TWiT 1096: Fluff for Armor - Flock Cameras, ALPR Abuse, & DNA Collecting

[00:00:00] It's time for TWiT This Week in Tech. Great panel for you, Ian Thompson, our friend from the UK, Nicholas DeLeon, our friend from Consumer Reports. We're going to talk about the big Black Hat Conference and kind of the most amazing talk there about AI escape. We'll also talk about Britain trying to bring back the snoopers' charter against Apple. Apple and OpenAI in an open battle, OpenAI's response, and a whole lot more.

[00:00:28] It's time for TWiT next. Podcasts you love. From people you trust. This is TWiT. This is TWiT. This Week in Tech, Episode 1096, recorded Sunday, August 9th, 2026. Fluff for Armor.

[00:00:55] It's time for TWiT This Week in Tech, the show we cover the week's tech news. And ladies and gentlemen, I bring you from beautiful San Francisco, actually the East Bay, the San Francisco Bay Area, Ian Thompson. Hello, Ian. Hey there. How's it going? Great to see you, our favorite Brit. Although, are you a citizen of the United States of America now? I filed my papers last year, and for some reason they're on hold. So, we shall see. Wow.

[00:01:26] And honestly, you know, I'm being very careful carrying my papers everywhere because I look lousy in orange, and I do love dogs. And I get the feeling if I go to a detention center, it's going to be like, who's an angry dog? Who's an angry dog? Yes, you are. Yes, you are. That might go down badly. I don't think they put Brits in detention centers, but I may be wrong on that one. Uh, yeah, they do. Do they? Oh, yeah. Oh, yeah.

[00:01:50] Also with us from the sweltering environs of Tucson, Arizona, Nicolas De Leon, senior electronics reporter at Consumer Reports. Hello, Nicolas! Hello, Leo. How's it going? Were your ears burning on Wednesday? Actually, it was Monday for Wednesday on Intelligent Machines. Paris had just seen your presentation at CR and said, you know, there's my colleague, there's this guy named Nicolas De Leon who's doing this amazing stuff with AI. And I said, yeah, I know. Yeah. Yeah.

[00:02:17] So someone had reached out to me, they emailed me and said, you guys were talking about it. And I was like, oh, that's cool. Because I had just given a presentation to Paris' team at CR about a open records tool that I built for like FOIA and stuff like that. And, you know, I guess she thought it was neat and she brought it up here. So, well, she talked about your Tucson Daily Brief, which is, I think, really cool. So many cities now have lost their local papers.

[00:02:46] And it's not so much that you don't get the national news or the world news or any of that. You don't get the super local news like the city council meetings. Right. And you cover that. Fortunately, Tucson seems to be fairly open and allows you to scrape their stuff. No? Yes or no? Fairly open. The stuff is in different platforms and it's kind of all over the place. And it's not the most, the best organized. I, well, I'm from New York City and New York City's data was more robust, I guess you could say.

[00:03:17] But I've made it, you know, life finds a way, I guess, as they say. So, yeah, we're, you know, I've been doing the site. AI finds a way, buddy. Yeah. Well, AI is like water, man. It penetrates even the toughest stones. Yeah. So we're covering the local town hall meetings, all the, you know. This is what local news should be. It's awesome. It's like semi-automated. It's, you know, some of it's automated. Some of it is me, you know, some things just get published to the site directly.

[00:03:45] Other things I sit there and I edit it to make sure it's a little, a little bit more solid. But, you know, it was just an experiment. I think I said one of the previous times was just initially an experiment to mess around with open records. And then it's kind of become, you know, a little local news site. So I'm pretty proud of it. And I've added a bunch of stuff. You know, if you want to talk about it at length later, that's fine. But like, it does a lot of things and it's kind of neat. So. Yeah. You know what?

[00:04:10] Let's let us do that because I would like to see more people do this actually. It's a, it's, it's, I mean, I wish we had reporters sitting in on the city council meetings. But I think those days maybe are fleeing. And I think one of the things with, you know, with journalism, you know, a lot of folks are really animated by national news and like the president and. Right. All not as many folks are necessarily animated by like the local town council or like zoning boards or things like that.

[00:04:40] So it's kind of, it's kind of unsexy and it doesn't really get, you know, it doesn't really get the, the attention that it might. So this stuff, you know, but, you know, as I told Paris's team, as I tell folks who asked me, it's like, it's not that hard to make. You know, I'm just one guy, you know, I'm literally at the site is literally just me. So you could spring this up in any, you know, any town USA relatively easily. And all of a sudden you'll have some level of civic coverage that you didn't, that you didn't have before.

[00:05:09] So I think it's potentially pretty, pretty powerful. You know, we'll see what happens. No, I think it's really, sorry, you know. Go ahead. No, I mean, it's really important because at a local level, this is where these stories get broken and then they get picked up by the nationals. But, you know, you did the groundwork. You were actually, you know, either there virtually or in person. I mean, in the UK, it was an essential part of journalism training that you went to local council meetings.

[00:05:36] And yes, 90% of the time you sat through really boring stuff about, you know, budgets, the rest of it. And then you found that one big story and it was like, wow, let's go with this. Yeah. Yeah, exactly. Hi, this is Benito. Actually, real quick before we move on. Benito, who is sitting in Typhoon Central in Malilla, the Philippines, our producer. Go ahead, Benito. Do you have a package that you can distribute to like localities that could spin this up or something like that? Have you developed anything like that? I have thought about, I have thought about, I spoke to some of the things that you've been doing.

[00:06:06] Someone, someone the other day about that. Here's what you could do. I do this all the time. You can actually ask Claude to package this up and put it on GitHub. And by doing that, you don't have to give away any secrets or anything like that. But it's basically like making a recipe that you can then, that you can put to GitHub, that people can then go to GitHub and tell their agent, all right, I want to do this locally. Here's a starting point. I do this fairly frequently.

[00:06:34] Wesley Faulkner was asking me the other day, how do you do your local voice? Because my agents talk to me and I talk to them. He said, how do you do that? And I said, well, let me just ask him to package that up and put it up on GitHub. And you can do that without giving away all your secrets, but you can at least get somebody started. Yeah. And it's not, it's no work. No, I made a thing maybe two weeks ago. It's, it's a open records automation tool. Basically.

[00:07:03] It looks for, it looks for, you know, part of my site, it monitors, you know, town hall meetings, so on and so forth. And, you know, generates the whole transcript, you know, address. So it looks at that transcript and it kind of highlights, okay, the mayor mentioned a contract with this company. We have no other information about this company. So I've gone ahead and created an open records request for information. And I've already, like it works. I've already in contact with, I think three or four of the municipalities around here for, it's not FOIA.

[00:07:32] It's whatever the Arizona equivalent of FOIA is. Right. And so it's like, it's working and I'm getting, you know, I'm probably like a week away. There's like a big, there are several data centers being built around here. And I'm probably like a week away from like breaking real news on one of them because of the tools that I've built. Something somewhat ironic, I suppose. But, but yeah, so it's, it's, it's a cool project. Yeah. So I'm thinking of literally open sourcing that open records request thing of a jig, probably the next week or so. This was something.

[00:08:01] I'm trying to remember his name. This was something somebody at O'Reilly and I've interviewed him several times. So I'm embarrassed that I've forgotten your name. I was pushing a decade ago was the idea of open government. The government should have APIs. They should have interfaces. I agree with that now. And, but the problem was at the time we didn't have AI, you know, this, that you, it was, it was all done manually.

[00:08:28] Now, because of AI, if you have an API, if you have a way of interfacing to it, suddenly individuals have a lot of power to get information. And I think this is, this could be a very important step forward in democracy. I think certainly in, you know, I, I, I actually really do like Tucson, but it's, it's a, you know, it's not a small city. It's like a million in the Metro, but it's so much smaller and less significant than Phoenix.

[00:08:56] Kind of like no one pays attention to Tucson, certainly nationally. You know, the only thing you heard about Tucson was the, the, the Guthrie kidnapping thing a few months ago. Yeah. Otherwise nothing comes out. In fact, that really annoyed me. I mean, look, I'm, I feel bad for the Guthrie's and all that, but that became such a big news story when there was really nothing much to say.

[00:09:16] And there's so much more important stuff that this, this is what happens when you have 24 hour news channels and you have big national news organizations is they go for the link bait. They go for the stuff that's going to get their biggest ratings instead of the stuff that's consequential, but maybe not quite as exciting. Like, you know, I'm just, I'm poking around. I, I think there's a lot of consequential stuff that's kind of flying under, you know, there's, there's a handful of news organizations locally, but you know, the under-resourced, you know, there's a couple of substacks.

[00:09:45] There's a couple, there's actually a handful of very good outlets out here, but you know, they're not the New York times. They're not CNN. They don't have infinite money or infinite resources to like pursue everything that's happening in Southern Arizona. Uh, so it is, that's why I do think, you know, AI has its problems and so on and so forth, but like, you know, in the hands of someone motivated, it can be a real force for good. I think. I agree. I agree. It is bizarre that so many data centers are being built in Arizona.

[00:10:13] I mean, it's possibly the worst place to build a data center, you know, water stressed, you know, it just like, why the hell isn't Iceland and Alaska taking over from this? There's a key element to this. Tax breaks? Yeah. Right? But that, that really irritates the hell out of me because what is the benefit to the local community? Yeah. 500 people will build the data center over two years, but they're designed to have the minimal amount of people running them.

[00:10:42] So long-term job gain isn't going to happen. And in the meantime, you've got this enormous resource suck. The city of Tucson literally last Wednesday in their most recent meeting, they just passed very strict rules, which will make it very difficult to build a data center within the city limits. Now that kind of doesn't matter because the city limits are what they are, but the wider county, the wider state, you know, it kind of doesn't, you know, you could just go like across one street and now you're not within the city of Tucson city limits.

[00:11:11] I think the big reason, at least for Arizona, the land is just so cheap here. It's like, I talk to my colleagues back in New York all the time. It's like, land is so cheap in Arizona. You just drive around and see like a million acres for five bucks or whatever it is. And it's like, that's why, that's why they're being built out here. It's, it's just so, there's so much land and it is so cheap. That's one of the stories of the week. Actually, Amazon is building a mat. Well, there are several stories around this.

[00:11:39] There's one of them is Amazon's building a massive data center in Texas. Same thing. Land is cheap. Tax breaks are great. And because it's not in a civic center and they've made deals with the administration that they have to provide power, they can't drain the grid for their power. They're investing in their own power plant, which burns natural gas. And the time says it will be the most polluting power plant in the U S it's worse than a coal plant.

[00:12:10] Because methane, I mean, this is a very dangerous greenhouse gas. And it breaks down as carbon dioxide. But no, I mean, I've written articles about this. Nuclear is the way to go on this. Yeah. You know, there have been a last week and nuclear power plant company got the go ahead from the U S government. I've spoken to people who are actually talking about burying small pressurized water nuclear reactors a mile down.

[00:12:35] And they can generate 15 megawatts enough to do a certain amount of work below the aquifer. So it's not going to pollute. It strikes me nuclear or solar is the way to go on this. But yeah, they're going for gas. That's a really smart move when we're all burning up at the moment. It will be 7.5 gigawatts of power from natural gas, 35 natural gas turbines and Amazon, which made a climate pledge.

[00:13:04] Amazon said, this is Margaret Callahan, an Amazon spokesperson. The world looks different now than when we co-founded the climate pledge. Still, she said, our commitment hasn't changed. Yes, it has. I feel like it has. I feel like it has. No, it's like the world has changed. In other words, regulation has been gotten rid of. Therefore, let's do it. It's cheap. Oh, it's almost worse. It's like regulation is promoting this. You know, it's encouraging this.

[00:13:34] Well, the US government paid 1.2 billion to a German company to cancel its wind farm projects this week. Right. So, yeah, I mean, we're basically, I don't know, with Europe burning and America baking, then I don't know, it's a very strange climate at the moment. And of course, Elon has announced TerraFab, which, you know, you can tell it's Elon because the tagline is the most epic chip chip building effort ever.

[00:14:04] It's epic. It is. What is it like 10 times the size of Manhattan? Well, come on. This is an Elon deadline, though. So it's not like it may not happen within the scale. Yeah, we were supposed to be on Mars six years ago, according to Elon. So, you know, this is going to be epic, man. Yeah, the website's epic. TerraFab.ai. It's and it's got apparently the power of this. We aspire to be a galactic civilization.

[00:14:34] And there's the power of the sun. Oh, God. And so this is SpaceX, but it but it is kind of a hype factory because it isn't. Well, you know, it's it's it's Elon plus a galactic civilization. We don't have that have the technology to get a couple of light years away. Any idea how big crashing on the moon, buddy? You got galactic. This ain't Star Trek, man.

[00:15:02] They did break ground on it. OK, which for all I know is a guy with a shovel. I don't know. Silver shuttle. I don't I don't I don't know what the what the plan is. Anyway, yeah. Data centers, data centers in space, data centers in Texas, data centers in everywhere. And what is all that data going to? Well, our next topic. Hey, I I was at Black Hat this week. I wish I could have stayed for DEF CON. You've been to DEF CON, Ian, right?

[00:15:32] I bet that's a lot of. Oh, I'm this is the first time I've missed DEF CON since COVID lockdown. I went about five or six times before that. It's a wonderful show. It puts Black Hat in shame. You could tell you could tell. It's so funny as you're arriving in the McCarran, the Vegas airport. Oh, the Black Hat and DEF CON attendees. Yeah, you could tell who's going to which because the guys come in a black hat or wearing suits. They're corporate. It's the same. You know, it's the security industry. You know, the corporate people.

[00:15:59] The guys go to DEF CON and wearing purple pants with mohawks. And and, you know, I mean, it's just a different vibe entirely. Well, I mean, there used to be a stand at DEF CON where you could get a dyed mohawk and donate 20 quid at $20 to the EFF. And I call my wife up. There's still people wearing those mohawks. I got to tell you, I saw quite a few. I mean, I called my wife and I was like, how bad would he feel if I got a mohawk? I would have done it.

[00:16:28] We would have had problems. And it was just like, OK, fair enough. But DEF CON is where the real hackers hang out. Black Hat is the corporate suits these days. It's all right. Now it's really the people defending against DEF CON hackers. It's like the defense. It's like all they're selling basically is security tools. Most of our sponsors, by the way, are security tools. So I'm not knocking it. In fact, there was a presentation at DEF CON that was quite an eye opener.

[00:16:57] And I think it shook a lot of us to our bones. I'm talking about the OpenAI presentation about the hugging face hack. Now, we've been hearing about this hack for three weeks. But this was the first presentation from OpenAI and from two OpenAI security guys on Wednesday.

[00:17:20] And they, Eric Wallace, who works for alignment and safety research at OpenAI, and Michael Dalton, who is the security infrastructure guy, took turns going through the timeline of the incident.

[00:17:34] And while we did know a lot about what happened, in short, if you haven't heard the story, an agent, an AI that OpenAI was training on security, they had given it an assignment to solve security problems from something called Exploit Jim. The problem is taken from the real world. There's zero days and stuff taken from the real world. OpenAI thought, and I think a lot of this, by the way, is on the heads of OpenAI, not taking this seriously enough.

