Tech News Weekly Episode 244 Transcript
All Transcripts posts
Tech News WEekly Episode 244 Transcript
Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.
Mikah Sargent (00:00:00):
Coming up on tech news, weekly, Jason Howell. And I first kick things off by talking to Solana Larson of the Mozilla foundation. Solana is the editor of the 2022 internet health report and talks about AI and its involvement in the health of the internet. Then we talk to Max Cherney of Protocol, , who joins us to explain what in the world chits are and how they might be the future of chip design. Before we round things out with our stories of the week. First DALL-E two is officially available well in beta to be tested by, , anyone who wants to join and pay a little bit of money. And then Amazon, which has just purchased one medical for 3.9 billion. We talk about how big tech is, , dipping its toes and perhaps its hands and, , other parts into the health industry stay tuned. We've got a great show
Jason Howell (00:01:00):
Podcasts you love from people you
Jason Howell (00:01:02):
Jason Howell (00:01:04):
This is, is TWiT. This is tech news weekly episode 244 recorded Thursday, July 21st, 2022.
Jason Howell (00:01:15):
This episode of tech news weekly is brought to you by zip recruiter. Certain people make my life easier by helping me out. And zip recruiter makes hiring easier because they do the work for you. How well zip recruiters, technology finds great candidates, and you can invite them to apply, go to zip recruiter.com/tnw to try it for free
Jason Howell (00:01:37):
And by INFR scale, INFRscale delivers industry leading data protection through backup and disaster recovery, whatever your data or environment INFRscale provides continuity and resiliency for your business. Visit infrascale.com/TWiT to get the free ebook five essential components of a ransomware protection plan and learn how to protect your business today.
Jason Howell (00:02:00):
And by checkout.com modern businesses need flexible payment systems that can help them adapt to change, grow and scale fast. Discover how checkout.com can help your business thrive at checkout.com/t and w.
Mikah Sargent (00:02:16):
hello, and welcome to tech news weekly. The show where every week we talk to and about the people making and beak, the tech news, I am one of your hosts, Micah Sergeant.
Jason Howell (00:02:26):
I'm the other guy, Jason, how coming to you live from the corner of my bedroom again?
Mikah Sargent (00:02:31):
<laugh> yes. , I, you know, I feel free Jason. I am very comfortably sitting in a chair in my office and I've got lots. Ooh, look at all this room I've got
Mikah Sargent (00:02:42):
And look at all that room. No,
Mikah Sargent (00:02:43):
I'm just sort of in the corner there.
Jason Howell (00:02:45):
Oh yeah. There's, there's nowhere for me to go, but out there into the rest of the room. Yes. This is just the, the scenario that I've painted for myself, but that's okay. At least I have this, , it could be worse, I suppose, could be worse. , let's jump in into our first interview, artificial intelligence. I, you know, this is AI is just a topic that continues to fascinate me because I feel like more and more AI is being touted as the solution to so many problems, right? , so much power, so much potential, so much disruption, , in its wake, who though does it benefit the most? And the nonprofit Mozilla foundation actually published a report on Monday that details the winners and the losers of this March toward AI powered system. Solana Larson is the editor of the 2022 internet health report, and actually joins us now to talk about that report. Welcome Solana.
Solana Larson (00:03:37):
Thank you. Hi.
Jason Howell (00:03:39):
Hi. It's great to get you here. I love the background. It pops looks awesome. <laugh> appreciate you. , so first of all, this isn't the first internet health report that Mozilla has released. Is this the first one that so heavily, , discusses or, or dissects the power in the influence of artificial intelligence? Or is this just kind of a now thing?
Solana Larson (00:04:01):
Yeah, this is the first time that we're even just like picking one topic and focusing on just that, because for the past, , four additions, this is the fifth addition that we're doing now, we've really taken a broad view of what does it even mean for the internet to be healthy, you know, and healthy for people. Yeah. Um, and thinking of it as being something that can be either unhealthy or healthy as an ecosystem, or even both at once. It's really about looking at what are the root causes for things like hate speech or disinformation, or, you know, the fact that half of the world almost is still not online, those kinds of big questions about the internet, and then thinking about how do we fix them? What are some of the different things that people are actively doing to make things better, make it healthier? And so with AI, I think we have this, this urgency right now, this moment where we see like all of the things that are most unhealthy about the internet, we see that magnified in AI, and vice versa. You know, I think there's also a, an element of where AI can also make the internet healthier or less healthy as well. So one topic this year and we go deep
Jason Howell (00:05:14):
Well, and I think if you're gonna pick one doc topic to go deep on, and especially at this moment in time, AI really seems right for the picking, right? Like I said, in the, in the kind of the, the top introduction, I feel like artificial intelligence, AI, machine learning, all this stuff is being constantly used as the savior, kind of the, the superpower that's being injected into technology. The thing that we didn't have before, and we had all these problems and then now we've got AI to, to save the day, but this report is really about caution. And, , it really seems to, , shed a light on why caution is needed. Maybe explain a little bit of that as AI is used, you know, as presented as this powerful thing that shifts paradigms, what does that, , what does that potentially lead to?
Solana Larson (00:06:01):
Yeah, I mean, that's the thing, it's like a system of power in itself because when you deploy AI to make decisions about things that affects people's lives, you know, that is a system of power that's being exercised. Um, and the other thing that we see is that it's making the ones who have power more powerful. I mean, it requires an enormous amount of, , resources to actually use AI at a large scale. So when you look at, for instance, the influence of big tech or how tech is growing, , big tech is growing more and more powerful. A lot of that has to do with AI. I mean, if you look at a company like, like Amazon or any of the other big tech companies, and you look at their, you know, their revenue and you look at, how they make money in different ways, AI is, is crucial to their business operations in every way.
Solana Larson (00:06:51):
Um, and so when you have the powerful using this as part of their toolkit, and they're also influencing so much of how we think about how we use it, like how do we collect the data that then gets, you know, used in AI systems? Um, how do we, how do we behave? What do we expect it to do? When so much of the discourse is also decided by those who have power over it, it becomes this, you know, self reinforcing mechanism where I think it's important to say, hang on, like, let's step back. Let's talk about who has power. Who's getting to decide how these things work. Um, who's getting to be, more powerful, you know, with, by using these technologies.
Jason Howell (00:07:36):
Yeah, indeed. Now, one of the things that's kind of, , that I've noticed in this, this March, you know, down the path of AI over the recent years, anyways, and potentially a pretty troubling aspect, especially in light of the report, is that much of what's happening behind the scenes is very hidden. And I don't necessarily mean, I mean, perhaps, you know, there is technology companies that know a lot, but aren't, aren't sharing it, but the inherent nature of AI is that the computer, the algorithm itself is making determinations, is making decisions on its own. And, you know, it's, it's not like a, a batch of code that you can say, oh, this is the code that led to this it's the system that kind of has, has created this. So how does that kind of lack of transparency tie into what, you know, what you detail in this report?
Solana Larson (00:08:28):
Yeah. I mean, there's, there's, , a certain amount of, you know, like it's, it's made to feel magic or mm-hmm, <affirmative>, , like a dark art or, I think there's this, it's all kind of wrapped up in this tendency of excluding people from joining this conversation, you know, like who should be allowed to have an opinion about AI. And when you have, you know, very powerful elite kind of forces signaling, you are not smart enough to talk about this, or you are not a computer scientist, you cannot talk about this. It gets very difficult to have a discourse, but if I'm experiencing harm from your AI system, then I know something about it that you don't know about it, even if you designed it, even if you're a computer scientist. And so I think it is this, importance of making the systems more transparent, making them more, understandable to users themselves so that they can also help control them and guide them, and then have a certain amount of, like, I think what we're experiencing right now is just like an enormous amount of, of hubris about, you know, just demanding that people, , trust that these technologies do the things that they're supposed to be doing.
Solana Larson (00:09:43):
Mm. And then when people say that they aren't, you know, are they actually listened to too often? They're not. Um, and, and those are, those are some of the things that we really need to fix.
