 This 10th year of Daily Tech News show is made possible by its listeners, thanks to all of you, including Paul Teeson, Ally Sanjabi, and Andrew Bradley. Coming up on DTSS is Instagram, the next Yahoo! Hope to continue Moore's Law, and how we should credit the chat GPT when it writes a journal article, which it is done four times. This is the Daily Tech News for Thursday, January 19th, 2023 in Los Angeles. I'm Tom Merritt. From lovely Cleveland, Ohio, I'm Rich Strafilino. From Austin, Texas, I'm Justin Robert Yogg. And I'm the show's producer, Roger Jay. My friends, it is a wonderful day when you can debate whether machines deserve authorship. Let's start with the quick hits. Amazon alerted customers Wednesday that it is ending Amazon Smile. That is a charity program. It will end by February 20th. Amazon Smile launched in 2013 with Amazon donating 0.5% of the price of eligible purchases to a user's charitable organization of their choice. Side note, it was often suggested as an alternative when you had a referral program active. So Amazon wouldn't have to pay the referral fee they could donate to charity and take a tax benefit. Amazon Smile's website claims more than one million charities have benefited from the program, though the average annual donation is less than $230. Bloomberg's Mark Gurman sources say the darndest things. And today they say Apple is working on new displays and a faster Apple TV to make further inroads in the smart home market after relaunching its larger HomePod speaker Tuesday. The displays could reportedly include a lower-end iPad form factor to control things like thermostats and lights, plus handle video and FaceTime chats, helping Apple better compete with Amazon's Echo Show products and Google's Nest Hub. Couple stories in the cryptocurrency sector to note. The National Bank of Australia is now the second major Australian bank to announce a stablecoin. Stablecoins are digital currencies tied to the value of real currency. So the value does not fluctuate. The stablecoin, if it's, say, tied to the Australian dollar, will always be worth one Australian dollar. The AUDN is planned to launch mid-year and will be used for things like sending money overseas or trading carbon credits. Meanwhile, the new CEO of FTX, John J. Ray III, told the Wall Street Journal he has set up a task force to look into the possibility of restarting the FTX exchange as it goes through the bankruptcy process. The idea would be to restart the business and then sell it to payoff creditors, including existing customers who have holdings in FTX. Ray said that he'll do whatever looks like it'll recover more value. Samsung requested the International Trade Commission open an investigation on the importation of third-party OLED displays for use by third-party and independent repair stores. The company claims the AMOLED displays it creates for mobile devices are protected by patents, something that third-party OLED display makers infringe on. Samsung requested the ITC block the import of these screens and request that several wholesale parts providers, named specifically in the request, like mobile centrics, injured gadgets, and DFW cell phone and parts be ordered to stop importing, selling, and using the displays. The ITC opened an investigation into the import activity on January 4th. Yeah, it's easier to stop the importing than it is to try to shut down the people making the stuff that violates the patents. Earlier this week Twitter announced an $84 annual subscription for Twitter Blue and today the company is extending that offer to Android users. Monthly users would need to pay $11 per month if they buy the Twitter Blue plan through Android, same as on iOS, but $8 if you purchase it through the web app. The new Twitter Blue plans are available in the United States, UK, Canada, Australia, New Zealand and Japan. All right. Y'all use Instagram, right? Justin use Instagram. Oh, yeah. I'm a big gram head. Rich, you on the gram? Yeah, I'm an IG fan. Okay. Instagram launched a new feature called Quiet Mode that silences notifications. Auto replies to anybody who DMs you while you're in quiet mode and let your contacts know, hey, you're in quiet mode. They're not ignoring you. It's you're just in quiet mode. Instagram says it will actually prompt teenage users to enable the feature at later hours in the day. Quiet mode is launching first in Australia, Canada, Ireland, New Zealand, the UK and the US with more countries to follow. Instagram also announced some added controls to control your time spent on the app. Take a break reminders and tools that let you pause, snooze and unfollow pages, groups and people. You might ask, why are they doing this? Well, one reason to stay ahead of regulators. A lot of regulators are looking at whether social media is harmful. So showing that you're doing things to reduce harm, especially for teens is going to help you with the regulators. Also because some Instagram users not responding to messages is considered rude. So a system