 Daily Tech News show is made possible by its listeners, thanks to all of you, including Tim Deputy, Brandon Brooks, and Hector Bones. Coming up on DTNS, guests are apparently paying to be on podcasts. Not us, but some podcasts. Are we okay with that? Plus, MIT scientists discover why your network congestion is bad and the things you need to know to navigate the continuing chip shortage. This is the Daily Tech News for August 4th, 2022. Today, maybe Thursday. I don't know. In Los Angeles, I'm Tom Merritt. From Studio Redwood, I'm Sarah Lane. It is Thursday. Deep in the heart of Texas, I'm Justin Robert Young. And I'm the show's producer on this Thursday, Roger Chang. The Thursdayest of Thursdays. Let's start with a few tech things you should know. Back in 2019, Google claimed quantum supremacy after its quantum computer called Sycamore performed a computation in 200 seconds that the company said would otherwise take a classic computer 10,000 years to complete. Wow. Sounds crazy, right? Well, now, Pan Zhang, a statistical physicist at the Institute of Theoretical Physics at the Chinese Academy of Sciences and colleagues have shown how to do that same computation in a few hours and believe that with a supercomputer, not necessarily a quantum computer, they could do it faster than how fast Sycamore did it. In a paper printed at physical review letters, they described how they recast to the problem as a 3D mathematical array called a tensor network, simulating the quantum computers method. Sycamore still required fewer operations and less power than a classical computer would. But Zhang suggests that quantum computing needs to find real-world applications to demonstrate quantum advantage. China's Alibaba reported revenue did not grow year over year in Q1 for the first time ever, but it didn't grow, but it didn't fall. An analyst expected it to fall. Alibaba attributed COVID lockdowns to the problem, but noted that recovery appearing in June is probably going to turn things around for them. Retail sales fell 1%, cloud computing rose 10%, and speaking of cloud computing and related news, Salesforce announced it was shutting down its offices in Hong Kong and accelerating its partnership with Alibaba, the exclusive provider of Salesforce software in greater China. Folks in South Korea, Japan, Hong Kong, and Taiwan can now reserve Valve's Steam Deck through the website Komodo. Valve says the first batch of reservations in these markets will be fulfilled by the end of the year, and Valve doesn't expect the addition of these markets to affect the delivery estimates for those who are already on hold with an existing reservation. Meta CEO Mark Zuckerberg announced international expansion of NFT support on Instagram following Instagram's initial NFT test launch in May with select creators in the US, users and businesses in more than 100 countries in Africa, Asia-Pacific, the Middle East, and the Americas can now share their NFTs on Instagram. Meta also announced that Coinbase Wallet and Dapper Wallet are now accepted as third-party wallet-compatible for use, and Instagram is also expanding its supported blockchains to include Flow. What if you don't own an NFT, though, Tom? Well, get in there and get one. I guess I should, yeah. Board Ape Club. Good time to buy. Why is market? It would be indeed. Reuters reports that Tencent, which acquired a 5% stake in Ubisoft in 2018, is reportedly trying to buy an additional 16% stake held by the Guillemot family who co-founded and still run Ubisoft. This would make Tencent Ubisoft's single largest shareholder. You might say, well, okay, I get where they're going with that. The other 80% of Ubisoft is publicly owned, and Tencent is reportedly looking to acquire shares in that pool as well. All right. Let's talk a little bit about podcasts. You're listening to one. You probably like them. Lately, we here at DTNS have been getting a more frequent kind of spam. Somebody asking us how much would it cost to be a guest on your show? To be clear, nothing. DTNS never gets paid to put someone on the show. That's a policy we have. We pay people, our contributors, those people that you see on our About Us page, and our guests come on for free. Our guests come on and we let them plug whatever they're doing. But we do not pay people. We do not get paid to put people on the show. Now, Bloomberg's Ashley Karman has an article up that says podcast guests are being paid up to $50,000 to appear on popular shows. Karman says the practice is particularly common in the wellness, cryptocurrency, and business arenas. A platform called Guestio, brokers paid guest appearances claiming podcasters have been paid $300,000 since 2020. Sometimes they are people paying to be guests, but Guestio also brokers podcasts paying the guests to appear. For example, Boxer Manny Pacquiao charges $15,000 for an appearance. Okay, that's fair for Manny Pacquiao to charge for that. And a lot of this is not different from what creators on YouTube and TikTok do. They get paid to talk about a product. The difference is those platforms already went through the big controversy over disclosure. Now, when somebody's on YouTube paid to talk about something, they disclose it. They say, hey, this company sent me this thing, or this company is sponsoring this thing where I'm talking about it. Podcasting, though older, has not gone through that controversy. And disclosures vary depending on