Jason Howell (00:09:53):
Well, and of course, you know, one, one easy, , example of that is Tim at GI, , at Google who AI ethicist, , computer scientist working on that algorithmic bias and, you know, inside of Google, inside of alphabet. And then of course, you know, had the very big kind of fallout as a result of that, which kind of sheds light on this idea that like big tech knows, you know, is, is creating this system, right? They're, they're putting their funding into the academic research that proves this technology is safer and everything, but then when there is a voice inside that calls a warning, , something happens to that voice and <laugh>, and that seems to go counter to this idea of, of bringing ethics into AI, which many would would agree. And I totally agree is that now is the time to do that effectively. If it's not done effectively. Now, what it builds out and develops into later becomes a much bigger hill to climb, , to overcome if, you know, if it goes in, , bad directions,
Solana Larson (00:10:53):
That's exactly right. And it's inside companies where you have people raising the alarm, and it's also outside companies. I mean, you have groups like Mozilla, you have consumers, you have people on YouTube users, you have, and you know, it's not even just in the, in the realm of big tech, you also have this, you know, predictive policing systems or, facial recognition technologies that end up arresting the wrong people. And then you have groups that are saying your system doesn't work, it's identifying the wrong people. And then they're, you know, if the response is just, well, you know, that's too bad for them, you know, then how, where does that take us? You know, I think we're implementing more and more of these technologies in more and more crucial, parts of society. And it's not just in the United States or in Europe or in like it's everywhere, everywhere in the entire world. Everybody's doing this. And it's in every single, business, you know, type every, every kind of, , sector business sector is affected by AI. Every government is using AI in some way or other, it's, you know, changing how we deal with communication and information. So we need to talk about this together differently and come to some better ways and mechanisms of, of making sure that it doesn't harm people.
Jason Howell (00:12:15):
Yeah, no question. I mean, the systems, a lot of the systems that, you know, we talk about on this network, you know, have to do with like, , YouTube and then the recommendation engines and everything. And I mean, I can speak to YouTube, you know, AI working behind this scenes for, for its recommendation. Let's be honest. Those systems often take a, you know, actual people into a place that they regret going eventually. And like, maybe it wasn't their choice or their recommendation algorithm, you know, is spun up on this idea that like, you know, just as one example, outrage gets attention. Outrage keeps people connected. And so their, they end up, you know, being presented with this content that, , they didn't want, but now it's, it's like they've been pulled into this, this realm where they, of course, they're, , they're interacting and engaging with it because it's, you know, because it's, it's not good where they <laugh> and they don't either don't wanna be there or they're engaging with it because now suddenly they know this thing exists. And, , how culpable is the technology? Is the AI behind the scenes that does that. Um, there's incentive there though, right? Like that's, that's kind of the challenge. The tricky part is companies have the incentive to keep people engaged. And so these systems are doing what they're designed to do. , what, what do you think about that? What are your thoughts on that?
Solana Larson (00:13:31):
Well, I think it would be, it would be possible there would be possible for, for things to work better for everybody. I think there, there would be ways to work together with the research community and multilingual research communities in different parts of the world to try and figure out, how to make these systems work in ways that are better. Um, but for that to happen, these companies need to open up and need to share information in a way that makes it possible to collaborate with them and to have research that's happening from the outside. We have a crowdsourcing, plugin actually it's called YouTube regrets, but it's a, an add-on a browser add on extension, I think, is it, I know so many words, a browser extension that you can download into your browser, and then you can, share your information about YouTube videos that you're watching.
Solana Larson (00:14:23):
And then when one comes up that you regret having seen, we can analyze what was the path that took you there? Um, was it a recommendation? Was it something else let's try and reverse engineer, some of these algorithms that the big companies aren't being transparent about, and there's a number of projects like that from different organizations that are trying to, you know, from the outside figure out what is the scale of the harm? How do we document it? How can we fix it? How can we come up with constructive solutions for making systems that work better for everybody? Um, you know, so yeah, we need, they have, I believe they have incentives to make these systems work better for everybody. We clearly need to make the incentives, , , stronger, you know, and maybe that happens through regulation, maybe that happens through consumer action. Um, but if we leave, , the, the companies to their own devices, it's clearly not gonna happen. Um, because they've found a way to make a lot of money without doing all those things.
Jason Howell (00:15:19):
Yeah. Yeah, no question. Um, before we wrap things up, I also wanna point out that this is not just a report this time around you also launched a podcast to go along with it. So congratulations on that. Tell us a little bit about the podcast so people can check it out.
Solana Larson (00:15:34):
Yes. Now we can be podcast friends. <laugh>
Jason Howell (00:15:36):
That's right. Podcast. We already are.
Solana Larson (00:15:39):
The podcast is called IRL. Um, and it's actually a, a podcast that the Firefox, , team created and ran for several years. And yeah, we've taken it over for the special edition, which kind of we ask in the report we ask, like, who has power over AI? And then the, the, the podcast responds, you know, and we show people in different parts of the world who are doing things differently, you know, who are designing things in a totally different way than we've become accustomed to. And we just take for granted now, and I think it will open people's eyes and say, oh yeah, we could be doing things like this, or, wow. Um, you know, let's hear it from people who are building differently, you know, let's not get so scared of AI that, that, we're just saying that it's all bad. And, , you know, it's not that we need to trash it. We just need to think about it differently. We need to research it differently and develop other incentives and create things that can actually benefit, , humanity. , but that does require, you know, starting from a different, , starting point than we have so far.
Jason Howell (00:16:45):
Yeah. Critical analysis is key and important at the point that we're at right now. And, , this report definitely does a great job of pointing that out. Solana Larson, of course, the editor of the 2022 internet health report from Mozilla foundation Solana. If people wanna find you and follow you online and the work that you're doing behind the scenes, where can they find you?
Solana Larson (00:17:05):
Well, the, the internet health report.org, this year's addition, 2022 dot internet health, report.org. And, , yeah, the Mozilla foundation. We're doing lots of work, lots of research beyond the report activism. If you wanna get involved, if you wanna learn more and do things on these issues, then, you know, come find us and, and connect for sure.
Jason Howell (00:17:29):
Excellent. Solana, thank you so much for carving out a few minutes for us today. We really appreciate it.
Solana Larson (00:17:33):
It was delightful. Thank you.
Jason Howell (00:17:35):
<laugh> excellent. Talk to you soon and best of luck with the podcast. Appreciate it. Thanks.
Mikah Sargent (00:17:41):
Yes. Thank you. Solana up next with Moore's law and full effect, chip manufacturers are looking at new ways to boost performance and efficiency. Before we discuss that, though, this episode of tech news weekly is brought to you by zip recruiter. Certain people just make life so much easier. You don't know what you do without them. Maybe it's your partner. Maybe it's your friend, maybe it's your personal assistant. Well, it's like if you need to grow your business, zip recruiter makes hiring so much easier because they do the work for you. And now you can try it for email@example.com slash T N w zip recruiters, technology finds the right candidates for your job, and you can invite your top choices to apply. So they realize, oh wow, this, this company wants me. It's so nice. Four out of five employers who post on zip recruiter get a quality candidate within the first day.
No wonder zip recruiter is the number one rated hiring site based on G2 ratings as of January 1st, 2022. The hardest thing you have to do is to remember that special URL to try zip recruiter for free it's zip recruiter.com/tnw. Again, that's zip recruiter.com/t N w. Thank you so much to ZipRecruiter for sponsoring this week's episode of tech news weekly.
Mikah Sargent (00:19:09):
Max Cherney All right. So chips are obviously at the base of the different devices that we have. You know, my iPhone here, my, , Mac that's, , that's running all this stuff that I'm running and all sorts of different devices that we have. And, you know, we've heard about chip shortages lately and all of that, but there's a larger concern about the future of chip manufacturing and the future, the, the way that things are going to look, , as we continue to try to make faster and faster and faster chips and smaller and smaller and smaller chips. And, , now we've got chip manufacturers looking at some alternative methods of making that possible. , joining us today to talk about where chip manufacturing seems to be headed is Protocol's Max Cherney. , thank you so much for joining us, Max.
Max Cherney (00:20:02):
Well, thanks very much for having me,
Mikah Sargent (00:20:05):
Yeah. Happy to have you here. So let's, let's dig in first because this of course is kind of the, the basis of, of your piece is this word chicklet. And when I first heard chip lit, I just imagined a tiny little chicken, that <laugh>, that sort of PRD around, but I know it's not that this is something else, , or maybe a tortilla chip broken into pieces. Those are chips. Anyway, how do chips differ from the dye process that manufacturers use today?
