TWiT 1098: Usain Volt - Meta, Addiction, & the Future of Online Engagement

[00:00:00] It's time for TWIT This Week in Tech. Leo Laporte, the Digital Jesuit is here. My car guy Sam Abul Samad. Of course, we're going to talk about flock cameras and why people all over the country are tearing them down. You might want to tear them down when you hear what they're doing. We'll also talk about the new Pixel 11. It's got AI, lots of it. And Meta, the new lawsuit that may cost Meta every penny it has. That's next on TWIT.

[00:00:29] Podcasts you love. From people you trust. This is TWIT. This is TWIT. This Week in Tech, Episode 1098, recorded Sunday, August 23rd, 2026. Usain Bolt.

[00:00:51] It's time for TWIT This Week in Tech, the show we cover the week's tech news. Ladies and gentlemen, I give you our esteemed panel this week. Fr. Leo Laporte, the Digital Jesuit visiting us from Las Vegas, Nevada. It's so funny that you go from Sin City to No Sin City and back and forth, right? It's basically the same thing. Really, honestly. It's not too different. You're in town, probably spend time with your family, which lives there, but also because I presume of DEFCON, yes?

[00:01:20] Correct. Originally, I was going to do the duo Black Hat and DEFCON this year, but we had a special apostolic visitation. So for about a month, I was in Malaysia, Thailand, Indonesia, Myanmar. How was that? Yeah. Hot. Super hot. It's really hot over there, Leo. Wow. We're going to be there in September or October. No, October, November. That will be rainy. Yeah, rainy and hot. Well, less hot.

[00:01:50] We're starting in Bangkok and we're going to take a little side trip to Angkor Wat. Then we're going to sail down through the Malacca Strait, visit Malaysia, Indonesia. We end up in Bali. So basically all the same places that I went to. Awesome. Except Myanmar. We are planning Myanmar. I don't know. You probably shouldn't. My daughter moved to Da Nang this year. Oh, nice. By the way, that is Sam Abul Samad. You know him from Wheel Bearings, the fabulous car podcast. Of course, he's a regular on our shows.

[00:02:19] He's VP of research at Telemetry, where he covers automotive topics. Your daughter moved to Vietnam. I'm jealous. I hear it's a wonderful place to live. She's loving it there. Yeah. She sent us some photos and video the other day. She was up at 530 sunrise and sent us some photos of the sunrise over the beach. And people, it was an exercise group out doing it, exercise dance group out on the beach across from her apartment.

[00:02:50] Yeah, we had a trip planned to visit Da Nang and Ho Long Bay and much of Vietnam. Then this thing called COVID hit and it kind of scotched those plans. I've heard of it. Yeah. We forgot a lot of it though. I haven't been to Vietnam for 26 years. I haven't been there since 99. It's, ooh. And the photos that people send me back, the places that I visited. What were you, 12? They look so different. When do I have 99? Oh, Leo. Yeah.

[00:03:19] My wife and I are hoping to go visit her next year. Oh yeah. You've got a place to stay. You're lucky. Yeah. Yeah. Da Nang is listed as one of the top 10 beaches in the world, according to Darren Oki. That sounds pretty nice. They also cater to digital nomads. Yes. There's more and more countries that are really leaning into that, including Thailand, Bangkok. There's a big digital nomad community. Yeah. And more importantly, Vietnamese coffee is the best. That's really all I hear about. Oh, oh. I will be going to the, to the Bay Area.

[00:03:48] I have to share the coffee that I got as a present from the Sultan of, um, where we know it was, we were in Indonesia of, uh, Jojo Karta, Jojo Karta. He gave us bags of Luwak coffee. That's that coffee that gets digested by a little civet. The civet coffee. The poop coffee, Leo. I think you'd like it. I've had it many years ago, but you know, ever since then I've gotten really into pour

[00:04:17] over and, uh, yeah. And I am the, I've kind of dialed in all my pour over. I have all the tools, including the total dissolved solid meter. In fact, this morning, I just want to tell you, I overground a coffee. I made it a little underground. It was a little too coarse and I got the extraction up to two TDS two. That was too high sweet spots. 1.5. I just, and I knew it the minute I tasted it, I said, over extract.

[00:04:44] I pretty much quit drinking coffee a number of years ago for my blood pressure. But, uh, this year I've been on a binge of making ice cream. Oh yeah. That's good for your blood. It's not a good substitution. My favorite local coffee, uh, ice cream shop closed up. And so we bought a Ninja creamy and I've been making a bunch of, uh, yesterday I made some fig ice cream. People love that Ninja creamy, but I know if that were in the house, I would weigh 400 pounds. Yeah.

[00:05:14] Well, I I've got my mom and my sister and all of my, my loved ones telling me I have to stop drinking coffee. And I have to remind them, do you remember where I live? It's like, that's all we have. And it's all yeah. That's a coffee is all coffee and pasta. That's all you need in life. Well, anyway, anyway, I, uh, enough of this silliness. We have tech news to talk about. Uh, Molly white was scheduled to join us. Uh, she got a scheduling problem. She'll be back next week. So that's great. But you know what? You two guys are all I need.

[00:05:44] I'm happy. Uh, so let's talk about meta because Tuesday, the trial began a landmark. Trial, uh, meta is being sued by, um, a number of States. And, uh, California, Colorado, Kentucky, and New Jersey. They are seeking get ready for this. Meta. I should point out is valued by the stock market around one and a half trillion dollars.

[00:06:11] These States are seeking 1.4 trillion dollars in penalty works for me. Yeah. So that would leave, uh, poor Mark Zuckerberg with a, you know, a couple of hundred million, you know, I don't know. Uh, poor guy, a couple hundred billion. I mean, I don't know. Um, meta also violated. They say the child online privacy protection app, COPPA by sucking up data about kids. Oh, really? What a shock. Um, they also say that consumer protection laws were violated.

[00:06:40] This was the path set forward. I think by the LA trial, which met also lost instead of going after meta for, um, uh, not protecting kids or whatever, they said it's a defective product. And so consumer protection laws cover it. And that seems to have been an effective pathway, uh, in both New Mexico and California,

[00:07:07] New Jersey attorney general, uh, said we allege in our complaint, she told NPR, what we're prepared to prove a trial is that they're deceiving consumers about the dangers of Facebook and Instagram dangers. They're putting profits over the health of a generation of young people. They're essentially taking the cigarette, uh, route. That's right. Because you, you started with the, the, the judgment in Los Angeles was relatively small, five, $6 million. Then you get to New Mexico. That's $567 million.

[00:07:36] Now you've got this one. And what they're arguing is not that, uh, as they had before that, oh, well, Facebook didn't protect privacy or Facebook had a breach. That was the regular route to go after one of these companies. Now they're saying you've encouraged compulsive usage. Your product it by its definition is, is not good. It's a bad product. It's, it's selling cigarettes. So if that is the tactic they're going to take, what they have to prove in court is that,

[00:08:04] uh, meta prioritized engagement over the safety of young users. They have to prove that, uh, meta actually knew that they were attracting a lot of users who were under the age of 13 on both Facebook and Instagram. And that they knew that the information that they were collecting was in violation of COPPA. Unfortunately for meta, their own internal documents seem to show that they not only did they know, but they saw that as a growth area. Um, this, that's going to be a hard thing to fight.

[00:08:31] Some of this, Frances Hoggins, uh, uh, uh, she was a whistleblower, uh, and they've gone after her big time. Uh, but those, those emails are damning, but actually go ahead, Sam, I'll let you talk. And then I'm going to say, I'm going to put myself in the very odd position of defending Mark Zuckerberg, but go ahead.

[00:08:51] Well, I was just going to, uh, you know, follow up on, on what Robert said, you know, in the absence of a lot of functional regulation and, and it's getting harder and harder to do any kind of real regulation in this country, because we've got one side of the political spectrum. Um, there's fundamentally opposed to all regulation.

[00:09:07] So in the absence of, you know, regulations like privacy laws, uh, or, you know, in the case of automated vehicles, any regulation related to automated driving product liability is kind of the only path forward. We have to protect the public from all kinds of products that are, you know, potentially harmful to, to society, you know, and this, you know, this is a perfect example.

[00:09:33] And I think taking that route, uh, you know, in going after companies like meta is, is really the only option that we have to protect ourselves from these companies. My attitude is a little different. Um, first of all, I don't think it's been proven despite Jonathan height and others who've written about this, that, uh, these apps are harmful to children.

[00:09:59] A, uh, you know, I, you know, this whole idea of social media addiction, I think is controversial. Uh, it isn't a drug. Uh, and I don't think it's, I don't think it's necessary to protect children from this kind of addiction. I know that that is a strange point of view for some, but I really think it's overstating the dangers. Now it look, I don't think meta is a good company. I don't want to make this clear.

[00:10:27] Their goal is to make money and they, uh, you know, as any capitalist company would decided the best way to make money is to make the most compelling product they could. I think if you use that standard, you could go after Hollywood for making movies that you love. You could go after liquor manufacturers. You could go after almost anybody. This is what you do in a business. This is what I do in a business.

[00:10:55] You think, I don't think about what shows are going to be most appealing to the audience. I would love to get you addicted. I can't. There's, there's, there's a difference. Yeah. I mean, it's, it's one thing to, you know, to want to make money. I mean, nothing, nothing fundamentally wrong with that. Ideally, it would be nice to be able to do that without actually harming your customers along the way, because, you know, as with, but I don't think the harm is like comparable to the harm of tobacco. We know smoking causes cancer.

[00:11:25] Well, it has not been demonstrated that using Instagram causes cancer or does anything. It does not cause cancer that we, that we're aware of. And the. This makes you stay up late. Addiction, addiction, you know, is questionable. You know, that's, that's a debatable one. Well, you have to prove harm if you're going to make this stick. And, and there are, there are definitely harms that have been shown like things like body dysmorphia. Well, Cosmopolitan did that as well.

[00:11:55] Teen Cosmo did that as well. Should they have paid. This is, but this is doing it at a scale that we've never seen in history before. Well, because the internet is at a scale we've never seen before. And that's why we need to, we need to step back and take a look at what we're doing because, because of the scale that we're, that social media is doing this at, you know, the amount of content that people are seeing across as many platforms as you're seeing is something we've, we've never seen in the past.

[00:12:23] And so, you know, it's one thing, you know, for somebody to be flipping through a magazine in the past, but this is, this is something totally different. And so I think we need to take a step back and take a look at, you know, what is happening. You know, certainly damage has happened before we are with, with print and with, with other forms of media, but not at this level. And so I think, I think it's, I think this is why we need to address it. We don't go after gambling, which is demonstrably addictive.

[00:12:53] I'm not saying, I'm not saying we shouldn't go after. But it's selective prosecution. And I think it's also anti big tech. Yeah. And let me tell you, I think it's time to start defending tech. That's my opinion. I think this is absurd. I'm with Leo on the fact that we haven't proven beyond a shadow of doubt. No, in fact, all the evidence is quite the contrary. Right. Every study is the opposite.

[00:13:21] The problem for, for meta is that we don't have to do that because the case is not about what meta did. The case is about what meta knew. And that's what the internal documents, that's what the smoking gun. I understand, but one of the issues here is that they're going to say meta didn't do enough with age verification. Do you want age verification when you use the internet? Already I'm being asked when I use AI, what's your age? It wasn't just that meta didn't do enough.

[00:13:47] It's that the internal documents show that meta knew in their high level meetings, they knew that their implementation was just a PR stunt. Because there's no good way to do age verification. In fact, that's one of the claims. They violated COPPA because it didn't do enough to block underage users, but there is no effective way to do that. And I'm with you. I don't think age verification works.

[00:14:12] However, the way that meta has been playing this game is that they've basically said, yes, yes, we'll do this. Yes, yes, we'll do this. And then they make a token engagement and that's it. You can't do that. That is not a responsible company. If you know that your product has at least the potential to have a harmful effect, which meta absolutely did because their own studies were saying we need to look at this.

[00:14:39] Then they have to make a good faith effort to show that they're trying to comply with the standards being set down by our legislative bodies. They didn't do that. So do I think that meta is the bad guy here that they're pure evil and they should be destroyed? Absolutely not. I think $1.4 trillion is a ridiculous amount. Do I think that meta has some culpability here? Absolutely. 100%. I would like them to be a better company. I don't disagree with you. It's obvious that they don't really care, but that's not illegal.

[00:15:09] It's not. It's not. No, you're right. You're right. If we were going to hold every company to the standard that we're holding. Most companies don't care. They would fold. They would all fold. Gambling companies don't care. Liquor companies don't care. Las Vegas doesn't care. They're all honey badgers, baby. That's called capitalism. Yeah. Yeah. True.

[00:15:36] And I think, as you said, virtually all companies don't care about their customers or about society, about what is better for the greater good. But, you know, unfortunately, that is the system that we live in. And that's why, you know, ideally we have some regulation to manage that, you know, to find a balance between, you know, between making money and what is best for society as a whole.

[00:16:07] And, you know, a good place to start would be to make gambling illegal. You know? Yeah. That's not going to happen. Yeah. I know. It's getting worse, right? Make alcohol illegal. People don't like big tech, right? They like to gamble and drink. They don't like big tech. And I see it as a little bit of a witch hunt against big tech. Big tech brought this on themselves. Okay. I understand.

[00:16:28] The other thing, but it concerns me because I think the outcome of this is some sort of reversion to the dark ages where, you know, we say, oh, it's technology. It's evil. It's bad. No, no, no. And I don't think that's the right answer either. I wish these companies were more responsible because they have brought it on themselves. But so have the gambling industry and the liquor industry and so forth.

[00:16:52] I also think that one of the straw men here is algorithms, algorithmic recommendations. And I would submit you take away algorithmic recommendations and you got something pretty darn boring. You've got a chronological feed, which we all say we want, but it isn't very sticky. It's not very fun. We kind of like the algorithmic feed. It's there because it works.

[00:17:21] No algorithms means, you know, everything's boring. I hate the algorithmic feed. That is one of the very first things that I'll disable if I can. On Facebook, you can turn it off, but it turns it back on again, right? Exactly. So, I mean, Reddit used to have the ability to turn that off. Now it's kind of integrated and you can't really shut it off. Because they know people spend more time. They're more entertained. They like it better when it's algorithmic. And I guess that's a different use case. I use social media to check in really quickly to see if there's anything I need to pay attention to. Yeah, me too.

[00:17:50] That's really it. Yeah. But, I mean, if you're living your life on social media, that's a different use case. And that's the use case where the algorithmic recommendations can really become physiologically addictive. It's no longer just a psychological thing. This is what happened was the promise of Facebook and Instagram originally was your friends in chronological order. And it was successful.

[00:18:19] And you and I liked it, Robert. We're present in both cases. We're old folks. Maybe you too, Sam. But they also noticed that they really were, and this TikTok, I think, was the first to really point this out, entertainment mediums, not social networks. And if you're going to be an entertainment medium, you got to pick. You want the stickiness. It bothers me. I go to X to keep up on AI.

[00:18:47] It's very, it's sadly, it's really the best place to keep up on AI. But when I refresh the feed, I'll see something from 25 hours ago instead of the most recent stuff. I want the most recent stuff. But we're the outliers. I go there to get information like you, Robert. Most people use these things as entertainment. I watch Lisa and friends. I watch everybody on Instagram. They're going, oh, that's funny. Oh, that's funny.

[00:19:15] They're not looking for posts from their friends in chronological order. However, they're looking for an entertainment medium. It's like TikTok, right? TikTok really, I think, pointed out that's what people really want. They're just doing what people want. So here's the thing. I'd ask open question, honest question. We can agree that that's bad, right? That kind of engagement that just goes next, next, next, next. That's not healthy. Well, no. We can agree that it's so indefinite. Here's what I would submit.

[00:19:44] I would say, I will stipulate. It's so fun that you have a hard time stopping. It's just like, you know, an amusement park. It's fun. But is it fun? Or because that's how addiction works. Addiction, it's fun until it's no longer fun. And then it's a necessity. Yeah. Yeah. Well, isn't that what we call fun? Well, yeah. But I mean, we're talking about microbursts. No, as a priest, there's a deeper satisfaction. Yeah. I'm trying not to push that too much. Yeah.

[00:20:15] You know, there are things that I do that are fun, you know, that, okay, now I stop and I go and do something else. You know, but when you feel a compulsion to keep doing it or if the amount of fun you're getting from it is wearing off, you know, referring to what Robert was saying, if it's wearing off and you have to do more and more to get that dopamine hit, then, you know, then it's gone beyond fun.

[00:20:45] And I think, I think that, you know. Whose job is that to protect you from it? And see, I'm with you on this deal because I think legislating that is difficult. When you get in a car, Sam, you know what's fun, Sam, when you get in a car, you step on it and you go really fast. Is that a bad idea? Yes. Absolutely. Should the car companies be forced to put governors on there that keep you at 55?