that explains why you're not responding will help you to feel like you don't have to respond and not feel burdened to use the app. Seems obvious to me though that Instagram, like many social networks, is in a fight to keep users from getting exhausted by the app they're using and quitting. So if you help them wean themselves, it's better than quitting altogether. I think they're going to quit altogether eventually. My premise is we're in the valley of social media. There's been a lot of talk about the declining use of Facebook. I feel like these things are indications that Instagram's trying to fend that off, but Instagram might be Yahoo in a world where Yahoo was about to meet Facebook for the first time. The next phase of technology platforms is brewing. Look at Be Real, which is saying, hey, we'll be like Instagram without the downsides. Ubo, Gas and Slay all saying, let's help teens have positive interactions with each other. The thing they all share is designing themselves so that you don't get exhausted using them. Justin, am I crazy to think of this metaphor? Your comparison to the side of the road, it's the metaphor police. No, in all seriousness, I think that your metaphor spiritually is correct in that Instagram and Facebook are long in the tooth. They are ripe for something that connects more with the younger audience for which these platforms were not designed to connect with. I do think that specifically the lesson that happened with Yahoo versus Google was that Yahoo allowed their product to degrade to such a significant capacity that Google's, while doing the exact same thing, was just markedly better. You could find more stuff with less ads that were cluttering at the top on Google than Yahoo. That to me is where the prescient conversation is. That being said, I do think that we are in a sea change era for social media because the landscape for which all of these were designed is rapidly changing. Ads are not selling for the same costs that they were before. The idea of maximizing the real estate to sell those very lucrative ads now is less than it was before. So, customer retention in a world where not every platform is on a rocket ship to Mars is something that they have to think more about along with what you said, trying to stay ahead of regulators, specifically in Facebook and Instagram's case, after they had a whistleblower that got a lot of attention in Congress. Rich, how are you feeling about this metaphor? I mean, I'm on record as saying Facebook is, or Metta, excuse me, is Yahooing extremely hard. So, in my heart, I want to agree with it. What's interesting about this is, have we seen this shift to digital wellness outside of any ad selling customer retention considerations a couple years ago from both Google and Apple kind of doing this on a platform level? And really, yeah, sure, you click on the I'm not driving so you can touch your phone while you're in the car. But outside of that, I feel like a lot of those features didn't really catch on in a big way that a lot of these companies touted them for. So, I'm very cautious whenever I see these kind of things. Again, Instagram kind of grafting on something else that if you squint kind of resembles up and coming social networks is really going to catch on, especially given that, yeah, people might be getting burned out, but they also have very established use cases for these apps. And I don't know, it seems like adding speed bumps, the utility would have to be extremely high to get anyone to change behavior on these apps. I'm always skeptical about that. Yeah, I mean, taking a page from what you were saying, Justin, about Google and how they succeeded better than Yahoo, if you were looking at Instagram and thinking they were going to be able to survive, they should be doing things that gas and Ubo and Breal are doing instead of just doing what TikTok is doing. And maybe they will. It's I think what struck me about Quiet Mode is it feels like, look, we know you don't want to use our app, but let's help you use it less. And then maybe it won't be so painful. And that feels like a bad product proposition. No, I disagree. Look, Instagram would be ripping off Breal and Ubo and gas if it had the same kind of upward trajectory that TikTok does. That's the reason why they ripped off that. That's the reason why they ripped off Snapchat because they wanted to make sure that they were not missing out on a certain categorical trend that was possibly threatening them because there was lightning acceleration upwards for those platforms. While they may be the future and I'm not saying that they're not, I do think that Instagram is trying to give creature comforts and specifically trying to stem any tide of parents taking this app off their children's phones. I do agree with you that they are taking away one very specific element of brand loyalty, which is that kids and maybe less now than they did maybe 10 years ago use Instagram as their primary messaging app, the primary way that they talk