 Coming up on DTNS, machine learning wants to help identify proteins and improve your meetings. But will you give up your privacy for the latter? Plus, the Supreme Court weighs in on the Computer Fraud and Abuse Act in a battle of privacy versus investigative freedom. This is the Daily Tech News for Monday, November 30th, 2020 in Los Angeles. I'm Tom Merritt and Lovely Cleveland, Ohio. I'm Richard Raffalino and I'm the show's producer, Roger Chang. Sir, Elaine has the day off today. We will miss her, but she'll be back tomorrow. In the meantime, we were talking quite a bit about food and what foods used to be considered acceptable and are not anymore. If you want that wider conversation, you got to get good day internet, patreon.com slash DTNS. Let's start with a few tech things you should know. All right. Well, FCC Chairman Ajit Pai announced that he will leave the agency when President-elect Joe Biden is sworn into office on January 20th. Pai was named a commissioner to the FCC in 2012 by President Obama before being named chairman by President Trump in January 2017. General Motors announced it will not acquire an 11 percent stake in the electric vehicle maker, Nikola, after initially announcing plans to do so back in September. Instead, GM is signing a non-binding memorandum of understanding, which is pretty much just as good, to supply Nikola with its hydrotech fuel cell system, which GM qualifies as a potential agreement. So they might not even do that. This comes as the US Securities and Exchange Commission has opened an inquiry into accusations of fraud by Nikola and the departure of Nikola's founder, Trevor Milton, as executive chairman. The fuel cell systems will be used for Nikola's planned heavy-duty class of commercial vehicles. Should they? All right. Well, Facebook confirmed it intends to acquire the Omnichannel customer relations startup Customer with a K, very pukish of them. The company provides businesses an aggregated view of where customer conversations are taking place across email, chat bot, websites, and mobile apps. Ten Crouches sources say the deal is going to be worth about $1 billion, and there is currently 175 million business users on Facebook for some context as why they might be interested in that. Analyst Gartner saw smartphone shipments continuing to bounce back closer to pre-pandemic levels in Q3, even as market share of individual companies moved around a little. Samsung and Huawei remain number one and two globally, while Samsung saw a modest 2% growth on the year to 80.8 million units. Huawei saw shipments fall 21.3% to 51.8 million, but they held on to second place. Unlike Apple, which dropped from third to fourth, Xiaomi overtook them for number three, growing shipments 34.9% to 44.4 million. Apple actually saw shipments decline slightly, 0.6%, though they did improve market share a bit. Overall, the entire market was down about 5.4% on the year, which is much better than the past two quarters of 20% declines. And developer Andrew Graff was able to use Windows Insider preview builds on Windows 10 for ARM to get a virtualized version running on Apple Silicon Macbook Pro. So Bootcamp may be dead on the Apple Silicon Macbooks, but Andrew Graff might have you covered here. He had used a custom patch, however, to execute guest code on the Mac CPU using the Quemu emulator, and the build wasn't exactly stable. So maybe hold off if you need stability. However, looking at Geekbench 5 scores running on that emulated Windows build, they posted a 61% higher single core and 76% higher multi-core scores than the Surface Pro X, which runs Windows on ARM natively. Developer Alexander Graff. Alright, let's talk a little more about using that machine learning to fold some proteins. Yeah, it's pretty cool stuff. Alphabet's DeepMind subsidiary. They're usually known more for making headlines about, you know, beating humans in video games like StarCraft 2 and Dota 2. Basically, all the twos they got you covered. But the company's AlphaFold system recently achieved a major milestone in predicting the structure of proteins. Traditional protein folding prediction is a resource hog, and with current hardware, it can take years to complete. That's why crowdsource projects like Folding at Home have been used to provide the equivalent of massive computer resources. Just aggregate all that compute and put it to some good use. AlphaFold made the process more efficient, meaning you don't need huge resources by training an attention-based neural network to predict protein folding structure. Attention-based means it can focus on certain inputs for better efficiency, essentially. It continually refines its predictions based on their history. AlphaFold was able to identify protein structure within a matter of days with accuracy within the width of an atom, which I'm told is very good. Correct identification of protein structure can help in understanding disease transmission, how allergies work, breaking down toxic waste, and a whole lot more. This is a really big deal. I know, you know, folding at home has been kind of, I don't even know, owning the space, but making people