[00:21:15] Not a 55, but arguably, yeah, they should have speed limiters on the vehicles. But they don't. You know, when you're on public roads, you know, there should be an upper limit. And, you know, that's actually something they're starting to do in Europe. You know, they're putting, I think it's 110 miles an hour is the speed limit. You know, I have a car in the driveway right now that can go significantly faster than that.

[00:21:38] And there is no, there's no reason why this car should go faster than that or why it should accelerate as fast as it does. I mean, this is a 6,000 pound SUV that goes from zero to 60 in three and a half seconds. This is ridiculous. It's a bad idea. There's no reason why any vehicle on the public roads should be that fast. I bought ludicrous mode when I bought the X. I did it a few times and I realized this is dangerous.

[00:22:06] I'm going to red out and I'm going to run into something and kill myself. So I didn't use it. But see, isn't the responsibility on the user? I mean, if we had, okay, maybe if we had good age verification methodologies, we could put a governor on Instagram that makes it unappealing to young people. I agree with that. I don't think age verification is the right solution.

[00:22:32] You know, because, I mean, one of the things you obviously lose with that is any degree of anonymity. You know, and there are real compelling reasons to have anonymity online. You know, it's been talked about endlessly. But, you know, being able to access information that, you know, maybe, you know, you don't want certain people to know that you're looking at.

[00:22:58] You know, and there's obviously negative things, but there's also a lot of positive things that can come from that. And so I think that, you know, we do need access to anonymity online. And age verification is fundamentally opposed to that. So I don't think age verification is the right solution. I'm not saying I know what the right solutions are, but I think what we're doing so far is not the right way. My feeling, I agree.

[00:23:26] I mean, look, there is maybe a societal issue. I don't think the answer is to make Instagram unappealing. So that, I mean, I think you've got to give some responsibility. It's like saying, you know, we should put something really bitter in alcohol so that people who want to get drunk have to choke it down. That, you know, I think you've got to allow some responsibility on the end user.

[00:23:50] The other side of this, which is what really concerns me, in the process of going to court, Metta appealed to the Ninth Circuit saying, Section 230, Your Honor, we should be protected. And the Ninth Circuit said, no, you have no protection. You've got to go to court and fight this. And this worries me, too, because Section 230 doesn't just protect big tech that can afford lawyers by the dozens.

[00:24:17] It also protects people like me who run forums and chat rooms. And if suddenly I'm forced to go to court to defend my Section 230 immunity, it basically means 230 is useless. Section 230 protects me from going to court. It dismisses it out of hand, saying, I'm not responsible for content somebody else posts on there.

[00:24:43] Section 230 addresses a different side of Metta's business. So 230 specifically calls out the publisher responsibility of Metta or you as Twit, which is content put there by other people. You cannot be held legally responsible as long as you respond in a reasonable manner when it's pointed out that you have content that is offending of some sort of statute.

[00:25:08] What they're saying about Metta in this case is that this is not a content issue. This is a platform issue. This is the way that the platform was created. So, therefore, it has nothing to do with the content that's being put into Facebook. It has everything to do with how that content is being used by the publisher. So that's why Section 230 protections don't come into play here. It's not a safe harbor issue. It's an actual attack on the way that Facebook and slash Metta does business.

[00:25:36] And for that matter, every other social platform that has an algorithmic feed. The algorithm in and of itself is – that is a form of publishing. That is something – that's not the users creating the content. That is something that the company has chosen to do in the way that they are feeding this information. They are selecting what their users are going to see.

[00:26:05] Well, EFF disagrees. They say this undermines Section 230 and it will force online platforms that host user speech – and they particularly singled out small platforms like mine – to fight lengthy and costly lawsuits before they're dismissed under Section 230. It pulls the legs out from under Section 230.

[00:26:29] According to EFF – now, I'm not a legal eagle, so I'm going to defer to the lawyers. This is an article written by EFF's legal division. Unfortunately, they say this opinion seems to be part of a larger trend reflecting the Ninth Circuit's increasing disdain for Section 230. And apparently, free speech rights more broadly.

[00:26:55] And it means just one more weakening of Section 230. And that's kind of my whole point here is, yeah, Meta's not a great company. I don't want to be put in the position of defending Mark Zuckerberg. But I do want to be put in the position – I do think I need to be put in the position of defending technology and social networks against lawsuits that will have a consequence that go far beyond punishing Meta.

[00:27:22] We had a really good discussion in the EFF room at DEF CON. And this was actually one of the things that we went back and forth with a large number of us. And the example I liked was someone saying, look, you have to focus on a specific issue of the lawsuit, the COPPA violations. The COPPA violations have nothing to do with content.

[00:27:43] They have everything to do with the fact that Meta was collecting information on individuals that they knew, they knew, according to their documents, they knew were minors under the age of 13. So that is 100% not a Section 230 issue. But it does raise the issue of how do you verify age? And there's no good way to do that without completely – in order to verify age, every single person using a platform, including the full internet Linux,

[00:28:13] your phone has to prove their age. And that is an incredible invasion of privacy. And that's what I mean by the – yeah, in its narrow interpretation, it makes sense. Oh, yeah, you should do that. But people need to consider what the larger implications are. Right. But now pull out and look at, again, those internal documents that show that Meta, they were making feeds, algorithmic feeds, specifically targeting those underage. So that's not an incidental thing.

[00:28:43] That's – this is a profit center. Right. So once you pass that line where you are now actively trying to get that engagement because you know it's a profit center, you can no longer plead that, oh, we're just publishing. You're not publishing. You are now manipulating to attract an audience that you know you're not supposed to. In that very narrow way, I think they are culpable. And maybe – this is Judge Yvonne Gonzalez-Rogers. We know that name.

[00:29:09] She was the one who spanked Apple in the Apple Epic lawsuit. She has ruled on many big tech cases. I think she's smart. But this is going to – to be effective has to be a very narrow decision that only impacts Meta. Yep. And I fear that it will have far broader reaching impact. I don't care if Meta goes out of business. I'm with you on that. I got no problem with that. No, but you're absolutely right.

[00:29:37] This is the kind of decision that people can start to say, oh, well, you know, there was this movie that had this three-second scene that I didn't like. And it harmed – Yeah, made me addicted. I had a – you know, or websites that use link bait. Look, I hate websites that use link bait. You know why they use link bait? Because you click it. Yeah. They manipulate you. Look, I really don't like it when car manufacturers make attractive vehicles because then I might buy them. And that's a harm to me. So – I think the minute they put fins on the Cadillac, they should have gone to jail.

[00:30:08] I mean, it's easy to show how absurd this can get if you push it. So that's why if it's a very narrow ruling that just says Meta, you knew and you went after it and you shouldn't have done that, okay, I don't have a problem with that. I just worry about the larger impact of a decision. And by the way, I find myself in a very uncomfortable company with the Wall Street Journal editorial board who I am not exactly a fan of.

[00:30:36] They are as far right as you can get. But they call it a 1.4 trillion state tort raid on Meta. They want to turn social media into the next big tobacco. And, of course, the Wall Street Journal says that's a terrible idea. Terrible. The problem is it keeps getting cast as a bellwether trial. It should not be. It should not be. Now we've got the means to go after all these guys. And guess who's next? Google.

[00:31:06] Yeah. All the people. Amazon. Yeah. All of them. All the AI companies. X. And all this does is it empowers people who say, well, my child was harmed by this. I mean, the harms we're talking about are, for the most part, pretty insignificant. Oh, like they stayed up late and didn't do their homework. I mean, I'm getting flashbacks to the decades long campaign against video games. Yes.

[00:31:37] I mean, so. This is, in my mind, this is the same thing. This is Tipper Gore saying we got to put labels on albums. We got to stop this rock and roll music from infesting our youth. Cut your hair and turn your music down, you punks. What they do over at my side of the pond, which is you put pictures of cancer-ridden lungs on the back of tobacco, which does not work. 100% does not work. But it is kind of funny to see it.

[00:32:06] We were in Rome. More people were smoking. It's like everybody smokes in Rome. Those warnings, Canada has them too. I've seen them all over, except in the U.S. We have fairly stayed warnings on our cigarettes. They don't. Yeah. If people want to smoke, they're going to smoke. They're going to smoke. Kids going to Instagram. And, you know, yes, admittedly, a small, small number of people, there's some real harm. They, you know, they harm themselves. They stop eating. They do.

[00:32:36] But that's, you know, you've got to think about the overall impact versus the outliers. You can't make rules about things based on harm to a few, right? When the Pinto gas tanks blew up, that's different, right? There was a very clear harm there. There was a clear harm. There was no uncertainty. There was a clear harm.

[00:33:00] Sticking the gas tank behind the rear axle, between the bumper and the rear axle, where in a rear collision, it gets crushed between the two. No, that's a juicy, juicy crumple zone. That's all that is. It's juicy. But it didn't put Ford out of business. But there's some things that you don't actually want to crumple. The penalties to that were significant, and Ford stopped doing that, but it didn't put them out of business, right? I don't think so. No, they're still around. And you know what? Good news.

[00:33:29] They're not putting gas tanks in the back of the vehicle anymore. Yeah. So that's okay. That was a good thing. So we'll see what happens. But just pay attention to this, because there are people out there who would very much like big tech to take a tumble. And I think that the problems we face with big tech are problems we face in general with late-stage capitalism, with greed.

[00:33:58] And yes, those are real problems. But I don't know if the solution is to take these companies down one by one, because you won't like the world that we end up in. All right, we're going to take a little break. When we come back, there's more to talk about. Did you get a Pixel 11? Anybody? Anybody order a new Pixel? I thought about it, but my 8 still works so well. I guess I can't justify it. I know. I'm still good with my 10 Pro XL. Yeah. Yeah.

[00:34:26] So Paul Thurrott did order an 11. So... Yeah. But we'll talk about it. I usually do every other year. Yeah. You know, I'm kind of an iPhone guy. I have to admit it. I don't want to be. I just don't... It's not... There's not... I wish there were more choice. I really do. We'll talk about that when we come back. You're watching This Week in Tech. Great to have Father Robert Ballasier, the digital Jesuit. Do you want to plug anything? I just see your blue sky handle.

[00:34:54] Actually, I do want to plug one thing. And that is my next-door neighbor here in Las Vegas. Long-time Twit Army listener. I had no idea. And he had been helping my parents when I'm not in town. So like when my father falls or something. And the funny thing is he actually once... When I was visiting, my mom says, Oh yeah, my son is here. He's a priest from the Vatican. And my neighbor says,

[00:35:24] Oh, I know a Jesuit priest from the Vatican who is on Twit TV. It's my favorite network. And she says, You mean Father Robert? Yeah. No, that's my son. So Scott Blalock, you are absolutely wonderful. He's gone through a lot of hardships in his life. He just survived cancer. And now he's actually packing everything into an RV. He's going to travel the US. So I mean, I love that. I love that for him. Scott, thank you very much for years of helping my parents.

[00:35:52] Oh, Scott, like all of the Twit Army, you're a good man. You're a good man. Also with the Sam Apple Salmon, who is, of course, the host of Wheel Bearings. Covers cars. We've got lots of car talk coming up. So we'll do that too. Our show this week brought to you by Superhuman. And their new Superhuman Go. Have you seen this? It's so cool.

[00:36:16] Between the endless stream in our lives of emails, documents to review, urgent decisions, staying on top of everything could feel like an impossible task. We have a contract that we're about to sign as a company. And we sent it to the lawyer. He said, I'll get back to you in a month. And it's like, no, no, we need, we need. A lot of us were promised that AI would make work easier. I told Lisa, just give it to Superhuman Go.

[00:36:44] Somewhere along the way, AI just feels like one more thing to manage along with everything else. That's why so many people are turning to Superhuman Go. I love the way this works. Superhuman Go is an AI chat that's always there when you need it, already aware of what you're doing. It doesn't ask you to start from zero. It sits right behind your, right next to your screen, just like Grammarly. It's the same people who give us Grammarly. So whether you're drafting an email mid-meeting or summarizing a long thread before a call

[00:37:13] or hunting for a firewall, three tasks deep, you don't have to switch context. That's always, you know, how that is when you cross through a door and you forgot what you came for. You don't forget what you're here for. Or Superhuman Go works seamlessly inside your existing tools to give you the help you need without ever making you lose your place. Lisa took the contract. It was open. It was right on her screen. She asked Superhuman Go, give us what you think. It was so great.

[00:37:42] Because Superhuman Go is from the makers of Grammarly, it works across the tools and sites you already use. We use Grammarly like crazy. So it was just a very natural extension of that. It's in your browser. It's in your inbox, your docs, your project tools. It's not intrusive. We're not talking, you know, clippy here. It's just there in the background for when you need it. Put simply, Superhuman Go helps you be the best version of yourself at work without the friction,

[00:38:09] without that context switching, without the busy work. You have the time and brain space to solve the problems that matter so you can focus on the work only you can do. That's the whole promise of this, right? Let Superhuman Go do the busy work. You'll do the thinking work. Go's AI Chat is available right there on the side of your browser, always ready. No new tabs, no switching apps, no losing your place.

[00:38:35] Use Go's AI Chat to draft emails and messages, to summarize long threads and documents, to search across all your apps without leaving the page you're on. Because we are big-time Grammarly users, it was just a really natural step for us to go to Superhuman Go. Prep for meetings in seconds. Follow up without starting from scratch. Lisa uses it when she's going to prepare for a sales call. It's super handy to say, okay, brief me on this call that's coming up.

[00:39:05] Superhuman Go keeps up and handles the repetitive stuff so you can move forward. This is AI that works with you, not on top of you. Check it out. Find out more at superhuman.com. That's superhuman.com. Thank you so much for supporting this week in tech. Robert, I know you're a big AI user. I am, indeed. I've fallen down the AI well, the rabbit hole.

[00:39:33] Help me, I can't get out. The new Pixel 11, we were talking about it, comes with AI. The New York Times asks the question, does anyone want that? There's a new camera mode called Magic Capture. I think that's the biggest feature. They say, Google said in the Trevor Noah presentation, people think camera phones deliver too much realism. They're too accurate.

[00:40:03] They want something a little more artistic. So, I mean, this to me looks just like Instagram filters, to be honest. It's just filter. I mean, everyone is anxious to call every product they have AI enabled. This is really just machine learning. And it's not particularly innovative machine learning. It's old machine learning from maybe 10 years ago. So, is it interesting? Sure. I mean, anytime I can do something local, I would prefer to do it that way rather than in the cloud.

[00:40:29] And yes, I'm sure some people will be able to use the engine to do some interesting, creative things. But AI so that my photos look better? No, no. No, I'm not going to buy that. If it comes with the phone, great. But that's not a driving feature. Google's putting Gemini everywhere. In fact, Lisa's new car that she just bought has Gemini as the assistant.

[00:40:53] When I set up the Gemini, little Gemini, the new speaker for my ex, Jennifer, she said, I keep getting ads. So, I set it up for her. And the first thing it says, do you want the AI assistant? I said no because I thought it might confuse her a little bit. Your new phone can order groceries? That seems like a recipe for disaster. Hail rides, book restaurant tables? I have not tried this.

[00:41:23] As much as I am an AI fanatic, I don't really trust AI to book restaurants for me. They keep trying to feed us that this is going to be Jarvis from Iron Man. And it's not there yet. It really isn't. And they really should stop because once they do get a product that's close to an actual virtual assistant, we're not going to believe them. Yeah. Oh, that's a good point. We're going to say, oh, yeah, we've seen this.

[00:41:48] I've had too many failures of all these systems to trust it, to allow it to do anything on its own. You know, whatever, you know, if I'm, you know, if I need to put stuff on the grocery list, I'm going to go in and do that myself. I'm going to add it to my grocery list. If I want to order a car, I will do that.

[00:42:12] I will pick up my phone, open up Uber or Lyft or Waymo or Zoox, you know, depending on where I am, and order a car ride. You know, I'm not going to rely on any of these assistants to do that because too often they get it wrong. You know, they misinterpret where I want to go or what it is I want, and it's just not worth the hassle.

[00:42:36] It's, you know, it's funny because, like I said, I spend more time with AI than I spend with my wife, to be frank, than any human. I mean, and I use it like crazy, but I use it in coding. I use it in sysadmin stuff. It's great for setting up stuff. For some reason, the more analog things, interacting with humans, self-driving, I'm much more reluctant to use it. I feel like you're right.

[00:43:05] It's not reliable enough. Why is that, Robert? I trust it to code. I don't trust it to make a book, a reservation. We trust it with things that we can fix at a later time. We don't trust it with things that are making decisions in real time. That's a good point. So I can refine the code. It can be refined. We can have process. It takes time. It takes a long time.

[00:43:34] It takes human AI interaction. I don't trust it to make phone calls for me. A decade ago, I had a fascinating conversation with somebody who was a head of automotive at NVIDIA at the time. And one of the things he was saying about AI was, AI is going to be great at providing superhuman performance at very specific tasks.

[00:44:01] But it's probably never going to be able to be really generalized. And I think what we're seeing now is exactly that. It is really good at very specific tasks like coding. Coding is something where there are boundaries, there's rules, there's constraints around. You can write a chunk of code, test it, know if it works or not.