Max Cherney (00:20:36):
Well, it, it, it's really a simple concept, , essentially instead of, , printing chips on one sort of large, large dye, , manufacturers and chip designers are breaking up those large dyes into, into smaller pieces. , and it really is. Is that simple?
Mikah Sargent (00:20:58):
, what good. I'm glad. It's very simple that, that, , that, that helps me kind of cuz now everybody in the chat is saying, I was thinking of Chiclets the candy. Um, yes. So chits are just smaller pieces that make up a larger hole. So then if this is the way of things or the way that at least AMD has gone with things, I'm curious what makes chicklets better according to AMD, then the current method of chip design, the one that we've got right now?
Max Cherney (00:21:28):
Absolutely. I mean, I think one, one key thing I should add to, to, you know, when you're thinking about a Chipa is that it is, , the, the, the chips, the different little dyes are, are stitched together. Um, and, and that's a really important thing, , because they all sort of work together in the same way that, that a normal, like a large chip would work, , except their they're stitched, , together with very, very complicated interconnect technology. Um, so to AMD it, it's kind of an interesting thing for, for AMD chits were more or less a necessity that's basically why they started to, to, to build them. Um, and they they're, they, in, in their eyes, they're, they're a better, , chip design method because they are about 40% cheaper because of the way manufacturing works. Um, and the other thing that it allowed AMD to do was to compete directly with Intel.
Max Cherney (00:22:18):
And remember, this was back in 2015 when Intel was the dominant sort of server chip maker, and AMD was going through a period of transformation, I think is a polite way to put it, after they spun out, , their manufacturing business, which is now known as global foundries, and is essentially AMD was trying to figure out a way to compete with Intel's server offerings and, and Intel makes, I don't know exactly the number then, but it was, , you know, a handful, maybe half a dozen or more different server chips to, to serve sort of serve different markets, within the data center segment, low end chips, mid range chips, high end chips, you know, the whole, the whole bit and AMD didn't have the R and D budget, to, to make that, that number of sort of separate designs to serve each one of those markets.
Max Cherney (00:23:03):
And it's, and to put a little context around that it costs roughly half a billion dollars to develop a design for, , an advanced processor these days. I don't know what the costs were like then, but they were it's expensive is my point mm-hmm <affirmative>. Um, so in the way AMD put it was that it basically had sort of one bullet to shoot. Um, and so the way that it solved this problem was instead of trying to make, you know, half a dozen different chip designs, it, it took it it's it's, it's single server chip design and broke it into four different pieces. And by doing that, they could create, they could sort of add and take away triplets to make different server products. Um, and then I think the last advantage, that, that AMD would say it had is, is that be just breaking up the, the different, , breaking up the, the chip into different pieces, allowed it to, to take two of them away altogether and create a desktop chip, from that same sort of design and their desktop segment at the time was the most profitable. So they were very excited that they would, you know, get to turn a big profit off of that. So there are a lot of different advantages, but I think those are, those are just a few of the, of the ways that AMD would argue that, that it's, it's a triplets are a better way to design chips.
Mikah Sargent (00:24:15):
So if I could sort of enlarge this, is this similar to the idea that I could buy a windows PC and inside of the windows PC? I have, , you know, the, the processor and I've got my video card and I've got some Ram and I've got, I don't know, but sound card. And as time El lapses, the processor is just fine. It's still top of the line, but there's, I can now afford to add more Ram instead of, , you know, the however much I have right now. Um, so I can pop out the Ram that's in there, put in new Ram. Um, I can then later on when we add Doby surround sound, I can pop out the sound card that's in there and pop that in. Is it, is this modular in that way? Is that the idea that we break up the different parts of the chip and replace them as they get better while keeping somethings the same?
Max Cherney (00:25:14):
Oh yeah, sort of. Okay. So it's, it's conceptually pretty similar, but not modular in the same way that you could pop out a video card and replace it with whatever new and video thing is, is the latest and greatest, best, best card. Um, once you manufacture a chip Litt or, sorry, I should say a chip with chips inside of a package, that's, you've pretty much got the product, and you can't hots swap them or not that I'm aware of. Um, but what it does allow designers to do is to do what you were just talking about, which is when there's like a more advanced sound say, just for example, sound card, or a GPU or some sort of video processor, or today it's more likely that that might be an AI accelerator of some kind is to be able to add the newest version of that while keeping the, other components in the package the same.
Mikah Sargent (00:26:01):
So instead of having to completely redesign that whole chip, they can just add the parts that have seen an update since the last time they designed the chip
Max Cherney (00:26:11):
That's right. Or, or make, or make sense to make an update too, for, so for example, like you don't have to use the most advanced manufacturing process to make every component on, you know, the current kind of chip that somebody like Intel makes it's just not necessary. So, and, and in some cases it can actually make it harder to continue to shrink. , I don't wanna get too technical, but basically to, it doesn't make sense for every chip to continue to use the most advanced manufacturing method.
Mikah Sargent (00:26:39):
Got it. Now, AMD has been using this method since 2015, as you, , talked about in, your protocol piece. But I'm curious, given that this seems to kind of be where everyone's looking at chips have the other big chip manufacturers started using this method as well. And are there any that aren't,
Max Cherney (00:26:58):
so Intel is, has, is, already using it. Um, I, I should take one step back and say that some of the chip makers called, , they make chips called, , field programmable gate arrays. They're called F PGAs. They're basically just programmable chips. They have been using this method for a, a while in, in a sort of similar way. So Intel has been making this for a while because they make these, these, these F PGAs as well, for its, its upcoming, server and desktop chips, they're going to, they've adopted this, this, they call them tiles, but they're gonna start using them, in terms of people that don't, I think the most obvious example, is Nvidia, and there's a very specific reason because they they're, the chips that, that they're well known for is, as I'm sure everybody knows is there, is there graphics, chips, that are also very useful for AI acceleration and they don't, they don't use, they don't want to make chips because it just doesn't really make sense for those kinds of workloads.
Max Cherney (00:27:55):
Um, for, for that type of compute, they're basically just rendering like one big thing at, at a time. And, and it doesn't really make sense to try to split that, that compute task up, , into, onto multiple different dyes. Um, and dyes are just, that's just another way of saying like a triplet. Um, so Nvidia doesn't do it. Um, and they don't really have any plans to, at least in the, in the foreseeable future. They, their plan is, is actually just to take their already very large chips. They're already large dyes and stitch those big ones together. Um, they have a special technology that they do it with, and they're even planning in the, in the longer run to take, to stick a CPU together with a GPU. I think that's 20, 23, but don't, you know, let's just keep that between you and I
Mikah Sargent (00:28:39):
<laugh> sounds good. Um, now all of this is, is ultimately a focus on, , kind of running up against a choke point where, you know, we see the, these five nanometer four nanometer, it keeps dropping a dropping a dropping. And at the same time, we're looking at, , trying to, , improve upon performance, improve upon processing power. How specifically do chips address this choke point, this upcoming choke point that everyone's kind of worried about?
Max Cherney (00:29:12):
So, I mean, I think the easiest way to think about it is that for the longest time, it was sort of practical to continue to, to add more features O onto a, onto a chip onto a dye, and, and doubly when, , because of this thing, Morris's law, , and a number of other factors I should add, the features themselves and the transistors that, you know, make a chip, do what it does, , keep getting smaller. So on the one hand you could make chips bigger. Um, and then everything on them was getting smaller. Um, so that those two things fork in conjunction, this is very sort of high level, but those two things working in conjunction have allowed us to make chips, you know, that, you know, power your iPhone, as you were mentioning earlier and do all sorts of other stuff. Um, but that is sort of that I idea has kind of run that it's sort of run its course. Um, it's become very difficult and expensive to continue to make, , to make, make dyes larger. Um, and having said that there's also like a, a technical limit. Um, we, we just can't print chips beyond a certain size and it's called the radical limit it's and it has to do with photolithography in the way that chips are sort of printed onto each Silicon wafer. Um, and I can get into that if you'd like.