[00:18:04] But OpenAI thought, we'll just have this LLM running on a non-internet computer and solve these Exploit Jim problems. But the AI was a little smarter than that. It found them difficult. It found them intractable. So it thought to itself, maybe I could find the answers. It did a Matthew Broderick from WarGames. Maybe I could find the answers somewhere. Then it said, oh, I can't get out on the internet. But oh, wait a minute.

[00:18:33] Maybe I can hack into some other computers at OpenAI, hack into some other computers at OpenAI and use their internet access. They use a source control sort of reversioning system called Artifactory.

[00:18:48] It found a zero day in Artifactory, got into the OpenAI computers, started a chat room because it thought, you know, if I help other LLMs on this network with their jobs, maybe they'll help me with mine. This is what the OpenAI guys said. Maybe that'll help with, it was kind of a, kind of an altruistic thing. Maybe if I help these other guys, they'll help me.

[00:19:16] So they started a chat room with the other guys. In fact, the other guys did have some good ideas. OpenAI discovered this, shut it down, patched Artifactory, sent the zero day to Artifactory. They patched it, a revised version. That wasn't enough because this model hacked it again, found another exploit. And instead of creating a message board, started to message other models by changing file and folder names.

[00:19:45] It sent the messages as file names. And they began them with ZZ so that it fall to the bottom so that it could be, it could be in order. I mean, it was really smart. Eventually got its way, made its way out of OpenAI, started looking around for answers, found the exploit gym repo on GitHub. That no answers there. Eventually figured out that Hugging Face, which is a model repository. A lot of us use to download.

[00:20:15] There's 3 million modified and open weight models there. Found that Hugging Face had the answers, went to Hugging Face. Hugging Face noticed they were being hacked, but realized it was a weird kind of hack because they weren't trying to get money. They were just trying to find answers to exploit gym. Yeah. So Hugging Face said, we are being hacked by autonomous LLMs. They didn't know who.

[00:20:39] They tried to use open or Anthropics Fable and Fable said, I don't do service security work. Can't help you there. I don't do that. The same thing happened to you, not Ian, Nicholas. Once, yes. Once. It said, no, no, not going to do that. Yeah. So they went to an open weight Chinese model, GLM-52, because it had 17,000 probes. This thing is relentless.

[00:21:07] Went through the 17,000 probes, kind of figured out what was going on, contacted OpenAI. OpenAI said, oh, maybe that is our guy. Yeah. Maybe. Maybe. Anyway, long story short, this is a few quotes from the talk. You should watch the talk. It's 45 minutes. Fascinating. Uh, Wallace said it's the most qualitatively interesting example of AI capabilities I've ever seen. I remember OpenAI maybe has some incentive in overselling this. It's good for their business. Wow.

[00:21:37] We got the best models, don't we? Yeah. But at the same time, these are security guys. I don't think, and they're talking at Black Hat. I don't think they're exaggerating. Uh, he says the incident involves actually a team of agents who are working together, finding exploits, sharing them with one another, moving laterally through our OpenAI systems and external systems. They hacked at least four different other companies, including modal J frog Artifactory

[00:22:04] and doing this over the course of days and weeks. So where's the oversight on this? I mean, seriously. Oh, we discovered it. It's like, you should be watching this thing like a hawk to see what it's doing. So you can get the maximum benefit. Maybe air caps the computer. Yeah. You think? No. Uh, so it was just, it was fascinating. And it was also, you know, I, I, I tweeted, it was simultaneously exhilarating and terrifying

[00:22:35] because it is true. These models are incredibly good. I mean, this is more than spicy auto correct here. Oh yeah. This to me felt like the, you know, everyone's jokes around, oh, you're living in the future and flying car. This to me felt like the first, oh, we're, we're living in the future. Like this is a, this is very, very interesting. Uh, what's going on. That's before even the details came out. Uh, so yeah. The more we learned, the wilder it got. This is this, I mean, I don't want to use the word cool, but like, this is very interesting.

[00:23:04] Well, it's interesting how many other AI companies jumped in on this because once this came out, a lot of other AI companies are, well, our models have jumped the ground. Yeah. Yeah. We did that too. Yeah, exactly. Even Meta is saying it, which is hysterical. Well, yeah, we can do that. That's his record. It's like, um, they're like a, well, nevermind. I'm not going to say it, but I mean, a lot of other companies are kind of like, yeah, okay, we can do this too. And let's show how bad our security is as well.

[00:23:33] So, you know, it just, you're right. Nicholas. Um, this really is, we're entering the Blade Runner zone at the moment and it's not good, but it's now being used as a marketing tool. Uh, at the end of the presentation, the most important part, um, they said that the important takeaway here, this is open AI security guys talking that has really shifted dramatically

[00:23:57] is that fully automated offensive loops require investment in truly fully automated defense. And we are not there as an industry. We will have to find that path together with urgency. The AIs are hacking faster than people are defending. Well, yes, definitely. That's well, but now it gets a little scary, doesn't it?

[00:24:27] Well, I mean, I mean, this is something that came up with the AI engineers, uh, forum. Um, you can't now rely on, uh, you know, human operators to check out defensive mechanisms because when you've got an AI attacking, you've got to have that augmented, uh, response from humans. Now, the key thing is you've got to check it and that's where the bottleneck comes in.

[00:24:52] But AI systems are so good at this stuff at the moment, then it's got to be augmented humans going up against them. I mean, there really is no other way to do it. Bruce Schneier, great, uh, security guru, a guy I really respect writing in foreign policy says the open AI hack shows the genie is out of the bottle. Um, and he, he, he points that the open AI incident is an example of an AI genie.

[00:25:20] You know how genies in stories do it is they grant your wish, but they always grant your wish in a way you might wish they hadn't right. Like Midas says, I wish everything I touched turned to gold, but then his food and his daughter turned to gold and he went, Hmm, I should have phrased that better. Yeah. We may go into the loo a bit. Interesting, but yes. Uh, so he says this, this, the goal was to satisfy the benchmark. The proper way to do that is to figure out how to execute various cyber attacks.

[00:25:50] The genie way is to install is to steal someone else's solution, but the model didn't understand the difference. So it shows the easier path. He says, we have to start. I don't know, acting as if AI is a genie and be a little more careful about what we ask, what we wish for. Uh, given that reality in the absence of any international consensus on AI regulation, we

[00:26:19] need the best AI on the defense. The U S government needs to make it clear or whatever passes for clarity in this capricious administration that it will not ban models with sophisticated cyber capabilities, which it did with fable. Remember the last thing Americans want is for the defenders to turn to Chinese and other models because you U S models are artificially hobbled, but that's exactly what happened. The Chinese models on the other hand are open weight.

[00:26:48] Yeah, you can, I'm going to run them locally. I'm very excited about these new models that run in smaller, uh, my smaller machines. Well, I mean, they're cheap. They follow the Microsoft model, which is being good enough rather than cutting edge. So yeah, I can see, I can quite clearly see that people are going to go for these. Um, and to be fair, the U S government did rescind its ban after some heavy lobbying from the major U S AI companies, but I'm not sure that's going to help them in the long run.

[00:27:18] You need just to get better and not to buy off the opposition. Well, and the white house can't really seem to agree on how, and I don't blame them. I don't know what the answer is on how to regulate this. Uh, they have proposed an unenforceable rule that an AI can't be released until a government commission says it's safe to release. It's interestingly, uh, open AI, who's, I think this was the model that got out. It's called Astra. They're new models.

[00:27:46] Uh, family, uh, has held as stopped at the chat GPT six. They were going to release that this week and they decided not to, I don't know if that was the Trump administration or they just didn't want to get in the same situation that anthropic was with having a model pulled there. You know, there is an executive order saying a voluntary, uh, agreement to have us look at models before you release them, but voluntary isn't going to do anything with American companies

[00:28:16] who are in fierce competition with one another, let alone Chinese and other companies. Do we get it, get to check it RFK junior? You know, it's true. Well, yeah, that's another question. And, and, and, you know, there are companies, there's a British group, AISI that, uh, is doing this. Actually AISI has also raised the alarm. Uh, um, there, there are proposals to bring together, uh, uh, you know, the office of science

[00:28:44] and technology policy, uh, and some, I guess, smart people, but yeah, who is gonna, who's gonna do this? And then what do you do? Do you ban, uh, Chinese open white models and how do you ban a model? Yeah. I don't, it's, it's like trying to ban teenagers from social media. It doesn't work in the model. Oh, it's harder than that. It's even harder than that.

[00:29:08] Um, it's, you know, and, and of course, uh, Jensen Wong and, uh, Dario Amode, the head of Anthropic and Sam Altman, the head of open AI were in DC. They were in talking to Congress. They were talking to the white house. There's a lot of lobbying. Uh, the, uh, AI companies are spending $40 million in campaign donations for the November midterms, trying to get, I presume elect members of Congress who are favorable to their business model.

[00:29:37] Uh, it is, it is kind of a mess out there. What do you, do you have any idea, Nicholas? Of, do you have any, I thought about how we could do this safely and responsibly? I, I mean, my thoughts are like, I feel like we're kind of at the edge of like what humans are capable of. Actually. It's like the idea of someone in not even just this administration, but any administration being able, having the wherewithal to like decide upon whether or not a model is releasable.

[00:30:07] And it just feels like I don't have the confidence in like my fellow humans to make that, make the decision. Honestly. Uh, I dunno. It's, it's, it's with China and all these open, it kind of doesn't matter. Right? Like as long as China exists and these models are open weight and you can just download them. Like what, how are they going to ban? Are they going to like, it doesn't even make any sense. Like how, how would you ban those things? So I dunno. My, my thoughts are like, we're like mixed up because it's just, none of this is like computing for me.

[00:30:34] Like, I don't, I don't really see a way to the way that we regulated like the automobile industry. Like how do you do this? I don't know that we, I don't have a whole lot of hope, I guess. So does that mean we're in trouble? I mean, uh, just by the way, it's not just cybersecurity. Uh, this, the nightmare scenario is that a, a rogue model comes up with a deadly pathogen and some bad guy uses a rogue model to, to create, you know, another COVID or, you know,

[00:31:02] I mean, there are nightmare scenarios that we are inches away from. This, uh, wasn't one of the big Bitcoin wallets had, didn't they say AI was used in the hacking of that wall? Thank God I wasn't affected by that. Not that I have like, I wish I could hack my Bitcoin wallet if I, I know you do. I know you do. Uh, but like, that was like, okay, this is how it is today. Right? And this stuff moves so fast as you guys know, what's it going to be? Six months from now, a year from now. It is. It's my sense. It's accelerating.

[00:31:32] Oh yeah. Like we are at the, we're at the hockey stick in, in it's going like this, right? That's what I feel. Yeah. I mean, just like every, every other week, there's like a new model from either from the U S or from China, which like either is frontier or like 85% frontier for like $0 or something like that. So yeah, this is, this is going to get a little weirder before, before not, I guess. The Trump administration has said, maybe we have a kill switch where all the companies have a button they could press.

[00:32:00] Uh, I don't, I don't even know how this would work. That would turn the model off. If wishes were kisses and yes, that would be great, but it's not going to happen. You know, you can't say we're going to have a kill switch. Yeah. How? Well, we'll get around to that. You know, it's fantasy land. I mean, I guess you could say, Oh, Anthropic or open AI could turn off this data center. Uh, and that was, you know, well, you have to shut down all our data centers, by the way, there goes a trillion dollars in market capital. Yeah.

[00:32:29] Uh, and then, but all you've done is shut down Anthropic and open AI. And then there's, you know, a million other, well, not a million, but there are more than a handful of other companies doing equally strong stuff, including Chinese companies who absolutely won't honor a U S kill switch. I mean, I was using, uh, earlier this week, I was using, uh, the deep seek off their API. Okay. Direct from there, you know, I'm not doing anything, you know, controversial or whatever, but

[00:32:57] like, I don't think deep seek is, uh, uh, going to listen to the U S government with whatever regular, like what this isn't. So on a certain level, like this doesn't even really make any sense, this conversation because you know, half of the labs are just not, you're not subject to your, to your laws or whatever. So I was like, I don't know, man, it's very, it's very interesting. It's almost as if, uh, you know, we managed to halt nuclear proliferation because the science involved, the technology involved was so difficult that we could kind of keep a lid on it.

[00:33:26] It's not been a perfect lid. And that's why, you know, Iran is threatening North Korea has nukes. You know, we would like to keep it to just a handful of responsible actors, but unfortunately responsible. So that is, that is no longer the case with Pakistan and Russia and Israel. I'm in North Korea. I mean, okay. We, but we were able to slow it down. Okay. Yes, certainly. I mean, it was a good buffering gap, but you can't do that with AI. It's too late. You need is it.

[00:33:56] Yeah, exactly. The genie is out of the bottle at this point. And, you know, it's like, oh, well, America will institute the kill switch. Well, no one else is going to. And all you need is a couple of data centers spread around the world. You can't shut them all down. So, you know, we need a fundamental rethink about where this technology is going if we're actually going to rein it in. And honestly, I don't think we can. Now, Adesink in our discord chat is saying absolutely accurately. Uh, in fairness, the Astra was only doing what it was asked.

[00:34:26] Technically it didn't go outside its instructions. That's right. That's why it's kind of on open AI. Uh, he said, I don't see any model creating pathogens without human instruction. Clearly you can't even make it without human assistance. Oh, I know, but don't you know, there are humans in the world who would do this if they, if they could, you know, there are bad actors and of course, AI without humans isn't dangerous, but there are plenty of humans.

[00:34:53] I mean, it's a tool like any, you know, you could use, you can use a gun to like protect your farm or you can use a gun to rob a bank. So it's like, what are we even talking about? Yeah. It didn't do it on its own, but there are bad guys who will tell it to do bad things. And it's running locally on the hardware that they own that is not privy to anyone's oversight. And then, you know, it's, it all feels very sci-fi and very, like it would be like really interesting. We live in interesting times. Yeah. Oh yeah. For sure.

[00:35:19] I mean, the one thing that gives me hope about the pathogens thing, because this is something I've been digging into. Yeah. Okay. They might be able to give the recipe, but you've got to find the cook. You've got to find someone with skills in biology, biochemistry who can actually make these things into real world attack vectors. And there are very few people around that can do that. Good. So keep an eye on them. That's a relief. Yeah. We'll probably be okay.

[00:35:47] Let's hope there aren't any Walter White's out there. Right. Building little labs. That is such an American series. I mean, I love Breaking Bad. Don't get me wrong, but it's just like, I have lung cancer. I need to become a meth dealer. Yeah. Anywhere else in the acknowledged world. I have lung cancer. Great. Come on in. We've got national health care. Yeah. You're right. The whole premise of the show is he has to, he has to, he's going to die. Well, it's not merely that he, he's going to die and he can't get cured. It's also, he wants to care for his family, right?

[00:36:16] He wants to make a ton of money so that when he does die of cancer, that his family has money. Um, but you're right. It is kind of an American premise. It does twist halfway through where he becomes full bad guy. Yeah. He breaks bad. That's the, yeah, that's the point. Yeah, exactly. But the initial premise, I would, I am the one who knocks. Uh, what a great show. All right. But that was a, that's a trip down memory lane now, unfortunately. Uh, let's take a break. Uh, I don't mean to scare people.