[00:44:30] But when there is more nuance to it, things become much more difficult. That's where it falls down. You don't ever see fully autonomous vehicles. Like there will never be a car that can drive down any city street. I think, you know, what I've said is I don't expect to see what we call level five vehicles. So the SAE levels of automation, the two key ones are level four and level five.

[00:44:56] Level five means that it's able to drive anywhere under any conditions. You know, anywhere the human can drive, the system can drive. Level four means that it can drive without human supervision or intervention within some constraints, within what we call an operational design domain. We kind of have level four now. We absolutely have level four today. My neighbor drove up from Southern California.

[00:45:22] Actually, his brother-in-law drove up from Southern California, put into his Tesla. He has a Model 3 or no Model Y. Put into his Tesla, drive me to this destination. Not only did it do the highway stuff, which most cars can do very well now, it did the city streets. I said it is stopped. But that's still a level two system. That's only level two? Level two means that it requires human supervision. And a human has to be ready to take over at any time. Yeah, but he did say, I don't know how I got here.

[00:45:52] And this is the fundamental flaw with level two. Level two defines, you know, this is what it can do. But I don't entirely trust level two systems. As an engineer who worked on active safety systems for many years, I don't trust those systems because I know what can go wrong. You know, level four, you know, does not require human supervision. And even there, they still make mistakes.

[00:46:20] Or there doesn't even have to be anybody in the vehicle. It can be a... That's what we're seeing now with Waymos, aren't we? We're seeing that with Zoox and Waymos. Yeah, you know, they're driving around, going to pick somebody up, and there's no one in the vehicle. So that's level four. That's level four, yes. And we're seeing it on city streets. We had Dan O'Dowd on Intelligent Machines on Wednesday.

[00:46:46] He's the guy who ran for Senate in California, not because he wanted to become a senator, but because he wanted to run ads against Tesla. And so people weren't taking the anti-Tesla ads. But once you're a political candidate, they have to. So you do it as a pack. Yeah. And he ran... I remember this. It was a couple of years ago. He ran these ads showing Teslas driving over mannequins, size of child, backing over people, doing all sorts of stuff.

[00:47:17] He's an interesting fella. He claims that software can be perfectible. And the proof is that he writes... His company writes a software for F-35 fighters and the A-380 plane. He writes software for, you know, mission-critical NASA. You know, situations where, A, the software has to run perfectly, cannot have any bugs, and B, it isn't even... Can't be maintained easily. You know, Green Hill Software... He says that's possible.

[00:47:47] Green Hill Software is an amazing company. Is that his company, Green Hill? Yeah. Green Hill Software, yeah. They do... They make some amazing stuff. He was really interesting, but I'm not sure I buy what he's saying. And their stuff is widely considered to be some of the safest software you can get. Right. He uses a subset of C. Yeah, they use a lot... They do a lot of embedded systems software. And it works really, really well. But, you know, I agree.

[00:48:15] And I have a lot of respect for Mr. O'Dowd. But, and the work that Green Hills does. I mean, we've, you know, in the past, when I was still an engineer, we looked at using their software. It wasn't the right tool for us. It's expensive. Yeah, it's very expensive. But, you know, for certain applications, it's really, really good.

[00:48:34] And, you know, I think, you know, what they've done is taken all of the best practices and taken them, you know, to the nth degree, you know, to try to make the software as safe and secure and reliable and robust as possible. And they've done an amazing job at that. Do I believe that there is no way to ever hack that software or that it'll ever crash?

[00:49:04] Not entirely. I think it's far less likely to crash, far less likely to be hacked than almost any other human-written software. But, you know, I can't say with absolute certainty that it's, you know, absolutely perfect. On a mission, I think, to stop full self-driving, Tesla's full self-driving. He, by the way, said something interesting. He said the Waymos are fine. He didn't have as much of a problem with the Waymos. They're not perfect either. You know, Waymos make mistakes.

[00:49:33] Zooks make mistakes. They're better than humans, though. I guess that's the question. Their sensor technology is better than Tesla. We know that. Yeah, their sensor technology is a lot better. Waymos and Zooks vehicles are a lot better than humans within certain driving environments. Right. Correct. In fact, they took the Waymos off the freeway. That's where they define, that's how they define the operational design domain. Right.

[00:49:56] You know, when they know it can work as safe or better than humans within certain constraints, that's where they will operate. You know, there are lots of places where Waymos do not operate because they are not confident that it can do better than humans. And, you know, when you look at the statistics, the crash statistics, you know, half of all traffic fatalities happen on rural roads. You know, half of them involve alcohol.

[00:50:25] Well, that, you know, the alcohol, getting that out of the picture, you know, obviously Waymos, you know, autonomous vehicles don't have that problem. But certainly half of, you know, half of the fatal crashes happen on rural roads. Waymos don't operate on those roads. You know, fatal crashes are a lot less likely in an urban environment, in a dense urban environment, because for one, your speeds tend to be a lot lower. So that's, you know, they are safer, again, within that constraint.

[00:50:56] Right. He was interesting. I feel like he's a little bit got a chip on his shoulder about Tesla, but you don't disagree. You don't disagree. I think, Sam, you've said that before. Yeah. You know, I think that Tesla's approach is fundamentally flawed. And, you know, I was glad to hear him call out, you know, the fact that, you know, they are testing, you know, they're putting beta versions of safety critical software in the hands of untrained consumers that do not know how to evaluate.

[00:51:26] Yeah, that was his main point. Yeah, that's something I've been calling out for more than a decade. And because we are in the road with Teslas, we are part of the beta test. Yeah. I mean, you know, usually... We're the NPCs in the beta test. Usually, you know, if you are going to get involved, say, you know, in experimental medical therapy, you know, you have the concept of informed consent. You know, you are the one that is taking some experimental treatment for cancer.

[00:51:53] You know, the doctors have to tell you, OK, here's the things that could go wrong. You know, we don't know if this is going to work. We think it can help. But here's the things that might happen. So you have that concept of informed consent for the participant. The problem with what Tesla is doing, and also to, you know, this also applies to Waymo and

[00:52:18] Zoox and everybody else that's out there, is that all the rest of us on the roads have not consented to be part of this experiment. Right, right. You know, pedestrians have not consented to be part of this experiment. And so, you know, we are at the mercy of their testing and what they're doing. And, you know, I think that, you know, Waymo's approach is much, much safer, much more robust.

[00:52:42] And they are much they have generally shown much less likelihood to get in, you know, get into situations and crashes where, you know, where other road users are going to be negatively impacted than, say, for example, you know, Tesla's, you know, there have been at least 50 crashes with the Tesla robo-taxi fleet in Austin since they launched it a year ago.

[00:53:09] And, you know, many of those have been the fault of the Tesla vehicles, not, you know, in the vast majority of the cases where there have been crashes with a Waymo vehicle or a Zoox vehicle, it's usually actually been the fault of another human driver that is interacting with that vehicle. But there was just an incident in L.A., I think, an incident about a week or so ago, kind

[00:53:34] of similar to what happened with Cruise in San Francisco, where a pedestrian was struck by a human-driven vehicle and knocked into the path of a Waymo. And the Waymo detects, saw the pedestrian and hit the brakes and slowed down. We don't have all the details yet about exactly when it detected and when it slowed down. But by the time that the pedestrian was knocked into the path or actually into the side of the

[00:54:02] Waymo, the vehicle, I think, was down to about five or six miles an hour. So, you know, it was slowing down. That was more of a problem, though, because Waymo kind of covered it up, right? Yeah, that's, yeah, that's, and that's an issue. That's one of the things that needs to be addressed with for all of these companies is we need to have, we need to have some rules, you know, mandatory data disclosure. Right. And, you know, one thing. My sense is Waymo's better about that than Tesla, though. Oh, yeah.

[00:54:30] Since 2023 or 2022, there's NHTSA, the National Highway Traffic Safety Administration issued a standing general order that any vehicle that is operating at level two or above that is involved in a crash. The company has to submit the data about that crash to NHTSA for and there's a database that you can download that's got all this crash data. And there's a lot of really interesting data in there, except for Tesla.

[00:55:00] Tesla submits all their data, but for reasons that, you know, we can only speculate about, NHTSA has allowed Tesla to redact most of the data from the publicly available database. So you go in there, you see, oh, there was a crash at this location on this date. And that's literally it. Everything else Tesla is declaring as company proprietary information. And so we don't really know what has happened in most of these crashes. I mean, that's what I want to see.

[00:55:30] I want to see if you're going to be testing on public streets, if you're going to be testing with our safety, you don't get you don't get proprietary information anymore. Yeah. That has to be public information. Yes, absolutely. So Waymo. So one of the things I'm kind of excited about is new looks for some of these self-driving vehicles. We talked earlier before the show about Zoox, which is now able to accept paying riders in Vegas. Did you have you ridden in Zoox?

[00:56:00] I've done it about a dozen times. It's actually it works really well. Really, really well. I'm really intrigued by them because they don't have a steering wheel. They don't have a front seat. It's living room style. They call it campfire style seating. It's set up so that when my Zoox pulls up, it changes to my color mode so that I know that that's my Zoox. That's something they added recently, by the way. How do they do that? What do you mean? It changes the car changes color? No, there's LED light strips and the windows. So you know what color you're looking for and that's your Zoox?

[00:56:30] Right. So like I was at a DEF CON and five Zoox would pull up and you're like, well, which one is mine? You don't want to look at the license plate for each one, but you had your own color scheme. So you just look for the color scheme. The Vegas airport at the Uber stand has a big sign saying, make sure your driver knows your name and you ask his name. Make sure you get in the, because people, I guess, are picking up the wrong people or posing as Uber drivers. I don't know. Yeah. It's all in the same area.

[00:56:57] Or people just getting in the wrong vehicle, things like that. There's no driver. So there's nobody to say your name. So I think that you have to do something. When I, when I went to the car, I was just gonna say, when I went for a ride with, with Zoox in January during CES, you know, uh, I was, you know, out in the pickup area at the Luxor and, you know, there were, you know, five or six Zoox vehicles that pulled in, you know, at the same time.

[00:57:23] And all they had was a little number on the back corner and you had to, you know, you had to go up and look at each one to figure out which one is yours. And I talked to Zoox about this afterwards. You know, I gave him some feedback on, on my experience. And one of the things they've added since then is these led light strips. So in the app, it shows you, you know, the color that's going to be on your car. So you'll see these various led lights. And the other, the other cool thing about the vehicle is that it has no front or back. It's, it's completely symmetrical. It goes either way. Yep. Yeah.

[00:57:53] In fact, it changes like the headlights turn into the taillights and the taillights turn into the headlights. Yeah. You just go from white to red or vice versa. And, and in the same ride, would it switch directions? It can. Oh yeah. Depending, depending on, you know, if it needs to maneuver somewhere, if it, if, you know, if it doesn't have space to maneuver, it can just back up and go back the opposite direction. It's got all the same sensors on all four corners. So it's got the same sensor pods at all four corners. It's got, and like everything is completely redundant.

[00:58:20] So you have, uh, there's a motor on each axle. So the, at the, the two ends of the vehicle, you have the exact same, um, drive unit, suspension, brakes, steering system. It's completely interchangeable. Uh, so, you know, this gets shipped to them from their supplier as a, as a module. They just bolted up to the body. Uh, so if steering fails on one axle, you still have it on the other axle. If drive fails on one axle, you still have it on the other axle.

[00:58:46] Same with brakes and the air conditioning and the compute, some, you know, uh, redundant compute. Everything is all completely redundant. This seems like a good idea. Like, so the Waymo's were Jaguar I-paces. Uh, in fact, Waymo's started to order now some, they still have a front seat and a backseat. So they're still kind of car-like, but they've ordered some new, they call it the Ojai, which is a play on a town in Southern California at Ojai. Yeah. Yeah.

[00:59:14] It's a little, it's a little minivan made by a company called Zeker, uh, which is part of Geely. Okay. Uh, yeah. It's, it's Chinese. Chinese. Yeah. I want to try one of the Zoox, uh, wheelchair accessible vehicles because they do have vehicles. You can specify wheelchair and it will be like a minivan. It's that has a ramp down. So, so those, um, are not autonomous. So what they will do is they will send you, uh, a wheelchair accessible, you know, if you

[00:59:40] request a wheelchair accessible vehicle, they'll send you a human driven vehicle with wheelchair access. It makes sense. At some point they will, I'm sure they will have an accessible version of, uh, uh,

[01:00:20] of the robo taxi. The ramp would emerge out from the floor. You roll in. And then one of the positions, they had to have no seats in there. They don't have anchors for the wheelchair. Nice. And you'd like it, Leo. When you, when you, when you get in there, you, everything is controlled via your phone or the control panel. Your Bluetooth, uh, uh, delivered music starts playing over the, the speaker systems. Uh, and it tells you exactly how many rides you've done, where you're going, how much energy it's going to take. It's kind of a geek ride.

[01:00:49] It's a little friendly. It's kind of fun. Yeah. There's a cup holder, a wireless charging pad, a touchscreen for each of the four seating positions. Uh, so you can pair your phone. Um, you can put it down on the wireless charging pad. So Zoox right now is only Vegas, right? Is it going to? No, they, they are doing, they've got their early riders, early rider program in San Francisco and Austin right now. Okay. And they want Miami. Other cities. Yeah. Why do they, so why do these self-driving cars start in San Francisco, which seems to

[01:01:18] me like that's second only to Boston as the worst possible place you can be a self-driving That's exactly why. Yeah. It's a very challenging environment. Okay. If you can drive there, you can drive anywhere. Yeah, exactly. Okay. If you can make it work reliably in San Francisco, chances are it's going to work fine in, in Phoenix seems easy. It's been wide streets. They're straight. Same with Las Vegas. Yeah. Vegas has a different challenge. It has to be able to identify drunk people, drunk people do things that they shouldn't do. Yeah.

[01:01:47] You know, in Vegas, nobody's ever moving fast enough on the strip except during that Grand Prix weekend, you know, that it's not that much of a problem. One thing I did appreciate is at one point I was going from Resorts World to Area 51 across the freeway and it was a big traffic jam. And so I'm, I'm kind of zonked out. Were you visiting the UFOs? What were you doing? Oh no, it's the, the shopping center. Oh, it's called, it's not actually Area 51. Oh, not Area 51, Area 15, right? Okay. There's a difference.

[01:02:18] It's a play on Area 51. Area 51 is a little further north. Yeah. Yeah. No, but it's, it, it started like turning my music up in the last five minutes because I get it, it figured I had fallen asleep and I'm like, oh, Hey, cool. Well, again, this is something that, that people should be aware of. If you're going to be riding in a robo taxi, there are sensors, there's cameras, there's smoke detectors and other stuff in the cabin. Yeah.

[01:02:44] You know, I mean, if somebody gets sick, you know, falls asleep, has a heart attack, you know, they want to be able to detect that. And, you know, so they, you're being watched, you're being recorded all the time. It's very upfront about that. That's in the app. Yeah. It says you are being recorded while you were in the vehicle. Yeah. And, you know, if you leave something behind, you know, then they want to be able to detect that, send it back to the depot. So they can take that out. Or if the vehicle needs to be cleaned, you know, whatever it might be. Pausing.

[01:03:14] We'll talk more about vehicles and other things in just a bit with Sam Abulsamit, our car, my car guy. And that's father, Robert Palisare. I, I threw Benito a little curve and started with you, Sam, the digital Jesuit. Great to have you both on the show. Actually, great to have a brand new sponsor. I'm really excited. This Week in Tech brought to you this week by Claude. Or as Paul Therat calls him, Claude.

[01:03:42] I am a Claude fan. I have said this many times on our Intelligent Machine show. My world changed November 24th, 2025, last November, when Opus 4.5 came out. I was using Claude code and I was blown away. Claude is built by Anthropic. You know the name. A public benefit corporation.

[01:04:10] And it is absolutely my go-to for the hardest problems in AI. I use a variety of AI models, some of them local, some of them in the cloud. But when, when I need the best, I always turn to Claude. And there's nothing like the Claude code harness. Claude is the AI for problem solvers. It's the collaborator that understands your entire workflow and thinks with you, not for you. Whether you're debugging code at midnight, building a financial model,

[01:04:40] or strategizing your next business move, Claude extends your thinking to tackle the problems that matter. And I've been using Claude pretty much nonstop and religiously. I have a Claude Max subscription, which I just, I love, because I get to use Claude whenever I need frontier quality brains. If you're a developer losing your day to tasks you want to hand off, including, as I mentioned, simple things like system administration.

[01:05:09] When, you know, a server is down, a Docker container is not launching, when my screen isn't blanking, I say, I launch Claude code. And I say, fix it. And it fixes it. You get your choice of models. Fable for the really tough stuff. This is, I don't think there's anything comes even close to Fable. It's the highest intelligence I can use. Opus 5 for when I want to do some coding is great. In fact, I use the two of them together.

