own set of IPv6 numbers and it starts to become a little bit more about them. But then you could say, well, but who was actually accessing the device, right? Was it me? Was it someone else? Was it someone from outside my network? Those kind of decisions play in, but it does, it would increase the accuracy a little bit. That's an interesting thought, Brandon. We also got Chris right now and saying there was talk of VPNs on Friday's show. So I wanted to remind users of Google Fi that you have a free VPN built into the service. Just turn it on. It's worked great for Chris since day one. He says if you're worried about using Google as your VPN provider, Google says they don't store data that goes through the VPN, but it may use the metadata to improve their service. The Verge has more information and we'll have a link to that in the show dose. Surf safe and have a great holiday, says Chris. Thank you, Chris. Thank you, Chris. Also shout out to our patrons at our master and grand master levels, including Andy Beach, Martin James, Bjorn Andre, Tim Ashman and Phillip Shane. We have new Patreon reward merchandise. In fact, I just got the final two mock-ups of the t-shirt and the sticker. There's also a mug and a poster with a DTNS logo created by Len Peralta to celebrate our six-year anniversary. If you're back at certain levels at Patreon.com slash DTNS for three months, you can get one of those items in your hand at the end of the three months. Get the details at Patreon.com slash DTNS slash merch. And if you have feedback for us, our email address is feedback at dailytechnewshow.com. We're also live Monday through Friday for 30 p.m. Eastern 2130 UTC. Put it on your calendar and find out more at dailytechnewshow.com slash live. Back tomorrow with Rich Straffolino. Talk to you then. This show is part of the Frog Pants Network. Get more at frogpants.com. Well, I hope you have enjoyed this program.