the episode. Sometimes they're clear, sometimes they're cleverly hid, and sometimes they're nonexistence. The USFTC says the same rules that apply to every platform like YouTube and TikTok apply to podcasting as well. You have to be clear about the source of a message, including whether somebody is paying to appear. So you might say, well, okay, at the point that it's disclosed, is it still a bad thing? And, you know, people differ in their opinions on this. John Lee Dumas, host of Entrepreneurs on Fire, told Bloomberg that guests who pay are better prepared, explaining they showed up on time, they delivered massive value, they had great giveaways and calls to action from my audience, and also the paid appearances are often part of an ad buy, and that can bring extra revenue to a show. Now, there's a difference between news and entertainment shows. For a show like we do, it's just not done to accept money for someone to show up, because you want to be clear that the people you are talking to don't have an interest in coloring their information one way or the other. That's a time-honored news tradition. NBC Nightly News, BBC World News, they generally, well, there are some exceptions, but they generally are not going to get paid to have somebody on the show. When you're talking about something that's a little less newsy, maybe it's more lifestyle-oriented, more independent, you know, less worried with, like, we want to try to get down to the facts, the lines start blurring. Justin, is there a case where, you know, like John Lee Dumas, where it's like, yeah, as long as we're clear about it, let's assume everybody's disclosing everything, maybe it's fine. I think it's totally fine. I think that we have a weird, a puritanical view of money as a barrier or preventer of authenticity, that it's going to be the antithesis of it. The reality for any kind of show that books regular guests, and I do think that if GTNS was a show that demanded a brand-new guest once every five days, you would eventually find yourself in a situation like I do on PX3, where I book a guest twice a week, where every once in a while the pile of unsolicited PR agents who are pitching you guests at all times, which I get, I don't have a gigantic show, but I get probably 30 pitches a day for guests, that every once in a while, you pick one. I had a guy, Evan Scribshaw, who is a betting, a political betting expert, his website, The Lines, paid a PR agent so they could reach out to people like me. If the money could go to me instead of the PR agent and I still like the guests, I don't know why that's bad. Yeah, I think the disclosure is really important and disclosures happen in many ways, especially in our new modern world. I think of, I don't know, if Future is paid to walk through a club in Atlanta for an hour and that he got paid to do it and gets his picture taken and people take pictures with him, whatever. It's like, okay, how different is this than that? And I think to your point, Tom, it depends on what you're talking about on the show. If I know that the future then gets paid to get on the podcast Drink Champs and kind of talk about the music industry and have some fun, that doesn't bother me. If he was on our show to talk about the future of crypto because he ends up being an investor in a crypto company and that's not disclosed in a way that makes us feel comfortable, that's a totally different story. Yeah, I think it has to do with, are you able to do what you're saying you're doing, right? And if I just want to like talk to a fun guy and they pay to be on the show, then sure, that might be fine. If I'm promising to you as the listener, I'm going to get to the bottom of what this person says and tell you whether it's true or not and they're paying me, well, I'm less incentivized. But those aren't the guests that, you know, according to the article, are really getting picked up. You're looking for expertises in areas that might be interesting or illuminating, right? And that's where the cryptocurrency and the wellness stuff is. Is you are looking for people that are trying to think differently and they're pitching themselves with those ideas. So you are effectively asking somebody to explain their worldview and, you know, greases the wheels a little bit. Yeah. If a new she gets to where to speak. Now, you should know when those guests come on to pitch their wellness products that the host, if the person is paying to be on, is probably not going to be as critical of them. And that's fine. You just should know this is going to be a friendly interview because they're paying to be on the show. I mean, maybe, but I guess it's more likely. It's more likely to be friendly. It kind of depends on if the podcast host or the guest is the bigger celebrity. But also it's like, look, in reality, not every interview is Frost Nixon. You know, you're not trying to get something out of somebody. Right. That's what I mean. I don't think that torches the value of it, but you should know like, oh, they may not have asked all the hard questions. And maybe that's fine. Maybe that doesn't even matter. Maybe I'm just curious what the thing is at all, which case you probably got it. Unless it's investment advice or, you know, it starts to get. Yeah. Well, let's talk about network congestion, shall we? Boy. I know network congestion is not fought by bandwidth caps, although some ISPs will tell you otherwise instead