[01:05:36] I have Fable plan, I have Opus 5 code. Claude code lives where you already work. You can use it in your terminal. You can use it in your IDE. You can use the desktop app. You can use web. You can use GitHub. I've used them all. It reads your code base. It's kind of amazing seeing it work and go, yeah, I get it. What do you want to do? Takes on real engineering work. You can use it to write tests. In fact, that's part of my Claude setup,

[01:06:04] setup is I say test-driven design. Write the tests first. I do red-green and it's so good at that. Write the tests first. It's great at refactoring. In fact, I'm doing that right now. I'm using Fable. I used Fable to redesign our Twit sales system, which is a huge MySQL system written in 2012. I pointed it at it. I said, here's the database. Here's the code. It rewrote it. 81,000 lines of Go code beautifully. Beautifully.

[01:06:33] And it's doing the debugging. It wrote the tests. It's amazing. Such great tests. Even watching pull requests, fixing CI failures while you're away. It's so good. I'll never forget one of the first programs I wrote with Claude. I think this is Opus 4.8 probably. I wrote an RSS reader. And it pushed it to GitHub. It even invoked CICL. It said, okay, I've built versions for Mac, Windows, and Linux.

[01:07:03] Which one do you want? And I went, you did what? I said, wow. Lay out what you want. Or if you're using Claude code or one of the Claude co-work apps, you just hold the space bar, talk it through, and Claude executes. It's a very natural way to work. Claude co-work is amazing. That is the desktop app that can handle recurring tasks for you, things like scanning Slack and email, create a prioritized action list so nothing slips through the cracks.

[01:07:31] And deep research, we use that all the time to research people. It does the best show prep I've ever had. I wish I'd had this my whole career. To research companies who want to advertise, it can pull together and does detailed cited analysis from dozens of sources. Claude design is how I design my website. It lets you quickly explore and build out ideas. While the new reflect beta gives you a look at how you're actually using Claude

[01:07:59] and encourages you to step back and decide what you still want to do yourself. I just, I mean, these guys, there's some secret sauce at Anthropic. It's kind of amazing. Claude Fable, which is Anthropic's most intelligent model to date, is built to carry a long, complicated project end-to-end instead of needing it broken up into separate tasks. It can one-shot the most amazing things. That's why companies like Stripe and Shopify and Pfizer and Rakuten and Notion

[01:08:28] all trust Anthropic to roll out AI in their businesses. It's why we're using Claude to rewrite our entire business workflow. For problems worth solving, get started with Claude at claude.ai slash technology. That's claude.ai slash technology. And check out Claude Pro, which includes access to all of the features mentioned in today's episode. Claude.ai slash technology. I bet you use a little bit of fable in your life, Mr. Robert, Father Robert. I do.

[01:08:58] This is my AI. There are many like it, but this one is mine. This one is mine. I say that all the time. And it's funny because I started in Claude code. That's what I use for months and months and months. I use a Genix framework called Harness, a Harness called Hermes. But honestly, every time I go back to Claude code, which I do daily, I'm kind of blown away by it just knows. It's gotten a lot better. It's so smart. Much faster than its competitors. I don't know.

[01:09:27] So, Leo, I got to ask you, are you now okay with Claude's fingerprinting? I'm not a fan. Look, I understand. They're doing it because the EU is making them do it. Yeah. And I understand that. I am a fan of, in fact, I think Google took the wrong approach. Google's been doing this for years, right? They call it SynthID. And the idea is that you could identify something that was created by AI.

[01:09:55] And I really like it that they put it on images and video, like they put a watermark. So that if you see a deep fake, you know it's a deep fake. Unless it's been cropped out. They stopped doing that. They said, we're going to have a hidden SynthID, but we're not going to have a visible watermark. It's like, no, that's the absolute wrong direction. And I feel the same way with Anthropics watermarking. If you're going to do it, make it visible. Make it easy for people to see it.

[01:10:26] Except that the approach they're taking, you know, kind of is kind of the opposite. They can't really do that with the type of solution that they've come up with. They feel like if they do that, people will reverse engineer it, which of course they're going to do anyway. And strip it out because there are people who really don't want you to know they're using AI to make deep fakes and things like that. And I do think that's problematic. I understand that what Anthropic did is because they were required to by the EU. And by the way, it's not just Anthropic.

[01:10:55] Everybody now is doing that. I wish it were more visible, but I also understand why they're reluctant to do that because the bad guys are ingenious. And yeah, so yeah, I'm not a fan of watermarking. I'm not. As far as I can tell, it's as invisible as you can make it. But that's kind of, I want it to be visible. Can a visible watermark be backwards engineered? Absolutely 100%.

[01:11:22] However, what it would stop is a lot of the very low effort AI slop that has been just clogging up our feeds. Yeah. So I'm okay with that. I would love getting ready. And you know what? I think if people saw, you know, Made by Claude in the lower right-hand corner or the upper right, that would be a good thing. I'd have no problem with that. I think that's what you really want.

[01:11:43] And content creators would learn how to use it because even if you're legitimately using AI to help tell a story, it shouldn't matter whether or not it's AI or not. The only ones who really don't want you to know are those who are, they're either trying to spread disinformation or it's an automated system where they're taking stories off of Reddit and trying to create them. Yeah, exactly. And I don't want either of those. Those are not just annoying. They're harmful.

[01:12:13] Yeah. Yeah. Maybe put it in the metadata. That would be okay with me. You just, you need a way to check it. The problem right now is... The problem is for it to be useful, it needs to be persistent. The metadata can be manipulated. There isn't really a very good solution for this. Yeah. I think that's really the real problem. The EU has mandated something impossible. Anthropic is working with camera manufacturers to build that technology in.

[01:12:42] So basically you can register original content. So if there's ever a question about whether or not it's AI or real, you can say, no, no, it's in the register. The original source, except with digital cameras, is anything ever really real? Well, no, that's kind of the point too. Yeah. That's the Neil. Is real? What's real? Is a painting real? What's real? Is this video you're looking at real? What's real? We had this exact conversation 10 years ago, Leo, when you were talking about photography. I have a like it.

[01:13:12] It's not actually my office behind me. I'm not sitting in my office. No, you're not. It's green screen. I have an M11P that has content credentials in it. It's turned on. So if you look at a picture of mine, you can check it. There are websites you can run it through that says, you know, this is original. It was taken with this camera by Leo. I think that's a good thing. I don't have a problem with any of that. In fact, I'd almost prefer it went that direction where real stuff was validated as opposed to unreal stuff.

[01:13:40] All of the cameras in my shop do that. And for a very specific reason, because we represent someone who we need to be able to prove authenticity. So, yeah. Now that Shroud of Turin thing, did you? No, I'm just kidding. Just kidding. If they'd had that, we would know, right? But no, no. What you don't understand, Leo, is that the Shroud of Turin is on the blockchain. Okay. Oh. So it's real. It is real. All right.

[01:14:09] Leo, before you go on to another topic. What is the status of the Shroud of Turin? Real or not real? Where do we stand today? No, it's real. It is a real cloth. There is something on the cloth that- Is it what is just his image? The question is, is it what it is purported to be? Exactly. That's the question. Is there something on the Shroud of Turin? Yes. Does it exist? Yes. Are there people who are saying that it's the face of Christ? Yes. I cannot prove that. It's- That is a matter of faith.

[01:14:38] I have to say, when I went to Jerusalem, which was an amazing experience, we did, you know, we went to all of the, you know, canonical sites, Golgotha and the stone and so forth. And, you know, there is a church built around Jesus' tomb and there's a shrine there. Can they prove that was Jesus' tomb? Absolutely not, because it was so long ago and there's no historical proof. But I'll tell you what is real.

[01:15:03] The people who are there, who are deeply devout, who are believing, and for them, it is a very powerful experience in that you feel the vibe there. And I think that's real. Yeah. That's palpable. It's different. And it's like the shards of the real cross of Christ. That used to be the big thing. It's like if you gathered all the shards of the real cross of Christ, you could solve the homeless population because that'd be enough wood to build millions of homes. Besides the Empire State Building. Yeah.

[01:15:32] No, that's, but, but that's not really the point, is it? It is. There's a Zen saying that you don't want to focus, when you have a finger pointing at the moon, the wise man looks at the moon, not at the finger. The less wise man is focused on the finger. It isn't about the finger. It's about the moon. Don't focus on the finger. You miss all that heavenly glory. I'm glad he did that. I didn't say it.

[01:16:02] Leo, before you go on to another topic, there is one other thing I wanted to address on the AV topic. Oh, yes, please. Which is, you know, something that Dan O'Dowd did not talk about is, you know, and what is actually probably one of the biggest potential software issues is the backend systems. You know, for a robotaxi to work or even a personally owned AV, you know, there has to be some backend systems. It has to be connected somewhere.

[01:16:31] You know, if you're going to have a car that is going to go park itself, if it's your own personal vehicle or if it's a robotaxi or an automated truck, there has to be backend systems that communicate with that vehicle to, you know, tell it where to go, to retrieve the vehicle, you know, to dispatch it. And those are actually probably the most likely to have security vulnerabilities. You know, you've, you know, for, and this is, this is one of the problems with security

[01:17:00] is you've got the entire value chain where there are potentials for security vulnerabilities to be introduced. Uh, you know, cause you know, not everything is done in house. You've got different vendors that are producing different, different elements of this. You've got tools that could be corrupted. Um, and so the, you know, the backend systems, um, that manage these vehicles are another part that, you know, we need really need to take a look at, you know, to make sure they

[01:17:29] are also secure, not just the function of the vehicle. It's one, it's one thing to make a vehicle that you can tell it, you know, okay, go to this location and it will drive itself through some environment and get there. But then you, like I said, you, you, you have to have some cloud infrastructure that actually manages all of that. That's kind of scary. So if the internet goes down now, those cars are connected to the cloud via cellular. Yeah. Must be. Right.

[01:17:57] And, and, you know, this is, this is one of the issues that in fact, we've seen with I did not know this actually. I mean, you, I'm sure, you know, we've, you've seen the stories of, you know, there's been, you know, 15 Waymo vehicles stuck on this block and blocking all the traffic. Yeah. Yeah. Well, that's yeah. During that, during that power outage last fall. Right. Um, that's exactly what happened. You know, you'd lost, you lost cellular connectivity in that area.

[01:18:24] And, uh, because those, those vehicles, at least right now are required to be monitored all the time. They have to have connectivity to their backend systems when they lose connection to, to their control center. You know, they're not being driven remotely. Um, but when they lose connectivity to their control center, they are supposed to pull over to a safe, safe spot, um, and stop and wait until connectivity is restored.

[01:18:52] Uh, and that, you know, if you, if you lose that connectivity, you know, whether it's a power outage or, you know, whatever else might happen, uh, just, you know, the, the cell network goes down and usually these things have redundant cellular connections. They have multiple, they'll have multiple connections. Um, but still, if that, if that goes away, then, you know, at the, at best, the vehicle

[01:19:15] can continue on to whatever it's, it's last destination was set as it should be able to do that because the, the actual control of the vehicle is handled locally. And the vehicle that's not being done in the cloud. So, you know, sensing, you know, the perception systems and telling it, you know, uh, you know, what's, what's around the vehicle and, and the maps and, you know, how to get from one point A to point B, all of that's done locally, but all of the management stuff is done remotely.

[01:19:45] Yeah. Wow. I had no idea. At least it won't run and run off the side of the road. It'll still drive. It fails safe, which is, that's good. Right. Well, actually we want, actually it's, it's supposed to be fail operational. So there's a, there's this distinction. There's fail safe and fail operational fail safe means that the system has to be able to detect something has gone wrong and provide an alert to the operator and the operator takes over. Uh, that's what we've had traditionally with driver assist systems in cars, like anti-lock

[01:20:14] brakes or electric power steering systems, whatever else. It fails safe. These things have to be fail operational. So if something goes wrong, there has to be a backup system that continues to safely operate the vehicle until at the very least it can get to what we call a minimum risk condition to, you know, basically come to a stop in a safe location. And so that's why you need all this redundancy of different types of sensors, multiple compute

[01:20:39] platforms, multiple, multiple braking systems, because you don't have that human behind the wheel that, you know, if your, uh, if your ABS fails or your, even your, your power brake assist or your power steering assist fails, the human can still operate the vehicle. It can, you might have to turn harder on the steering wheel to steer it, but you can still do that. You can still bring the vehicle to a stop.

[01:21:03] Um, but you know, when it's all being done electronically, then you need a secondary system that can still stop the vehicle, can still steer the vehicle. That's why, you know, like what I was describing earlier with the Zoox were having all that redundancy of steering, braking propulsion and sensing is essential for these types of vehicles. Interesting. Now that's a little cautionary tale.

[01:21:28] Um, I remember the years of, uh, yelling throughout the house, I'm on the internet. Stay off the phone. This is why I drive a 36 year old car that has no connectivity. No, no driver. Yeah, no, I get it. Yeah. Uh, let's talk about Apple. We were talking about the EU. They've made a deal with the EU. Remember the European commission didn't like it that Apple was charging 30%. They wanted to negotiate that down. Boy, they sure got a break.

[01:21:58] It's 26%. Uh, and okay. Well, uh, John Gruber writing in the daring fireball blog said it's a total near total victory for Apple. Uh, the main commission drops. If, if, if you, uh, if you're an app in the app store using the Apple payment system, the commission drops from 30 to 26%.

[01:22:25] Uh, and all apps still have to pay a 5% core technology commission, even if they're outside the app store. So Apple gets a little, little VIG, no matter where you're buying your app from. Now the point that's made, and I think the EU makes this point, I've made this point is the people who pay for these commissions and fees by Apple aren't really the developer so much as you that, that, you know, when I subscribed to a, uh, Claude max, it's 200 bucks.

[01:22:55] If I go directly through Anthropic, it's, I can't remember what, I think it was 248 bucks. If I went through Apple and for the longest time, I thought that's how much it costs until I wised up and realized 48 bucks of that was going to Apple. And so that's the actually really only 40 bucks, but I guess, uh, no, maybe, no, I don't know. Maybe sales tax, California sales tax. Math is hard. Anyway, uh, I don't know if there's, they don't charge sales tax. Maybe they do. I don't know.

[01:23:23] I mean, they're probably taking a little bit of a hit because they don't want to transfer for the entire thing. They're taking out that's right. Cause, cause 30% of 200 would be 60 bucks. So they're taking a little bit. Some of it's taken by Anthropic, but some of it's taken by me. And that's the point of these fees is really, it ends up costing the consumer more. Um, it's an interesting breakdown that the EU came up with this number.

[01:23:48] Uh, but ultimately I think it's a, it's a, it's, it's the best thing that could have happened from Apple's point of view. Now they have to negotiate the same thing in the United States. Uh, we shall see Epic of course wants Apple to give up a lot more. Well, the, that 26% is for apps that use Apple in app payments, right? If you're using some other payment system, then the fees are lower. Right.

[01:24:15] So there's that 5% core technology commission, the, uh, Apple store distribution. This is how the EU calculated 10% of the app revenue, the implied value of the in-app payment APIs, 5%, the implied value of an in-app purchase using Apple's payment processing, 6%, which is roughly what a credit card costs. So that makes sense. Uh, that's where that number came, came up with. And yes, uh, I don't know what the cost would be if you use a third party store.

[01:24:45] I think it's just the 5% core technology commission that Apple takes. Yeah. Which is not unfair. Uh, I think, I mean, I mean, you're on Apple's platform. What are you getting? What are you getting for that? What's, what's the value add by going through the Apple store? I mean, aside from, yes, you're on the platform. Yes. You get to be in the, you get to be in the, uh, Apple ecosystem, but I think you're right. If you're looking at a traffic or Netflix or Kindle, Amazon's Kindle books, it's kind of

[01:25:15] rent seeking. Like here's the, here's your 5% being in the app store and using in-app purchases, Apple in-app purchases. That's different. It's the lowest friction path. Yeah. And you should pay for that. You're right. There, there is no, there, you aren't really getting any benefit. I really, I think that the way that should have been done is that for anything outside of the app store or outside of Apple's in-app purchases, what they should do is Apple should

[01:25:42] just go back to charging some fee for developers to, you know, to, to use Xcode or, you know, other, you know, to use their tools and then get nothing from the app. Well, they do get a $99 a year developer fee, which is nothing. I mean, that's pretty low. Yeah. I mean, you know, maybe that should be a little bit higher and then nothing for, you know, nothing from the apps because they, they really are not contributing anything. Well, here's the, the, the example.

[01:26:09] So your Netflix, let's say you have a Netflix app on the, on the iPhone and people want to subscribe to Netflix. They do it through the app. They do it through Netflix payment processing, Stripe or whatever. Netflix is going to pay 6% to Stripe, but that's, you know, it's a credit card, but why should Apple get anything? They shouldn't, but they're going to be the opposite of Alex Lindsay. I'm, I don't, I don't agree that Apple should get a dime for that. It's rent seeking.