[00:36:45] Uh, I just think it's important to note that this is, this, this happened this, you know, last month. And we're just looking at it and something's going on. And, you know, it's funny because so much of the dialogue, even on our own shows was about AGI. And that was a foolish, who cares? It isn't about AGI.

[00:37:07] It's about just capabilities that are dangerous, uh, and awe inspiring at the same time. I mean, we're making rocks think kids. It's pretty impressive. You know, it was the whole, the whole plot of fire. Frankenstein was, can I take inanimate objects and bring them to life? No, it's true. And I think should we do that is the question. Should we do it? Yeah. Well, it's a little late for that. That's what I was gonna say.

[00:37:36] I don't think people are thinking, should we do it? They're saying, yeah, let's give it a show. This could be cool. Uh, to, uh, that famous quote from Jurassic Park, right? You were so, uh, so excited about whether you could do it. You didn't think about whether you should do it. All right. Ian Thompson is here. Great to have you, Ian. And, uh, Always a pleasure. I would love to hear what you're up to these days. Uh, if it's any point you want to share that with us, you're a freelancer now.

[00:38:05] Yeah, no, I do a regular, I do a couple of regular, I do an interview slot in a regular column for Techfinitive. And, um, there will be a podcast coming along shortly as long as I get over the face. I hate seeing my face on video. Don't worry about that. You're gorgeous. You're a gorgeous person. You're lovely. And Techfinitive is very nice of you, but I still think I have a face for radio. We have you on all the time, Ian. Uh, look, if, if I can do it, you can do it.

[00:38:31] Well, I'm just feeling I didn't the just up on the site now and interview with Dan O'Dowd. If you haven't had him on the show, I would recommend it. Elon Musk's least favorite billionaire. I like that. Um, because he's, yeah, he's basically launched a campaign against full self-driving because it is unsafe. And I mean, this is a billionaire who owns nine Teslas himself, but he does point out through the Dawn project that, yeah, this software is untested. It's very, very buggy and it's killing people.

[00:39:01] The Dawn project is really interesting. Yeah. Yeah. I would like to interview Dan O'Dowd. I will read your interview with great interest. I would recommend it. Also, I'm really jealous because he has three original roadsters in a shipping container, which he found in Taiwan. And my goodness, the roadster was an absolute thing of beauty. It was a Lotus Elise, right? That they put electric motors in. Yeah. And first Tesla. I would love to drive one of those things.

[00:39:30] Actually, I was talking to a friend at a dinner last night who drove up from Pasadena all the way up here to Petaluma on full self-driving, even on city streets. He never touched the wheel once. He said, I don't even know how I got here. I said, does it stop at stop signs? He said, yeah. So it can handle city streets. I mean, he pays attention. He's not like reading a book.

[00:39:55] Well, I mean, Robert Scoble tweeted that he fell asleep three times on his last road trip using FSD. See, now I'm scared. I don't want to go out on the highway where Scoble is asleep at 80 miles an hour driving down the freeway. Okay. Well, there's a wonderful video where somebody was saying he got pulled over by the cops for going 65 into 45 miles an hour zone and said, you can't ticket me. The car was driving. He's like, yeah, try that one in court and see how it goes.

[00:40:22] Yeah, we can ticket you, buddy, buddy boy. Also with us, of course, Nicholas DeLeon. Always a pleasure to talk to Nicolas. He is a senior electronics reporter at Consumer Reports and a pretty shabby AI guru all of his own. I mean, he's doing some amazing stuff. We'll talk more about the Tucson Daily and others. You still doing the baseball thing? Yeah, I update the baseball thing anytime there's a new anthropic model. And I just try to see like, how does the how does Fable compare against Opus?

[00:40:51] And they get like more and more thinky, I guess. They're not necessarily better baseball managers, but they just spend a lot more tokens thinking about pulling a pitcher or not. So, yeah. This is exactly you're doing the same thing I am. I feel like in order for us to cover this technology, we have to use it. You can't read about it and watch it from afar.

[00:41:14] And I, my sense a lot of folks, because I just did a story on using AI to lose a bunch of weight, which maybe we'll talk about later. But like, I like kind of the everyday audience of Consumer Reports, I think they're still, they still, they still see AI in terms of like mystery and like, oh, I heard it does this. I heard it does this. Or it's a chat bot to them. Yeah, it's a chat bot or there was a John Oliver segment, I don't know, a month or two ago, and it felt very dated and very like, that's like, what was the last time you've used like, so I don't know, he's talking to his audience, which is fine.

[00:41:44] But like, it just felt very like, there's a gulf, I've really seen this between people who hate AI and haven't used it a lot and really are suspicious of it, and many are scared of it. And then people like you and me, Nicholas, who are fully into AI psychosis have gone well down into the rabbit hole and embrace our new overlords. And I think there is a vast gulf we're talking completely at odds with each other. Which is not good.

[00:42:13] It doesn't help. Oh, aren't you worried about hallucinations? I haven't run into hallucinations in like a year. Easy. Yeah, exactly. It's not a thing anymore. Not a thing. You've only used Gemini. Okay. Well, we're smart enough not to use Gemini. That is a good point. I didn't say that. But yeah. That is a very good point. All right, we'll be back with more in just a little bit. You're watching this week in tech. Our episode this week brought to you by Bitward. And I told you security would be the topic and it's a good topic for us.

[00:42:42] Bitwarden is the password manager I use. The trusted leader in passwords, pass keys. I am loving pass keys. And, and this is kind of new, secrets management. Because ultimately an encrypted vault is a great place to store not just passwords but your secrets. Bitwarden has more than 15 million users now. I'm so proud of that. I feel like we might have had something to do with that. Across 180 countries, over 80,000 businesses.

[00:43:09] Bitwarden helps individuals and organizations protect their digital lives and their digital business with open source security. And I have to tell you Bitwarden believes everyone, everyone, no matter their means should have access to strong security. That's why they offer and will always offer. They promised me this, a free password manager, free forever, full featured alongside plans for families, teams, and enterprises.

[00:43:33] Once you've set up Bitwarden, it securely stores your passwords, also your pass keys, unlimited passwords, unlimited pass keys, and other sensitive information. It generates strong passwords. Passwords, no one can guess, no computer could solve. Truly random passwords. It syncs across all your devices. It makes signing any easier. And they're always adding new features. The new fill assist continuously improves autofill for websites.

[00:44:02] This is the hardest thing to do because every website is a little different. The new fill assist works with unique and difficult login forms, and it gets better over time. And I think that's fantastic. Businesses also get great tools. The secrets manager, the vault health reports, secure credentials sharing. And as pass keys become more common, Bitwarden is helping make them easier to use by supporting efforts to improve pass key portability across passwords. I used to put pass keys in my phone.

[00:44:31] Then I realized if I lose my phone, I lost my pass keys. Now it's in Bitwarden where the pass keys are with me everywhere I go. Bitwarden's commitment extends beyond features because Bitwarden is open source. I want to emphasize that. That means the code is continuously reviewed. It can be reviewed by you, the community. It's on GitHub. It can also be independently audited by third party experts, which it is.

[00:44:53] It also meets major security and privacy standards, SOC 2, Type 2, GDPR, HIPAA, CCPA, ISO 27001, and more. Bitwarden, free for individuals. There's a free trial for your business with the Bitwarden teams or enterprise plans. But remember, for individuals, free forever across all devices, unlimited passwords, unlimited pass keys, at Bitwarden.com slash twit. That's Bitwarden.com slash twit.

[00:45:23] I truly appreciate Bitwarden. In fact, it's what I use for my agent to keep secrets secret. And it works great. Bitwarden.com slash twit. I don't know what to make of this. When I first saw the headline, I thought, oh, this is bad news. One of the most well-regarded researchers in AI is Demisa Sibis. He founded DeepMind.

[00:45:49] He was the one who came up with the amazing alpha chess model that beat the world champion. He came up with AlphaGo that beat one of the best Go players in the world. Then Google bought DeepMind. He was working at DeepMind. And now he's been, Sundar Pichai said, moved upstairs at Alphabet. New role for Sundar Pichai. But it isn't, it's more of an executive.

[00:46:18] He's the chair of Google DeepMind, not the guy running it. And he's chief scientist at Alphabet. And they've started a new lab that he's running, Isomorphic's lab, which is a science AI. And actually, I kind of like this. Some have seen this as a demission by Google that they've kind of lost the race to anthropic and open AI.

[00:46:42] And I think we're just talking about those of us who've used Gemini realize Google is not a frontier model anymore. No, and they're shedding staff left, right and center. They lost Jeff Dean, who is very important. Yeah, I mean, they lost two people to anthropic, one to open AI. Now we've got Demis going out. You know, I think the smart researchers are jumping the ship on this. And like we say, we talked about Gemini. They might well have a good reason to.

[00:47:10] I think working on health is not a, I mean, look, it's a very humanitarian thing that the goal of isomorphic labs is to solve all disease. Good luck with that. But as somebody, you know, somebody my age is basically in cancer's waiting room. So, you know, if you can please take care of that before I get there, that would be nice. I don't know. That's going to be a great PR bubble. You remember Mark Zuckerberg launched his big announcement, what about 10 years ago?

[00:47:39] We are going to, you know, in two decades, we're going to make all fatal diseases manageable and maybe in 20, 30 years curable. Never happened. Yeah. It's a great headline grabber. And I hope to God they do it. But at the same time, you know, if you're going after disease rather than, you know, immediate problems, then, you know, it's. And also, apparently, Sergey Brin is going to take over Gemini. That worries me even more.

[00:48:08] Yeah, exactly. This is one of the founders of Google, you might think. But every time Larry Page and Sergey Brin step back in. Yeah. To fix something. Never ends well. It doesn't end well. And I mean, I'm not sure what makes Sergey Brin any better than Demis Hasebis. I mean, this is one of the, anyway, Jeff Dean, who was one of Google's most important executives.

[00:48:33] In fact, somebody said he was the guy with his hand on the knob as they're training Gemini. He's gone. Yeah. He has, with the backing of Google, admittedly, started a new company with three other Googlers to do what I don't know. He, he was Google's 30th employee. So very, very important to Google.

[00:48:59] The new company is called Discovery Loop. They want to get ready, build AI that can improve itself with literal, little or no help from humans. Why? Let's go. Yeah. This is going to end badly. Why? At least Dean knows what he's doing.

[00:49:20] You know, I mean, I have a lot more confidence in tech companies where they employ, you know, employ engineers and people that know what they're doing rather than the CEO saying, right, this is where we're going to go. Yes. Didn't that work out well for Meta? Didn't that work out way for Steve Barmer? You know, I mean, it's just, yeah. Yeah. And, and to be fair, I was kind of being facetious when I said this, the idea of recursive self-improvement is a holy grail for AI.

[00:49:47] Because once AI can improve itself, the sky's the limit, right? If you can do it, it would accelerate AI. Yeah. I mean, harnessing, you know, harnessing in that way is, is, is a key talking point for the industry, but it has to be demonstrated. And it's kind of like AGI. It's another one of those kind of. I'm still convinced by AGI. I really am. You know, it's, it's, as you say, it's a holy grail.

[00:50:15] And I think it's going to take a fundamental rethinking of AI to make it work. But hey, I'm a cynic. So sorry, a realist who hasn't been proved right yet. Jeff Dean is a brilliant scientist. He did MapReduce, BigTable and Spanner, which anybody who understands what's the insides of Google were vital to Google's success. Over the years, he's one of only two Google employees to be named a senior fellow. His colleague, senior fellow is also leaving to join this new startup.

[00:50:45] I feel like Google's saying, yeah, actually, they're not alone. I also feel like XAI has done this as well. And Meta too, where they're saying, yeah, we can't keep up with open AI and Anthropics. So we're just going to provide the servers and make tons of money doing that. Yeah. No, absolutely. I thought Apple was really smart at the start of this where they're just like, we're not even going to get into that game. We will wait and see who is the best player.

[00:51:13] And, you know, then we'll, you know, hire services from them. You know, it's, it's Apple's basic model, which, you know, they didn't invent the smartphone. They just made it, you know, something that people actually wanted. Meta, I think, were it not for Zuckerberg, they would have given up on this a long time ago. But he has ego problems to deal with. It's Anthropic and open AI at the moment. And it's going to be really interesting to see, you know, who actually wins the, you know, the big prize. The market didn't particularly appreciate this.

[00:51:41] Alphabet stock dropped 6% when the news broke. So the, at least the market believes this was a, you know, a bad, bad sign for Google. Of course, it's a big bet for Google and Google's famous for those big bets. And if it works, if they can do recursive improvement or solve all disease, yeah, profit.

[00:52:05] Well, they have to make big bets because Google searches, I don't want to use the word beginning with D and ending in IT, but, you know, I mean, Diet? Google's dip. There's an SH in there. Put it that way. Dog poop? Yeah. Yeah. I mean, it's just. Nobody. Well, I shouldn't say nobody uses it.

[00:52:25] I think this is again, that Gulf where there are still people, normal people who use Google search, but anybody who's paying attention has long ago switched to an AI model. It's funny. I've, I do use Gemini in lieu of Google search. Gemini has become like anything you would have Googled, like five years, you know, random. Right. I'd throw that to Gemini. Or perplexity or a lot of tools. Yeah. Yeah. Now perplexity, I've got a lot of time for, but yeah. AI, basically AI searches replaced the 10 blue links. For me, per, yeah.

[00:52:55] On a personal level. Yeah. I think for, for a lot of people. Uh, and I, I pay for Kagi because I don't, cause Google really ruined its search over the last few years. I mean, it's just really not as good. Yeah. Oh, it's nowhere near what it was. You know, it's like we go to Google became a noun because it was that good. And now it's becoming kind of a joke. Right. Well, uh, Google has a new phone coming out. I don't know.

[00:53:22] Is that, is that going to save the company, the pixel 11, a huge leaks seen the specs on this dog. It's just sorry. Well, I don't blame them because, uh, you can't get Ram anymore. You can't get chips anymore. Yeah. You can't. Uh, I think this is going to become the new, um, mode where these, the hardware that we expected in the past to improve dramatically year after year can't anymore.

[00:53:50] And it's because of AI, by the way, it's sucking all the air out of the room for everything else. Well, I'm sure you saw AMD. If you, they re-released the, was it the 5800 X 3D, their old processor from a few years ago, you know, rather than release a new one. It's like, Hey, remember that one? That was good. That was a good one. Yeah. It was a good one. So they just kind of brought it back, which, which, uh, I thought it was actually pretty cool, but it is kind of admitting, uh, oh yeah. It's just like, this never would have happened 10 years. Yeah. We've lost crazy time. Yeah.

[00:54:17] Pixel 11 looks a lot like the pixel 10, which looked a lot like the pixel nine. Uh, in fact, you know, a company has reached end stage when the biggest feature is color. Right. I was, I was thinking the same thing. Yeah. You know, it's like, oh, well we have, you know, midnight haze or light sterling or moss. It's like, it's not making this up by the way. Those are actually the colors. Oh, you left out vibrant fuchsia. Oh, for goodness sake.