to their friends. Same with Snapchat. So if they're saying do that less, then you are by definition creating less brand loyalty. That being said, I don't think that it's the kids that want to be on the app less. It's the parents that are doing it. This is a gift to them. That's interesting because there's lightning and there's lightning, right? You can argue that Ubo Gas, Slay and Be Real all have had lightning growth. I think Slay got like 17 million users in a month, but that's still dwarfed by what Snapchat eventually did, what Tic-Tac did. And it's for teens at Instagram going after teens with a feature specifically in the case of Ubo Gas and Slay would get them in hot water because it's a feature in those cases limited to teens. And Instagram, anytime it does something with teens has to navigate a bunch of people saying you shouldn't be targeting teens. Including Instagram parents who grew up on Instagram. Yes, which is interesting to think of now. Yeah. All right. Let's change topics to Moore's Law. Justin, what is Moore's Law? I'm glad you asked, Sherman. Moore's Law is usually the mean of doubling the transistors in a microchip every couple of years, leading to continued increases in computing power. And like podcasting, it's been declared dead many times. However, the latest roadblock does look bad. And this could be the end for real. We're making one nanometer chips, but silicon atoms are 0.2 nanometers wide. So you can't make them smaller than that. And even before you get that small, some electrical properties break down. But there's always a group of scientists somewhere, these beautiful boffins who say that they can save Moore's Law. One such group at MIT thinks that they've done it thanks to 2D crystalline wafers. You know what the heck are 2D crystalline wafers, Rich? Well, but let's just head off the emails. They aren't technically 2D, but they are a single crystal material that are only one atom six. So as close to 2D as you can get outside of like a flatland platonic situation, which I don't think scientists are working on, they conduct electrons more efficiently than silicon. So you get faster processing, which if you're Moore's Law, hey, that sounds great. The challenge though has been how to put 2D crystalline materials onto standard silicon wafers. There are a few ways you can do it. You could flake it off bigger material, which takes a lot of time, especially since not each flake is one atom thick. The whole way through, it's extremely manual process, really hard to industrialize that. Another way is to grow the crystals, which has been done on sapphire, but there's no existing infrastructure for sapphire-based wafers. Preferably, you would grow them on a silicon wafer, since that's what chip makers use. But when you try that, the crystals grow in random orientations. There's always a gotcha. That causes something called grain boundaries. The crystals on silicon don't all grow in the same orientation. So when the grains bump up against each other, it acts as an electrical barrier. Not great for all of that processing. But those clever MIT folks found a way to stop that. Yeah. You got to make your crystals get right, line up right. The MIT team coated the silicon with silicon dioxide, which they then patterned as they put down the coating with tiny pockets that can trap the crystal seeds. When you spread the crystal material gas across the wafer, the atoms settle right into the pockets in a nice, orderly fashion to create a one atom thick crystal with none of that grain problem. They were also able to use two different types of crystal and material to create two layers that were still each one atom thick. That would allow you to do some flexible materials and other things. Showing this method works with silicon means that existing manufacturers can now adapt it, if they want, to make more efficient chips and giving this technology one of the best shots at being able to keep Moore's law going. Leaving us with the question of, well, right now they all claim one nanometer, two nanometer, four nanometer process. What crazy branding word would they come up for crystalline wafers? We'll have to wait and see. The other thing that's really interesting is we're kind of, I've been seeing a little bit more coverage of kind of these ultra-thin, like atom thick, single, yeah, single, yeah, thick materials. In fact, on our weekend edition of DTNS where me and Amos cover photo news, we were covering a photo sensor that was made out of molydenum disulfide, create a basically a sub one megapixel camera sensor kind of out of the same thing. So obviously a different material, it doesn't address the Moore's law use case, but definitely something that we're seeing a growing interest in in science that is super exciting. So, folks, if you want a good band name or just to know what the MIT people are doing with this single crystalline growth, non-apotaxial single crystalline growth, opening at Amos this year. Your first album is Crane Boundary. Rolls off