[01:26:39] They've already, yeah. They've already made their profit on selling you the device. Right. And, and all the subscriptions nowadays that Apple. Yeah. Yeah. So anyway, the EU, I think basically gave in and said, okay, you know, you're going to give us a little break. Thank you. We'll take it. I don't know. I mean, the logic that they're using is that, you know, they are giving developers access to, to an audience, but we know from the early days of smartphones that those smartphones

[01:27:07] that didn't have an, that did not have content, did not have apps died. You know, look at, look at Microsoft and it's Metro. Look at all of its, its efforts to do a mobile operating system. So yes, you do get access to a marketplace that you wouldn't have access to otherwise, but also that marketplace would be useless without you. I hate these situations where they've forgotten.

[01:27:39] That's a good point. Imagine if you couldn't get Netflix on an Apple device, that would be bad for Apple. Yeah. I mean, it'd be bad for Netflix too, but that's the point. It's a mutual thing, right? Yeah. Okay. So 5%, maybe that's a little bit much. Anyway, we'll, we'll watch with interest because as you know, what's going on right now is Apple has to bring a proposal back to judge Yvonne Gonzalez Rogers.

[01:28:07] She's busy, you know, who rejected their initial proposal of, I think it was 27% and said, no, no, you guys, you got to give me something better. So 25. Okay. 25 last offer. The thinking is maybe Apple goes low as 15, but we'll see. And remember there is a cut if you are some Netflix, for instance, and you have regular after a year and so forth, but it's still going to be at least 5%. I don't know. We'll see.

[01:28:36] We'll see what judge John Gonzalez Rogers says. Cause she has been very clear. She spanked Apple. She said, you lied to me. She actually said Apple's in contempt of court because you lied to me. And, and so she's been very harsh with them. So we'll see. Anyway, judges really don't like being lied to. That's a, that's kind of funny. Yeah.

[01:29:05] Apple, by the way, did lay off some people this week in an area that I'm guess I'm not too surprised. A hundred, a hundred people off the vision pro group and a hundred people from Siri, more than 200 jobs. I did immediately text Alex Lindsay. He said, no, that's not me. I'm in a different division. I said, oh, that's a relief. Um, the cuts, according to Bloomberg, Mark Gurman of Bloomberg include largely shutting down the vision pro gaming team.

[01:29:35] Well, that's, I mean, is there a vision pro gaming? Is there a game? There are no games and reducing the size of the team that makes vision pro immersive content. That's a little more concerning because honestly, maybe it's just me. What do you have without that? Unless they have, you know, done deals with creators. Well, they have that camera now, the black magic camera. So yeah. Yeah. But I mean, and maybe, you know, maybe that's part of what Alex is doing, you know, is working

[01:30:04] with outside. He is. He's developing relations. Yeah. Not just for immersive content. To get up, to get third parties, to create that content so that Apple doesn't have to do it. He may have more of a job. Yeah. Yeah. So I don't know if this, is this bad news for the vision pro? Is it, is it kind of, is the writing on the wall? Let me put it this way. Leo, do you remember maybe six, eight months ago, I told you about the vision pro that I found in St. Peter's. Someone had left it behind and I turned it into the lost and found. It's still there.

[01:30:35] Oh, geez. So it's, it was so, I mean, whoever lost it, they're like, eh, whatever. No. It's not a sticky device. Wait a minute. That's so bizarre. So the whole thing, the $3,500 face computer. In its case. In its case. Oh, so they were carrying it around. They were carrying it around. They were, for some reason, I guess they were using it in St. Peter's. They left it on a bench and this wandered away. I turned it into lost and found. It's still there. Maybe they didn't know where they left it. That's possible. But I mean.

[01:31:04] There's the assumption. If you left something in Rome that you're never going to see it again. You can find it, right? There's a finding app for the last time you used it. Find my, find my vision pro. That's a good point. They should know exactly where it is. They should know exactly. Except that battery is only good for two hours. Yeah, but it will remember where the last place it saw it. And I presume there's a lot of iPhone users wandering through St. Peter's. So it knows where the last place it was. So the first thing I'd do is I'd call. I didn't know there was a Vatican lost and found.

[01:31:33] What else is in there? Do you go and check every once in a while? Oh, absolutely. It's in the. So the Swiss guard have an office near the West gate. I'm just checking to see if the vision pro is still there. No, because you can go. Because they sell things with like Swiss guard hats and sweaters and key chains and stuff. And so as long as you're wandering through, you go, hey, does that vision pro that I turned on, is that still there? Oh yeah. That's sitting on the shelf. That's hysterical. Oh, did you actually find it and turn it in? I did. Yeah, it was me.

[01:32:03] You're a very honest man, as I would expect. I mean, I. Let me ask you though. Do they sell this in the Swiss guard merchandise store? They should. I'm holding up. So my friend Jeff Atwood, who does a club show with us called Off by One, because he is definitely off, made these for me. Oh, I love it. It's a bust with my head on it. And you're in a toga.

[01:32:31] And he decided if he was going to do that, he got my favorite t-shirts, which are the comfort colors. I really like them. Although they're somewhat ruined by the bust of me on the front. He said, I just ended up getting, I didn't know which colors, so I got you all of them. So I had like a hundred of these. A hundred of these. I don't know what I'm going to do with them all. Give them away on Club Twit. Give them away.

[01:33:00] I think I will have to have a giveaway, but that means I'll have to put them in envelopes and mail them. I guess I can do that. I know we should. We should have club. Club members. Would you? Do you want a bust of Leo on your T-shirt? Wait until the next time you have stamps.com as a sponsor, and then you can have them set it up for you. Every color of the rainbow. Everybody that comes to the house to visit, you know, comes over. That's basically it. You get, everybody who visits our house gets a copy of Salt Hank's cookbook and a T-shirt of me.

[01:33:30] It's like when people come here, you know, anybody that leaves the house has to leave with a water bottle that I've gotten from somewhere. You know, from some media program, you know, like, you know, I've got here a Mach-E one. Oh, that's nice. A Mustang. A Mustang. A Mustang. Yeah. That's pretty sweet. And my wife has warned me not to bring home any more water bottles. I'm that way with T-shirts. Anyone who visits me in Rome will go home with an indulgence. So, I mean.

[01:33:59] Is the gates still open? No, no. The gates are long closed. Long closed. Sorry. Did you know about that, Sam? The Jubilee Gate? It was open. You can walk through it. All your sins are washed away. 25 years. Every 25 years. Yeah. Oh, which, by the way, Leo, this is the very first time in the history of the Catholic Church that the doors were opened by one Pope and closed by another. Oh. It's never, ever happened before, which is kind of weird. Yeah, that is odd. Huh.

[01:34:30] Are we liking Pope Leo? I'm liking him. Yes. Yes. And actually, the funny thing about the AI encyclical is that I have been getting so many contacts from journalists that I've met over the years. Everybody thinks you wrote it. That's why. Which is weird. No, no. But, you know, I can comment on it. But the question I get asked more often than any other is, why does Leo hate AI so much? I'm like, did you?

[01:35:00] You didn't actually read the document. No, he didn't. That's why I liked it. He specifically celebrated it. He said, yes, there's cause for concern and there's ways that it can be abused. Of course. And I think that there's also a lot of use cases that are very bad. Yeah, we were talking about that. Don't call restaurants with your AI.

[01:35:28] Which, by the way, Leo, just FYI, I did put your name because they had requested people for future forums on AI. Oh, I'd love to go. Oh, I'll be there. If you hear from someone, that's where it came from. I will go there with bells on and I will wear my T-shirt. I think that's, it's not a buonorati, but it's pretty close.

[01:35:55] Actually, if you wanted Leo to take that, what you would do is you would add the lettering for Villanova. He goes crazy. Anyone wearing a Villanova hat? Yeah. Did he graduate from there? He graduated from Villanova. I'm holding this up so it can be the thumbnail for this episode. The new merch? The new merch. No, you know what? That's a good idea. I'll ask Jeff what he wants to do. He's the one who bought them and sent them to me. But I bet you he'll want to give it away.

[01:36:19] Maybe we'll do a charity thing, take donations for charity and we'll send them out to club members, something like that. I think that's a good idea. Or set up a booth at the Sonoma County Fair, you know, twit booth, you know, hand out T-shirts. Get your busts here, you busts a Leo. Get your busts. I might have a few bobbleheads too in the back room. And homemade gem. Yes.

[01:36:43] Speaking of Find My, very interesting kind of detective work by 404 Media. They had been approached by rare booksellers who said, we don't know, but somebody's buying a bunch of rare books and we think it might be going to AIs for training. Can you help us? And 404 said, yeah, we're going to place an AirTag in the spine of it.

[01:37:10] And the AirTag revealed that those rare books, at least in that case, were going to an Amazon AI training facility in your neighborhood in Las Vegas. I know exactly where it is. And the team there, by the way, the logo for the team is of an alligator devouring books. The team there. Not a good sign. Yeah. Tears the books, the paper from their spine, scans it, and then destroys the books.

[01:37:40] Is that an alligator? No, that's a dinosaur. It's something that's chomping. Chomping the books. So they definitely take pride in their work. Now, I'm probably on the wrong side of the law here, but I said, so what? Big deal. These are old books. Nobody wants them. They're going to end up in a dump somewhere. At least they're going to end up in AI, helping us all get better AI, right?

[01:38:08] I mean, they're going for books that are out of copyright protection. That's right. And they don't have AI content. So they're training on real human content. And so Jeff Jarvis said, what are you crazy? These are rare books. They're, they're, they're scholars want these books. Well, scholars should go get the books. The reason they have to destroy them, by the way, I pointed out is because the court says, if you don't, it's, it's, it's not fair use.

[01:38:36] The publishers don't want those books to go back on the market. So the publishers are happy if the books are destroyed. They do have to take them apart to scan them efficiently. You know, you can't, they have to cut off. These are old, you know, out of print books. The publishers aren't going to get any benefit from the resale of these books anyway. But that's the point. They don't want, they want you to buy a new one. Well, for it to be fair use, it has to be transformative. And transforming it from text to electronic is transformative.

[01:39:06] So that's within the bounds of the law. This was the Anthropic case where the judge said, it's fine that Anthropic bought these books, destroyed them. That's fair use. But when Anthropic used a pirate database of digitized books, that was illegal. And then Anthropic had to pay, I think, one and a half billion dollars to authors for doing that. So the judge basically are doing what the courts told them to do. Yeah. I mean, I personally, I think rare books have to be treasured.

[01:39:32] I think that they're, you know, they're going to, they're going to last longer than any electronic copy of anything. But legally, they're not doing anything wrong. They're, they're buying it at fair market value. It's not under copyright protection. So they can ingest it into their AI training systems. And they're destroying it to be in compliance with the law. So if you don't like what they're doing, either buy the copy first or change the law. Yeah. In July, this is from 404.

[01:40:02] One bookseller told 404, they received a very large order of a thousand books. This is on Biblio, a rare book marketplace. By the way, the use of the word rare implies that they're actually valuable. They may just, I mean, my books are probably rare too, but they're not valuable. I've got a lot of rare books that are only rare because no one read them. Yeah. Then, I mean, and then I've got rare books that, no, no, they're generally treasures. And they are kept in an archive and we would never sell them to Amazon or Anthropic or anything.

[01:40:31] No, the Vatican Library is world famous for doing that, right? Exactly. Hi, Leo. This is Benito. So there's also books out there that no one knows that they're valuable until 100 years later. And if those get destroyed now, we won't know in 100 years if that book is going to be useful to us or not. Which is why there are copies of three of Brian Brushwood's books in the Vatican Library. No! Someone may have snuck into the stacks. There's a copy of Scam School in the Vatican Library?

[01:41:00] There is. There is. I did that eight years ago. That's just funny. Wow. Some poor monk in 150 years is going to come across that. What is this? What? He'd be very puzzled. That's really funny. Did Brian know that? I'm not sure. I just did it for giggles. I'm like, yeah, you know, I've got a book you could add. Why not? That is a hoot. Along with a copy of the Anarchist Cookbook, by the way.

[01:41:29] Well, that belongs there. I think that's a historical document. That is historical. Yeah, absolutely. Amazon is... You might have something to say about this one, Sam. Amazon is aiming for delivery drones to reach 500 U.S. neighborhoods by the end of the year. Can I specify that it gets dropped in my pool? Did you see the video? I was just going to bring up that video. There's a very funny video of a woman who was so excited because her Amazon delivery was coming via drone.

[01:41:59] So she whoops at her phone. This is great. And it drops it in the hot tub. Oh, dear. Amazon is going to launch Prime Air. That's what they're calling it. Delivery drone service in Chicago, Cleveland, Atlanta, Syracuse, Boise, Idaho. More to come at the end of the year. If you... I guess every Prime Air site. So you have to go to the...

[01:42:28] I guess maybe it's a checkbox. I don't know. Well, it offers delivery drone service within an area of 175 square miles. Wow. But drones are going to fly a long way. Delivery times range. And this is the thing that kind of bugs me. And this is an example of, just like Instagram, consumer demand, companies giving consumers what they want. But is it a good idea? Delivery times range from half an hour to an hour.

[01:42:54] I got to have that Xbox in half an hour. Yeah, you really don't. Yeah. I mean, most of the stuff they're going to be delivering, people don't need in half an hour. I mean, there are absolutely legitimate use cases for this, like delivering medicines. You know, that... That makes sense. Yeah. I mean, I saw the movie Balto, right?

[01:43:17] Where they strap the medicine to the husky who goes through the wilds of Alaska and delivers the medicine and the child is saved. Yeah. You can do that with a drone. You know, somebody needs some medication, needs some prescription medication. Right. And, you know, for whatever reason, you know, they're not able to get into a car and drive to a pharmacy. Or, you know, maybe they live in a rural area. You know, it's 175 mile radius. They live in a rural area.

[01:43:45] And it's not practical for them to get to a pharmacy, especially if they're sick and they live alone. Makes sense. Yeah. Yeah. I mean, there are absolutely legitimate use cases for this. But I think probably 99.9% of the crap that's going to be delivered by these drones... My cigs. ...could have waited until the next day. I need my cigs. I need my front looks. Okay. I'm hungry. So these are going to be operating in the same space that Flock is operating its fleet of drones.

[01:44:12] So you've got two companies that want to flood the air with drones just above the non-regulated space. At some point, there's going to be a collision. Albert in our YouTube chat says, how long before they start shooting them out of the sky? People just... Yeah. Well, you know, and I don't know about these particular drones that Amazon's using, but up till now, they have... Somebody just popped... Darren Oakey just popped up another picture in the chat.

[01:44:42] Darren is trouble. Let me pull that up. This is in our Discord chat. You know, if these things are as noisy as most drones... Oh, you guys look good in those t-shirts. That's good. I don't want these things buzzing around over my house all day long. You realize you guys are wearing each other, right? You realize you're wearing each other. Yeah, yeah, yeah. Robert's wearing me. I'm wearing Sam. Sam's wearing Robert. Yeah. That's a little disturbing. I like it, but that's disturbing.

[01:45:13] AI. Is there anything it can't do? By the way, I didn't know this, but Walmart's been using drone delivery. Walmart has celebrated a milestone. They have more than 1 million drone deliveries, 66 stores in four states. They want to expand that to 270 stores by next year. The sky's got to be full of delivery drones.

[01:45:37] Walmart has a more difficult path to it, though, because before every delivery, they have another drone that carries an old greeter that says hello, and then it drops off the box. Yeah. Stand by. Keep the pool cover open. We got a delivery for you. Also, the side effect. Yeah. I mean, you're... Sorry. The side effect of all the buzzing. You're going to hear drones all day, every day. Oh. Yeah. That is... Just like if you're living next to a data center. Yeah, exactly.

[01:46:05] It is a noisy world we live in. To what you said, Robert, these things are flying in what is currently unregulated space, below 400 feet. And, you know, unless these things, you know, have, you know, sensors on them, or they're using something like ADS-B, you know, to provide, you know, they're broadcasting their location so that they can avoid each other. You're probably going to start seeing collisions of these things. Yeah.

[01:46:30] Or, like in my neighborhood, you know, I live in an older neighborhood that is full of trees, and, you know, I can definitely see these things getting caught up in, you know, these old mature trees. Here you go. Ready? Ready? Here comes the drone. Here comes the drone. Here comes the pool. Here comes the package. Wait a minute. No, don't do it. No. I mean, you heard the sound of that thing. You know, this is loud.

[01:47:01] Yeah. I know. I've flown drones. I know. They're really loud. In fact, they got to work on quieter drones because you can't sneak up on people's, you know, like you used to. You need to drive. They can learn from owls. Owls have the most stealthy flight characteristics in the avian world. They do. But then again, they don't hover. So, and they, or, you know, vultures. Yeah.