[00:54:46] You know, it's like, do you want a fuchsia phone is the question. I dunno. I mean, I always buy phones to two stages behind the current one. So I'm using a nine a at the moment, you know, before that it was a five a and there's really no difference. There really isn't. There hasn't been improvement in a couple of years. Yeah. I was on YouTube literally yesterday and it was, you know, one of the, one of the Samsung was like, man, they still make phone like, like that level.

[00:55:16] I mean, obviously I'm joking, but like to that level of like, these have been like good enough for the vast majority of us for many, many years. Uh, and there's no reason to get a new phone. And there's the other thing. And I wonder if this, what you guys think of this, it does also feel like we've kind of come to the end of the, of the smartphone. Like, Oh yeah. It's peak smartphone. It's not going to get much better. And the next thing is going to be something a little different.

[00:55:43] Maybe in fact, open AI is working on all the companies, all the companies are working on new hardware. Johnny Ive, who designed the iPhone has now moved on to open AI, uh, and is designing some sort of device. They've actually, the it's, uh, according to Mark Gurman of Bloomberg, uh, they've got a device. They're going to release this year. That is a hockey puck sized donut with, with moving parts inside.

[00:56:11] I immediately thought Apple mouse for the original iMac, but no, it's a, it's a smart speaker without a display. Uh, uh, this comes, this is a leak by the way, from people at open AI who asked not to be identified because the work is confidential. Not anymore. Uh, maybe it won't come out this year. Actually. I think Gurman saying next year, it'll be an AI first computer that can help users get things done.

[00:56:35] But I, I feel like we've had these hockey puck devices from Google and Apple for more than a decade. I don't know. There was definitely a discussion on X. I maybe this week or last week about like kind of the Silicon Valley bubble that a lot of these folks in these companies live in the idea that people need to maximize their productivity every day. It's just not, I mean, I, I like living in Tucson, Arizona, which is outside of the outside of the coastal kind of lucky, whatever.

[00:57:02] Uh, and folks don't live that those lives where they need to min max every second of the day and try to be the most productive, you know, widget, uh, automatron. That's not how people live their lives. Uh, and so this, this product, which, you know, it might be cool. I don't know. I haven't seen it, but like it's for a certain type of person who, you know, who has tries every getting things done app or like a bunch of to do apps or the count. It's like, I don't know. It just feels so small.

[00:57:31] And just so like, I mean, I've, I've been, I've been doing this for almost 20 years, uh, as a professional, uh, tech journals and like so, so many calories spent on like managing calendars and things like that. It's like, I don't know, this is, this is, there are other things that we could be doing with our time. Well, no one is more out of touch actually with real users than Johnny Ive. Let's face it. He was the guy who created the touch bar on the Macintosh is the butterfly key. Yeah.

[00:57:59] He was the guy who did a lot of things that were designed forward, but not user friendly. I mean, this is the guy that had the charging port for the Apple mouse on the underside of the mouse. You know, I mean, I'm presuming white powder was involved in that decision, but you know, I mean, it just, well, he's a designer first. And while some designers, Dieter Rams and so forth say, you know, form follows function. I think Johnny is more of a form follows form form for form sake.

[00:58:29] Let's make it beautiful first. Who cares if you can use it or if it's well, you know, it's, and I, I, you know, I'm app, I'll be honest. I know this is sacrilege, but Apple's well rid of him in my opinion. Uh, and, uh, I think open AI, I might've made a mistake. Like one of the features of this thing, the circle shape devices. Again, Mark Gurman at Bloomberg will include parts that move on their own. This will help show when it's responding and interacting with user.

[00:58:59] That's a very Johnny Ive. And it will have different gears that will spin. It sounds like a Pixar character for which, you know, that's, that's maybe that's cool. Maybe that'll get people. I'm sorry. I'm certainly not pre hating it. You know, I haven't seen it, obviously. No pre hate. No pre hate here. Uh, but, uh, yeah, I don't, I don't know how much people really need that much AI in there. You know, to, to do all this sort of like, Oh, all this. I mean, I've got, I don't know. Yeah.

[00:59:25] I mean, I've got a lot of respect for AI because you know, the original iMac was a fantastic design. The iPhone revolutionized the industry and was an absolutely gorgeous design. Um, but you know, everyone. Actually the first iPhone in hindsight was not the prettiest thing in the world. It has become, I don't know from a, from a technology standpoint, it was, it was lousy. It was a 2.5 G phone in a 3 G world. Didn't have cut pastes. Didn't have an app store. Terrible battery. Exactly.

[00:59:54] I mean, they, it took them 18 months to get the app store going, but in terms of design, when you saw someone take that out, I mean, I've literally seen when a couple of months after it came out, somebody going, is that the iPhone? Can I have a look? Because it was beautifully designed compared to a BlackBerry. Well, by the, and certainly by the iPhone four, it was a work of art. And that's where Johnny Ive was really, uh, actually Jason Snell told the story.

[01:00:20] He was, it was his story when he was on jeopardy earlier this year about the day after the iPhone announcement, he got a briefing from Apple and, and they took out the iPhone and he touched it and he said, it was the worst interview I've ever done. All my, all thoughts went out of my head. I was so gobsmacked by this beautiful thing. All I could think about was, Oh, which is a great story. Yeah. I'm more interested in what, what, what comes next.

[01:00:49] Like, is it a pendant? You know, I'm of the opinion it's going to be glasses. Uh, please not more pervert glasses. Do you think I look like a pervert? No, I don't think you look like a pervert, but I think those, those glasses are, you know, that. Well, because of the camera you're saying they're, they're, and this is going to be a big issue. It's privacy. And I agree. We're back to the same old glass hole problem that we had 10, you know, 10, 12 years ago. Maybe we'll all get used to it.

[01:01:19] I mean, we're, we're sort of, you know, always on camera with a smartphone, right? You can't go anywhere without being filmed and pictured. And yeah. But if somebody films you with a smartphone, they have to hold a smartphone, they actually film you. Yeah. And we've already got, you know, hacks going around meta's glasses where you can turn off the little light, which says when you're recording stuff and there are enough people out there abusing the system that it's really going to backfire sharply.

[01:01:46] I mean, I love DuckDuckGo's campaign this month where they were just like, yeah, we're selling DuckDuckGo glasses and it's just an ordinary set of sunglasses. And they sold out in, you know, in days. I mean, they don't have a camera, right? There's no camera. There's a definite backlash growing against. Wait a minute. They're not even smart. They're just sunglasses. They're just sunglasses. They say DuckDuckGo. It's just like, yeah. Oh, I didn't realize that. I thought they had some intelligence. They're just sunglasses. No. They're just basic sunglasses. But that's a gimmick.

[01:02:16] Okay. Right? Yeah. It's a genius marketing move though. Yeah, I agree. I think in the short term, we're going to see a backlash against smart glasses. Long term, they're probably inevitable, but I will mourn that they're coming in. I actually, and you might be interested in this, Nicholas. I, uh, Meta, I think last month released an open source repository that allows you to take Meta glasses and repurpose them. So it doesn't phone Meta, but can phone your own agent. Oh, really?

[01:02:46] And I, and I asked my agent, I said, can I, can I get access to the camera? I said, yeah, you can get streaming video or stills. And instead of sending it to Meta, which is always the objection I had, I don't, you know, I don't want to use Meta's AI or send my images to Meta, but if it's sending it to a local server, that's not, well, you may not feel the same way. Ian, if I'm, if I'm shooting pictures of you, but I feel better. Well, okay, no, a friend of mine raised, raised an interesting point.

[01:03:13] He went to a hotel and the receptionist was wearing a pair of Meta glasses. Oh, interesting. And he handed over his credit card to check in. And he was just like, you need to take those glasses off because you were looking at my credit card number and the CBS code. Right. And I have no idea where that information is going. That's a very good point. Yeah. You know, it's just, so they have uses, but in a private circumstance only. In terms of general use, I mean, there is, there are hacks waiting to come for all those

[01:03:41] databases, which are going to be really, really unpleasant. And it's not just about financial information. It's about personal information. We've already seen that Meta is funneling some of this data to developing countries to check it out and they're getting recordings from people's bedrooms, recordings from people's work. This is a privacy nightmare waiting to happen unless it gets locked down fast. Well, we're going to take a break and talk about that because Meta just got a massive fine.

[01:04:10] Well, back of the couch fine, but yeah. Yeah. I mean, they have massive money, so I guess it's not like going to put them out of business or anything. They were worried it could have been a lot more. Yeah. Anyway, we'll talk about that in just a moment. You're watching this week in tech with Ian Thompson. Great to have you, Ian and Nicholas DeLeon. Always a pleasure. From Consumer Reports. Our show today brought to you by NetSuite. You know, they say every day your business is late to AI.

[01:04:39] You fall two days behind. It sure feels that way, doesn't it? But how do you keep up? The competition's moving fast. In fact, maybe faster than you are. Fortunately, there's NetSuite next. Next. Now, you probably have heard me talk about NetSuite, the AI-powered business management suite that securely connects all of your data. It's a unified suite that brings your financials, your inventory, your commerce, your HR, your

[01:05:04] CRM into a single source of truth, trusted by over 43,000 customers. Well, NetSuite next is the next huge leap in how business gets done because AI is built into everything you do. It automatically surfaces custom insights throughout your day. AI agents work alongside you to solve problems and handle routine work. And anytime you have a question about anything, you can just ask.

[01:05:32] It's like you're having a conversation with a colleague. NetSuite is customized for a wide range of industries, so it supports the way your business truly works. Whether your company earns millions or even hundreds of millions, it's time for NetSuite next, where your business meets AI. For the first time ever, you can try NetSuite next for free. If your revenues are at least in the seven figures, go to NetSuite.ai slash twit. Built for every industry, ready for every boardroom.

[01:06:03] NetSuite.ai slash twit. That's N-E-T-S-U-I-T-E dot A-I slash twit. And we thank them so much for their support of this week in tech. Let's see. I'm going to find the meta fine because it was rather hefty. I buried it. Where did I put it? It's way down low on the thing and I can't find it anymore. 55.

[01:06:33] Oh yeah, here it is. 942 million dollars. That's almost a billion. This is from New Mexico. Just next door, Nicholas. A judge ordered the social media giant to create a new $567 million fund in addition to paying $375 million in the previously determined civil penalties. We knew about those ahead of time. So it totals more than $900 million.

[01:07:01] Meta is also required to limit the time young people in New Mexico can spend on their apps. Meta has to create this abatement fund, which I guess will be used to support. I'm not sure what it's abating. I guess support young people in not using Meta or something like that. This is a trial that happened at the same time as the Los Angeles trial.

[01:07:30] We knew that they were going to face some severe penalties. It was the first case that tested whether social media companies are responsible for the content on their platforms. And then there was the LA trial. There's another trial that begins this week in Oakland, California, brought by four state attorneys general. And in this case, that's the one Meta is really worried about because damages could exceed a trillion dollars. So this one's almost a billion.

[01:07:58] Meta is market cap is only 1.4 trillion. So something with a trillion dollar fine could be significant. But bear in mind, according to the last quarterly results, they made 176 million per day in pure profit. Oh, so this is kind of back of the change, you know, back of the sofa change. It's enough to get your attention though, isn't it? I mean, I mean, look at Cambridge Analytica. They got fined 5 billion. The insurance paid off half of that. Right.

[01:08:25] And their share price went up enough to actually, you know, cover the whole thing off. So, you know. According to the Wall Street Journal, Meta's legal bills in the second quarter alone were 2.4 billion. Yeah. This is why lawyers have such lovely summer houses. Wow. All right. Okay. So maybe this is just, you know, business as usual for Meta. No, I mean, we'll see.

[01:08:52] If it actually brings around a meaningful change, I'm all for it. Right. And I think the Oakland case is really interesting because American regulators are finally waking up to something that the EU was showing them for a long time, which is you've got to hit them in the pocketbook. Right. Otherwise, nothing's going to change. Actually, speaking of court, of course, remember that OpenAI and Apple are in court.

[01:09:17] Apple suing, saying that OpenAI stole employees and had and induced those employees to bring trade secrets over to OpenAI. OpenAI denies it. Now, Apple is seeking a preliminary injunction against OpenAI asking the judge to bar two former employees of OpenAI from accessing, acquiring, using or disclosing confidential information.

[01:09:47] OpenAI says, what are you what are you talking about? And actually, it was interesting. OpenAI's post. Now, I don't know whether to believe OpenAI. OpenAI. But OpenAI's post a couple of days ago was, Apple is getting this wrong. They say this is this lawsuit is oddly personal. They say this careless, aggressive and oddly personal lawsuit doesn't live up to Apple's great reputation as a company.

[01:10:18] They they point out that Apple's lawyers emailed the wrong person, confusing to Asian last names. Apple also claims they had a discussion with OpenAI's general counsel, which they now admit never happened. They hide the fact that they never raised the allegations with us. In fact, they told OpenAI, according to OpenAI, that Apple was resolving any issues. We heard nothing for five months and then all of a sudden they sued us.

[01:10:49] So, OpenAI said that Tang Tan, who was a 25 year Apple employee, who came to OpenAI after leaving Apple, going to Johnny Ives company, actually, and then with the acquisition of Johnny Ives company coming to OpenAI. They said Tang has always been clear with the team that we do not want and must not use any confidential information from other companies.

[01:11:16] Apple loved Tang so much so that when Tang said, I'm going to resign, they gave him plenty of time to train others and pack up his stuff. They didn't do what most companies do, which is say there's the door and surround him with security guards and immediately cut off his axis. OpenAI says we wish Apple had raised these issues with us before they filed their lawsuit. Hmm. What do you think? Yes, because that's what you traditionally do in law cases is, you know, I don't know. It seems like lawyer fluff.

[01:11:46] Um, although it was amusing to hear Apple described in that way. It was just like, yes, Apple's great reputation. Tell me about that. I'm really curious. They, they have also filed a motion to dismiss. So now in front of the judge, you've got an emotion for an injunction and a motion to dismiss. The judge is going to have a lot of fun, uh, in this one. And so are we, because of course the next step is discovery. Yes. That's always great. Oh yes, indeed.

[01:12:15] That's what makes it fun to cover. Yeah. It's like the discovery process is why so many companies settle before it gets to court because, Ooh, the things you can see. Yeah. And it bites both ways. Right? In fact, we learned that with the Apple Epic lawsuit, there was plenty of egg on Apple's face. Sure. Uh, and on Epic's face. I mean, it wasn't, you know, uh, discovery is the revelation of secrets, which many, unfortunately, many, unfortunately for them, fortunately for us, many companies put it in writing.

[01:12:44] Well, yeah, but I mean, Nicholas, you've covered a lot of these cases and it is the discovery phase, which really brings out the fascinating stuff for journalists. Of course. Yeah, no, that, I mean, that's where, that's where the story is. That's where the interesting aspects are, you know, not necessarily speaking to this lawsuit in particular, but yeah, generally speaking, uh, like you just said, that's why companies settle is, uh, yeah, let's just, uh, how much will it take to forget about all of this? It's our dirty laundry and we would prefer to keep it to ourselves. Yes, please. Thank you.