aware of the complexities and the resource-intensive nature of folding proteins to put this into a kind of a different beat here for DeepMind that we haven't really heard a lot about before, Tom. Pretty interesting stuff. Yeah, we are around 204, 205 ranked in Folding at Home with the Good Day Internet team, as a matter of fact. So, you know, a lot of folks in the audience are very active in that. Still a useful project to be involved in. It's not like this AlphaFold is just going to replace Folding at Home tomorrow. But a great example of where machine learning can move us beyond the limitations of the hardware, right? We talk about quantum computing a lot as being able to change the limitations of the hardware with different hardware. This is using a different kind of machine learning. As I like to explain, artificial intelligence is almost meaningless. Machine learning is a very big and wide tent under artificial intelligence that at least is meaningful, but can mean a lot of different things. And here, you have neural network learning, attention-based neural network learning. So a very specific type of artificial intelligence that allowed them to say, well, look, if we just have it teach itself, it can very quickly get to the right answer without having to brute force it, without having to use all of that machine power like a Folding at Home has to use, where it coordinates millions of computers around the world to provide the ability to brute force and just keep trying combinations until you find the one that work. This is an example of machine learning being able to better guess which one might work next. Yeah, and it's really interesting to see a situation where I think we're so used to so many compute projects, where I think a lot of times we think about cloud computing, you just like infinitely throw resources, where we can very much reduce time as a consideration. I think for the longest time, we've kind of thought of protein folding as one of those tasks that's kind of, it's just so complex that no matter how much you throw it, it's still going to take so much incredible amount of time. And to be able to, what we don't know about, what we haven't seen from AlphaFold yet, is actually like what the hardware requirements are, what the cost is actually going to be for research institutions, at least as far as I've seen. This may be in some of the more in-the-weeds stuff. But the fact that they can eliminate that time is something like, this is exciting to me, because previously we just didn't think that that was something that was on the horizon, at least from my perspective, certainly. Yeah, and yeah, it'd be nice if we could just look at the proteins and go, oh, that's how they fold. I think it's not quite that easy at this point. The US Supreme Court heard oral arguments Monday in Van Buren versus the United States, not the president, different Van Buren. Regarding the application of the US Computer Fraud and Abuse Act at issue is a former Georgia police officer. We've talked about this on DTNS, if this sounds familiar, a former Georgia police officer who accepted money to look up a woman's license plate in a law enforcement database in 2017. No one is arguing that it was wrong for him to do that. What they're arguing is which laws apply to it being wrong. There are some laws about bribery and taking money and all of that apply here for sure. But the question is, does the Computer Fraud and Abuse Act also apply? And do the penalties you would apply because of the Computer Fraud and Abuse Act therefore need to be applied to this case? Now, the Computer Fraud and Abuse Act in the United States says it is a crime to knowingly access a computer, and I quote, without authorization or exceeding authorized access. That's the important part of the act here. Van Buren was authorized to access this database. An issue is whether he exceeded authorized access by using it for this private unauthorized purpose. Doesn't even matter whether it was a crime to actually share the data or not, this is unauthorized. He was not allowed to look up this woman's license plate for a private person. The Supreme Court is going to focus on what unauthorized means in the text of the Computer Fraud and Abuse Act. So this isn't the Supreme Court deciding if he was wrong or not. This is the Supreme Court looking at, okay, how should we interpret unauthorized in this case? And then once they decide that, then everything else follows. Van Buren's case likens the use to looking up sports scores on a work computer. So Van Buren's team is saying, look, using your work laptop to check the game score might not be authorized technically, but it's not a crime. The crime is in what he did with the information, not in the accessing it, and therefore it's not a violation of the CFAA. Might not have been right, but it's not a violation, not fraud and abuse. There's other problems with it, but that's not the problem. Groups like the ACLU and the Electronic Frontier Foundation side with Van Buren, arguing that a broad interpretation of the law would make security research, investigative journalism, possibly