[01:47:26] We were out at the lake this afternoon, you know, just, you know, watching a couple of vultures just soar around, you know, because they've got these huge wings relative to the size of their body. And they just, most of the time they never, except when they're taking off, they never flap their wings. They're just soaring around. And, you know, it's amazing. Lazily looking for carcasses. Yeah. Or, or they could wait the extra five minutes and have an actual delivery person deliver it. I'm just, I'm just going to spit the all that. That's true.

[01:47:50] Diesel in our club says that Walmart finally is going to start taking Apple pay in a week. Yeah, that's true. Tomorrow. I saw that. What? They gave in. Did you know, I did not believe this until I actually encountered it in the wild, Leo. There are people who actually believed that Apple pay wasn't a credit card, that it was Apple paying for stuff. I thought it was a joke. I really did. Well, Uncle Apple going to pay for this one.

[01:48:20] People who are complaining. I have to pay everything I put on the card, but it's Apple pay. They really didn't. What? When they got the bill, they went, what? I thought it was a joke. And then it made me very sad. Well, hey, Leo, Lowe's and Home Depot just started taking tap to pay like about a year ago. I live on tap to pay, man. I don't, I love that everywhere I go. And yeah, I never get a bill. It's amazing. Big wad of cash. I just, you know, what is cash?

[01:48:50] You know who I carry the cash for? I carry the watch for the stores. I carry the cash for the people. When I tip, I always tip in cash. That's true. That's true. Cash is king. Is what Lisa says. But so far, I haven't gotten a single bill for my Apple watch for years. It's, it's gone straight to Lisa. Although this, this time back, Leo, this was the first time this ever has ever happened to me. I, I tipped 15% because I always tip 15%. It's not fine dining. It was like a far restaurant. And the waiter said, oh, you know, normally the tips 25%.

[01:49:20] And I was like, wait, what? Oh, that's not nice. I did not realize that this was that out of control in the United States. Who goes to a priest that says you didn't give me a big enough tip? What? I didn't, I didn't. We don't do that in Europe. No, they don't tip at all. In fact, Darren, our Australian says tipping is stupid. It is. Tipping is stupid. Right? That's, that's, that's a uniquely North American thing. I like to tip. It makes me feel good. No, I like to tip, but I mean. I would rather the people just get paid sufficiently.

[01:49:49] Well, that would be nice. That would be nice. All right. We're going to take a little break so I can get paid. Ladies and gentlemen, you're watching this week in Tech Father Robert. So you have some money to give to Lisa to pay off your Apple pay? My Apple pay. She hasn't seen the spark bill yet. I don't know. No, no, Leo. I'll know when she gets that one. Yeah. Yeah. She thought it was just a couple of Mac minis. Yeah. Hey, I feel okay. She just bought a car. She just came home.

[01:50:19] I saw it in the camera. Nice bright red. What color? Bright red, of course. She only buys red. But didn't she get like a CRV or something? She got a Honda Prelude. The new Prelude. The blue that they have on the Prelude is also a fantastic color. She says it looks just like a Porsche. When I, if I get a new BMW, I'll get a blue one. Because that's, that's the BMW colors. Get, get the Portamal blue. Portamal. Portamal blue is an excellent color. Algarve in Portugal, Portamal. All right.

[01:50:48] I'll, I'll just charge it to you, Sam. Charge it to Apple pay. Apple pay. You could, you could buy anything. Our show today brought to you by Doppel. Oh, this is good. This you need. AI. One of the side effects of AI is social engineering. We're just talking about deep fakes, right? People are using AI to make social engineering attacks completely convincing.

[01:51:13] From phishing emails to fake websites to impersonation attempts. Actually, we, we had some fun with this. They made Anthony using a tiny sample. This is the amazing thing of my voice. Like 20 seconds made this. Hey Burke, this is definitely not Leo asking you to buy gift cards. That is amazing, right? Sounds just like me. And imagine if your employee got your voice on the phone saying, hey, quick, send a check.

[01:51:42] So you got the phishing emails that look real. They've got the websites that look real. And now you've got voice fake, deep fakes. It's becoming harder and harder to tell what's real from what's designed to deceive. And that's why organizations need more than a collection of point solutions. They need a unified approach to stopping attacks before they reach their people. Doppel is the way.

[01:52:09] Doppel is an AI native social engineering defense platform. Doppel strengthens human risk management by training employees to recognize deception, provides digital risk protection across every channel, and delivers agentic email security that doesn't just score the inbox, but takes down the attacker infrastructure behind the message. Wow. Doppel protects against the entire social engineering attack chain with one comprehensive platform.

[01:52:36] Digital risk protection detects threats across multiple channels, links alerts into a real-time threat graph. These insights also power phishing simulations, security awareness training, helping strengthen employee defenses through the next generation training and testing. Email security. It inspects every message, traces it back. This is mind-boggling. Traces it back to the attacker infrastructure behind it and helps take that infrastructure down.

[01:53:04] So the campaign can't target your organization again. Doppel offers best-in-class integrations and partnerships, making it easy to work alongside your existing security stacks. Join hundreds of companies already using Doppel to protect their brand and people from social engineering attacks. Doppel. Outpacing what's next in social engineering. Learn more at doppel.com. That's D-O-P-P-E-L dot com. Let me thank them so much for their support.

[01:53:33] So NVIDIA has an interesting problem. They make, of course, the graphics processors that everybody is dying to get but can't. All the data centers use them. Here's the problem. They may actually have too many soon because none of these data centers are being built fast enough to buy up all the chips.

[01:53:54] They're worried there could be a glut, which is why NVIDIA is now committing half a trillion dollars to building data centers. They're trying to create the demand. This is an Ohio data center, which NVIDIA is going to back with $105 billion. It's an interesting issue. There's this kind of overhang that NVIDIA is facing.

[01:54:20] Now, I remember, Sam, we've been talking about NVIDIA for a long time. Of course, famous at first for graphics cards for video games. Then Bitcoin became hot. Very famous for Bitcoin graphics processors. They sold a lot of those. You and I talked a lot about NVIDIA in automotive for AI. And then finally, of course, the AI boom has really been great for NVIDIA.

[01:54:46] These guys, I got that sense talking to you maybe five or six years ago, really are firing on all cylinders. Just amazing. Yeah. I mean, they've been able to – it's like a surfer catching every wave. They've caught all the waves. They've caught all the waves so far.

[01:55:05] But this one may be a little bit different, ultimately, just because the size of this wave is so huge. And right now, they are making so much money. But the question is, do we actually need all this data center capacity? And the thing is, NVIDIA's chips are amazing. They are incredibly powerful.

[01:55:35] But, you know, where they really shine is in doing the training. Once you get to the inference, you know, once you've got a trained model and you're doing – you're just using it for inference, you don't necessarily need that same kind of power. You can potentially transition over to other types of chips that are much more power efficient.

[01:56:02] Because, you know, these chips consume a huge amount of electricity. You know, and that's one of the big complaints that people have about all the data centers being built is, you know, they're requiring so much electricity that it is forcing electric rates to go up in a lot of places where these data centers are being built.

[01:56:24] And it's also – you know, if they're not able to get the local utilities to build up the capacity fast enough, you know, you've got like XAI, for example, you know, installing 35 – you know, it's about 35 or so gas turbines to power as data centers in Memphis, which are, you know, adjacent to a predominantly low-income neighborhood. These things are noisy.

[01:56:53] They are putting out a lot of pollution. And, you know, the data centers are loud enough. And then you've got these gas turbines running 24 hours a day to power these things. You know, and if we can do something to, you know, reduce the power consumption – and I think this is something that a lot of people are looking at. How can we reduce the power consumption?

[01:57:15] The Blackwell chips are definitely more power efficient than – sorry, the Rubin chips are more – one of the things that we've seen in the last three years, you know, since this AI data center built up, is the fact that they are talking about these things.

[01:57:44] They're labeling them based on the amount of power they're consuming. Not necessarily what they're actually capable of doing, but just how much power they're consuming. Talking about a gigawatt data center. It's consuming a gigawatt of power. That's enough to supply hundreds of thousands of homes every year for one data center. And, you know, I think if we think about that, you know, how much electricity we are using for these things, it's just mind-boggling.

[01:58:13] And I think that's really problematic given all the other issues we have with, you know, with the environment and energy use. China's doing the same thing, but they're doing it with sustainable. Energy. That's one way of solving it. They're putting in a lot of solar. Yeah. But this is the wrong time to back off on solar. Not allowed to have nice things in America.

[01:58:37] I mean, right now there is a four to eight-year backlog on large turbines for gas turbine generation. They're very extremely difficult to put together, very expensive, and very difficult to ship. So, even if all these data centers wanted to generate their own power, it's at least half a decade before they'll have the ability to. And those gas, they use a ridiculous amount of water. You're looking at two to three years for transformers. Yep.

[01:59:07] Yep. Do you think, I mean, Jensen's looked like a genius up to now. Do you think he's stumbled? Should, is he making a mistake? I think the mistake is that right now he is a vendor selling shovels at the gold rush. And that's a very safe place to be because you're taking on virtually no risk and you're making profit hand over fist. He is now getting into the business of running data centers.

[01:59:36] That's not speculative anymore. That's no longer selling just the tools. He's now selling the raw compute. I don't think that's a good business. And when you're giving the prospectors the money to buy your shovels, that's where things start to get problematic. Also, and the Wall Street Journal calls attention to this, there's a lot of off-the-books investment. A ridiculous amount. Yeah.

[02:00:00] So, this is the pyramid of debt that is on balance sheet, but below it is the huge amount of money, almost a trillion dollars. $1.65 trillion off the books. Minimum. Minimum. Yeah. Because- So, tell me why that's important. What does that mean? As you were saying earlier, Robert, these are all separate entities that they're creating to build these data centers. Special purpose vehicles. Yeah.

[02:00:25] And so, you know, the companies, companies like OpenAI, Anthropic, Meta, and others, they are signing agreements to use the capacity of these data centers. But until the data centers actually exist, you know, they don't have to count that commitment as debt or as financial commitments on their balance sheet. Well, it's not necessarily even a commitment, as Elizabeth Lopato points out on The Verge.

[02:00:54] Didn't OpenAI promise, or NVIDIA promise $100 billion to OpenAI? Which then they said, nah, never mind. So, it's book-to-market economics. And the problem is the last time we saw this done at this scale, it was Enron. Enron was the company that got to it. And it wasn't even close to the scale. Yeah. And I mean, Enron was a novice compared to what we're seeing right now. So, should people be nervous? You've got all of this. Exceptionally nervous.

[02:01:20] All it will take is for one of the big five to go down because all of their circular deals with each other are dependent on the cash flow between each of those entities. But these aren't stupid people. Why are they doing this? Because they have too much money. It makes money now. Yeah. They have too much money and they need to put it somewhere. Yeah. I mean, the VCs have so much money that they're putting into these AI companies. And, you know, companies like NVIDIA are- But is it real money or is it funny money?

[02:01:47] I mean, if it's real money, that's okay, right? Well, it's real money that they're putting in with the expectation of exploding that amount of money in the future years. Right. You know, but in order, you know, they're putting real money into companies like Anthropic and OpenAI who are then committing to building these data centers with even more money. Money which has to be borrowed by these special purpose entities that Robert referred to.

[02:02:15] And that's that off-the-books debt, which doesn't show up on the balance sheet of these companies until those data centers are built. You know, but in the process, somebody is already borrowing all this money. You know, companies like Larry Ellison, you know, Oracle has all- You know, I think they were the biggest or second biggest, I think, on that list of the off-the-books debt. So it's like a pyramid scheme kind of. Mm-hmm. Mm-hmm.

[02:02:44] Here's the off-balance sheet obligations. Look at Alphabet. $900 billion. Yeah. They are the largest. Then Meta, Microsoft, Oracle. Oracle's got a mere $300 billion. Oh, okay. Oh, well, they're good to go. Yeah, they're fine. Amazon and NVIDIA is actually- Plus the other $110 billion that Larry's got to spend on Warner Brothers. Right. So why are they- The reason they're putting this money in, the reason this money is available,

[02:03:13] the reason people are loaning this money and investing this money is because there's a feeling that there is a lot of money at the other side. There's a huge upside to this based on what? AI? No, based on the current market. Just look at what happened with SpaceX. You have a company that we know doesn't make money, that's going to lose money hand over fist, and yet people bought into it. That's all they need. They're going to get paid. These VCs are going to get paid.

[02:03:40] All they have to do is maintain the path because nothing in the recent market history has showed them otherwise. Who's going to get hurt by this? They only have to wait until they can cash out after the IPO. Is it retail investors? Is it institutional investors? Yeah, absolutely. Retail investors are going to be caught holding the bag for everything. All the investment funds. My 401k, my retirement that I'm going to count on in my golden years, which are going to be in any minute now. You might want to consider converting that to cash. Yeah. You know what I did? Or at least controlling what's in the index.

[02:04:09] You guys all scared me. So six months ago I did, and then I went, no, and I bought back in, and I'm glad I did because it's gone up, up, up. Exactly. So people see it go up, up, up, even though there's nothing underneath that. Why should SpaceX go up when they're losing billions of dollars a year? It's on the promise that maybe in the future they might do something that makes profit. Actually, what I end up buying is index funds and market baskets and mutual funds because I don't want to invest in any particular tech company.

[02:04:38] But all of those funds have bought into the Magnificent Seven, have bought into the, yeah. Yeah, I get it. Yeah. Yeah, but what do I do as some poor schmuck who's just trying to keep my retirement alive? You're right, because if you are not buying into the wave, then it's almost like you're losing. You feel like you're losing out. Yeah. A safe investment when someone else has 2X, 3X. With 4% inflation, my money's shrinking by 4% every year.

[02:05:06] I guess I could, I don't know, what's, I don't know, I could buy gold. I don't know if that's a good investment either. I mean, I can get upset at what I see in the market, but the problem is nobody knows how it runs anymore. It's not running on PE. It's not running on basic fundamentals. So, hype is just as good a metric as any other, I guess. Okay. We're not financial experts and do not take any advice. Yeah. Definitely don't take financial advice from us. Really, really seriously, don't do that.

[02:05:36] Don't do that. Yeah. Mike B says, I need a better money manager. Better than me? Beans. Magic beans, Leo. It's all about magic beans. That's the hard part of all of this is, you know, you can't time the market. And so, you know, then in your heart, you probably should get out. But when? Yeah. Yeah.

[02:06:02] You're lucky, Robert, because you have a big institution behind you that probably won't go belly up. And we're extremely conservative about investments. I bet you are. So, but I mean, like Louis XIV melted down all his gold furniture so that he could pay for his chores. I don't have any gold furniture. I only invest in tulips, Leo. Tulips. Tulips. The bulbs. They're going to be big. All right. Time for a break.

[02:06:28] Our show today, you're listening to Robert Ballasier, the digital Jesuit, visiting us from Las Vegas, where he was just at DEF CON. Great to see you, Robert. How's your family? Everybody good? Everyone is good. Everyone's healthy. And I just took him on a road trip to the San Francisco. So that was fun. Aren't you a good son? And Sam Abul Samad, whose brilliant wife, got me lifting kettlebells and swinging clubs.

[02:06:57] And thank her for that. I will. I thank her every day. Yes. She's wonderful. He is, of course, the host of the Wheelbearings podcast. We're glad you're both here. We're glad you're here, especially our club members who make all of this possible. Thank you. Our show this week brought to you by, oh, another new sponsor. Built. B-I-L-T. Here's a question. Are you getting rewarded for paying your rent or your mortgage right now? We were just talking about finance, right?

[02:07:27] Are you? You're not getting rewarded for paying your rent or your mortgage. If the answer is no, well, you should check out Built. Becoming a Built member means earning points on every housing payment, no matter where you live. Built points can be redeemed for flights, hotels, lift rides, Amazon.com. Purchases, future housing payments, and more. An unexpected benefit, Built gives every member access to its agentic neighborhood concierge.

[02:07:56] It's right in the Built app. It doesn't just give you information about your home and neighborhood. It executes. It's connected to your home and over 50,000 merchant partners. So you can actually tell Built, pay my rent, make a dinner reservation, book a fitness class, schedule a lift, and more. It makes things happen while automatically applying your member benefits along the way.

[02:08:19] And for a crowd that thinks a lot about what agentic AI actually looks like in practice, this is an example worth paying attention to. First, download the Built app, B-I-L-T. You'll find it right there on the App Store. Second, link to any credit or debit card. Any credit or debit card, you link it in the app so you don't even need to get a new card. Your purchases at your local spots become even more rewarding. Third, now this is the fun part. This is where the fun begins.

[02:08:48] Ask the neighborhood concierge in the app pretty much anything. It'll tell you about exclusive member benefits at restaurants, fitness studios, on your lift rides, and more around your home. And if you're a homeowner, Built works not just for rent but for mortgages too. Your housing payment. Isn't that one of your biggest monthly expenses? Why wouldn't you want to earn points? You're paying it anyway. Download the Built app and join the membership for where you live. At joinbuilt.com slash twit.