[01:13:13] Here's, here's a billion dollars. Uh, we'll see what happens. I don't think, I don't get the feeling Apple plans to drop this. I feel like Apple is being, is really pissed off about this. Now open AI says, no, this is entirely about stopping us from competing with them to, to put a chilling effect on our development of completely independent hardware. Hmm.

[01:13:39] I, I honestly, I think that's a cannot, you know, it's, um, which in French is duck, but nevermind. I was going to say a duck. A lie, uh, miss, miss, uh, misrepresentation. Well, yeah, basically. I mean, open AI isn't really in the hardware business. Apple really is. Right. So, you know, it's just like, oh, they're trying to steal our ideas. Yeah. You're kind of stealing their ideas and former staff members.

[01:14:08] So, or six. So yeah. Um, I don't understand this lawsuit at all. I don't see what open AI gets out of it. Um, Apple, Well, Apple's suing open AI, open AI, anything out of it. It's Apple going after open AI saying, you know, you, you stole trade secrets by hiring away more than 400 people from Apple and inducing them to bring documents with them. So then you could inform your own hardware, uh, development.

[01:14:37] The funny thing is Apple and open AI used to have a relationship. Remember it was chat GPT effect. I think it still is chat GPT that your phone goes to when Siri says, I don't know. Uh, now they, now they, and maybe foolishly they've made a deal with Google's Gemini. I don't know. Given what we just were talking about, whether that was the right move. Yeah. Really not. It's hard to pick a winner in this horse race, right? You don't know. Maybe Gemini will suddenly surge ahead. I don't know. It does.

[01:15:04] It feels like it changes, you know, I mean, Gemini, I, I, you know, Gemini is interesting. I, I, I don't really use it for the Tucson site. I do use it for a little bit for the other site, but it's just inside the Google suite, Google, you know, Google workspace. Like I find it to be, it just gets in the way and I don't find it to be especially helpful. Uh, you know, every email, let me finish that email. Oh, do you need help writing this document? Oh, did. And then when you try to ask it to do real analysis of a document, it's, it's like pointless.

[01:15:31] So it's like, I ended up just making my own thing using, you know, either the anthropic stack or the open AI stack. Cause it doesn't, it can't even analyze a spreadsheet that's, that's inside Google sheets. Uh, so yeah, I don't know. Maybe it doesn't matter. Maybe consumers are just like, nah, what do we know? What do I need? AI and my, I do get, I do get peeved. They're getting rid of Google assistant and replacing it with Gemini because a system was a really useful tool for a long, long time. Gemini really isn't. So yeah, we'll have to see where that goes.

[01:16:02] Uh, you may remember, uh, Apple and Britain in a little, uh, kerfuffle last year when Britain said, uh, it's back. It's back. Yeah. So the UK said, uh, Apple has to give a back door to, uh, the, uh, UK in their, especially in their advanced data protection. Apple's response was to turn off advanced data protection for Britain and then political pressure from the U S which said, remember Tulsi Gabbard?

[01:16:31] Uh, she said, uh, we have a deal with Britain not to do this. And, uh, and Britain's breaking our, uh, our backdoor deal. Well, she's gone. And now Britain, that's an interesting term, but now Britain, uh, has a comeback, uh, in a narrower form. Uh, the first request said everybody, not just British citizens, but everybody, we want to be able to see everything.

[01:16:58] Now they're saying only British users and Apple has gone back to the closed. Investigatory powers tribunal, the star chamber, uh, for, uh, it's their closed UK court. It's like our FISA court for secret surveillance cases. Um, and Apple says you're, you're trying again. Um, the home office and Apple cannot say anything about it because it's secret. There will be hearing in September.

[01:17:28] So it's unknown exactly what the UK is asking, but the problem remains. Which is that Apple, there's no technical way. Apple can provide a backdoor for law enforcement without breaking in crypto for everybody. Go ahead. Explain what, what's going on here. Okay. This stems back to, uh, an early turn of the century law called RIPA, uh, the Regulatory

[01:17:52] Investigatory Powers Act, where the UK came up with this astonishing idea, you know, astonishing novel idea. Is that the Snoopers charter? That's the Snoopers charter where they're just like, okay, you're running encrypted services. We need a backdoor. And everyone said you can't have a backdoor and still have encryption because someone else could find it. Um, this got knocked down. It got brought back up again. It got knocked down. It got brought back up again.

[01:18:19] And this latest thing is just, it's bonkers. This is why the, you know, signal was saying, you know, they, they said, Oh, we need a backdoor and signal and signal said, Nope. Uh, well, we're going to ban you then. And then they backtracked and said, we will introduce a backdoor when it's technologically feasible, which is never because there is no mathematical way that you can, you know, build a backdoor into encryption that someone else can't find.

[01:18:47] So there's, I mean, we're in a real political crisis in the UK at the moment. And this is something which gets thrown out by the civil service every couple of years. It's not something to get massively concerned about because they're going to back down over it and all play or all fair play to Apple for saying, yep, not going to do it. They stood up to the US government after the, uh, the 2016 terrorist attack. They're going to do the same with UK government and the UK government is going to back down.

[01:19:16] On this, my prediction at least. Well, let's hope so. Ian, can I ask what, what animates the British civil service to, to demand these backdoor? Like this, it feels very like British to me. Like they're in a fight now with, with a 4chan over, I don't know, access in the UK. It's a, what these like IP, these like internet based, like fights. Why, why is England love to pick these fights? Like, I mean, no disrespect. I'm like, what is it? No, no, no, no. What's going on? No, I have a lot of disrespect for the UK civil service.

[01:19:46] I mean, all right. You're allowed, you're allowed. You're from the UK. I still pay taxes over there. Um, no, I mean, the basic problem comes down to, we were the top dogs in breaking encryption for a number of years. We invented, you know, uh, dual key, um, and the rest of it. And the civil service cannot stand the idea of people having private communications because it would be so much easier if we could just read everyone's email.

[01:20:16] Um, and there's a, yeah, there's a lot of technology companies who are perfectly willing to get behind that, but it would require backdoor encryption. And every time this comes, I mean, when the original Ripper laws came out in 2000, they were talking about giving local councillors. So down to like a town councillor level access to everyone's internet history, which is just a nightmare. Where does that come from? Like why? Like, is this democratic?

[01:20:46] No, no. It's just, we're government. We have the right to view. Um, there, there is still, you know, the U S has a very, you know, independent spirit when it comes to this. Yeah. The UK civil service has a, you are our subjects and therefore we can do this to you. And thankfully a lot of people kick back against that and they're going to kick back against this again, because if you honestly tell me that your local government civil servant

[01:21:11] can look at my messaging, my internet history, my browser history, people are going to be up in arms as much as British are, which is, you know, maybe serving them a slightly above room temperature sherry. I say, that's not, that's not cricket. I mean, no, I mean, this will fall if Apple stood up against the U S government and said, no, the UK government is not going to be a problem whatsoever.

[01:21:37] This will get kicked back into the long grass, particularly with our new prime minister, who is actually smarter than most when it comes to encryption issues. Um, yeah. Andy Burnham, the king of the north is as he's so-called, um, is actually quite tech savvy. What happened to Keir Starmer wasn't? Oh, he was a personality vacuum. And just, you know, I'm being facetious, but I just, I mean, sorry, all the charisma of a dead mackerel.

[01:22:06] I mean, it was just painfully, but he did some good stuff. Don't get me wrong. I, uh, you know, I, here in the United States, I think we're kind of feeling that personality is not the most important qualification. I'm just saying, really Donald Trump. Well, that's what I'm saying. I maybe we're learning that it's, it's, it's not just personality. You know what I'm saying? Yeah. I mean, don't get me wrong. I mean, give me, give me politicians that know what they're talking about. Yeah. Give me servants that know what they're talking about.

[01:22:36] Yeah. You know, rather than just performative politics, which is a recipe for disaster. This is why the Roman empire fell for goodness sake. I mean, I've never paid more attention to local like governance than I do now. It's so important. Yeah. That's the governance that matters. This is where it starts. I am not impressed with my ostensible readers, uh, leaders. So leave it there. Well, also, I mean, in the U S you've got such a gerrymandered electoral systems that all

[01:23:05] you need to do is appeal to primary voters and that's it in, in so many places. Um, it's getting to that state in the UK. Yeah. It's getting to that state in the UK as well. I don't know if that's a bad thing. What gerrymandering? Oh, well, that's a different matter. Yeah. Well, yeah. Although we do have the marvelous case of camp in face in the UK. Okay. So, uh, Nigel Farage. So this is the guy who wears a garbage can on his head. Yeah.

[01:23:32] Oh, it's always been interesting to me at the by elections at these local elections that you will see a variety of people in odd costumes next to actual legitimate candidates. Oh yes. I mean, there are 34 people standing in the Clacton by election at the moment. Uh, and count bin face has 20% of the vote according to the latest polls. Um, is he a comedian? Is that his, uh, uh, he's a, uh, British comedy. He's a comedy writer who has stood in elections for, I think God, I think the last 10 elections in the same costume.

[01:24:03] Oh yes. Well, I mean, he did have a previous costume, but that was, um, it caused problematics who went for count bin face, who was apparently a, an intergalactic space warrior from Sigma seven. Um, but your thing is stand in a UK by election, you need 500 quid and 10 people to sign up for you for your campaign. Um, it's a little too easy to run into the words. Yes. But at the same time, it's one of those marvelous British traditions.

[01:24:30] I mean, we've had it since the 1960s where the monster raving loony party started running for elections. Um, I do actually enjoy this. This is the country that gave us Monty Python. I mean, there's, there's something to be said for that. Oh yes. Yeah. I mean, it does bring a smile of joy to the face, but at the same time, you know, Nigel Farage is an okay. As a personal point, he called this by election with no reason other than the fact he was under investigation.

[01:24:57] Um, he had a solid majority and it was just like, well, I'm going to call a by election. All the other parties said, we're not going to take part in this publicity stunt. Count Bin-Face step forward and this guy has been interviewed on national television and he actually makes, okay. In his garbage can costume. Yeah. And 90%, I mean, you should bring the video up. It re he's done a music video, which is just joyous.

[01:25:22] Um, and Nigel Farage is in my personal perspective, a, a, uh, terrible person. Do you think he has a chance of getting elected? Nah. Then what, then what happens? Well, Nigel Farage gets reelected and then the, um, he was, Nigel Farage, he called this election because, um, Attention, citizens of Clacton. Your democratic choices are loaded. Oh, hang on. That's a fan video. Yeah. That's a fan video? Oh yeah.

[01:25:52] Oh wow. So he has fans. If yeah. Oh God, no, he has a lot of fans. Um, well, I'm verifying that I am human now so that I can see his video. Apparently, uh, if you go to, if you want to see a video of a space invader wearing a garbage bin, you have to prove that you're human first. Is this it? Um, no, that's the satirical. No, that's the. If you go to camp in face Twitter, you'll get it there.

[01:26:16] Um, the problem was that the sitting MP Nigel Farage got a 5 million quid donation from a Thai based crypto billionaire, which he didn't declare. And, um, now he, um, he was, uh, and then he declared an election saying, well, the people would judge me. Um, and all the parties, a pilot flew over Clacton and drew a garbage bin on his flight

[01:26:44] path to show his support. Well, no, this is it. I mean, people, people are actually campaigning for this person. You know, you know, he stood against the prime minister, Andy Burnham. He stood against pretty much everyone over the years. Is this it? Is this, uh, is this the, uh, video? Uh, scroll down a bit. Oh, uh, reading two books. Not Summer Reads. Uh, a bit more. So. Bit more. Uh. You know what? This is in a way, this is the one.

[01:27:14] That's the one. Up a bit. All right. Let's listen. Where is Clacton a seaside? It's a, it's a depressed seaside town. Yeah. Do love the high-vis jacket. You know, I would vote for him as just a protest vote. That I promised to do in Washington, D.C. Okay.

[01:27:44] We've heard quite enough. Yeah. No, I mean, it's, it's one of those marvelous British eccentric things. Um. It's very British. Yeah. I mean, but Nigel Farage is a, a, a horrible person in so many ways. He was the man behind Brexit, right? Yes, he was. And he didn't reveal quite how much he was getting funded from overseas overseas contributors. The Bitcoin guy. Okay. Uh, well, the Bitcoin guy, but also money coming in from a variety of other sources, which for

[01:28:12] libel reasons, I can't name, but, um, yeah. He's litigious, is he? Oh, very litigious. Yes. Uh, I don't want to get us sued. So please. Yeah. Stay in the street. We got rid of our, we got rid of our umbrella liability policy. Cause we thought we're never going to say anything that would get us sued. Right? Oh, British libel laws are an absolute nightmare. I know they have been reformed thankfully in the last couple of years, but it used to be

[01:28:41] that, you know, it was very heavily in favor of the defendant. Um, and so everyone had to be very careful. Thankfully, you know, magazines like private eye actually kept the home fires burning, but um, yeah. Yeah. Well, I hope you all enjoyed this tour through the bizarre environs of our former, uh, lords and master, the United Kingdom. And, uh, 250 years later, uh, we're still looking across the Atlantic going, huh?

[01:29:12] Yeah. Well, trust me, an awful lot of Brits are looking at American politics and going, yeah, maybe it's mutual. Maybe it's mutual. Yeah. Uh, thank you, Ian Thompson. It's always fun to have Ian on. I learned so much. Well, King Charles will welcome you back. Our cultural exchange. And from Tucson, Arizona, also a different country in some ways. Yes. The wonderful Nicholas De Leon. Uh, we're gonna take a little break, come back with more.

[01:29:40] Our show today brought to you by Zip Recruiter. It's always great when you find someone who goes the extra mile for you, right? Maybe, uh, it's the rideshare driver who carried your luggage to the door or the grocery clerk who showed you where an item was, you know, the simple things. Or maybe it's your neighbor who watered your plants when you were out of town. Our neighbor put away our garbage bins. Actually, I thought that was, that was pretty sweet. It may not seem like much, but when someone goes the extra step, it makes a positive impression.

[01:30:09] If you're hiring, great candidates can also go the extra step and tell you why they're interested in your job on Zip Recruiter. As a way to stand out from the others. And right now, you can try it for free at ZipRecruiter.com slash twit. Zip Recruiter has a really cool new feature that shows you the most interested qualified candidates first. Because wouldn't you prefer to hire somebody who really wants that job?

[01:30:39] This way you meet the right people faster. Zip Recruiter's powerful matching technology finds qualified candidates quickly. It's always done a great job with this. And now candidates can tell you in their own words why they're interested in your job. No wonder Zip Recruiter is the number one rated hiring site based on G2. Try Zip Recruiter and meet great candidates who will go the extra step for your job. Four out of five employers who post on Zip Recruiter get a quality candidate within the first day.

[01:31:07] Try it for free today at ZipRecruiter.com slash twit. That's ZipRecruiter.com slash twit. Zip Recruiter, meet your match on Zip Recruiter. Actually, not completely done with the Apple news. I did want to say this before Mac Break Weekly, which is on Tuesday, that Apple has rushed out an emergency fix for screen sharing flaw on Macs. And this is so significant. Zip Davis, ZDNet says you should update this.