illegal, certainly harder, and even make violating terms of service potentially illegal if the case was taken in certain directions. They argue that a privacy law should deal with privacy issues, not the Computer Fraud and Abuse Act. So there are going for a strict interpretation. This is about computer fraud and abuse. This may be a privacy violation, but it's not computer fraud violation. You should have a privacy law applied to this, not this law. It's what you might call an originalist interpretation. However, privacy groups like the Electronic Privacy Information Center, or EPIC, argue in an amicus brief of their own that the law is meant to protect not just outside attackers, but unauthorized use by insiders. A Senate report from 1996, they point out, described the CFAA as designed to increase protection for the privacy and confidentiality of consumer information. So they're saying if someone in a company is authorized to access your personal information, but they do it not for the purpose they're authorized to do it for, that should be illegal. That's a privacy violation that should fall under the CFAA. And EPIC argues that Van Buren was accessing a government database, not a public internet-facing website, so it should not impact activities that involve terms of use and scraping things from the internet, etc. Concerns about research and journalism can be therefore addressed in separate cases. The Supreme Court's decision in Van Buren will likely impact another pending case for the court called LinkedIn versus Haikyuu, where LinkedIn issued a cease and desist under the CFAA against Haikyuu, which was scraping public LinkedIn data to make its own human resources tool. LinkedIn has petitioned the Supreme Court to take up that case as well. So the Supreme Court's going to decide on this, and that will actually be a precedent in the LinkedIn Haikyuu case about public scraping. So as far as EPIC's contention, we're going to get that. The court will rule in this Van Buren case later this year. They heard the arguments today later this year, or maybe even early next year, we'll get the decision. Yeah, it is. It's a tough, I don't even want to call it balancing act, because they're going to, if they come down with a decision, they don't just punt it back to the lower court ruling or something like that. Come on one of these two sides. I am a little bit more, I guess, sympathetic to the ACLU argument in terms of this. Again, from my perspective is a hacking-based law. We should have privacy law dealing with privacy issues. Not to say that I want people with authorized access to internal databases or government databases or Facebook employees, or people with access to Alexa recordings and stuff like that, to be able to use those in unauthorized ways. But I feel like the law would be better with a more specific privacy focus as opposed to a more blanket approach. That's what I think is better, is that what I think the law actually calls for. That's what the Supreme Court obviously will decide here. Yeah, I may be revealing a little bit of my legal interpretation tendencies when I say this, but I agree with you as well. I think the law was written for a purpose, and we shouldn't start using it for other purposes. Instead, we should pass other laws, and I know that's hard. I know it's more tempting to just get a court to decide the thing the way you want to, and then you don't have to work on the hard business of passing another law. But I feel like the privacy violation from within that Epic is arguing is a good argument, and I could see the Supreme Court saying, yeah, no, that would be abuse. I do think there's a way the Supreme Court could rule against Van Buren and still set a line that says, but it does shouldn't affect security research. It shouldn't affect public scraping. Those are separate issues. I have no inkling of which way the court is going to rule on this. Interesting note, though. I did listen to some of the oral arguments, and Justice Thomas did pipe up. He rarely speaks if people don't understand why that's significant. He almost never says a word. So obviously it was of interest to him. All right. Well, over on the other side of the pond, the European Commission is considering expanding its geo-blocking regulation to include audio-visual content, aka video streaming. The Commission is holding discussions about the audio-visual sector as part of its upcoming media and audio-visual action plan, catchy name. The EC knows that an average European consumer has access to 14% of films on streaming services compared to content available to all EU member states. So if you take the catalogs of everything available to all EU member states, the average person gets 14% of that. This, of course, varies widely by country, is another problem that the EC is seeing, with, for example, 1.3% of content available in Greece compared to 43% available in Germany. The Commission notes that other types of copyrighted content, music, ebooks, video games, for example, offer a much more homogenous catalogs across EU states with something like 90% shared when you compare it to overall availability versus