by algorithm. So your computer doesn't know the optimal speed to send packets. Your computer does its best, but maybe the packets are going too slow. Then you're not making the best use of the network. Maybe they're too fast and you might instead overwhelm the network and that causes dropped packets. And that causes resends and overloads the whole system even more. So what to do? Congestion algorithms look at how many packets are getting dropped and need to be resented. That's used as a signal of the network's overall health. And the algorithm adjusts to how fast a computer sends data packets so that it optimally uses a network's capacity. But this doesn't always help. Yeah. The folks at MIT wanted to find out what the best algorithm for network congestion was. So they studied it and put out a paper called Starvation in End to End Congestion Control, showing that while these algorithms do help overall to ease congestion over the entire network, there will always, and this is the weird thing, in every algorithm they tested, always be a scenario where at least one sender receives no bandwidth when congestion is being managed. If you've ever been sitting next to somebody with the exact same cell provider as you, you both have the same number of bars, but their connection works and yours doesn't, there are multiple explanations for that. But one of them could be this, that this algorithm is like, you're the one that gets to lose so that everybody else gets better bandwidth. Now, the problem is because of packet delays being the measure. Packet delays aren't always caused by congestion. Sometimes they're caused by other things like delays in acknowledgement from the sending source or intentional packet queues that delay it on purpose because they're queuing up stuff. Stuff like this is often lumped under the term jitter. If you know the term jitter, that's what we're talking about. Algorithms can't tell the difference between congestion and jitter. Jitter causes the algorithms to make poor decisions and eventually that leads to the starving of somebody's bandwidth, albeit unintentionally. So the authors of the study figured, well, let's just test these algorithms and find out which one doesn't cause the problem. And that's what they did, right? Yeah, yes, they did. They created a computer model that accounted for it. Jitter and tested every congestion algorithm that they could find. Every one of them was delay convergent. So in other words, at some point, somebody's bandwidth got starved. If that was you, you're probably mad that day. The only solution they found was to allow for more variance in delay, meaning not combating delay quite as strictly. The more variance they allowed, the less starvation overall. They hope to find an algorithm that can eliminate it entirely. So that's what we're working on at this point. The team will present their research at the ACM Special Interest Group on Data Communications Conference. Yeah, this is not going to have immediate practical effect on you. It's certainly not the only reason that you ever have bandwidth go out. There also might be a capacity problem on the pole near you if you're on 4G service or something like that. But it is a great way to help actually fix the problem of bandwidth congestion rather than some marketing oriented way like bandwidth caps. I think it's also just a great thing to know. You know, my wife and I were in Europe for the past three weeks and this happened constantly for us that we would both have the exact same phone, the exact same reception, the exact same network. And yet one of us was able to post the vacation picture to Instagram and the other one wasn't. And luckily it was usually me that won, if she just had a very algorithmically unlucky phone apparently. You know, it's a... Like I said, it could also be a capacity where you always got connected to the nearest cell service pole before she did and there just wasn't a good connection for her. It's funny that it does tend to be one person over the other. Eileen always gets connected when we're places and I don't. It's just the luck of the draw. Or is it the jitter? You're a little jitter bug, Tom. I am. I am. Well, folks, if you're like, who's Eileen? Eileen's my wife, if you don't know. And she's going to be on the show next week. All next week is DTNS Experiment Week. We're swapping out our normal DTNS shows and we're trying out some new ideas. Last year we launched shows like Barbecue and Tech, The Tech John on Experiment Week and they went on to become their own shows. This year we've got Rob Dunwood and Rod Simmons with DTNS Reaction to Samsung's upcoming fold announcement. Nicole Lee will be doing a tech culture show from an Asian American perspective and that's the one my wife will be on. It all starts next week, Monday, August 8th, right here on the DTNS Feeds. Previously on The Chip Shortage. When lockdown swept the world, companies bet that folks would stop buying as many things, especially cars. As it turns out, folks bought a lot more things and returned to buying cars a lot faster than expected. The disruption to the normal patterns of shipping caused all kinds of problems and logistics, like too many cargo containers in one port and not enough in another. Shortages in components cause companies to try and stockpile parts buying more than they normally would, a.k.a. the toilet paper effect. Weather anomalies like the ice storms