[02:09:18] That's J-O-I-N-B-I-L-T dot com slash twit. Use our link so they know we sent you. Joinbuilt.com slash twit. J-O-I-N-B-I-L-T dot com slash twit. Now I get the name Built. I get it. We are going crazy, by the way, in the AI world. In many, many ways, obviously.

[02:09:44] But certainly the Chinese models are gaining a lot of attention. It seems like every week a new one comes out. There's one right now everybody's speculating about. It's free. It was free for a few more days. It was free this week on Open Router. And it was free on Noose, which I use for my subscriptions. No one knows where it's from. It's called Ox OX Alpha. And it seems pretty good.

[02:10:14] And the consensus is it is probably the latest from a Chinese AI company, probably Z.ai, which does GLM. The GLM models are really good. And this one's free for a week. A lot of people are using it. And GLM 5.3, which came out, I think, a few weeks ago, is doing really well. Bloomberg says GPO, Z.ai, is iterating fast.

[02:10:41] And this is concerning to the government. The White House apparently drafting plans to require countries to pick a side in the U.S.-China AI competition. This is according to Reuters. They're saying to companies and countries, you should not be using Chinese models. Yeah, because we've got a lot of really good sway and goodwill that we spread around the world. So I'm sure that's all our lead. Yeah, I mean, everybody trusts us right now.

[02:11:11] You know, I mean, it's really a new way to exercise soft power. I mean, China's done this with their Belt and Road Initiative, where they help people build dams and roads in third world countries. Gaining friendship. It's soft power. Now this is it. And I can't say it's a bad thing. And we currently are in an environment where our government does not believe in soft power. They only believe in hard power. They believe in transactional relationships. And that's not going to invite anyone to actually help us with this.

[02:11:39] And a fascinating comment from Canadian Prime Minister Mark Carney the other day after the trade talks broke down. And he said, yeah, you know, the U.S. is now signing agreements in pencil. I kind of like Mark Carney. He's really outspoken. And I kind of enjoy his point of view, even though it is at our expense. I'm a French Canadian. I can say that one. Let's talk flock.

[02:12:12] There's so much to say. Three flock cameras within a block of my house right now. Yeah. There is in our little town, there are flock cameras, two of which are pointed at the public swimming pool. That sounds like a good idea. I don't think there's any license plates going by. So flock is, we normally, I think of it as an ALPR, automated license plate reader, which look, you know, you drive on a more than license plates. That's how they sold it. That's how they sold it at the beginning. They said we're only tracking license plates.

[02:12:41] And to be fair, you don't have a right to use the public streets to drive on them. You have to get a license. You have to take tests. You have to show your competence. Unless you're an automated driving system. Right. Then all bets are off. Yeah. But despite what sovereign citizens might say, you know, it's a privilege, not a right. And I guess it's reasonable to say law enforcement should have automated license plate readers.

[02:13:09] It's helpful in tracking lost vehicles, tracking stolen, you know, kidnapped people. There's all sorts of law enforcement uses for it. The problem is it seems to be law enforcement, well, we've seen a number of cases, uses it for other reasons, including stalking girlfriends. Polices are being told by a usage policy for the flight flock license plate reader not to talk about the cameras.

[02:13:40] This is from Bruce Schneier. Do not mention ALPR usage to the occupants of the vehicle. Do not mention ALPR usage in your report, in your police report or your complaint unless absolutely necessary. Why? Why are they trying to hide this? What are ALPR readers good for? What are they not good for? And what are they doing beyond reading license plates? Because they're violating the Constitution of the United States.

[02:14:07] And they are actually collecting a lot more than just license plate data. Fourth Amendment. Fourth Amendment, yeah. So here's the problem. In order for law enforcement to be able to do an investigation on you, they first have to have probable cause, and then they have to obtain a warrant in order to look into your affairs and see if there's actually any crimes being committed. Law enforcement can now bypass that by buying a subscription from flock, because it's not a government agency that's collecting information on the individual.

[02:14:34] It's a private company, and the government agency is just buying that data from a private company. It completely bypasses the Fourth Amendment. And law enforcement knows this. That's why they're trying not to mention flock when they're writing out these reports, because they don't want that question of whether or not that evidence should be permissible. Well, when you put— Okay, one of the things we know is good for law enforcement is to put officers on foot walking through neighborhoods,

[02:15:00] looking around, kind of like a flock camera would, making observations. They don't arrest somebody to this probable cause, but they're always looking. How is this different from that? Because this goes forward and back in time. This is data that's always stored. This is why flock tried to make a big hoo-ha, but, oh, no, no, we're only collecting license plates. Until now, it's been revealed. No, no, they collect everything. They collect images of bumper stickers on your car. Identities. Identities. They're collecting identities.

[02:15:30] They're collecting clothing. They're collecting any distinguishing factors that can be used in biometric location later on. They've just released a new product to law enforcement. It's basically an AI prompt. You can say, find me everyone who interacted with this red car between this date and this date. And because they've got cameras everywhere, they can not just find the people, but they can find the addresses that those people ended up at.

[02:15:59] That's too much. I've also seen some communities installing Bluetooth sniffers on their flock cameras, so they capture every— You know, when you're driving down the street with your phone, it's advertising its Bluetooth presence, including your phone's name. Yeah, they're starting to consolidate all kinds of information from different sources.

[02:16:19] And they had a deal earlier—it was either earlier this year or late last year—with Ring to start consolidating, you know, trying to fuse information from Ring cameras with the flock database. So, you know, as Robert said, you know, this is collecting way too much information. You know, so it's way more than just the license plates. You know, there's bumper stickers.

[02:16:42] You know, if there's a dent on the car, you know, they're putting together all kinds of little bits that go way beyond what they have publicly said. And also, sometimes the information is just plain wrong. You know, I put in a story down below. I don't think you guys talked about it previously. This was actually about a month ago.

[02:17:06] Two friends of mine who are both auto riders got picked up by police because of incorrect flock data. They both happened to be driving press vehicles from Jaguar Land Rover. And the way it works with the press fleets, usually the manufacturers that own these vehicles, they usually put on manufacturer plates from whatever state they happen to be based in. JLR is based in New Jersey. JLR North America is based in New Jersey.

[02:17:36] So they have New Jersey manufacturer plates. And the manufacturer plates usually are a little bit different from the others. In New Jersey—in Michigan, they look like regular plates except there's the middle character is an M. In New Jersey— For a manufacturer. Yeah. In New Jersey, they have—the manufacturer plates are 3-4 and DTM. And then there's a two-digit number that's much smaller in the middle, in between the 34 and the DTM.

[02:18:03] And somebody had incorrectly entered information in California into the flock system or I think actually into a system for reporting stolen or missing license plates. And they missed the middle two digits, and they just had the 34 DTM.

[02:18:26] And so the flock cameras, the flock system, will capture partial information because a lot of times police, you know, when they put information, they may have a partial plate. So they put in the 34 DTM. And so as soon as the flock cameras noticed vehicles with the New Jersey manufacturer plates on them, it flagged it for the local police.

[02:18:50] And so my friend Joel Fetter, who lives near Minneapolis, he had been—it turns out he had been tracked for four days by flock cameras before the police finally— Why did they think they were—was it mistakenly reported as stolen? The license plate was reported as stolen. But by whom? The license plate was reported as stolen.

[02:19:12] When somebody managing the press fleet in California had incorrectly put in information into the license plate system. So the manufacturer— No, it wasn't stolen. It was reported as missing. Sometimes when they're doing marketing photo shoots, they'll take the license plates off and put on a different license plate. And so they had misplaced it. It wasn't stolen. They lost the license plate, not the car. Yeah.

[02:19:38] But, you know, so they flocked, misinterpreted the whole thing as a stolen vehicle. And so Joel and his wife were going up running some errands. I'm watching Joel's video on here, and the police were fairly understanding about this. Yeah, but this whole thing took over an hour before they got it sorted out. And that has a relatively happy ending. Yeah. Joel is white. Imagine if this had been another one of my friends whose skin tone is a different color. Right.

[02:20:08] It might not have turned out quite so well. He was surrounded by three officers. Four police cars. Four police cars. Yeah. We don't have to imagine it. Boxed in. It's already happened. Three days after this, another friend of mine, Tim Esterdahl, who lives in Nebraska, was driving another JLR press vehicle. Exactly the same situation. But those two small middle digits were different digits from the ones that were on Joel's car and the ones on the original missing license plate.

[02:20:38] And he got pulled over, and he was held for about an hour and a half while they sorted out the situation. I mean, this is ridiculous that this is happening. You know, with and, you know, the information should have been corrected as soon as they sorted it out with with Joel. But it wasn't, you know, and three days later, it happened again in a different state. There have been more than 30 cases, Leo. 30 cases of misidentified vehicles that led to arrests.

[02:21:08] In a couple of those cases, people sat in jail because even though the officers knew that the data didn't match, what their terminal was telling them that this was a suspect. So, again, there's a question. The big question for me here is accountability. Who is responsible when it gets it wrong? Because the police officers are saying it's flock and flock is saying it's the police officers, which means that someone who gets caught up in one of these mistakes, they have no recourse.

[02:21:35] Grand jury declined to indict a man who was caught on camera destroying a flock poll. There is a backlash. The people are starting to see that these flock cameras are everywhere. We're living in a surveillance society. And a grand jury in Ohio refused, which, by the way, very rare a grand jury will refuse an indictment. The prosecution. You can indict a hand sandwich. That's the saying.

[02:22:04] If a prosecution wanted to, they could indict a hand sandwich. In this case, the grand jury said no. He removed the bolts on the ALPR poll, toppled it over and destroyed the surveillance camera and its solar panel. And nobody's unhappy about that except the police. And a lot of communities have been canceling their flock contracts, although some have been following that up by signing contracts with Axon, the company that makes tasers. Because they have their own ALPR system.

[02:22:34] The problem is that canceling the flock contract just means they're canceling the service. That doesn't necessarily remove the camera. Right. And so there is a solution for that. Fortunately, because the flock cameras are being paid for with taxpayer dollars, we have a responsibility to make sure these things are properly taken care of. And they're all outside, so they're exposed to a lot of UV radiation. And you don't want them to get sunburned.

[02:22:56] So somebody has actually come up with something with the flock sock that you can get to protect your local flock cameras from UV radiation. Oh, yeah, because they're so dangerous. You know, the sun is so bad, as you know. You don't want them to get sunburned. Exactly. Yeah. It's a carbon fiber sock. You don't have to take the pole down. Just I'm sure this is illegal.

[02:23:22] No, you know, Leo, actually, on a completely unrelated note at DEF CON, we may have had access to some flock hardware. And we discovered that in the newer models, they're installing cut filters so that if red or green lasers are shined thrown through the lens, it will protect the sensor from being burned out. Unfortunately, it does not really stop blue lasers. So if you have a high intensity blue laser, you can burn out the sensor almost immediately. Yeah, I mean, definitely don't play around with blue lasers. But don't do that. Don't do that. Don't do that.

[02:23:50] I'm sure this is illegal and we are not advocating anything. We specifically said not to play with blue lasers near a flock camera. If you're wondering where the flock camera near you is, this is a map called DEFLOC, which records 64. The CEO of FLOC called this a terroristic organization. No, it's just public information. Yeah. Yeah. Don't you think you have the right to know what flock cameras are in your neighborhood?

[02:24:18] A flock camera has no reasonable expectation of privacy if it's on a public road. I'm just saying. Yeah. Fortunately, my community does not have flock cameras. And neither does Ann Arbor, which is neighboring us. There are flock. The other thing to keep in mind is not all contracts with flock are from communities. Some of them are on private property.

[02:24:41] So even though Ypsilanti Township and the city of Ypsilanti don't have any flock cameras, there are a couple of major retailers nearby me here that have a bunch of flock cameras in their parking lots. Wow. Of course, they might not need it now that they've got their flock drones. So what's that all about? Flock drones. That's their next big growth market.

[02:25:04] They've realized that they can do plate reading from a distance with drones that don't have to worry about things like permits or being allowed to by the city. They can just overfly an unregulated space and have pretty much immunity from any sort of backlash because you can't even see it. Wow. It's a lot harder to protect your flock drones with flock socks, though, if they're up in the air like that.

[02:25:28] But that's why I order stuff from Amazon constantly, just to create a line of drones that defends that area from flock drones. Defense orders. I'm ordering. These are defensive. Yeah. And by the way, not all cameras on poles are flock cameras. There are red light cameras. Right. And there's also Axon, as I mentioned. Oh, Axon makes them too. But red light cameras aren't necessarily, or are they, tracking every license plate that goes by?

[02:25:56] Usually they take a picture when you go. The resolution's not high enough to really do that. It's kind of chunky. Well, I mean, they're high enough to read. They're high enough. It's high enough to read the license plate. You got to know who to send the ticket to. A single, but that's a single view. It's just a single shot. When the light changes, they take a picture in case anybody goes through the intersection. Okay. Well, we'll keep an eye on this one.

[02:26:18] It is interesting, though, because I thought that we were at that point that we were so desensitized to cameras in public because it's in everything. You just always assume something's going to be recorded. That flock has managed to pretty much energize people from the left and the right, no matter what your political leaning might be, to say, this is not okay. No, we do not like this. Yeah. We don't mind it against our enemies, but here at home. Not here. Not here. This is here at home.

[02:26:47] This is Palantir for domestic use. The two things that seem to be uniting people on the left and the right are flock cameras and data centers. Data centers. Yeah. Everybody hates them both. Yeah. Actually, you could run a campaign today. You could just say, all I care about is no more data centers and no more flock cameras. You'd win. You'd probably win. You'd win. Well, it's funny.

[02:27:12] Here in Michigan, the Republican candidate for the U.S. Senate seat is Mike Rogers. And a couple of days ago, he came out backing a data center moratorium for a year. Because it's good for the voters. Right. However, it also turns out that he personally has invested about $2.5 million in. Data centers. And like Blackstone real estate trusts that are in building data centers.

[02:27:42] He's like an optimist pessimist. He thinks the glass is both half full and half empty. It's like an engineer. He thinks the glass is twice as big as it needs to be. Yes. I love that. I love that. All right. One more break. A few more stories. Now that we've got you all head up over flock, I guess, you know, we're going to have to keep covering this story. This is a growing story. And it kind of snuck up on us.

[02:28:07] These, I mean, between ring cameras everywhere and flock cameras. There's 120,000 of these things in the country. Or more. Well, you know what it is, Leo? So, the Technorati have been, we have been raging against the growing omnipresent surveillance state for, what, two decades now? And every time they've kind of hemmed and hawed about it. So, I think we just assumed everyone's going to be okay with everything. Yeah. So, the fact that they're finally, this finally crosses a line, I kind of enjoy it.

[02:28:37] I mean, you know, you talked earlier about, you know, the backlash against big tech. I think the reason why we're seeing this backlash is because of leaders in big tech, people like Zuckerberg, Peter Thiel, Alex Karp, and, you know, and so many others, you know, that are, many of them, you know, some of them are openly advocating for the end of democracy. The end of freedom. Right. Yeah. Yeah.

[02:29:04] And what's the name of that ridiculous philosopher, you know, that they all follow? I can't remember. No, no, no, no. This Silicon Valley guy. Oh. The wannabe Ayn Rand. Yeah. What's his face? I can't remember his name. I know who you're talking about. Not Peter Thiel. No. Well, Thiel is a follower of this guy. Right. I know who you're talking about. I can't remember his name right now. Yeah. Well, Flock says they've 10 in 2025.

[02:29:34] It helped locate 10,000 plus missing people. 42% of Flock customers at his cities recovered more than half of all stolen vehicles. That one in five cases were cleared involving Flock evidence, including abducted children, home invasions. It wouldn't work in Las Vegas because like 90% of the cars aren't registered. Curtis Yarvin.

[02:30:04] Yeah. That's the guy I'm thinking of. Okay. None of that matters because you've passed the line, the ick line. You've done something that has collectively made the public say, no, I don't want that. I don't know why that stands out for me, but I don't want that. And that's Flock right now. Flock can say anything they want. And people would just say, we can't trust you anymore. You cross that line and there's no way back. Right. All right. We'll take a break.

[02:30:34] Come back and talk about something nice. Thank you. Something light, something like a drone dropping a package into a swing pool. You're watching This Week in Tech with Father Robert Ballasair, the digital Jesuit and Sam Abul Salmon. My car guy, Molly White, who was scheduled this week, has moved to next week. Couldn't make it this week, but we look forward to having her on next week on the show. This Week in Tech brought to you by Adaptive, the first security awareness platform built

[02:31:04] to stop AI-powered social engineering. Here's the shift. Attackers don't need malware anymore. They just need trust. A cloned voice, a convincing deepfake on Zoom, or an AI-written phish that looks like it came from your IT team. Guess what? It didn't. Adaptive prepares your organization with simulations across email, SMS, and voice. Deepfakes, vishing, that's voice phishing, and AI-generated phishing, including scenarios

[02:31:33] that can mirror your own brand and executives. And when employees report something suspicious, Adaptive can help you triage it fast. So security teams aren't buried in false alarms. If you need training fast with Adaptive's AI content creator, you can turn a breaking threat, something that just happened, an incident report, or a compliance doc. Somebody said, we've got to do this now into interactive multilingual modules in minutes. You don't need a design team.