your average European consumer. So it looks like, again, this audio or excuse me, this geo-blocking regulation is relatively new for the EU, so it's only 18 months old. So it's good to see them reviewing this. Tom, why do you think we're seeing this big disparity between things like music and video was digitized around the same time? You would think the rights issues around those would be relatively the same, right? Yeah, it's a historic vestige that music was easier to release in multiple places. It was also easier to get a hold of things a little bit in the physical distribution days. And so books and music and whatnot, they're also not as time sensitive as movies and TV have been in the release. The history in all of this is that we used to not be able to release, say, a movie everywhere at once, because you had to ship it. By ship it, I mean you had to put it on a ship. There wasn't even airplanes back when movies were first being distributed. So it made sense that, well, I can never release in New York, but that doesn't mean I'm going to be able to have the release at the same time in Tokyo unless I do a bunch of massive coordination. So we'll just have different release dates, and also we'll have different people distribute them. And so it started to make sense to make business arrangements based on that, to say, oh, we'll have a different company handle the distribution in Japan versus India, versus Germany versus the United States. And those old ways of doing things are still baked in here, especially in the European Union where they are all separate countries, even though they do have these combined marketplaces. So for some reason, music books just didn't, and certainly video games, because it didn't exist back in those older times, just didn't carry forward those business deals that said, well, we want to make a good deal in Germany and then we'll make a separate deal in the Netherlands because maybe we'll make a better deal there and we'll make a different deal in Greece. And that's because of those business deals. The short version is it's because they have different deals with different people in every market that they don't want to just make it available everywhere else. So it comes down to, do you want to short circuit that by having a legislature, in this case, the European Commission, change the law and say, yeah, those deals may have been great, but you can't do them anymore. Let's just make everything available to everyone. Or do you want to wait for the slower market approach where a company like Netflix is pushing harder and harder, harder to just release everything everywhere at the same time and eventually we'll get there? Yeah, I was going to say it's interesting that Netflix is kind of relatively unmoored from the concerns of people that are involved in the theatrical distribution of films and certainly it makes it much easier for their marketing purposes to just say, instead of, hey, this comes out and this is available in Germany, the Netherlands, and Spain, but it's not available in some other Italy or whatever. So I could definitely, they have an interesting position and certainly they have the market share to push it. So we will see if they can put their foot on the gas there for that. I mean, again, if I'm living in Greece, I am all for them being aggressive. Well, I don't know, you say it's 1.9% share, but maybe Greece has like amazing, Greece-oriented content and they're all like, yeah, we don't even care about that other stuff, you know? Folks, if you want to join in conversation and tell us how great the Netflix selection in Greece is, join in our Discord where you can talk to us by linking to a Patreon account at Patreon.com slash DTNS. Just choose to link it to your Discord account on Patreon.com slash DTNS. Microsoft has applied for a patent called Meeting Insight Computing System that would use a system of cameras and sensors to record data in conference rooms, things like body language, facial expressions, the amount of time participants contribute to a meeting. Your speech pattern, quote, consistent with boredom and fatigue, they'll tell of your board. Then they'll add that to whether attendees were texting or browsing the internet and that will combine with factors like, was the meeting efficient? Was the room temperature really hot and uncomfortable and an overall quality score will be created for those meetings? The system could eventually predict the likelihood of holding a high quality meeting based on time of day, who's involved, et cetera, and flag potential challenges ahead of time, maybe suggesting an alternative time or a different venue or maybe different people, different length of time, I don't know. The system could be used for in-person, virtual, or hybrid meetings. Microsoft has been working on the system since 2018 and told the BBC the application of a patent does not necessarily indicate that the technology described will be implemented in a product. This is especially interesting because Microsoft is taking a bunch of heat over a productivity score that it includes in Microsoft 365. Now this is a different thing. This