the touch. Yeah. All right, folks, if you want to join in the conversation, maybe you're like, I don't know about non-apotaxial. I've got another single crystalline growth I prefer. Well, come on over to our Discord. You can join that by linking to a Patreon account at patreon.com slash DTNS. Nature has a news article, Nature the Publication, but not the peer-reviewed part, the not the peer-reviewed scientific paper part. This is the news gathering part of Nature, has an article called ChatGPT listed as author on research papers. Many scientists disapprove. I imagine them frowning. It notes that a large language model called ChatGPT, we've covered it a bunch on the show, although in one case it's GPT-3 getting credited, not ChatGPT, but in four cases, either ChatGPT or GPT-3 has been credited as an author on a published or pre-print published paper. So let's run through the four examples. First, Jerry, start us off. A pre-print paper, one not yet published, on MedRVix about using ChatGPT for medical education list chat GPT as the third of 12 authors on the paper. One of the press's co-founders told Nature that ChatGPT's author credit was sneaked in, quote, much as people have listed pets, fictional people, et cetera, as authors to journal articles in the past. But that's a checking issue rather than a policy issue, unquote. Yeah, then there was ChatGPT is one of two co-authors on an editorial published in Nurse Education in Practice. The journal's editor-in-chief told Nature the credit was made in error and would be corrected. Oh, for two. ChatGPT-3, the broader platform behind ChatGPT is one of the three authors in an article called, Can GPT-3 Write an Academic Paper Buy Itself with Minimal Human Input Question Mark? That article will soon be published in a journal that accepts GPT-3 as a co-author. ChatGPT is credited, along with Alex Zavorankov, for a perspective piece in Onsko Science on the pros and cons of the drug Rapa Mycin. It was done as a demonstration of the capabilities of ChatGPT, but it was also peer reviewed. Zavorankov told Nature he tried to use ChatGPT to write a more technical paper but said, and this is a quote, it does very often return the statements that are not necessarily true. And if you ask it several times the same question, it will give you different answers. All right. So most publishers told Nature that a generative algorithm like ChatGPT does not qualify for an author credit. In the one case where it's going to get the credit, the GPT-3 one, they didn't respond. They didn't explain themselves to Nature. Holden Thorpe's editor and chief told Nature that AI cannot be credited as an author, but if it is used it must be cited or it would be considered plagiarism. Taylor and Francis' director of publishing ethics and integrity believes the use of large language models should be cited in the acknowledgments section. So you don't get an author credit, but you say in the acknowledgments, hey, we use ChatGPT-3 for this part, that part, etc. What do you folks think? How should we credit large language models or LLMs or should we credit them at all, Justin? You know, there was once a time that we had integrity in this country where hard-working men and women would copy edit for magazines and journals and newspapers. And then like John Henry falling to Inky Poo, it was indeed the machines that came for those people's jobs and the dreaded spell and grammar check removed the ability for them to be the sole arbiters of grammar and spelling. And that's why we should always list the word processor that you wrote your own in every single author acknowledgement. Obviously ChatGPT and GPT-3 are steps beyond that, but in reality the only thing that this indicates if these journal articles went without detection is how robotic and monotonous most journal writing is and the fact that it could be replicated by the by and large, very simplistic writing style of ChatGPT-3. You know, I think the acknowledgement way probably is going in the immediate future, right? Because I do agree that this kind of generative AI is just going to be rolled into so many things, whether it's ChatGPT creating stuff wholesale or something like an advanced version of Grammarly, which is already out there, that kind of gives you suggestions, oh, this will firm up your argument or something like that. I think in the near term, the acknowledgement is the way to go before that kind of fades into the technological background, especially because if there is a situation where you're asking, hey, ChatGPT, write me out, you know, the pros and cons of the drug Rapamyacin and something that you had not thought of, like almost like a consultation, right? You're consulting with a colleague on this that brings you on a new train of thought not necessarily because this can't write on a technical level that it would be able to necessarily be credited for, you know, original writing. I do think there is a way that you could use this conversation to be like, hey, this did actually help