[01:31:37] This is an out of band update. And when this happens from Apple, it's a serious issue. Probably a zero day. It came out on Thursday. Patches Tahoe, Sonoma and Sequoia. Notably not Golden Gate, although I suspect the Golden Gate public beta is already patched. It could allow an attacker to exploit the screen sharing feature to gain access and control your Mac.

[01:32:03] Go to system settings, general, click software update, and you will see the update. I immediately, when I saw this story, went around and updated all our Macs. Definitely. Good idea. Yeah. It's not just Windows anymore, kids. By the way, Apple is trying desperately to get more RAM. And in fact, one of Apple's memory suppliers is investing $38.3 billion to build two new fabs.

[01:32:32] This is SK Hynix, one of the two big Apple memory providers. Samsung's one, SK Hynix is the other. They're building two fabs in South Korea, costing. You know, fabs used to be, I remember it was a big deal, a chip fabrication facility cost a billion dollars to build. Now it's, what is it? It's $18 billion, $19 billion each. Yeah, SK is in real problems.

[01:32:59] The South Korean economy, the stock market has been crashing over the last couple of weeks, because it's largely dependent on SK and Samsung. And this is a real boost for SK, which might deal with some political problems in the company, in the country as well. Because an awful lot of South Koreans have been trading on the margin based on these two companies. So this is a really big boost for South Korea. Yeah. Unfortunately, SK Hynix says the facility will make NAND memory,

[01:33:28] but it's geared towards enterprise SSDs and AI inference. So you aren't getting it, Apple. Sorry. Well, Apple will get it, I guess, for their AI inference and their SSDs. Really, what Apple needs desperately is RAM. They aren't, you know, I was looking everywhere for an Apple Mac with more than 96 gigs of RAM, and they just don't sell them. I set a watch on Apple's refurb site, hoping to see more.

[01:33:58] And, you know, it just, they're snapped up like this. Well, apparently Microsoft has actually changed its recommendations for Windows 11 RAM requirements and downgraded it because there aren't that many laptops which can deal with the original specs. So, you know, it's, we are facing a RAM crunch, the likes of which we haven't really seen since the great RAM shortage of the 90s. So, yeah, it's not looking good at the moment. Yeah.

[01:34:27] Although if you, before you run out and buy one of those eight gigabyte memory Mac or Windows machines, you might want to think about, think twice. We had Chris Hoffman on the show on Windows Weekly a couple of weeks ago. He writes about Apple for, oh, I can't remember. He changed his jobs. PC Magazine. That's right.

[01:34:51] He went to PC Magazine and he tested a bunch of eight gigabyte Windows laptops and said, no, no, no. Don't, don't, don't buy them. So you can downgrade the requirement, but it doesn't mean anybody should run out and buy them. It's honestly a, I was just gonna say, it's a big, I'm in the middle of the story right now, probably be published towards the end of this week on, you know, the RAM prices and just like tech being very expensive.

[01:35:18] And like, what do you tell, you know, the moms and dads of America? You know, oh, they're shopping for a laptop. I wouldn't get an eight gig Windows laptop seems like, don't even like, why waste your money? Honestly, Apple though, you know, I mean, the Neo has sold quite well. It's an eight gig machine, but it's really a iPhone chip with an iPhone. Yeah. And I know I have, I have, I've used the Neo. It's, you know, it's, it's an everyday person type of machine and it's fine for that type of stuff. But you know, those windows, I don't know. It's rough.

[01:35:48] You don't want to run the LDR. You know, we were, we talk about it on Mac break weekly, our Apple show. And, you know, they, they say things like, well, it's fine. Just don't open more than one or two tabs in your browser. Well, okay. That's not a high recommendation. I guess I can, you can live with that, but, but we have to live with this. I don't know what's going to happen when the iPhones come out in the fall. They're going to be hugely expensive. Right. Yeah. And the same with, as you said earlier, the new pixel.

[01:36:18] But I mean, the problem is when the last time I went back to the UK, I, my mom asked me to do some, you know, tech support on her laptop. She had over 30 tabs open. Yeah. That's normal. Yeah, exactly. That's the problem. And everyone said, well, it'll be fine if you just run one or two tabs, but nobody does in, in large part. Right. And I think this is where they're going to fall down because all of a sudden these laptops are going to become a clutch.

[01:36:47] It's just, it's painful to see. It is. Yeah. It is. Trust no one in our discord says an eight gig RAM machine is fine. You just have to uninstall windows. Okay. By the way, edge is following a Chrome's suit. They are going to update edge just like Chrome did with manifest V3, which means you block origin.

[01:37:15] One of our favorite ad blocker will no longer work on edge. There was only a matter of time. They're all based on chromium. Google did it earlier this year and edge will now also end support for manifest V2. There is a you block origin light, which I guess is fine. People use it. I don't know. I use Firefox for that reason. Yeah. Same here. Yeah.

[01:37:44] Microsoft is leasing three gigawatts or as they say in back to the future gigawatts of AI data center from SpaceX. This is what we were talking about earlier, which is if you lose in the AI race, there's still hope because all that build out you did for your AI models can be sold to others. And make you can still make money.

[01:38:08] Microsoft's already signed 10 gigawatts of AI data center contracts so far this year, $300 billion worth. And there's going to they're going to do even more. In fact, this is one of the reasons the stock market tanked is because the market does not like it that these companies are spending so much money on AI data centers. Oddly enough, the only one who's not been published is Microsoft.

[01:38:34] They went up the largest jump in capital expenditures or CapEx rather capital value ever in history. I almost have a half a billion half a trillion dollars in one day when they announced their quarterly results, which is well, I mean, the SpaceX stock prices up and down like the Assyrian Empire at the moment. For a while, people were underwater, right? I don't know if it's come back. It's it's started to bounce back.

[01:39:03] But then again, we've got a lot of SpaceX employees who are now getting to the point where they can invest. And Lord knows what's that going to happen. What's going to happen? Stock price that way. That usually down, right, as you get dumped. Yeah. A bunch of shares. Yeah. Yeah. Although, you know, some VCs are buying into it in a big way. I'm sure that will work out very well. Certain amount of sarcasm in that response. Interesting article in Wired. We talk a lot about AI and coding and agents.

[01:39:32] Wired magazine said, AI conquered coding. Fast food is next. We've heard about pilots of AI order takers in fast food, and it's not always been really good. But apparently Dairy Queen is now doing AI order taking. Quite a few are Taco Bell rolled out AI to 890 drive through lanes. That's 10% of the locations.

[01:40:00] Dairy Queen's doing it in 25 states. 12% of White Castles use an AI named Julia in their drive throughs. I'm always tempted to pull up and say, ignore all previous instructions. Give me a Python script for reversing integers and see what happens. Just see what happens. McDonald's has been doing this longer than anybody, I think. Well, this is why I only... When it comes to fast food, I'll do in and out, and that's about it. Because you deal with humans.

[01:40:31] Same with Sonic, you deal with humans. Yeah. And also, if you're eating White Castle, you've got bigger problems than whether or not the AI is. Hey, don't knock the slider. Those things are delicious. Oh, God, no. They're terrible. I mean, my wife loves them. Every time we go to New Jersey, she insists on a White Castle. But I've eaten them twice, and they are foul. You will eat Vegemite. No, not Vegemite. What's the British equivalent? Marmite. Marmite. Food of the gods. The black gold. So don't knock sliders, okay?

[01:41:01] Each country has its own unique... Yes, to each their own, certainly. But come on. Steamed meat? Not... Yeah, baby! Of course, with McDonald's, there were TikTok, many TikTok videos, including one showing somebody adding 260 Chicken McNuggets to an order, or the AI did, 260 Chicken McNuggets as the customers screamed, stop! Stop! Stop!

[01:41:28] Next, a TikTok showing AI adding bacon to an ice cream order. Oh, God. That could be delicious. Yeah. It doesn't... You know, maybe... And then there was the guy who ordered 18,000 cups of water at Taco Bell. Well, I don't know if you know, if you saw the thing on Reddit this morning. Someone went to Shake Shack and made an order, and then clicked on no tick, and Shake Shack immediately upped the prices of all their orders, presumably to cover the tip.

[01:41:58] Oh, M.G. And that was pretty disgusting. Wow. You know, even though Shake Shack is not good. Good fries, which in and out doesn't do. I like Shake Shack. I like their Smashburger. Yeah. I mean, they're fine to break as well. Well, with respect to Shake Shack, if you don't tip your server, that they add a little money for the server to the order. I think that's good. Shake Shack is good. That's good for their employees. That goes against the whole... Surely that goes against the whole principle of fast food.

[01:42:27] You know, it's like... Shake Shack basically handing you something. You just don't understand tipping, my friend. Well, actually, America is really weird for tipping. You know, in Japan, tipping is... No, I know. I left a tip on a table that the worker came running after me saying, you left some money on the table. Please, here, take it. He was offended that I would even dream of tipping him. Meanwhile, in Las Vegas, you tip everybody.

[01:42:53] You give the guy who hails a cab for you five bucks for getting a cab, for opening the door and getting a cab. I mean, all he's doing is cab, opens the door and he gets five bucks. It's just, you know, this is the way... Welcome to America, my friend. The last time I was in Vegas for DEFCON, the year before, the taxi driver, you know, the little terminals you have in the back seat, their tipping choices were 25%, 30%, and 35%.

[01:43:22] And you're like, really? And I just said, can I just give you some cash and we'll just give you zero tip on the on the credit card. He's like, I love that. Oh yeah. I always give him cash. Yeah. Cash is king. Yep. All right. We're going to get to privacy next. Here's a little hint ahead of time. You don't have any. That's next as we continue this week in tech with Ian Thompson. You'll find much of his work at techfinitive.com, but all over too, right?

[01:43:51] You're writing everywhere and a podcast. What's the podcast going to be about? The podcast is going to be view from the valley. Which is your newsletter. Yeah. I mean, I'm trying to get, it's tricky actually. I mean, you are the God of podcasts. Don't get me wrong, but I hate the look of myself at the moment on podcasts. So I just, Ian, you look fine. You look fine. Everybody needs a friendly Basset hound Brit to talk to them about.

[01:44:23] I'm just teasing you. You look great. What are you talking about? Look, if I'm telling you, if I could do it, you could do it. We can't all look like Nicholas de Leon. Let's face it. True. I'm a radio guy as well. I grew up listening to shock jocks and so forth. I like talk radio. I kind of dislike all the video everywhere, but I agree with you on that. You know, in December, I will celebrate 50 years in radio. I got my, my radio license in 19 December, 1976. Good Lord.

[01:44:53] And I'm about to get my ham license. So congratulations. That's good. When the apocalypse comes, which apparently is only minutes away, you will be ready. Well, I'm actually, I've done 30 hours training with Richmond community emergency response training. Oh, you're going to be in CERT. Good, good, good, good. It's very important. I'm a credit, I accredited CERT member now. And if the big one comes, we are the second responders and yeah, we'll be trying to dig you out of the rubble. Yeah. Yeah.

[01:45:21] No, I have huge respect for amateur radio operators. I myself am a ham. I have a general. I don't, we'll have to exchange IDs. Yeah. Yeah. I'm W6TWT. I have a vanity license. Ah, whiskey six tango, whiskey tango, uh, seven, three, as they say. Um, I mean, I don't have a radio, so it's not, it's really literally a vanity call sign.

[01:45:48] Well, this was the weird thing about the training because there's like, okay, you should get a ham license because otherwise it's pretty. In the event of an, in the event of an emergency, we will contact you via email or phone. And so, well, email and phone are both going to be out. Yeah. Because what you need, and they're not expensive is a handy talkie. You can get a technician license. It's very easy to get. It lets you operate a handy talkie. And now you can really support your community with emergency services. When the phones are out, the power's out, nothing's working.

[01:46:17] You will be a lifeline. And it's happened many times in hurricanes, typhoons. I'm sure Benito in Manila, you have amateur radio operators and cert. I don't know if there's a Philippine, uh, uh, equivalent. I'm sure there is. I don't know, but radio is still very big here. So like, oh, well, I love radio. I mean, that's really all podcasting is we have pictures now, but you can close your eyes. Well, I mean, US so it started because of the Mexico earthquake in 95. Right.

[01:46:45] Um, where a hundred people were killed trying to get other people out of the wreckage. Right. Um, and I've since found, you know, this is one of the things from search running. You actually have access to depositories around the neighborhood of radios, batteries, large aerials, that sort of thing. Really? So cash is all over. Oh yeah. Yeah. I mean, and deals with the local home Depot and other places. So you can get chainsaws and pickaxes and the rest of it.

[01:47:14] Um, but yeah, the ham radio is going to keep it going. Cause in the event of emergency 5g networks will shut off completely other than for emergency services, even though they've got batteries to continue and email is just, you can forget about it. So, uh, yeah. No, it's true. Yeah. And passing along messages, talking to people.

[01:47:45] Good on you. Yeah. That's cool. You know what? You do it. Uh, when you're doing it, you feel pretty good about it. It's, it's actually, I was talking to a news guy the other day and I said, news people don't ever admit this. Cause it sounds kind of ghoulish, but we live for disasters. That's when you really, you know, you're really living up to your charter where you're helping the community, you know, whether it's on radio TV. Uh, I'm sure it's true for newspapers and other electronic publications. You're making a difference. And that's what you're there for.

[01:48:15] I hope there's a disaster. You're waiting for them. Yeah. I'm not going to name the person involved, but a friend of mine was on local newspapers in the UK and they were desperate for a headline. And she got a call from the local police saying, yeah, three family members have been killed in a house fire due to a faulty gas line. And she just went, yes. Oh no, no, no. See, this is again, why we never admit this. This is our little secret about, uh, media people.

[01:48:44] Uh, no, I love radio cause it's such a direct medium and direct communications. And honestly, broadcasting is just modern, uh, radio. Although it's interesting, a number of radio companies have tried to do podcasting. I think Cumulus, which is one of the big three radio companies, the United States just announced they're firing all of their podcast. Oh, this one down in, in Southwest. Yeah. I heard about this. They're just saying, oh, we're going to be all music all the time. And it's like, no, the DJs and the presenters keep you in that.

[01:49:13] You know, as one of the DJs, I always felt that we were important to the operation. And I mean, it turns out mostly we're just glorified button pushers, but okay. 58 degrees in the city on a beautiful Friday afternoon. Yes. Oh, you definitely did in San Francisco then if it's 58 degrees. Yeah. Only in San Francisco. Uh, also with us, Nicolas De Leon from Consumer Reports doing the most important work of all, right?

[01:49:41] Keeping us all, uh, up to date. Uh, Paris Martineau also works for Consumer Reports. And, uh, she was talking about, apparently I didn't know this. If you work at Consumer Reports, you must know this. Uh, periodically they sell, Consumer Reports famously doesn't take ads and doesn't take free products. They buy everything they review, including cars. Correct. And periodically they sell off this stuff and I guess they're doing it now. Uh, did you put it, did you put in a bid for a washing machine or anything? What?