in Texas, droughts in Taiwan and fires in Japan complicated an already difficult situation. And then of course, the war started in Ukraine, disrupting supply chains even more. Inflation started changing consumption patterns again, which brings us to now, Tom, where are we? Companies are dealing with uneven component supplies, but full warehouses, hey, Best Buy and Target both said their warehouses are brimming, but they're expecting dwindling consumer demand and they're not the only one. Let's focus on two cases that point to where the chip market is going to be heading. First, automakers are going to be more involved in chip design, meaning better electronics in your car, but also higher costs. TSMC's CEO, C.C. Wei, told Reuters that he never once got a call from an auto executive until the shortage got bad. And then they were his best friends. Now car makers and chip makers are working side by side. This gives automakers more visibility into the supply chain, but it's going to cost them because they're participating. They're not just taking, they're not just placing orders. They're involved in the R&D process. Emergency measures that they put in place to deal with the shortage are becoming permanent. So parts suppliers like Bosch and Denso are spending an increasing amount of chip production, increasing amount on chip production. Companies like GM and Stellantis are taking part in chip design. Ford has struck a new deal with Global Foundries directly. Chipmaker called Skywater Tech is asking automakers to buy the equipment they're using to make the chips or pay for some of the R&D. Japan's Renesys Electronics and Dutch NXP Semiconductor are co-locating engineers to help automakers design new architecture from the ground up. That's going to give you better efficiency in your car. It's going to have a single chip that can control everything. It brings them closer to what Tesla's already been doing. Tesla already designs its own core chips, and it means there's going to be more chips in your cars. Reuters says the number of chips per car in 2026 is expected to be double what it was in 2020. So automakers have reached a new normal of their own with better but more expensive chip design. Sarah, tell us about case number two. Oh, Tom, I'm glad to do so. Reuters reports on a case that seems like a more straightforward effect. Consumer demand has dropped, leaving inventories to stack up. Reuters also reports that Samsung scaled back production in Vietnam where it has six factories and two major hubs that put out around half of Samsung's smartphones. So they're doing a lot of inventory there. One campus in the north of the country can make around 100 million devices per year out of Samsung's total of 270 million. A worker at that campus told Reuters they're not only cutting overtime for workers but going down from six-day work weeks to four, sometimes three. Samsung told Reuters it isn't reducing its annual production target, but Reuters couldn't confirm if that meant it was shifting output to its Indian or South Korean plants or perhaps both. Managers told workers at the plant that the inventories are high, hence the reduction in output. So it may be that Samsung suddenly realizes it just has enough units to meet demand and demand is stagnant. Listeners to DTS know that you heard about the chip shortage early here. You probably heard about it before a lot of your friends. The takeaway on this story that we're doing here is that we may be headed for a glut of electronics in some sectors and higher costs in others, and sometimes you're going to have both. I think that this makes a lot of sense and probably is going to work out pretty well, at least for consumers in the areas where there are gluts, especially if we do go into a recession and people need to stretch their dollars even further than they already have with inflation. But what I'm fascinated about is the cars. Cars have certainly become more and more and more dependent on electronics, be it in cameras or various different displays. But to really double down on that, I think is fascinating. Yeah, and I think the silver lining for cars here is because they were forced to do it in order to get chips at all, to figure out how to make chips and to navigate the cost of them, I think you're going to see better designs. And if you've ever sat in a brand new car, I know not many of us do very often, but if you ever have and you're like, this is an impressive car, but the electronics seem kind of old. That's why carmakers just bought off the shelf old chips for the most part. Like we need a display controller. It doesn't have to be fancy. Now they are like literally co-locating engineers to be let's design the cabin electronics from the ground up. Let's design a chip that works the best for what we do and that way we can put a long-term order on it and make sure we have a supply of it. I mean, I will say as somebody who purchased a car over the last two years, more than I would like to admit because of how the electronics looked and functioned, I think this is a good idea for the modern consumer. Yeah, it's one of the reasons Tesla's are so impressive because they've done that work from the ground up. Well, speaking of the modern world, it seems like just a couple of weeks ago we were talking with Joel Telling, who was a guest on our show about the world of 3D printing. Where it's going? Because it was two weeks