inform my thinking on this kind of going, having something to go back and forth, even if it's effectively, you know, an AI black box wall that you're bouncing a tennis ball out of. Like, I do think as an acknowledgement, I think that's valuable, even if it's only to show, hey, these can be valuable tools for thinking about these kind of things or as kind of, I guess, an interactive reference, even if nothing else. I go back and forth on this because on the one hand, I think we're over anthropomorphizing this. I think putting chat GPT in your author field is a stunt. I tend to on the gut level think, yeah, throw it in the acknowledgements, let people know, hey, I use this tool because this tool is so new and because that's interesting to people and because people may say, oh, that that changes how I think about this, especially these ones where you're like, I asked chat GPT to write the whole article. I mean, sure, okay, let's acknowledge it. It's kind of obvious from the entire article that you used it. But the more we talked here, Justin, you kind of persuaded me. I feel like if the whole reason you don't give chat GPT an authorship, according to these publications is that the author has to be responsible for what they say, then the person who used chat GPT is responsible for what it says. It's responsible for making sure it's accurate, right? It's why Alex Zhavoronkov did not use it for a more technical paper because it wasn't accurate enough. Then does it matter if they use chat GPT? They're the ones responsible. They take responsibility for what they used. If they used a calculator, a word processor or a large language model to produce it, it's their words at that point. Yes. Yes. And I don't think that there's any a benefit for anybody writing these articles. I think if anything you'll be looked down upon for listing it, and people will assume that you got a lot more help than you did, anybody who's actually used these models understands that it writes like a freshman on Adderall who had two hours to finish an assignment. It is not great writing. It might be helpful to stimulate your brain so you can change it and you can make it better, but it is not at least now a standalone solution. I will very quickly play the devil's advocate and that I don't think it should be required to be in the acknowledgments, but I do think that the fact that it would be listed in there could be helpful to get over necessarily that stigma of if this can be used. I'm not saying it can necessarily right now be used, but if it can be used as a productive way to augment your writing, like again, I think long term this all fades into the background, but in the short term, I actually do think there could be some value if you are getting value out of it. If you're just doing it because you'll know this get picked up by a couple blogs, you'll get your name out there, then it is a silly start. Well, I will say that if we list that technology, then we should also list other underlying technologies that help make that paper what it is, including food, sunlight, and sleep. Yeah, I think it's whoever's taking responsibility should get noted so you can go to them. I don't know that it fundamentally changes things if I know that ChatGPT was used, but it's Bob or Pat that I have to go to if something was wrong, and it doesn't matter how they got there, right? I guess maybe methodology. It should go in methodology as, oh, as part of the method of creating this paper, the way I describe the tools I use to the microscopes and all that, maybe it belongs there. But if you're looking in the acknowledgments of a book, every time you'll read that, oh, I really am appreciative of this library for being this fantastic body of metal. There are ways to acknowledge stuff and materialize it as they're still taking responsibility for their words, but is that just pure artistic flourish? A little bit, sure. No, it's fair. I don't think we're 100% settled on this. I think some amount of info first, a certain amount of time though, is going to be necessary. Breaking news, by the way, Sarah Lane letting us know this in our Discord. Thank you, Sarah, for being on the ball all the time. Netflix, Reed Hastings, stepping down as co-CEO. He will become Executive Chairman, so he remains at the company, and Greg Peters, who is now COO, is promoted to the role of co-CEO with Ted Serendos. Wow. That's a biggie. It's a biggie. I hesitate to analyze why it happened, because they just announced it along with their earnings, and Netflix beat Q4 subscriber growth target, so mixed financial results, but it doesn't sound like the earnings were overall horrible. Maybe Reed Hastings is like, hey, I want to go spend more time making solo albums. Maybe. Maybe it's a new project. This is a narrative transition for a thing, a lot of technology, and Reed Hastings at this point would be able to do whatever he wanted. I've noticed Serendos fielding more opportunities and more questions than Hastings lately, so yeah, maybe. There's