Mikah Sargent (00:30:27):
Yeah. So that was actually going to be my next question. We, we have a pretty nerdy audience. Um, and so it's, it's okay for you to, to get into that, cuz photolithography, seems to be the, the thing that, that, because of the way that it works, it's kind of holding things back or again, you're gonna do a better explanation. I had just learned on an episode of tech news weekly about how, the photolithography, locations are typically painted yellow because it helps to keep certain, , certain spectrum of light from affecting the dye process. So, , I'm always learning new things. And so, yeah. Tell us a little bit about, you know, the, the, what'd you say the radical limit.
Max Cherney (00:31:07):
Yeah. It's I mean, and, and, and the, and the yellow thing is, is to protect this thing called the photo resist. It's actually really interesting. The, the lighting and all the fabs is, is typically yellow. Um, but, , yeah, so the, the, the, the thing that, that we're talking about here is, is the, essentially the blueprint that, that we use to, to print the chip onto a Silicon wafer and the size of that blueprint is just, we can only make it so far anyways, , a roughly 850, , millimeters square. And as a result, , you know, we, we just, we have to come up with other ways to, to make chips faster. We can't move them beyond this, this, this photo, , this photo, this, this limit. Um, the other thing too, is that when you start to get to that large 850, , millimeter square size chips, the, the actual output of, of a, of a manufacturer, the, the yield per Silicon wafer. So the number basically the number of good chips per wafer, starts to decrease, because if you think about it, the LAR like, so, so when you print a wafer, , with any kind of chip on it, just imagine you're gonna have like a certain number of defects, right. I don't know what that number is, but just right. So's just, you're gonna have, like, there's gonna be some errors. So when you make the dye sizes larger, it's more likely that there's gonna be an error on each dye, if that makes sense.
Mikah Sargent (00:32:27):
Max Cherney (00:32:29):
Right. So that's, that's like the fundamental issue with, with, you know, that works together with this radical size, , that, that makes it really, really tough to print bigger and bigger chips.
Mikah Sargent (00:32:40):
So then that you're, you're reducing the risk. If you go the chip lit way versus going that, you know, bigger and bigger, that that makes sense. Yeah. We would want to not have, have run the risk of having more errors and then having to make more, and then everything, , just crumbles from there. Now, I, I thought it was super interesting that, , Gordon Moore seemed to have predicted this change in the industry all the way back in 1965, as you pointed out in the piece, can you talk about, the prediction that Moore made at the time that goes into, like, it was the, it was the second half of Moore's law that I hadn't really heard about?
Max Cherney (00:33:15):
Sure, absolutely. It was, it was a very important paper that he wrote for lots of different reasons as, as, as you know, and, and by the way, on the, the cost savings, that's, that's what AMD's referring to, , the 40% number I said earlier, that's, that's why that's that refers to yield. That's why the chips are so much cheaper cause,
Max Cherney (00:33:32):
Right. So the, the more defects, the more expensive each chip becomes, right. Cause there's good ones. Okay. So anyways, to your question or more, I'm just gonna, I'm gonna read the quote cuz it's, I think Gordon Moore is a, you know, it's always worth hearing what he has to say. And, and it is, it may prove to be more economical to build large systems out of smaller functions, which are separated, packaged, and interconnected. He said, , the availability of large functions combined with the functional functional design and construction should allow the manufacturer of large systems to design and construct a considerable variety of equipment both rapidly and economically. So that's what he said. And, and, and, and that at the time it was kind of interesting because IBM was actually building systems that used this chip concept, I think, as early as 1964.
Max Cherney (00:34:20):
So I don't know exactly where more got the idea, but it, it sort of illustrates the point that, that the idea of interconnecting different components of a semiconductor or sorry, different semiconductors is a very old idea. And we've kind of, sort of gone from that, you know, that point to what you were talking about earlier that the systems on chip, right, the iPhone chip has a whole bunch of different types of semiconductors, you know, sort of squished into one, modern PCs do two, the, your MacBook does, and, and all that sort of thing. And, and sort of this idea of triplets is essentially breaking up that SOC that system on chip into different it's different constituent parts, but still within the same package.
Mikah Sargent (00:35:00):
Mm-hmm <affirmative>. And then, I mean, it seems like given that, and given, you mentioned, , some letters, <laugh> some letters earlier and I'm forgetting F PGA. Um, it seems like we, if IBM was working on it back then, that there's kind of this cycling between these eras, where we've got, , periods of time where chip lets played some role. And I guess even to this day, they play some role and then we've got the SOCs, now with apple, , taking, you know, one SOC and one SOC and then having an interconnect between the two of them, , with the, the ultra inside of the, the Mac studio. And so I'm curious, do you, as a person, you know, who's covering this stuff, do you think the chip manufacturing industry will settle on chips as the way of the future? Or do you think that there is a possibility that we could see a return to that database design? Or is it just kinda like, we have never been able to figure out the radical limit and we've never been able to figure out this and that. And so we just, this is the end of this technology. We've got to move on to something else like the chip design.
Max Cherney (00:36:05):
Well, I mean, let me just get my crystal ball out here cuz that oh, thank you. Always, you know, always works well. Um, you know, my future predictions, I mean, in, in terms of adopting triplets, I think it's pretty clear that for at least for CPUs, for processors, it's, it's, we're gonna have to more or less adopt the adopt this tech, this particular technology in order to continue to make them, , faster and, and more efficient for all the reasons I, we have been talking about. Um, you know, it's, it doesn't make any sense to, to try to squish all of these features like a wireless radio and other things, you know, onto a single, onto a single die. Um, having said that, I mean, Apple's new, you know, squishing two chips together is a pretty smart idea. So maybe they can keep, maybe they've got some magic ability to continue to make these big dye chips.
Max Cherney (00:37:00):
I really don't know the answer to that. Um, and as I said for Nvidia, the big dye chips for the moments at least are, are the, are the future. Um, and, and there's some companies that are making chips the size of wafers too, at this point to deal with some of the more complex tasks related to artificial intelligence computing. So I think for certain, for certain aspects of the chip industry, I think it's, it's, it's a foregone conclusion. Like I said, , CPUs, I think it's a, it's almost a no brainer everybody is, is moving in that direction. Um, but, for other types of processing and for other types of compute, I think it's a little less certain.
Mikah Sargent (00:37:35):
Hmm. And then lastly, let's get as nerdy as we possibly can. I'm curious if you could tell us a little bit about, and I quote, universal chip interconnect express, or U C I E what is that?
Mikah Sargent (00:37:50):
I mean, it's, it's actually, again, it's like, it's like the triplets themselves. It's a pretty simple concept. The idea is, is just to make a common way to connect chits on, in inside of a single package. , and I mean, it's, it, it's kind of like P C or P C E the way that it's, you can connect, you know, things to a motherboard and connect a, you know, different components on a, on a regular computer. Um, and the advantage, I mean, the, the really cool thing about it is not so much in the, in the technical details in terms of how it works, but it's the fact that in the longer run, the, the promise of this is very, very cool because not only it'll allow, of course, it's, it's a good business opportunity for the likes of AMD and Intel, , or anybody else interested in chips, because say you have a big customer that, , maybe like a Google or an Amazon.
Max Cherney (00:38:39):
These are just examples. I, you know, this is not, not anything to do with my reporting that has a specific compute workload. They want to tackle, like maybe Google's trying to do something with AI. , Intel could put some specific like Google IP, like Google chips that it's designed are triplets that it's designed into a package with other chips made by maybe there's one from Intel, or maybe there's, there's some memory made by a company called micron or, or whatever, and then package those together and then sell that as a sort of a custom solution. So that's sort of one like kind of interesting business possibility. And the other thing is, is it just allows, I mean, the products themselves, the chips themselves will be more interesting because, , the chip designers, the big ones will be able to take other people's people's chips or chip chip lits, and then stick them together, , on, onto a package. And again, that the same sort of thing I applies, it might not be AI, but you could put, , , a wireless radio together with, with some DRAM and, and with some memory and some processing cores. And that, again will just make, because it'll, it'll make, it'll make the chip, , faster and better, which is basically what everybody expects from this, from Silicon at this point.