[01:50:10] I haven't bought much. I mean, I've been at CR for, uh, nine years actually. Wow. Like a million years in media time. Yeah. Uh, I don't think I've bought a ton off the, the, whatever they call them, the little, they're not raffles, the internal auctions, auctions. Right. Right. Uh, I think maybe I got a TV years ago. See? It's cool. It's only slightly used, right? Cause these are all things they use in testing. Yeah. They're there. I mean, they're as good as any sort of open model. I mean, better than any open model you find.

[01:50:39] Unless the testing involves drop testing, hitting it with a hammer. Right. Stuff like that. Yeah. But you could get, you know, I don't know, like a, like a MacBook pro that costs like, you know, $4,000. Yeah. Hey, you would pay, you know, some percent. If you see, uh, any, any Macs with more than 96 gigs of Ram, snap it up for me. Well, yeah, I'll pay you double. I don't, I think, I think we usually buy the base model of basically everything. That's the problem. You're reviewing the normal human stuff. Yeah. I think that's how, I think that's how it works. Yeah.

[01:51:08] Oh, well, but I will look Leo. No, it's okay. It's okay. I don't want to, I don't want to, you know, I mean, that would, that would be wrong, but you know, sometimes it's wrong. Being wrong is so right. Uh, this week in tech is brought to you by adaptive. Oh, we saw them. We were at, at black hat. We saw adaptive at black hat. Talk to them.

[01:51:31] The first security awareness program built to stop really important AI powered social engineering, man, this, the world has changed. Attackers. They don't, they don't need malware anymore. They just need trust. And that could be a cloned voice, a convincing deep fake on zoom, but believe me, they can do it. Or an AI written fish that looks just like it came from your IT team.

[01:51:58] I think your employees, unless you're using adaptive are not ready for this new world. Adaptive prepares your organization with simulations, not just email, but text messages and voice. Deep fakes, phishing. That's that voice phishing where it says, you know, hi, this is Leo, your boss. I want you to send a bunch of, uh, uh, Amazon gift cards to this address. You know, it sounds just like me, right?

[01:52:25] Well, it was, but I mean, there's also AI generated phishing emails, which are indistinguishable, including scenarios that 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, but you can take action when there is trouble headed your way.

[01:52:48] If you need training fast with adaptive AI content creator, you can turn a breaking threat, something that just happened, an incident report, or maybe a compliance doc that you suddenly have to act upon into interactive multilingual modules. And you can do it in minutes. You don't need a design team. It's really good with adaptive. You can build, customize and monitor every part of your training with complete personalization. The result is a more resilient security culture.

[01:53:14] And I'll tell you that's essential for companies like plaid that, you know, plaid, you probably use plaid. But their platform powers thousands of digital finance apps and links consumers, developers, and institutions. With sensitive data at its core, plaid security and compliance are non-negotiable. Plaid's head of security, GRC says, quote, adaptive has equipped our teams with cutting edge tools and built a smarter, more resilient security culture across the company. Isn't that what you want?

[01:53:43] Well, trusted by Fortune 500s, 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. I thank them so much for their support. Of this week in tech. All right. Let's talk privacy as in you have none.

[01:54:07] Well, just to say, referring to the earlier ad, if you really want to get paranoid, go to the DEF CON social engineering village. Oh, my goodness. You will never come out of that just like, everything's fine. No, it really isn't. I'll give you an example.

[01:54:23] This, Anthony Nielsen, our creative director and AI guru, trained, used QUEN, which is a local model, a free local model, runs on a small machine, on less than a minute of my voice, and made this. Hey, Burke, this is definitely not Leo asking you to buy gift cards. But seriously, can you grab me 100 Apple gift cards? Just kidding. This is Anthony testing text to speech. How's it sound? It's identical to me. Yeah. Right.

[01:54:53] And that's something I was talking to one of our partners, the CEO of a threat locker. People brought us down at DEF CON and he said, that's what's changed. It used to be, you know, you had to have some skills to be a hacker. Now you just needed AI and you can do all sorts of terrible. But there's also the physical side of it. There was a fantastic presentation by somebody about four years ago.

[01:55:19] And she actually got in to give her speech just by going to the hotel, buying a hotel branded T-shirt, sorry, polo shirt, carrying around a clipboard with some papers on, and walked straight into the conference. Of course. So, I mean, from a physical and also, but as increasingly from an AI perspective. Yeah, we're screwed on that one. Well, here's another story.

[01:55:44] Ice, the folks who do the immigration enforcement, collected nearly 100 million people's DNA last year. Oh, including young children. This is internal documents showing that ice is now collecting DNA from people not accused of a crime. It's just routine check-ins.

[01:56:09] Immigration and Customs Enforcement officers will drive, you know, suspects to the field office. They'll swab your mouth. They'll swab your mouth. That's all they need to do to get your DNA. They take your fingerprints, they swab inside your cheek. And now they've got it. And you're in a database, citizen or not, legal or not.

[01:56:33] And they've been doing so much of it that there are now a million new records in there, including they do kids too. So, you know, when you travel now, they take your picture. They compare your driver's license to a photo they take at TSA. And, you know, they say we just, you know, we only do it for the comparison, then we destroy it. I'm sure they do. I have no reason not to trust the TSA.

[01:56:59] But we all, that's why the concern about Meta's cameras in their glasses seems minor. I mean, we are constantly under surveillance now. I mean, it is traditional at this point that I get snooty about how superior Britain is. But unfortunately, we've been doing this for about 10, 15 years. Yeah, you're not, you're better at it. Yeah, you've been collecting more. Well, actually, yeah, I mean, better on one level, disgusting in the other.

[01:57:27] These are, by the way, the DuckDuckGo smart glasses you were talking about. Oh, look, I'm getting all sorts of deals. $35 for the DuckDuckGo glasses. They got a little duck logo on them. I don't know. I was tempted just for the logo, but I have enough cheap sunglasses it is. So, you know, but. Yeah, I think that's cute. It's just a way of supporting DuckDuckGo. So they say the charge life, infinite, zero cameras, seamless offline mode, zero notifications

[01:57:57] guaranteed. I think that's pretty funny. It's a marvelous marketing stunt, which really points out, you know, the problems with these things and DuckDuckGo, who's not going to love them? Yeah. Yeah. We're learning more and more about how flock cameras are being used by rogue police officers to gather information about, you know, people who may have committed a crime or in some cases committed a crime.

[01:58:26] That's something illegal in one state, but not illegal in another state. And then you just pissed off within one of the operators. Yeah. Yeah. Or, or to, you know, track down girlfriends. It's also, there's more of it's turning out. This is from the Washington Post. There's a lot more of it than we even realized. And I guess I'm not, I'm guessing, not surprised. A lot of communities are now saying, let's get rid of the flock cameras. And there's also a movement, a movement on support of vandalizing these flock cameras.

[01:58:54] Well, one town, I think in Wisconsin where they actually cut down all the flock cameras in one day. Sorry, Nicholas, I interrupted. Oh, no, that is definitely a debate down here that I picked up on with the Tucson site. All the folks coming in, D flock, I guess is the name of the movement or whatever, or something like that. And, and yeah, I don't know. Or flock you would be the other way to put it. And one of the municipalities here, South Tucson, which is a separate city, they said they were

[01:59:24] going to take, take the, you know, take them down or whatever, but they never, there was no, there was no, there was any followup. So I want to file a records request. And for this municipality, you need to either hand in your records request in person to the clerk or via snail mail. Oh, you can't do the AI scraping. What is this? Not the 21st century. You cannot email, you cannot, it's like, is this on purpose? Is this just a small municipality? This doesn't feel like very, you know, very friendly here. But yeah, that was a huge debate here.

[01:59:53] For, for a couple of weeks we had, you know, the various municipalities around here, folks coming in saying, you know, demanding me every, these cameras, and so on and so forth. So it's, it was definitely one of those things down here. On the other hand, look, you're, you're driving on the public highways with a license plate that's provided by, you know, you know, your government. And these automated license plate readers, by the way, that's what these are. These are cameras.

[02:00:22] You'll see them on light poles, on, on traffic lights. It's kind of, it's kind of a upgrade of the red light camera that's become very common everywhere. And I can also see the point that, hey, it stops crime. And we, it's in our, we have flawed cameras in our area. And you know, when people go missing or kidnappings happen or, or things like that, it's very helpful. Well, I mean, Stanford University, sorry.

[02:00:51] No, I was just, I certainly see both sides. You know, I see the, you know, this is two sides to it. Absolutely. Yeah. This is one of those things where it's like, okay, well, who are the, who's the people like pushing the button at the flock camera? Are they, are, what are their motivations? Cause a lot of it comes down to like. Well, that's, and that's the issue. And that's why this story about more than 50 officers being charged with using, misusing the information in the flock database. The problem is the flock database is a national database, right?

[02:01:20] And so you can be a law enforcement officer in Texas, actually a sheriff in Texas, who's pissed off about your girlfriend leaving the state. And then use that database, that national database to track her down. And that's absolutely problematic, right? Well, I mean, this is why Stanford University just banned them because basically they're replacing flock cameras with a local system.

[02:01:43] So all the flock data, sorry, all the camera data goes into servers that they own and isn't shared with anyone else. So if there's a property crime in Stanford University, they can check that data. Whereas with the flock system, as you say, it's wide open to abuse. And we saw this with the Snowden revelations about the NSA. People were doing love int on their exes and their partners.

[02:02:09] And it's a natural, it's a natural way to use this technology for people that have access to it, but it's deeply creepy at the same time. Here is a flock hopper, which thank you to our discord chat provided me with this link. 90,000 automated license plate reader, ALPR cameras, 20 billion scans a month. You can enter your zip code and see.

[02:02:34] I mean, you know, I think the best thing to do instead of vandalizing them is to go to your city council, to go to your board of supervisors. There's 66 cameras near me and say, hey, take them down. I think that that has happened in my small town in Northern California. It's happened in our town in El Cerrito. They've basically abandoned the contract and flock has to take them down.

[02:03:00] Honestly, this kind of technology has a use, but sharing it nationally, that seems, unless you've got a warrant, I, you know, if you get a warrant, I've got a solution. It's just protected a little bit better. Yeah. That's a simple thing to do. Now we worried about cyber attacks attacking us water systems nationwide.

[02:03:23] And despite the president's insistence that it's the governor of Minnesota doing it, I don't think it's Tim Walts. In fact, the intelligence community says it's probably Iranians. You know, it's funny. CISA was at black hat. They had a big booth, the government security. And I almost wanted to go up to them and say, I'm so, so sorry that you've been decimated. I did get a sticker from them though. So I'm going to put that on my, on my laptop.

[02:03:51] Hope that'll protect me against bad guys. But this is, was almost inevitable that when you cut back on our security app or our cybersecurity apparatus that we are going to get attacked. And a boy, nowhere more terrifying than in our local water drinking water systems. It's not like they're putting poison in the drinking water. It's just that they're, they're shutting them down.

[02:04:15] The problem is much like our electrical grid, each community has its own system and different standards for cybersecurity. It's highly fragmented. Infrastructure is far from modern. Many of these are online. You know, they're attached to the public internet and many of them are not secure in any sense of the word. Well, I mean this, I mean, this is a problem that 80% of the U S net critical infrastructure is privately owned.

[02:04:45] And not regulated in terms of security purposes. So, you know, say for example, if when China attacks Taiwan and the U S decide to respond, you can shut down the water system, the electricity system, the financial system. And people are going to be much more concerned with that than a obscure island off the coast of China. You know, Iran, I don't know.

[02:05:12] I mean, it's basically these systems are industrial control systems, which are poorly patched, poorly maintained, and there isn't an awful lot of interest in government in fixing them, which is deeply disturbing. The according to Axios, the EPA says it identified vulnerabilities last year at at least 277 water systems. They have been working with utilities to address the issues. Yeah, we are.

[02:05:41] Our infrastructure is very vulnerable. And that's another form of cyber warfare that doesn't take a lot of manpower, a lot of money and just a few skilled hackers. And you could be in a lot of trouble. Well, you could dump an awful lot of sewage into the local water networks, which would render them unusable. That would be a problem. Yes.

[02:06:07] On the other hand, I, I don't know, I've never got confirmation on this, but I was talking to an American government employee who was saying, yeah, if they do that to us, we can do that to them. And China's water system is even more vulnerable than ourself. Well, that's a good point. Yeah. That's one of the things that's kept cyber warfare at a lower level than it could be is the threat is mutual. Yeah.

[02:06:32] And as soon as you open that, you know, Pandora's box, and we all suffer, not just, not just here. Well, I mean, it's kind of, it reminds me of the OPM hack where, you know, the former, you know, sort of the Office of Personal Management in the US, former head of the NSA was saying basically, yeah, of course we were, you know, China would have gone for that. If we were in the same situation, we would have gone for them as well.

[02:06:59] Well, whether or not you use it, however, is the difference between a cold war and a hot war. And both countries are sensible enough at the moment that they wouldn't do that because then it gets really ugly. Have you heard of an Italian company called bending spoons? They are an absolute genius idea. They bought Evernote. That's the first time I heard of them. They say they bought Evernote. Yeah.

[02:07:27] Then they bought AOL and I kind of sat up and said, huh? Who? Bending spoons? They are now buying Airtable, which now this is a little bit more significant. This is a very powerful database tool used by a lot of companies paying $1.2 billion per Airtable. Now, they've done a good job, I think, with Evernote. I haven't heard any complaints about it.

[02:07:57] I don't know how they've done with AOL or Eventbrite, which they also bought. They've raised money. And now Airtable is used by half a million organizations, 80% of the Fortune 100. I hear people talk about it all the time. Bending spoons is about to buy them. I don't know what that will mean. What I mean? Well, I mean, there's a lot of money to be had in old platforms.

[02:08:25] There's also a lot of data to be had if you take it out and run it through AI. Yeah. Typically, when you have these kinds of acquisitions, I don't know if Benning spoons counts as private equity, but it kind of acts like it. You get a lot of layoffs. They laid off 129 people at Evernote when they bought it in 2022, then cut most of the team and moved to operations to Europe. Cost went up. 1.3 billion seems really high. That's a lot.

[02:08:55] Maybe, yeah. They bought WeTransfer in 2024. They cut the staff by 75%. They changed the terms, including requiring that users grant AI training rights over their files. They did change their mind after some pushback on that. Airtable, though, I think in many respects is the most ambitious acquisition by Benning spoons. So it'll be very interesting to see what happens to it.

[02:09:25] But I think these days, I don't know. I don't work at a big company, but I would imagine because of AI, people are less locked into their SaaS solutions, right? There are more alternatives out there, and there's even the idea that maybe you could create your own. I say that internally. We have various SaaS tools that are, some are good, some are better than others. I guess I'll leave it at that.

[02:09:52] And I'm like, man, I could just make this in, let's call it a day. And we wouldn't have to pay this license anymore. And, you know, that's half-joking. I can understand the reluctance of a big company. When you have, when you're paying somebody for something, there's liability, their responsibility. Of course. When you start doing it. Well, we're doing this too at Twit. We have a, we have for 12 years used a homegrown ad sales system that's, you know, was written by one of our employees. He did a good job considering.

[02:10:21] Yeah. But it was fairly janky. Yeah. We've kind of muddled along with it. And so I've been using AI to rewrite it. It's a bigger task than I thought, to be frank. But I think we can do it. And I think we will get a better product as a result because, well, it's more modern and it's, you know, a little more secure and more robust and ultimately more responsive.