also, we are entering into an era of consolidation for a lot of these services, so I also wonder, and who knows whether or not Netflix is going to be a buyer. There's been rumors that maybe, especially with Iger coming back, you could be a seller to somebody like Disney, so interesting. A lot of hints being sprinkled. I can't believe they canceled another one. Yeah, Reed Hastings added to the other list of canceled Netflix. When I'm renewed. All right, Rich, let's get into fonts for a quick moment. Yes, please. Let's get into something we can all agree are amazing. Fonts, US Secretary of State, Anthony Blinken sent a cable to all US embassies directing the State Department to shift from sending documents in Times New Roman to the San Serif typeface Calibri, specifically 14-point Calibri, as of February 6th. The Washington Post John Hudson reports that the shift is due to the fact that typefaces with serifs, those little flourishes on your extenders, your terminals, your, I'm sorry, your ascenders and descenders and terminals, like Times New Roman, introduce accessibility issues for people who use assistive technology, things like screen readers. Times New Roman has had a good run at the State Department. It became the standard typeface for memos in 2004. No word on how Helvetica, though, is taking the snub. The Post quotes London designer Jack Luhlen, who cautions, there is no one-size-fits-all typeface as an accessibility solution. And a senior research associate at the Engineering Design Center at the University of Cambridge, Ian Hosking, said that spacing between lines and contrast and color between text and background make a bigger difference than just the change in typeface. Still, though, for accessibility, I'm a laudable move. Why not comedy sans? I mean, look, we just read the back part of that. Is it really about accessibility or is this font nerds font-nerding? Well, no. So on background, on background, many staffers are not, are very upset that Calibri. It lacks a certain elegance. Because font nerds will get mad at the slightest things. And serifs versus sans serifs are effectively a holy war to certain people. Well, and to add some extra intrigue to this, so the State Department said one of the reasons they just chose Calibri is because it's the default on Microsoft Office. Well, last year, 2021, excuse me, two years ago, Microsoft announced that they are actually going to be transitioning away from Calibri. They have like five candidate fonts that they're testing out right now. So we may be covering the change to one of these other fonts. I look forward to the house investigation. And what typeface they will use to print the documents for the house investigation. Feedback at DailyTechNewsShow.com in whatever typeface you want to use. Send us your thoughts, font nerds and others. Thank you, Justin, Robert Young for being with us today. I know you got lots of cool irons in the fire. What you got going on? I was about to do a total separate episode for Friday's Politics Politics Politics. And then I read one story in Politico that blew my mind. And the rabbit hole that it set me down is what the Friday episode is going to be about. Praz, the third member of the Fuji's, next to Lauren Hill and Wyclef Jean, is implicated in a multimillion-dollar attempt to effectively bribe not one but two presidential administrations to drop an investigation into the largest theft of money in world history. Billions of dollars stolen from the Malaysian Sovereign Wealth Fund. It implicates the Chinese communist government of various different United States banks. It's the reason why the Fuji's tour celebrating the 25th anniversary of the score got canceled in 2021. If you want all the details, go ahead and subscribe to Politics Politics Politics right now and get the Friday episode when that goes live on Friday. But 100% worth it just to understand why the Fuji's tour got canceled, right? A load. And then... It's baffling, including the fact that the money that also went to bribe two presidential administrations funded The Wolf of Wall Street, a movie about financial fraud. Beautiful. I was going to say, you cannot write a script about that we'll get funded by the same thing that caught... Anyway, yes, there's a metaphor there too. It's great. Well, another thing you can't make up is our thanks and appreciation for our brand new boss, Graham, who just started backing us on Patreon. Thank you, Graham. Graham's the best. And if you are not a patron yet, you could be the best right now. Just go to patreon.com slash DTNS because you're going to get an extended show. We're going to keep it going. Talk about that Netflix news a little more, I think. You can also catch the show live Monday through Friday, 4 p.m. Eastern 2100 UTC. Find out more at dailytechnewshow.com slash live back tomorrow with Chris Ashley and Lenny Peralta and talk to you then. This show is part of the Frog Pants Network. Get more at frogpants.com. Diamond Club hopes you have enjoyed this program.