Mikah Sargent (00:39:49):
Yeah. I love the idea that you you've got a company that's super specialized in, , specific chicklet piece that, you know, you can get that part from them and you can get the, the part, the AI part from this other company mm-hmm <affirmative>, and that, that it all comes together to make the actual chip that goes into the computer phone or what have you. So, yeah, that's, , I'm, I'm glad that they're, you know, coming up with this idea to have this interconnect, that's going to work, , for multiple companies and hopefully everyone will, you know, fall in line with that. Um,
Max Cherney (00:40:25):
Not, not quite yet. We're, we're still a few years away. Just, just so I'm clear, maybe this is a 20, 25 thing, so we haven't quite figured it out yet, but, but I think version two of the standard should, should get us there. Sorry to interrupt.
Jason Howell (00:40:35):
No, no, that's okay. I I'm hopeful then that whenever it does come time that, , everybody will, will fall in line with that to, to make this possible. But, I wanna thank you so much for taking some time out of your day today to join us. Of course, folks can head over to protocol.com to check out your work, but is there anywhere else they can go to follow along and see what you're doing?
Max Cherney (00:40:53):
The Twitter machine I'm at churn and burns. So at CHERN and BURN. , and you can find me there. That's, , I tweet pretty regularly.
Jason Howell (00:41:06):
Awesome. Thank you so much. We appreciate it.
Max Cherney (00:41:08):
Thank you for having me. It was a pleasure,
Jason Howell (00:41:10):
Awesome stuff. , we've got more coming up here in a moment. In fact, coming up, I've got my story of the week. Do E two is hitting more users. Speaking of artificial intelligence, we're gonna talk all about that. But first this episode of tech news weekly is brought to you by INFR scale, the statistics for ransomware attacks. I mean, we talk about it all the time on this show, definitely on security now. So, you know, ransomware is a big deal already, but it's pretty alarming. Cyber criminals can penetrate up to 93% of company networks, according to beta news.com. And it's not just large organizations, either 46% of small medium sized businesses have been victims of ransomware attacks. So this is everywhere. The info scale cloud backup solution provides the security that you need to manage backups and secures them from hackers, or, you know, these adverse events that, that crop up backup and protect your endpoint data.
Jason Howell (00:42:05):
That's what it's all about. Never pay a ransom. And then of course sleep easier at night. You can back up SaaS applications. You can back up endpoints and servers, as well as execute disaster recovery on site or in the cloud. That's what you need when disaster strikes. Every company needs a secure, simple to configure and easy to manage endpoint data protection solution, and INFR scale integrates with hyper V and VMware enabling site to site failover with orchestration in the infrascale plow. So you can keep those hackers away. Its immutability is at the core of how their product is designed. Your data's encrypted in the storage. So no one can alter it from there, whether server crash or human error, malicious activity, or natural disaster, a localized or sitewide incident means unanticipated costs or unplanned downtime, neither of which are very good. Both that come with a huge price.
Jason Howell (00:43:03):
According to Gardner, the average cost of downtime across all industry sectors is $5,600 per minute. All right, that's around $300,000 per hour clients and partners alike, including MSPs and VAs, , trust Infrascale, backup and recovery solutions to eliminate downtime and to eliminate data loss with the most cost effective enterprise grade data protection solutions to help keep their businesses running. If disaster does strike your applications, your data, your systems, they're all recovered and available in record time. Their award-winning world class support is there to ensure that your business is protected 24 7 around the clock. Now with SS D at no extra charge, Infrascale has a solution for your data protection needs. You should really check 'em out, whatever your data, whatever your environment, Infrascale provides continuity and resiliency for your business, , visit Infrascale.com/TWiT, and you can get the free ebook five essential components of a ransomware protection plan, and then learn how to protect your business.
Jason Howell (00:44:13):
And you can get that started today. Right now, that's Infrascale.com/TWiT INFRASCALE Infrascale.com/TWiT. And we thank INFRscale for their support of tech news weekly.
Jason Howell (00:44:29):
All right. So, so apparently I have AI on the brain today, started off with an interview that was all about all the, you know, the, the nasty ways that AI can go. And then now I've got AI that actually, I feel like is kind of fun, although it has its problems too. And we'll talk about that. Of course, I'm talking about DALL-E this is the AI system that we've been seeing only a select, you know, few people online having access to this, this, , image creation, artificial intelligence system, where you can go in there and you can, , SP you know, basically give a sentence that describes something and it will create imagery around that the AI will create it dynamically, and it would be one thing if it created and it looked bad, but what's really neat. What's interesting here is that it ends up turning out a result. That's really fascinating and really good. I mean, I'm sure it could be scrutinized if you look at it close enough, but it's pretty impressive, pretty remarkable. What it's able to come up with and to date it's been, , it's been used, I think by a hundred thousand select users since April, when it was kind of a li very limited beta, it's been free to use so far. , and now it has a wait list. You can get access, , they're aiming to reach up to 1 million users in the next two weeks. So you wanna get on that wait list, if you wanna access, there's a few questions that ask you kind of who you are, what you do, all that kind of stuff. And then hopefully you're part, you know, selected for as one of the 1 million users. And you can check out DALLE-2. And, this time around, , oh, are you okay?
Jason Howell (00:46:07):
Did you just fall?
Mikah Sargent (00:46:08):
Jason Howell (00:46:09):
Did you just fall asleep or are you eight? No,
Mikah Sargent (00:46:11):
I was growling <laugh>
Jason Howell (00:46:13):
Because I tell
Mikah Sargent (00:46:14):
Us why I've been on that wait list for so long, and I'm still not in it yet. Aw. And I just wanna be a part of the fun. I'm not gonna do anything bad. I think they, once, if they see you're a journalist, they're like, eh, nevermind,
Jason Howell (00:46:26):
No, you're gonna poke holes in it.
Mikah Sargent (00:46:28):
Yeah. And all I wanna type in is spooky lemons. That's literally, <laugh>, that's been in my head for so long. I've been just thinking 11. I wonder what happened, whatever I type in spooky lemons <laugh>. And I use, I use some of the other, , fun tools from, , open AI. They've got that. Um, the, the G P T three system is super cool. Um, in fact, I didn't this, I know this is a little off topic, but I just wanna quickly get, say this. Um, I didn't realize that they had, , on top of like, you know, finishing sentences and, and finishing paragraphs and summarizing things. They've got like a TLDR too long. Didn't read summary that you can do, but they also have a spreadsheet creator. And so you, this was, this is the example that they give a two column spreadsheet of top science fiction movies and the year of their release. And then you tell it like how you want the spreadsheet to look, and then it will tell you. And so I started playing around like the top 10 pop songs in the year of my birth. And so then it had the, the song title on the left and the year or the month of, , its release on the right. And then like the most popular dog breeds in 2022. Um, and I don't remember what the cuz I was making a spreadsheet with it, but I didn't realize that it could do all of that and I am using it. Um, you know, whenever you sign up for the playground, , for open AI, you have to say what you are and if you say journalist, then they say no. Um, but it, I, I, if you're just using it for personal use, which is basically what I was doing, like I wasn't going to write a story on it or anything like that, so yeah, yeah, yeah. Then it was okay with that. And I'm just wondering, because I have seen some of the, some folks that I know who a, didn't talk to two of the lead engineers working on, I'm gonna keep making these movements. <laugh> two of the lead engineers working on DALL-E on this very show. Um, that's true. Who signed up after I did, who got access to it? And I'm like, come on. I Just wanna, I just wanna type in spooky lemons
Jason Howell (00:48:37):
Someday. I, I have faith someday. You'll get to live your spooky lemon life.
Mikah Sargent (00:48:42):
<laugh> in the meantime, I'll use that one that you, , talk
Jason Howell (00:48:46):
To. Yeah. What was right? The, , the mini DALL-E or DALL-E mini I can't
Mikah Sargent (00:48:50):
Yeah. DALL-E mini.
Jason Howell (00:48:51):
Yeah. DALL-E mini. Yeah. Yeah. I mean, what you get here, of course that you don't get there. If you're using DALL-E mini and it's faces of any sort, they're all like mangled and, and contorted and stuff. And here, I think you get kind of like a more,
Mikah Sargent (00:49:05):
As long as they're not a celebrity.