is a thing that's actually in Microsoft 365. It's not a patent. It's an opt-in feature that aggregates information about email use, network connectivity, et cetera, similar but not the same thing. And security and privacy advocates this week were pointing out that one of the default settings let managers check data on individual employees which would let them see who sent the least emails, who collaborated the least on documents, other individualized metrics. Organizations choose who can see that individual data with an option to anonymize or remove it entirely but they don't have to take that option. User data is aggregated over a 28-day period. Microsoft rich saying like, well, the point of the productivity score was not to spy on people but we, I guess, leave that up to the companies because they might want to, for a benign reason, look at individual scores and compare them. Yeah, it seems like this is Microsoft not realizing the lowest common denominator of a potential micromanager or something like that. Where they're like, in principle, this actually reminds me a lot of something that we see in enterprise networking is client-side monitoring is considered increasingly important where it doesn't matter what your internet speed is or what your jitter is or what your ping time is. It matters what the experience is on the endpoint, especially with the meeting stuff. Like that to me is like, do people seem like they are engaged in this meeting? Like that's all we care about. We couldn't care less about the content or anything else like that. That's the metric we're kind of basing all of this off of. So I can see from, this actually reminds me of Microsoft like not getting people a weird way about why they might not be upset about it. And it warms my heart that in an age where we're seeing antitrust investigations of everyone but Microsoft, that they can still be old Microsoft. But I can see the sense in this but I'm surprised that no one also said, what's the worst possible use? Now, to be fair, Microsoft has obviously been playing this for a long time. They released a productivity score back in October. They've been working on this meeting since 2018. I don't think they realized the emphasis that we would be having on a work from home culture during the middle of a pandemic, for example, where we are simultaneously seeing a lot of different organizations, taking very invasive workplace, like genuine workplace surveillance. I want to see where your mouse cursor is and that kind of stuff for some organizations. I don't even know how widespread it is. Anecdotally we've heard stories about this. So I think this is a feature that Microsoft came up with. Again, they're a productivity couple. They really understand productivity. I have faith is even the wrong word. I have confidence that Microsoft kind of knows what they're doing in terms of, hey, we're going to gamify this a little bit. But it seems like in this context it seems a lot more problematic that Microsoft maybe couldn't have planned for when they initially had planned these features. A few things to keep in mind though. You have no, there's no legal obligation for the workplace to give you privacy over your work. They can't spy into your personal life, but you're also not supposed to use your workplace tools for your personal life. So this is a different animal than privacy violations by an Amazon Echo or Facebook or something like that. This is Microsoft saying, well, in the workplace they have the right to see this stuff. And I think there's a defense to Microsoft saying, yes, they can look at individuals, but that's not a problem with the tool. If it's being done, and there's no evidence it's being done, there's a fear that it could be used for that, that is up to your management to decide. And they could certainly spy on you in lots of worse ways than this if they wanted to. They could push spyware onto your machines and companies have done that in the past. So I think Microsoft's coming in from the standpoint of like, hey, the problem is with your corporate organization if they take advantage of that. This isn't a public tool. This isn't an ad-supported tool. People are starting to use the privacy arguments that they use in the public sphere and they're bringing them over into the work sphere. And I'm not saying Microsoft's right or wrong, but I think that's why they're sort of taken off foot here is they're like, hold on. Yes, that's a thing at msn.com with our public-facing users. We didn't think that was an issue internally because lots of worse stuff can happen already internally. And you have the right to look at this information. So I'll be curious how the productivity score conversation plays out and what, if anything, Microsoft changes. That said, however that plays out, I think the idea of using some machine learning to help improve meetings is a pretty good one. And I also think that the idea of like a, if you were able to fully anonymize this and just able to get a metric that you could trust and that you could compare like across industries and say like, listen, we have a productivity score of six. We're