[02:32:02] You just need Adaptive. With Adaptive, you can build, customize, and monitor every part of your training with complete personalization. The result is a more resilient security culture, which is essential for companies like Plaid. I use Plaid. You probably do too. It powers thousands of digital finance apps and links consumers, developers, and institutions. With sensitive data at its core, Plaid's security and compliance are non-negotiable. That's why they use Adaptive.

[02:32:29] Plaid's head of security, GRC, says Adaptive has equipped our teams with cutting-edge tools and built a smarter, more resilient security culture across the company. Trusted by Fortune 500s and backed by NVIDIA and OpenAI, Adaptive is building the defenses we need for the AI era. Learn more at AdaptiveSecurity.com. That's AdaptiveSecurity.com. We thank them so much for their support of this week in tech.

[02:33:01] Did you see? This is... A happy story, Leo. It has to be a happy story. There's nothing happier than when robots hurt themselves. This is the fast robot at the Olympic... Oh, he just ran into the... Oh, that's got... And blew up. Oh, wow. Sparks. The robot Olympics are going on in China right now. By the way, the robots beat the... Look at the little wobbly there in the middle. The 100-yard dash. I'm sorry.

[02:33:31] You know what? We're now on record at laughing at robot mayhem. Just for those... Which means we're going to get it. We're the first up against the wall after the robot revolt. So, Leo, when... If AI robots want to board a passenger plane, do they have to remove their lithium batteries before they get on board? That's a good question. See, these are the... They're going to throw over 100 watt hours.

[02:33:56] At the Chinese Robot Olympics, a robot beat Usain Bolt's record for the 100-yard dash by several tenths of a second. Well, that's like a car beat him. It's a machine. Yeah, right. I know. So what? Right? It's a robot. Yeah. Was it named Usain Bolt? No. No. It was not. Okay. That's saying Volt. I like that one. I like that. Yeah, this robot, not obviously one of the victors in that 100-yard dash. But it was fast. I don't know. I think it's kind of scary, actually.

[02:34:25] I don't want robots to run faster than humans. But, Leo, do you remember like 16, 17 years ago when we started watching those DARPA challenges and we were laughing at them falling over? Yeah. I mean, the fact that they've done this in just over a decade, that's okay. It's impressive. Even if you don't like it, it's impressive. A lot of the same people that were building those early DARPA Challenge vehicles are the ones leading these companies today.

[02:34:52] You know, Chris Urmson, Brian Selesky from Carnegie Mellon. You know, they led the Waymo program for quite a few years until they went off and did their own thing. You know, many of the others that were involved in those programs are still involved today. And they've learned so much. Yeah. We've come a long way, baby. Meanwhile.

[02:35:20] Chris Urmson's got, you know, his company is running autonomous trucks around Texas and New Mexico now. Yeah. But can you play Crisis on it? That's all it matters. That's the real question. You can't get, I'll tell you what. Not enough GPUs, you probably could. You can't get them in the United States. They are banned in the U.S. Just like BYDG robots. Yeah. FCC last month added advanced robotic devices to its list of banned devices because of their national security risks.

[02:35:50] Robot vacuums included. Lawnmowers included. Yeah. I hope you enjoy your... Anything that generates a map. Anything with a camera. Yep. New models of farm-made. Humanoids. Quadrupeds. To qualify for entry, a robot needs to be assembled in the U.S. with 65% or more of its components by value produced in the U.S. And, of course... So that's zero. That's zero. We have none of those. That's zero. I think the Boston robotics stuff is actually made here. They got bought, though.

[02:36:20] All of their own? All of the assemblies? I think so, yeah. Hyundai owns them, but I think they are made in the U.S. Hyundai's a South Korean company, so not on the banned list. Because I can't see a lot of the electronics from Boston Dynamics coming from U.S. fabs. Wow, that's the problem. In fact... Well, the chips probably aren't, but... But that's the important part. Yeah. I mean, an aluminum arm isn't going to report back home. It's the electronics that are going to. So that's the important parts.

[02:36:50] And the software. Right. Yeah. And this is problematic for U.S. robotic companies like Boston Dynamics because they can't get the parts they need because those are banned as well. On a corollary, on this Asia trip that I just did, one of our primary vehicles was a BYD. And that is a really nice vehicle. It was an extremely nice vehicle. And inexpensive, right? And crazy.

[02:37:17] Half the price of a Tesla, twice the luxury of a Tesla, the range of a Tesla. I'm thinking, wait, is this the reason why they're banned? They're going to be selling across the river in the next couple of months. Canada. Yeah. But you can't bring it into the U.S. It's like 120% tariff or something. Well, I mean, there's the tariff, but also there's a bill in Congress, sadly, co-sponsored by one of my state senators, one of my senators. Yeah, because he's protecting the big three.

[02:37:47] You know, they would ban those vehicles from even being brought into the U.S. So you can't even register it in the U.S.? Not just that, but like the U.S. ambassador to Canada said that they are going to institute a policy to not even allow them to cross the border. Not just not register them. Like a Canadian that happens to own one would not be allowed to drive it into the U.S. Wow.

[02:38:11] Meanwhile, Dan O'Dowd, who we had on Wednesday, we talked about it, who is the Dawn Project founder and the guy crusading against Tesla, also warned us about the vulnerability of our grid, our water systems. Here's a story for Wired. Brian Barrett, Andy Greenberg, writing, China is strapping digital bombs to civilian infrastructure. We don't need a digital bomb.

[02:38:38] We need the infrastructure that we've been implementing for the last 40 years. This has been a DEF CON talk for, what, 20 years now? The SCADA, the supervisory control and data acquisition devices that basically run every water plant, every power plant, are so woefully poor when it comes to security that at DEF CON, we can violate these within a minute. Like 60 seconds, get it and we'll be in the device. Using probably technology written by U.S. and Israeli security companies.

[02:39:08] It was called Stuxnet. It was used to destroy SCADA centrifuges in Iran, but it leaked. It leaked. It leaked all over the place. Now everyone has access to it. And those devices, they were never designed for the internet. They predate the internet. No, they were air-capped. Remember, they had to do candy drops to get them on the SCADA devices. And then they tried to slap on security at the end, but we know that doesn't work. That never works.

[02:39:38] They're just, they should be, every single one of them should be ripped out. Just put them all in a sandbox so that they can't, oh, never mind. So you've heard about Assault Typhoon, which is the Chinese infestation of our phone networks, which by the way, nobody can do anything about because if you fixed it, not only would it cost a lot, you'd have to rip out all this equipment. You'd have to rip out everything. And the telecommunication system would be down during that process.

[02:40:06] Now they're calling it, it's not Assault Typhoon, it's Volt Typhoon. Same idea. A Chinese cyber attack on our infrastructure, water and electricity. And as Dan O'Dowd pointed out, if they were to pull the trigger on that, we would be starving within days. I mean, literally without food. A blackout, you could do, you could trigger a nationwide blackout by attacking nine of the 10 most important substations in the country.

[02:40:35] And not cyber attacking, like with a gun or with a Molotov cocktail. If you took out nine, we wouldn't be able to restore power across the United States for months. That's how Oral was longer than that. In the face of Assault Typhoon, T-Mobile did do something about this. This is a report from Bloomberg. Cyber security staff at T-Mobile chopped a cable. This is the cable. They cut it.

[02:41:02] As they were looking for Assault Typhoon, when they found it, there's a frayed piece of yellow Ethernet cable on display at T-Mobile's Seattle headquarters. A security staffer cut it during an intense nationwide hunt for Chinese spies, Assault Typhoon. They weren't able to get rid of Assault Typhoon entirely from their network. But that's one way to do it. Just cut the Ethernet. That has a history.

[02:41:30] Do you remember when the I Love You worm was first released back in the, was it 90s? Yeah. You had one network administrator who panicked and just took an axe to the communications trunk coming into the university. That was one of my favorite stories. It worked. It worked. Yeah. That's a little harder to repair than a single severed Ethernet cable.

[02:41:55] But meanwhile, robots are running into walls at the Chinese Olympics and bursting into flame. And that's just what I like. I know we did. We're doing it again because... Because it's funny. It's funny. That is so funny. It's that backpedaling. The backpedaling just sells it for me. So to describe it, it's at a 100-yard dash. It's running down the track.

[02:42:19] For some reason, it loses track of the course, runs into a wall, stumbles backwards, and bursts into a shower of sparks. Oh. Poor little Chinese. Could you delete this part? It can't even stay in its lane. No. I don't want future AIs to see me laughing at this. I know. We have to cut this out. Oh, no. We have to cut that out. I just play it again as a moose-bouche after all of the disaster stories and disaster scenarios we've been talking about today.

[02:42:49] I don't know. I will say, though, that's progress. That is still progress. They will learn what caused that failure state, and they will fix it. Don't run into the walls. I guess. That would be. Well, that's what you're going to see chasing after you when you're running away. Yeah. Faster than you can run. Faster than Usain Bolt can run. Or we could go really dark and you strap a bomb to it and have to run it in the guy next to you. That's true. You don't have to be faster than the robot. Yeah.

[02:43:19] Sam Aboul-Sam, it's always a pleasure. Thank you so much for being here. He is the VP of Research at Telemetry, where he covers automotive tech. No one knows more than Sam does. You can get a little bit of that in his podcast, Wheelbearings, with Robbie and Nicole. Wheelbearings.media. No more Nicole. Nicole is retired. Yeah. Aw. She's walked away from the whole automotive journalism stuff. I love Robbie, though. In fact, we've got to get him on the show. Yeah. Absolutely. Roberto Baldwin. And he could talk about his many cover bands.

[02:43:49] Yeah. He just did a show last night with the Talking Heads Band. Oh, nice. Oh, nice. Robbie's great. I love Robbie. Yeah, he is awesome. Send me his, send Benito his contact. I think you're back. I will do that. We've gotten out of touch with him, and I just really always enjoy having him on. He'd be great. And in addition to Wheelbearings, I've got a couple other podcasts. Oh, tell me about it. I've got my Telemetry Transportation Daily, which is like a two-minute daily hit on some interesting news item in the transportation space.

[02:44:17] And we also have the Green Cars podcast that we're doing, where we interview interesting people in the transportation space. This past week's show, we had Scott Case, who's the CEO of Recurrent, a really interesting company that gathers data on EV batteries, both for to allow people to track the health

[02:44:44] of their battery and their EVs, but also to help figure out the value of EVs for resale. And they're doing a lot of really interesting stuff. It was a fascinating conversation. And this week, coming up on Wednesday, we are going to have Alan Clark, who is the VP of Advanced Development Projects at Ford.

[02:45:11] He is the leader of Ford's Skunk Works project that has been working on low-cost EVs. And so their first product is the new Ford Fathom, a compact electric pickup truck. So it's about a maverick-sized pickup. It's just their slate competitor, right? Yeah. And it's going to be, I mean, they're out and about testing now. They should be starting deliveries to consumers probably about this time next year, I think,

[02:45:39] or late summer, sometime in the fall. The plant is being retooled right now. They're going to start pilot production there early in the new year. I thought they'd given up on electric trucks. They gave up on the Lightning because electric full-size trucks are kind of the wrong product. People who buy F-150s are not buying electric F-150s. Well, a lot of people buy trucks thinking, oh, yeah, I'm going to do a bunch of towing.

[02:46:09] But the reality is the way people actually use their trucks, the Lightning was actually the best version of the F-150. And for what most people actually use an F-150 for, they would be much better off with a Lightning than a gas one. But for the people that are towing, it's the wrong vehicle for towing a big trailer. Right. But for smaller vehicles like the Fathom, which is about the size of a Maverick, it's going to be a great solution.

[02:46:38] And they're doing a whole bunch of other products off that same architecture as well. So check that out on Wednesday, Alan Clark, to talk about what the Skunk Works was, what they're trying to do, and how they're developing their new EVs. I think I met Alan, didn't I? When I was buying the Mach-E. No, that was Darren Palmer. That's right, Darren. Yeah. Alan has not been out and about very much since he joined Ford a few years ago.

[02:47:07] But that's starting to change now as they get closer to the launch of the Fathom. I'm glad to see American car companies not give up on EVs. I love my EV. Yeah. I mean, my wife has said she will not go back to a gas car. And she loves driving her EV6. She is a big. Recently, her best friend bought a Chevy Blazer EV. Loves it. Nice. Although, the dealer didn't do a great job of explaining a bunch of stuff.

[02:47:36] And so she came over here and I spent a little time. Sam will tell you. Yeah. Here's how you open the charge port door. And here's how you plug in the charger. Right. Right. Dealers. Green Cars, the podcast. And of course, The Telemetry, a daily podcast. Thank you. And wheel bearings. Both. And wheel bearings podcast. And then all the stuff that we do at Telemetry as well. Have microphone, we'll travel. Yeah. Sam and Bill Samet. Thank you, Sam.

[02:48:05] Father Robert Ballas here. When do you go back to Italy? September 10th. The morning of September 10th. I'm jumping on a plane to get home back to Italy. Right after the Big Apple event. You're going to Cupertino the day before. Yes. Of course. Of course. No, actually, I don't know how much of this I can. I can say this. I am doing creative work for a big content producer in the Bay Area. Yeah. I can say that.

[02:48:35] Not me. That you will see on the big screen at some point. What? Yes. So I get to do some fun stuff. What a tease that is. I cannot say any names. I did have to sign a bunch of NDAs. Oh, I know. Oh, I know. But when will we see this stuff? This. No, no. I think the release is Christmas. I think the release is Christmas. Probably. Yeah. And what kind of work are you doing?

[02:49:04] Are you praying over it or what? All sorts of work. You know. Stuff. Voicing a priest in a new animated movie. Things and stuff. Stuff. Yeah. Yeah. Robert, you're amazing. If you've just think of my background, my ethnic background, and you can probably figure out the movie. I'm just going to put it that way. Oh, how exciting. I don't know. You can look it up. You'll find it. No street balutes. I don't know. What is it?

[02:49:32] Actually, what I really want, though, is I really want a Leo t-shirt. Well, I was just going to say, you'll find me on the streets selling Leo t-shirts. I got a lot of them. Those actually do look really good, Leo. They do look good. Thanks to Jeff Atwood for making these. Actually, the whole story is he, of course, is one of the creators of Discourse, which is a forum software we use. And on their executive page, all of the executives are portrayed as busts like this.

[02:50:01] So he decided to make me an honorary bustee. That's pretty good. Did you know that at the Pixar Studios in Oakland, there's a bust of William Shakespeare's And if you flip the head of the bust back, there's a switch. Like Batman. You click that switch, it opens up a secret door where Steve Jobs and a bunch of Pixar animators used to have their Friday booze and cards nights. Now that's cool. Yeah.

[02:50:30] That's very cool. See? He knows a lot. He knows where all the secrets are. That's Father Robert Ballas there. Anything else you want to mention? No, I'm heading back to the Rome. I've been on the road basically for almost coming up in three months. So I'm excited to just get back and try to remember what I was doing before I left. Get back to your little cell and your hard bed. And the next time you'll be coming through is when? When? I don't know.

[02:50:59] Soon, I hope. I love Rome. I would move to Rome if I could. If I could. I just found out about this on our campus. I had no idea it existed. There is a special place with two much nicer rooms that have a private common area like a little kitchen. It's owns patio that looks out over Rome. It's a special guest place that we can reserve that I'm just saying. You're just mean to me.

[02:51:28] It's got two rooms attached to it. All right. I'll be there. Count me in. Thank you, Robert. Thank you, Sam. Thanks to all of you for joining us. We do Twit every Sunday from 2 to 5 p.m. Pacific, 5 to 8 Eastern, 2100 UTC. You can watch us do it if you want. We do it live in front of a non-studio audience. Of course, club members get special access in the Discord. If you're not a club member, this would be a good time to join. We need you.

[02:51:56] Twit.tv slash club twit. Support the programming we do. If you enjoy it, it would sure be nice to have you. Lots of benefits, including the Discord access and ad-free versions of all the shows. Twit.tv slash club twit for more information. You can also watch, even if you're not a club member, you can watch us live on YouTube, Twitch, X, Facebook, LinkedIn, and Kik. We're everywhere. After the fact, download audio or video of the show from our website, twit.tv.

[02:52:25] There's a YouTube channel with a video, of course. If you go to youtube.com slash twit, you can find all of the shows and all the channels. Or subscribe in your favorite podcast client. You'll get it automatically as soon as we're done. And that's nice. You'll have it just in time for your Monday morning commute. Thank you all for being here. Thanks to our producer, Benito Gonzalez. Appreciate your hard work, Benito. Our editor, Kevin King. Our executive producer, Lisa Laporte.

[02:52:52] And I will see you all on Tuesday for MacBreak Weekly. Meanwhile, another twit is in the can. Bye-bye. This is amazing.