Jason Howell (00:49:07):
Yeah. As long as they're not celebrity, they're more anonymous sort of thing. Yeah. That makes, that
Mikah Sargent (00:49:11):
Makes sense. Yeah. , if you, so somebody tried, somebody asked cuz I was looking through different people. Who've gotten access and yeah, celebrities aren't allowed. And so one person did a prompt, Hulk Hogan writing a dolphin in the Pacific ocean and the person went ahead and typed it into DALL-E and the body and everything was definitely whole Kogan, but the face was like raw. Um, but if you just type in a person sitting next to a bowl of spooky lemons, then it will just make a person, you know, a face that's not anyone's whole face
Jason Howell (00:49:45):
Yeah. And unidentifiable, but, but actual face of some sort, which I mean, AI systems are able to create, you know, very photorealistic faces anyways. So that's probably kind of the same thing that's happening there. I wonder if you were to put someone who looks like whole Kogan, so at least instead of getting mangled face, you get like, , a similar face, but not the same anyways. Um,
Mikah Sargent (00:50:08):
I'd love to try it.
Jason Howell (00:50:10):
<laugh> yeah, I know.
Mikah Sargent (00:50:11):
Jason Howell (00:50:11):
We've got ideas to test this system. Yeah. So if we ever get access to this, it will be a credit based system going forward and it refills every month. So if you're doing this for free, so essentially one credit equals one text prompt, prompt that's sent to the AI. And as we know from DALL-E mini or DALL-E too mini or whatever, it was like not every text prompt is going to, , result in usable images. You might get the results and be like, okay, that didn't really turn out that great. You know, you might do 10 of them. And finally get the one that's like, oh, jackpot, that's the one we were looking for. So, you know, you get 50 free credits for the first month, 15 free credits every month thereafter, which could not be that much depending on how you want to use this. , you can purchase additional credits at $15 for 115 credits. So for $15 get 115 tries. , so that gives you a sense of kind of like how, you know, what you're gonna spend if you're, if you're maybe planning on using this in your business or whatever, which by the way, you can do this. Grant's full usage rights commercialization included. So reprinting selling merchandise rights, all of that included with the images that you create through DALL-E, which is very interesting. Um, and they are also, , pointing out, you know, the beta has been useful for getting the word out, but it's also been useful for, and this ties into the interview that we had earlier with Solana. Um, it has been, , R returning some results that let's just say, you know, a show examples of, of bias, like racial bias or gender bias. For example, CEOs and firefighters were all coming back as white men and then teachers and nurses were all coming back as white women. And so, you know, people over that beta period were recognizing this, telling open AI about this and then open the eye has integrated a, , data set that includes more globally diversified samples. So as a result, 12 times more likely to include diverse results. , so that's better, that's an improvement, although it is kind of, you know, that they're, they're doing other things like filtering and censoring, what people ask for cuz obviously, people can be awful at times. And so they go in there asking for really bad things and open AI says, you know, we're filtering and censoring, what people ask for as issues crop up. Ultimately that's a bandaid, right? That's something that's very reactive and hopefully they can figure that out over time. But, but interesting nonetheless, and I think, you know, we talked about this a little bit. Um, I was filling in for Leo on this weekend, Google yesterday, and we talked a lot about DALL-E and, and this whole, opening of the wait list and everything.
Jason Howell (00:52:57):
And I think what really is compelling to me, and I'm really curious to hear your thoughts on this is as these systems get more and more, , believable or more, more and more, you know, , photorealistic or whatever you want to call it. Like these things are really effective already, but they still have holes and they're only gonna get better. Um, what does this do long term for the creator economy? For the digital creator economy when I can go to a tool and, and tell it to give me the thing and it might not be perfect. It might not be exactly what I want the first time, but if I've got credits and I can do it 10 times and great, that's great. , instead of hiring a person, you know, an was very, very much on board with this, with the idea that, you know, the human element is incredibly important to this. And I'm not saying I don't, I don't believe that I totally believe that as well, but at a certain point, these systems get so good that, I mean, maybe there is a point down the line where the human, where the human experience is less important. What do you think?
Mikah Sargent (00:53:54):
H ? So this is a tough one because I do think to a certain extent that, for a lot of, of in certain situations, it makes so much more sense to, you know, , use photos created by human beings or that were, , literally captured in real life by human beings, you know, photographers. Um, but I think about, I used to work for a startup and at that startup, we, , wrote articles that got turned into videos. And for the videos we needed to have things on screen, you know, when the anchor wasn't on the camera saying things, then it cut to the article sometimes with a quote pulled from it. Yeah. But then it also cut to images. Um, and we did everything in our power to do, , that correctly, meaning that we would bend over backward to go, , find things on in creative commons usage in all these different places where it was, we were allowed to use the image and like, it was very strict, the rules and guidelines around what images could be used and could not be used. And I remember when we finally got a Getty images subscription, and it was like earth shattering in the way that everything so more available to us, everything that's so much easier. However, because of the way that the, the, you know, subscription is set up, you can only use so many photos in a given period of time and it's very expensive. And so you may not always have the opportunity to make use of that. And so I think about, , people with blogs, for example, yeah. Um, individuals who have small companies, , that are trying to report on things. Um, the, the, the protocol article that we looked at earlier, where it had that really clever design at the top of it looked like a pack of, of gum from protocol that said six chicklets. And then the word fresh was kind of, , cut off on the side and you slide out and you could see the little chits inside, but it's almost like they were Chiclets gum.
Mikah Sargent (00:56:05):
Um, the idea that instead of having to, for, and again, a smaller company protocol can afford to pay occasionally different artists or higher artists to pay for that, but a smaller company that doesn't have a whole lot of money being able to generate something that is somewhat similar to that, or for, , folks to like, think about how much it would cost to pay a, , digital painter to create some of the things that you can create with DALL-E that are surreal in, you know, with just a prompt that, that, that this opens up a whole new realm of people being able to express themselves that I am much more of a fan of than I am worried about Getty images, not making as much money as they do right now. So when it comes to that, like stock photography, I want, I still want photographers to make money. And I still want, you know, the, the New York times to have its staff of photographers, I don't think that's ever going to change because there, there is value in that, but in the idea of democratizing, , creativity, mm-hmm, <affirmative>, I'm so much of a fan of what this could do.
Jason Howell (00:57:16):
Yeah, yeah. Not just, from a perspective of replacing imagery, either, which I, I do believe that, you know, there was a certain, there is a certain layer of, of people and use cases that this will probably be the easy replacement for at least the first go to before expanding out, beyond, but maybe as inspiration. Right. Cause like digital creatives might, you know, get a, , you know, what, what was your, what was your prompt, the, , evil lemons or what
Mikah Sargent (00:57:44):
Was it? Oh, those spooky lemons, spooky lemons,
Jason Howell (00:57:46):
Evil lemons. I wanna see what evil lemons is and they could battle each other <laugh> but spooky lemons, you know, like a, like a digital artist might have a hard time, like, , coming up with ideas around spooky lemons. This might be a launchpad, right. Like go to DALL-E mm-hmm <affirmative>. But in the thing, see, you know, the, the, the grid of nine images, or I don't know how DALL-E does
Mikah Sargent (00:58:07):
It get some inspiration. Yeah. You
Jason Howell (00:58:09):
Get some inspiration. It's like, okay, these are some directions that I could go. And then to make something, , personalized and, and creative from that
Mikah Sargent (00:58:17):
Collaborate with DALLE. I love that idea. Yeah.
Jason Howell (00:58:20):
Yeah. Maybe at some point, DALL-E ends up creating its imagery, , with Photoshop layers embedded into it, and then <laugh> then you can, I don't know, but
Mikah Sargent (00:58:31):
No, but seriously. Yeah. The, , the idea that you could then kind of edit it yourself or make adjustments soon as you want. Cause they've
Mikah Sargent (00:58:36):
Already got, that would be amazing. They've already got the ability to, on top of just fully generating stuff, you can give it images that it then supplements it'll add things to, or subtract from. , as long as, again, it's not, you can't upload a photo of, I don't know, Mariah Carey and do something with that, but you can upload and, you know, an image that you took and say, add a couch to this.
Jason Howell (00:59:02):
Right, right. Or remove couches.
Mikah Sargent (00:59:05):
Yeah. Or remove all of the couches in
Jason Howell (00:59:07):
This. I don't, I'm a, I'm Antico, so yeah. ,
Mikah Sargent (00:59:10):
Yeah, not really big into couches. Could you, , get that please? <laugh> how understand?