a publisher or whatever. We have a productivity score of six. Other publishers in our vertical, in our company size, in our market cap, whatever you want to call it, at the average is seven. Like what are we, that too is useful information to a business and I think is an interesting avenue. It seems like though anytime Microsoft gives something a score, Windows experience score, productivity score, they seem to run into trouble. Microsoft just don't cry. Right now, anything that you use the word privacy and people freak out and they're like, what's to stop it from invading my home, right? Doesn't matter what logic goes out the window. And when you're scoring people, people get weird, people don't like it. Game of five. All right, well, to finish this off here, last week we discussed a large metal monolith was discovered in the Utah desert, very ominous, lots of internet chatter about that. Now the Bureau of Land Management, the Utah's Bureau of Land Management, says the illegally installed structure has been removed by an unknown party. Google Maps data showed that the object was originally set in place between August 2015 and October 2016. Alas, the monolith will only remain in our hearts from now on. Yeah, so we talked about the fact that somebody used Google Maps and some skills they'd learned from playing the game Gesser to figure out where it was. What we didn't talk about is that people then went and took Instagram photos there and then somebody went and took a photo one night and the next night somebody else showed back up and it was almost gone, the top of it, the prism on the top of the monolith was still there and some of the tools that it or the plates that had been used to put it in the ground were still there. But the rest of the monolith itself, seemingly overnight, had been spirited away. Oh monolith, we will remember you always. All right, shall we check out the mailbag? I think we must, Tom. It is our duty, Solomon Sworn. Sean wrote in and he said, X and 10. They're so frustrating. It occurred to me this morning that in the past two weeks when Robert Herron and Patrick Norton both talked about the LG CX, it is probably more accurately the C10. I have a 2019 LG C9 OLED TV. I was confused when they went to CX thinking it was a new product line. It wasn't until this morning that X and 10 occurred to me. I'm so grateful that Microsoft went with 10 as did Samsung. Switching to Roman numerals is frustrating. It's made worse when it's a one-time thing. Just figured I would share this CX or C10 realization with you all. Yes, Sean. I think you speak for millions when you say this because Robert and Patrick, they know their stuff. So CX must be the common usage. But Sarah Winton looked up and Wirecutter was insisting that it should be C10 because it does follow the C9. But this reminds me of so many other arguments like is it macOS X or is it macOS 10? And it's 10 according to Apple. That doesn't stop people from saying X. And I personally agree that just use X when you mean X unless you're an ancient Roman. Indeed. Indeed. I should note though, this is also very common in the camera industry. There was actually a Venn diagram that came out that basically every major camera maker makes either a model that has X, a model revision that features X or both. So it's virtually impossible to kind of... That's not even going whether that X stands for 10 or whatever. So sometimes people use X and you think it might mean 10 and they're like, no, that's X. So yeah. It's crazy. Well, thanks. And remember, you can always send in any feedback at feedback at DailyTechNewShow.com. Of course, we have to give a shout out to our patrons at our master and grandmaster levels including Philip Sheim, Erwin Stur and Pat Sheeran. Thank you so much for your fine, fine support. Really appreciated. And thank you from the bottom of our hearts. Yeah. And the top. We want to send you a holiday card. Okay. If you're a patron and you've given us your address by December 10th, we will. So that's December 9th. You need to have that address. December 10th is when I'm going to go in and download the database. You can check if you've got your address in Patreon by going to patreon.com slash pledges, finding DTNS, and looking in the right hand column. If you see it there and it's correct, you can leave it. If you don't see it next to DailyTechNewShow, maybe you see it next to other patrons, but not us. You want to click update and add it to us. That is if you want the holiday card. If you don't want the holiday card, that's fine. I don't need your address because I'm not going to send you a card, but I can't send you a card if you don't give me your address. So do that. And it's got some customized Len Peralta art on it. Always a good time. Get that right now if you're not a patron by becoming one at DailyTechNewShow.com slash Patreon. All right, we are live Monday through Friday at 4.30 p.m. Eastern, 2030 UTC. Find out more at DailyTechNewShow.com slash live. We'll be back tomorrow with Dr. Kiki Kiki Sanford. We can't wait. See you then. This show is part of the broadband network. Get more at FrogPants.com.