ago. Protocol has a great write up on how Hollywood is also embracing 3D printing in an interview with Dream Smith studio founder at Jacob Sniman, who tells the publication, he's wanted to use 3D printers to design props digitally instead of with clay and wire mesh for years. But the technology just wasn't ready for a long time. The printers he could afford weren't able to make the big enough props that he actually needed to make. Today's 3D printing is alive and well in Hollywood, however, with larger printers made by companies like Form Labs, there are others that are now in the budget for shops like Dream Smith. Sniman has been using Form Labs as 3D printers for props on both seasons of the HBO Max sci-fi series Raised by Wolves, RIP Raised by Wolves. It has been canceled, but needed a lot of 3D printing to make stuff come to life. For example, instead of creating a silicone mask of a human's face, that requires the actor to be there in person. It's cumbersome and time-consuming. 3D capture studios can make a digital mold and then the folks at Dream Smith can use those files to make molds that are accurate going forward. When I first saw this story, I thought it was going to be about props more than about modifications to the face, you know, molds and masks and things. So this is fascinating that instead of having to sit in a chair every time, and obviously they have to sit in a chair to get the mask applied, but instead of having to sit in a chair for each thing, they can just do a scan and have it on file. It makes perfect sense. And then 3D print whatever the mask is before they make them sit in the chair and put it on. And gives you such freedom for digital design. Yeah, right. You could design a bunch of different things, even try a few, see which one works best. Yeah. And have the time and energy to do that, which you didn't have before. That's great. That's great. Yeah. 3D printing sounds better and better all the time, even if you don't work in Hollywood. Well, let's check out the mail bag, shall we? Let's do it. So Doug wrote in about our conversation a couple days ago about risk five. Doug says, I've seen risk five being explored and developed by the big embedded semiconductor companies for a couple of use cases that are always apparent. One is for controls on really low end devices, smart sensors, smart motors, actuator drivers. These are often a lot less than $1 devices. They don't need much compute, computing power and are often used various legacy designs instead of arm anyway. Risk five avoids any licensing costs. Let's get them some compute performance improvements. And the hope is that with one open industry hardware standard, they'll get the volume needed to spur the creation of software tools from vendors instead of having to write all those tools in-house. The market isn't very sexy says Doug, but it's still worth many billions of dollars. Another use case is for the development of security cores and higher end system on a chips. These chips do use arm cores for main compute. They're not going to change because they benefit from the massive software industry that supports that, but there is talk of making the security cores for risk five in risk five for exactly the same reasons as people use for security software. Lots of eyes on it being able to share best practices, et cetera. Doug finishes up with saying not everything has to beat Intel in order to be profitable. Yeah, we had a good conversation with John Dvorak a couple of days ago about risk five. If you didn't catch that, go back and listen to that if you're interested in this sort of thing. And thank you, Doug, for sharing the additional insights on this is really helpful. Absolutely. If you have insights that you'd like to share with us, we would like to hear them. Feedback at Daily Tech News Show is where to send that email. Email us early and often. Also, Justin Robert Young, you better be here earlier and often because we missed you a lot over the last few weeks. You know, you had a great vacation, but let folks know where they can keep up with your work. Now that you're back in the saddle. Well, it is less than a hundred days until the midterms here in America. The can the campaigns are heating up. The primaries are winding down, which means you need to be listening to politics, politics, politics as you care about who you are going to vote for and whether or not the candidate that you love or hate is running a good or bad campaign. This is the place to find it politics, politics, politics, politics, politics wherever you find your podcast. Excellent. We also want to extend a special thanks to Mike Escusha. Mike, you are one of our top lifetime supporters for DTNS. You already know that, but now everyone else does too. Thank you for all the years of support. We really, really value the folks who stick with us over the years, like Mike does. Thank you. Thank you, Mike. Speaking of patrons, stick around for the extended show Good Day Internet. Will we talk about food? Probably. But we talk about all sorts of stuff. You can also catch this show. DTNS is live Monday through Friday at 4 p.m. Eastern, 2100 UTC. Find out more at dailytechnewshow.com slash live. And we're back doing it all again tomorrow, talking about Deep Mind's Alpha Fold with Dr. Nicky Ackermans. This show is part of the Frog Pants Network. Get more at frogpants.com. Diamond Club hopes you have enjoyed this program.