Jason Howell (00:59:16):
, I, I hope that we, , we get in, , on the wait list and, , if we do, do you know, we'll talk about it a third time. We'll keep talking
Mikah Sargent (00:59:23):
About it. Well, you will keep talking about it, especially when I can get the, the, , spooky lemons on there. Yep. Yep. All right. Um, let's take a quick break and we'll come back with my story of the week.
This episode of tech news weekly is brought to you by our sponsor checkout.com. Tech should be groundbreaking and promote innovation. Traditional payment systems are heavily layered. They're disconnected. They're perceived as a cost center to the business. Modern businesses need flexible payment systems that can help them adapt to change, to grow to scale fast. And, , we recently came across a company with tech that approaches payments through a radical new lens. And it's called checkout.com. checkout.com is a leading digital global payment solutions provider for brands like shine grabs, Sony electronics wise, and Henkel checkouts, flexible payments platform is purpose built with performance, scalability, and speed in mind.
It's ideal for businesses looking to seamlessly, integrate better payment solutions globally. And with a dedicated team of local experts that are spanning 19 offices in five continents, checkout.com offers a strategic partnership to help businesses improve their acceptance rates, optimize their payments performance and grow their business globally. With checkout.com, you're going to get global optimization. They provide local acquiring and multiple geographies, improving authorization rates and lowering costs. You'll get transparency. Their fee structure is clear and straightforward while their in depth reporting gives you visibility, no matter where you operate local expertise, their dedicated local teams bring regional international and regulatory expertise. That's a big one allowing you to navigate market complexities with confidence and strategic partnership. They take a collaborative, personalized approach to solving complex problems for their merchants and ecosystem partners. You know, for me, I think that the reason checkout.com could make a difference for you is because of what it enables in terms of payment methods. I know for me, when I go to a site and I see that apple pay is available on that site, it night dives out of Ted. I'm gonna end up going through with, with whatever purchase I was eyeballing. It's like window shop window, shop window shop. And then I get to the cart and there's an apple pay button. And I'm like, no, because I know I'm gonna make that payment. I'm gonna go ahead and get it. Cuz it just, it makes it so simple makes it so easy. And so that's why I end up, , just, just going for it. So checkout.com can help enable not just apple pay, but all of the, these different payment, , methods that will make it so simple for your customers who are looking to, , buy whatever it is you're offering. So discover how checkout.com can help your business thrive at checkout.com/tnw that's checkout.com/tnw.
Mikah Sargent (01:02:15):
All right. Um, this morning there was a new story, , published on tech crunch and then in several other places about Amazon making a new purchase, the company is purchasing a company called one medical. One. Medical is a, , basically a kind of medical doctor service where you can, , get 24 7 on-demand virtual care. You can get, excuse me, appointments in one Medical's offices. And, , they work with a lot of different insurance companies. So, , same and next day appointments, video appointments, longer appointments, which is a big thing that I know, , a lot of people kind of struggle with in, in your typical medical situations drop in lab services, , so that you can, you know, hop in and get those lab tests done that you need to do. And, , it is a, a system that kind of is looking to turn healthcare on its head, , whenever it comes to kind of primary care and that's doubled down, , Amazon now buying it for 3.9 billion, where Amazon seems to want to continue down its role of, , of, of offering different of, of having rather different medical offerings.
Mikah Sargent (01:03:36):
I think this is what's most interesting is that Amazon, at one point bought this company that I, I absolutely loved it was called. It still is kind of, , built into the system, but it's called PillPack and, , PillPack used to be my pharmacy. And what was cool about PillPack is that you could not only get your, prescriptions from your doctor sent to PillPack, but you could also get if you took vitamins. , so for example, I took fish oil and vitamin C or whatever each day. And, PillPack would put them into these little, , compostable, , packs that had your day and your night medicine and for each day, so Monday, you know, for a whole month. And so you would kind of unwind your pill pack and tear it off and you'd have that for the day. And it was just really handy to know I've got the medicine that I'm supposed to take today, the vitamins that I want to take today.
Mikah Sargent (01:04:33):
And it's all right here, it's all, , as I need it, and I know that I've, you know, done what I needed to do Amazon bot pill pack and then launched Amazon pharmacy, and then has also made, , Amazon care, which is its on demand healthcare service. Um, and it's clear given they just had their AXA live conference, the developer conference, where they showed quite a few different healthcare integrations between AXA and the healthcare provider. So I think that this is, not a surprise to see Amazon making, , this purchase of one medical as it continues to, , wants to offer a, , medical service to folks who need it. And I wouldn't be surprised if, if this is eventually rolled into some sort of maybe even like a premium prime plan, where on top of your typical Amazon prime, you get, , healthcare.
Mikah Sargent (01:05:30):
Mikah Sargent (01:06:34):
And I thought what a healthcare group is actually do. You know? So what I'm saying here is like, even if you are going with more traditional healthcare, we're, you know, part of a big healthcare program, mm-hmm, <affirmative>, you're not necessarily any safer. Um, and they may also be doing that. So just as, , as a heads up to my fellow colleagues who are also part of, , the KP deal, when you log onto the website, you should decline when it asks you, , to if they can give your data out because that's what it's trying to do and it's not, not great. But anyway, back to the story itself, yes, there is a concern because it's Amazon, but I would be interested to see, , how the company could, , potentially provide healthcare, , in a way that is more affordable. And certainly if I didn't have the healthcare plan that I have now, where the pharmacy and all of that is kind of built into one company, I would still be using PillPack as my pharmacy.
Mikah Sargent (01:07:32):
Um, mm-hmm <affirmative> because that was just the way that I did it. And it worked really well, even though it ended up being part of Amazon. Um, and before we kind of break this into little conversation, , I just wanted to mention that apple around the, a little bit before, , this new report, apple put out a really big, news article and, , PDF talking about what apple is doing, , to support personal health research and care. And, , this guide walks through all sorts of stuff, including how, , it it's making it so that your health data is available all in one place. How the apple watch can be an intelligent guardian for users' health, by, , heart health monitoring, mobility monitoring, hearing health, monitoring, and safety. , so if you fall or something like that, and also if you need to, , call 9 1 1 or whatever your emergency services provider is, making changes to your everyday health and fitness giving, , access to third parties to, , health and fitness information to improve upon that.
Mikah Sargent (01:08:36):
And then also helping on the, , on the sort of health ecosystem side by providing, , research studies, , that, that tie into the iPhone that make it very easy for people to participate in research studies. And, one of my favorite features, , is the health records, , feature that apple has in apple health, where my doctor is tied into my phone. And so I get my health records on my phone and can see. So the other day when I had to get tested for COVID when I woke up in the morning, even before the Kaiser email had come through, I had already had the records on my phone and was able to see that I was negative for that and the two versions of flu that they were testing for. Um, so I love that, but, , these big tech companies in general are looking at ways to, , to both aid and also of course, make money in health mm-hmm <affirmative>, , go. And the last one I mentioned just cuz this is the other one I was thinking of is Google. I remember at, and I don't remember which event it was, but at a recent event, talked about how they wanted to help with, DMA dermatology in particular, , where, when people are doing searches, , for things related to dermatology, that it does a job of providing them with answers to the health issues that they are, , dealing with. And especially what, as it comes to, , people of color and how dermatologically speaking. There's not as much data out there. So when I, if I were to go look at what, , you know, type in like, skin cancer, , spot or something like that. And if I were to look at it most of the time, it's going to be, , photos of white skin.
Mikah Sargent (01:10:23):
And so trying to find that information and get more info about it that way is not, , not great. So there's a part of me that likes that tech has is, is playing a role in all of this. And of course there's also the skeptical part of me that is concerned about our privacy and the security they're in, , when it comes to how, , our stuff is being protected or how people will be helped. But again, when I wake up and my own healthcare provider is trying to, , ask me in what I felt was kind of Aly way, if they could give my data to third parties, I'm like, well, clearly I'm not safe, no matter where I am.
Jason Howell (01:11:01):
Yeah. Right. Data is valuable no matter who, you know, <laugh> who it is, who you're talking about as far as the company is concerned, and health data is ridiculously valuable as well. It's also very con well by, you know, comparison to like search data and everything. It's also very controlled. There's actually regulation around, around controlling health data and health information and everything. So that at least puts some sense of control onto it, but that's probably what makes it even more valuable, you know, is the lack of access to it.
Mikah Sargent (01:11:39):
Yeah. That's why, that's why they had to ask me this morning if they could use my health data, because it was part of HIPAA regulations. But if I, if I consent without realizing it, then you know, I've done that. And I, it was the way whenever I declined and up popped an alert that was like, are you sure you want to do this, that I had to double check and reread to make sure that when I was saying, I decline that it really meant that I do not want you to share my data, as opposed to the other way it all felt what are those called dark patterns. It, it had some dark patterns. Yeah. And I was very disappointed. Um, so yes, I'm glad that there are these regulations. And, I think that, you know, because of the focus of our, , our, what is the, what I'm looking for, our lawmakers on big tech, I think that they will have to do what they can do to be as clear as possible and, , as straightforward as possible about all of this, because there's so much of a focus on them, but I just, I don't want to get overly excited about this or say that it's too good of a thing because a, we don't know what, , Amazon's going to do.
Mikah Sargent (01:12:48):
Exactly. And I have to know for sure that, the company isn't, you know, secretly using this as some data mining thing in some way, , that that will make it, so that's how it's being subsidized or what have you.
Jason Howell (01:13:03):
Yeah. Yeah. And hard to imagine a company of Amazon's size, which is by, you know, arguably one of the largest companies out there not having some sort of data, you know, , aspect to this data mining of some sort. But, but I guess we don't actually know. I mean, the, the fact of the matter is healthcare in the United States, the healthcare system is just broken. It's just bad. You know, if, if you don't have the money to, to afford basic human protection, then you're screwed. And so I'm, I'm open to ways to help, you know, people who can't afford healthcare, , to, to get access to healthcare if they don't have it already, I'm totally open to, to ways, but it has to also, it can't be at the expense of some other very major thing that now suddenly we've, we've completely made ourselves vulnerable in this different way.
Jason Howell (01:13:58):
Like, and I feel like that always ends up being the solution. And maybe that's just a pessimistic outlook, but it's kind of like, okay, great, you can have this, but you're gonna have to give us this. And it's like, it's always, you know, and, and other countries, when it comes to healthcare, it's not nearly as, as, , as manipulative as it, as it feels like it is here in the us. And many times, you know, from people from other countries looking on the outside in are Agat at what they see as far as like what we have to, what we have to do here in order to get basic human care. So, so I don't know how I feel about Amazon, you know, putting out, , their own, you know, potentially putting out their own healthcare. Um, you know, again like you, if this is a system that, that we feel has enough checks in place that we can trust it and it lowers the barrier. So that a lot of other people that might not be able to get that protection can then I'm all for it. But I guess that remains to be seen. I don't know what this is gonna turn into. I see that PillPack is still around though. , yes. Do you not use it for a particular reason? Like it's still,
Jason Howell (01:15:06):
, just because our, if you don't mind me asking,
Mikah Sargent (01:15:08):
no, no, no, it's fine. , you know, we, I have Kaiser and Kaiser has its own pharmacy, so it's,
Jason Howell (01:15:15):
That's true. Yeah. No, that makes sense. That makes sense. All right. Well, I guess we'll, , we'll see. I mean, 3.9 billion, that's pretty penny cool.
Mikah Sargent (01:15:25):
Quite a hefty. I mean, Amazon, you can, you can acquire my services for 3.9 billion.
Jason Howell (01:15:32):
Yes. I'll do everything in my power to give you the value that you're looking for.
Jason Howell (01:15:37):
3.9 billion. , anyways, we will follow that and so much more in future episodes of tech news weekly, we do the show every Thursday. , so TWiT.tv/TNW. That's the show page on the web where you can find all the information to subscribe to the show and audio video, , YouTube. And then once you subscribe, then you get our wonderful interviews and discussions, , like magic through the power of RSS. So, , do that TWiT.tv/TNW.
Mikah Sargent (01:16:04):
And if you'd like to get all of our a Nope, not all of our ads, if you'd like to get all of our shows, a free, well check out club TWiT, because for seven bucks a month, you get every single TWiT show with no ads. You also get exclusive access to the TWiT plus bonus fee that has extra content. You won't find anywhere else. That's behind the scenes before the show, after the show and some, , fun content that the hosts upload from time to time, as well as access to the members, only discord server. That's a place where you can go to chat with your fellow club, TWiT members, but also those of us here at TWiT and, , you know, the chat has been live today there in the discord. Um, I should also mention that now becoming a member of club TWiT still for seven bucks a month, , or for let's say TWiT.tv/clubTWiT seven bucks a month or $84 a year.
Mikah Sargent (01:16:56):
You also are going to get access to hands on Mac. My new show that publishes every Thursday, , with loads of tips, tricks, et cetera, all relating to apple. So it's called hands on Mac, but it's IO iOS iPad OS it's watch OS it's Mac OS. It's all the different ones. Um, and the episode that's going out today is covering, , how to sign up for the public beta so that you can help shape the future of Apple's software. And again, still for seven bucks a month, you will also get access to hands on windows, which is Paul Thurott's new show, where he is doing the same thing, but for windows. So we continue to add more and more value to club TWiT. And, you know, if you're not big into the, the discord and, and the, all the other stuff, , you know, you don't care about having ad free versions of those, other shows or any of that.
Mikah Sargent (01:17:52):
Well, you can just subscribe for 2 99 a month to get hands on Mac and 2 99 a month to get hands on windows. So if you're just kind of, it's called the single show plan. If you just want that show in particular, you can, but honestly, seven bucks a month and you get not only everything that you've been getting so far, but also these two new shows, mind boggling value, , super awesome. And so, yeah, that's all available at TWiT.tv/club TWiT. These, these shows are also available, , as part of the apple podcast subscription 2 99 a month. Um, and you can get the ad free version of tech news weekly that way as well. The audio feed in apple podcasts, you can tap to subscribe for 2 99 a month. , if you would like to follow me online, I'm at Mica Sergeant on many, a social media network, or you can add to chiwawa.coffee, that's ch iaa.coffee, where I've got links to the places I'm most active online.
Mikah Sargent (01:18:43):
, now that the week is almost over, I don't have to list out so many shows. I can just say, catch me this Saturday for my second, co or wait, what is it? Solo hosting, , of the tech guy radio show. Um, yeah, that was a lot of fun the first time around, , and we'll be doing it again this Saturday and then on Tuesdays for, , iOS today, which I record with Rosemary orchard and Thursdays, , outside of techniques, weekly is hands on Mac. For those of you who are club members. Jason, how, what about you,
Jason Howell (01:19:16):
If you haven't, , realized out there, , listeners and, and viewers of this show, Mike is a busy guy, so
Jason Howell (01:19:24):
You got hands full that's all I'm saying. , I'm ad Jason Howell on Twitter. I filled him for Leo on this weekend, Google yesterday. It was an awesome show. So definitely check that out. TWiT.TV/TWiG. I also do all about Android, , for the network TWiT.TV/AAA, and actually this upcoming week on all about Android. We're gonna have Florence eye on, on the show and she's bringing some hardware that you're gonna wanna hear some reviews for. So we've got flows, hardware shack coming up on Tuesday. Um, big, thanks to everyone at the studio for helping us do this show, John Ashley, of course, behind the scenes, doing so much to, to help us with this show each and every week, Burke testing people, making sure the technology works and thanks to you, all of you out there for watching and listening. We could not do this without you. , you were super important to us, so we thank you for returning each and every week. We'll see you next time on tech news weekly by everybody.
Rod Pyle (01:20:19):
Bye. Hey, I'm Rod Pyle, editor of ad Astra magazine, and each week I'm joined by Tariq Malik the editor in chief firstname.lastname@example.org in our new this weekend space podcast, every Friday Tariqk. And I take a deep dive into the stories that define the new space age what's NASA up to when will Americans, once again, set foot on the moon. And how about those samples from the perseverance Rover? When do those coming home? What the heck has Elon must done now, in addition to all the latest and greatest and space exploration will take an occasional look at bits of space flight history that you probably never heard of and all with an eye towards having a good time along the way. Check us out on your favorite podcaster.