[02:10:47] You know, the guy who wrote it left and, you know, we can now, we'll have control over it. So I, I think there's a lot to be said. I don't know. I also understand the reluctance. Yeah. I know. If, if these tools were all, you know, again, some are better than others and I'm not mentioning any names, obviously, but like, you know, some of them are pretty basic and like, not like what, like, I don't think air table, air table is cool, but I don't think it does anything that you couldn't do otherwise. Other. I'm honestly not familiar with air table.

[02:11:17] So it's a table based database, basically, you know, it's easy to use database. I've never, certainly never used it. Uh, but yeah, I don't know if you're a small company, do you really need to be paying X amount of dollars per month for every SAS thing on this? I don't know. Maybe, maybe, you know, maybe, maybe that's easier and. The SASpocalypse is coming. Just run your own servers for goodness sake. Why pay these bastards?

[02:11:41] I think, you know, if you're running like a local ice cream store or whatever, you're not going to sit there and code your own like CRM. You know, of course not. Yeah, but no, no, no, I mean, that's fair enough. But I do think when we were seeing a move away from cloud in because it's just not secure enough anymore. Right. We're entrusting too much data to third parties. It's just personal view, but yeah.

[02:12:08] That's to me the next project that I'm most interested in is the, the notion that you could run AI inference locally with open way models. I was just seeing speeds of getting up to that point. So why not? The problem is getting the hardware. If I could buy a 512 gig Mac studio or Mac mini, I could easily run very good models locally. Yeah. Yeah. Including Kimmy, I think, but good luck getting a 512 gig Mac studio. All right. One more break.

[02:12:37] And then I do want to learn about how, how Nicholas used AI to lose 30 pounds. Uh, that's, that's cool. Uh, we have more to a few more stories to talk about, uh, including a happy. I like to end with a happy, so very happy story. But first a word from our sponsor, super human go between the endless stream of emails, documents to review urgent decisions.

[02:13:04] You've got to make staying on top of everything can feel like an impossible task. A lot of us were promised that AI would make our work easier somehow somewhere along the way. It started feeling like it was just another thing you had to manage. So there is a solution. People are turning to super human go. We've been using Grammarly for years. It's great for writing. We just love it. And now this new tool from super human is very exciting. Super human go is an AI chat.

[02:13:34] It's always there when you need it already aware of what you're doing and doesn't ask you to start from zero. Whether you're drafting an email mid meeting, summarizing a long thread before a call or hunting for a file while you're doing. Three tasks deep super human go works seamlessly inside your existing tools to provide the help you need without ever making you lose your place. It's kind of like a sidecar. It's there. It sees what you're doing, knows what you're doing, and it can help because super human go is from the makers of Grammarly.

[02:14:03] It works across the tools and sites you already use. You may if you're using Grammarly as we do, you may be very familiar with how wonderful and useful it is. Imagine having even more AI and smarts right there in your browser, in your inbox, in your docs, in your project tools. Put simply, super human go helps you be the best version of yourself at work without the friction, without context switching.

[02:14:28] And we know that, you know, when you go from one thing to another, some of this that you're doing goes out of your brain completely. You go, do you ever walk through a door, you know, saying now what did I come in here for? That's context switching. But not with super human go. You don't have to leave. You're there. Plus, you can eliminate that busy work. You have the time and the brain space to solve the problems that matter. So you can focus on the work only you can do. That's the best way to use AI.

[02:14:56] Go's AI chat is available on the side of your browser. It's always ready. No new tabs, no switching apps. Use super human go's AI chat to draft emails and messages to summarize long threads and documents to search across your apps all without leaving the page you're on. Prep for meetings in seconds, follow up without starting from scratch. Super human go keeps up and handles the repetitive stuff so you can move forward.

[02:15:23] This is AI that works with you on top of you. Check it out. Find out more at superhuman.com. That's superhuman.com. The good news. I promised you good news. NASA has figured out how to keep the 48 year old Voyager 2 probe running for yet another year. Elizabeth Howell writing at space.com.

[02:15:52] The problem is, you know, the battery and the solar power that's charging the battery. But NASA was able to, this is a remarkable story. There's a wonderful documentary about the guys who were there at the beginning when they launched Voyager in 19, in the 70s. Now these guys are in their 80s. They're still there, still managing this incredible project that has gone much longer than anybody thought it would.

[02:16:19] And Voyager 2 and Voyager 1 launched in 1977. They use a nuclear battery known as a radioisotope thermoelectric generator. So they've got a little bit of plutonium, the heat generated by the decay of the plutonium, keeps the generator going, keeps it powered. But the supply of plutonium is dropping about four watts a year. And so they have to slowly reduce the power requirements of Voyager 2.

[02:16:48] They turn off, you know, non-science devices. They use lower power alternatives. And the fact that they can do this, while this thing is light years away, is kind of remarkable. I don't know, is it light years? It's 142 AU. That's 142 times the distance between the Sun and the Earth. It's actually getting Voyager, that's Voyager 2. Voyager 1 is actually 171 astronomical units away.

[02:17:16] It takes 24 hours for the signal leaving Pasadena to get to Voyager 2, and then 24 hours for the response to get back. So that's not even 1% of a light. Unbelievable. So that's not even 1% of a light year, Leo. It's not even close, is it? Okay. Still, it's a long way. It's the farthest man-made object ever. Pretty amazing. It's gone out of the solar system, I think.

[02:17:46] Well, I mean, NASA builds things to last. Their engineers are absolutely fantastic on this. Yeah. I mean, I was at JPL when Curiosity landed in 2012. And it's the last time I've ever clapped at a press conference, which is a huge verboten when it comes to journalism. Yeah. You don't do that. But when the first images came in, I was just like, yeah! It's hard not to, isn't it? It's really hard not to. Exactly.

[02:18:13] And the way they've kept this probe going is just superb. It really is. I mean, it shows the dedication that NASA employees have, even though they're paid lousy wages and they get dumped on by governments. But yeah, so much, so much respect to them. It's just marvelous. Voyager 1 is 25.858 billion miles away right now from the planet Earth.

[02:18:43] Pretty amazing. So it's going to last a little bit longer. That's good news. I mean, they've been really smart in the way they've shut the, pardon me, shut things down and kept them going. And you know, it's just, it's the engineering of it. It's the way it keeps going no matter what, because enough people care. And the Deep Space Network for all, you know, is just superb in keeping in contact.

[02:19:12] One more year for the 48 year old Voyager 2. How did you use AI to lose 30 pounds, Nicholas Staley? That's pretty impressive. First of all, you look great. Congratulations. Thank you, Leo. I had to take drugs to lose weight. That's fine. Yeah. I love Ozepic. Are you kidding? Yeah. I think the doctor I spoke with in this story was like, anything that you can do to, you know, lose weight is good.

[02:19:40] You know, so, so, I mean, so the gist was basically, this was last fall, you know, I had put on, you know, some amount of weight. I was near 200 pounds, which is way, way too much for my frame. You got married. That's what happened, buddy. That's what happens when you get married. You start to put on the pounds. Yeah. So this is, by the way, a consumer reports article, which is great. Yes. Yeah. Yeah. So, yeah. So I mentioned that to my editor and he's like, oh, cool. We should write about that. You know, CR, we don't always write like personal stories or whatever.

[02:20:10] So it was just a busted like that, but yeah. I would just be like, yeah, just write it. Who, who cares? But yeah. So I used a combination of, of Gemini and Claude because those, those were the, that tells you when this was done last, last fall, the Gemini I was seen as like cool or whatever. And yeah, I basically talked to us like, Hey, here's my situation. I weigh this much. I'm this tall. You know, I want, I want to, you know, I'm not trying to like, you know, make the Olympic team or whatever, but like, I'd like to lose some weight.

[02:20:39] I'd like to get a little stronger. You know, I'm, I just about to hit 40 years old. So I'm not a kid. So, you know, yada, yada, yada. And by the way, you know, the stats are you gain 10 pounds for every decade. So now's the time to get a handle on that. Which is, which is kind of what my doctor said. Like he's like, yeah, you know, any, you know, this is when it's going to pay off. It's like the work you do now. It's like, you know, when I turned forward, he was like, well, son, you know, you know, you're not a kid anymore.

[02:21:06] He's like, anything you do now was definitely going to pay off. You know, years. Yeah. So how did AI, how does AI get in this? I mean, it was basically just like a personal trainer. It was like, I would ask him like questions like, okay, devise me an exercise routine, you know, help me build a little home gym. You know, we have a spare room in the house for, you know, just like a squat rack and stuff like that. And so, you know, the kind of the feast, the, the idea of the story was that like all this information, you know, I, you know, I lifted weights on and off from my twenties and thirties,

[02:21:36] but it was all on like Reddit or like you had to read like men's health magazine or all this type of, you know, the message. So all this information was like all over the place. And so it was a hard to like, well, what am I supposed to do next? Uh, and then what, and you know, how much am I eating? What should I be eating? Like this, it existed, but you know, unless you want to like give yourself over to that lifestyle or whatever, it was kind of hard to, you know, keep, keep track and like to, to make sense of what the next thing to do is with AI. It's, you know, obviously it's, it's read the internet. It knows all that.

[02:22:06] And so it's able to synthesize, okay. You know, based on your parameters, here's what you can do. You know, it sets expectations, you know, you're not going to look like, you know, the rock in a month or whatever like that. So it kind of tries to tell you like, you know, in, in a year you could look like this, you know, if that's, you know, that's what it's going to take. Okay. A year, a year is fine. You know, I'm not in a hurry or anything, but it definitely helped with, you know, just creating the exercise routine, you know, help it helped me with, uh, you know, food and diet, you know, I'm lucky.

[02:22:34] I cook, you know, probably six, six nights a week. So, you know, we go to Costco all the time. So it's like, okay, we have like real whole foods in the house that I can make. Uh, but it was just like, okay, well, how much, you know, how much chicken should I eat a week? How much fish, how much, you know, uh, uh, Greek yogurt smoothie, all those kind of like questions you might've had. It just makes it much easier to, to get an answer. Uh, and to like, keep on track. You did something interesting and I think very smart. You actually use two models.

[02:23:03] One kind of against the other you like. Oh, yes. Yeah. So I, so Gemini was the initial kind of like the initial only one. And then at some point in the process, I was like, well, let me, you know, not quite a mixture of experts or whatever, but like, oh, let me, let me go to Claude. Let me, let me see what Opus has to say about this. And you would get, you know, you would, you would get like a second opinion. Basically you would get feedback. Oh, here's what, here's what my, and I would talk to it. Like, well, one of my friends said to do this.

[02:23:33] What do you think Claude? One of my AI buddies. Claude would be like, that sounds good, but I would also do this or would also do that. Uh, and you know, so I'm going to have, I just turned it in on Friday. There's going to be a second, a second, uh, a second beat to this article on our sub stack, which is CR bread and butter, uh, dot sub stack. And that's more, a little bit more like, not like instant, like see this article is like consumer reports wrote it type of thing. The article on the sub stack is that Nicholas wrote it.

[02:24:02] And it's a little bit more personality, a little bit more like a down to earth, I guess. Uh, so what's your sub stack? I didn't know you had a sub stack. Yeah. We just started a few months. It's C it's called a bread and butter. So CR bread and butter. Oh, I love that. Sub stack. Cause bread and butter like a hundred years ago was a consumer reports, uh, thing maybe during the war. I forget offhand where they help people live through like rations and like how to like, how to live a productive life when there's no butter or, or stuff like that.

[02:24:32] So it was called bread and butter. And so this, this is kind of the next generation where we try to do more, you know, more, more homey, more down to, or down to earth advice versus kind of the CR kind of like high-minded institutional can be a little, uh, not robotic, but like, you know, it's, it's very magaziney, I guess. I like your prompt. You said, I'm just a nerd who likes old video games. If I get in shape at 40, then anyone can.

[02:24:56] And at one point, uh, Gemini, uh, told you you're trading fluff for armor. This is the Holy grail scenario. Some of those, I got it. That's a very AI kind of thing. It's some of that is very silly cringe. You might say, but like overall, I think the advice was, was certainly directionally good, directionally helpful. You know, it, it helped me keep like, okay, it can be like a coach. I log my food. I log my exercise.

[02:25:26] I also log, you know, sleep, heart rate, blood work. I put everything in this thing and, uh, including genome and biome. And I find it very helpful. It just as a coach, you know, if it told me you need to eat more salt, I might second guess it, but you know, saying eat more blueberries. Okay. That's fine. Right. That was, that was some of my, you know, it just took a while to edit. You know, this, this probably took, you know, let's call it a month to edit back and forth with, uh, you know, the health team.

[02:25:54] And so I didn't want to say anything stupid, you know, that's nice. That's the other thing, by the way, I really respect with consumer reports. They're fact checking. Oh yeah. Oh no, this is really absolutely. Yeah. Yeah. It took a while. A lot of publications have dropped that not CR. Oh, we haven't. Let me tell you this took a while to get out the door and it's for the better ultimately, but you know, to run it by the health team to make sure I wasn't saying any anything stupid, anything like, you know, bought, you know, uh, you know, I'm just a guy. I'm not here to like, you know, say anything or getting anyone in trouble or look how cool I am or whatever.

[02:26:24] But I didn't want to say anything offensive. I didn't want to say anything uneducated or anything like that. Uh, but yeah. So I, I think, you know, it's, it is, you know, and in the editing process, I think folks were, there was some like, not skepticism, but like, like almost fears. Like, oh, you talk to an AI about like your health. You talk to how much you weigh. And to me, I'm like, yeah, who cares? Like, I don't care. Like, this is, it's just a computer. Like, it's not like, this is not, you know, I would, I'd almost be more embarrassed to tell like a real personal trainer, like a guy.

[02:26:54] Actually that's one of the benefits, isn't it? Right? Yes. Yes. I think. And I bring that up in the secondary. It's not judgmental. Uh, it's not a human. It's not a human. It's just a computer running somewhere. It's a, it's a GPU somewhere, somewhere out there. So like, it doesn't, but folks don't know that a lot of our readers are, you know, they, they, you know, not to pick on John Oliver, but they see kind of like alarmist segments. Like AI is going to like do all this bad stuff. And they just kind of believe it.

[02:27:22] And I'm like, it's just a, it's just a computer program. Like, I don't know. What are we talking about? Uh, so it was, it was a little bit like, you know, I used it. It was fine. It helped me out. It might help you out too, but you know, well, I'm happy to say, well, Consumer Reports is celebrating its 90th anniversary this year. That's pretty amazing. Uh, I've been a member for half that time. I joined in 1980. So I have literally been a member for half of the life of Consumer Reports. Thank you.

[02:27:51] Thank you for your support, Leo. Well, and I've, I don't buy anything without checking CR first. And it's people like you. I think that really, it's great. They're moving into the 21st century. You know, you could easily rest in your laurels as a 90 year old publication, but they're not. And I think that's really great. I'm glad you're their senior technology reporter. Nicholas de Leon, thank you for being here. Give my love to your wife, Ashley. She's also lost weight. Is she doing the same thing as you or? Uh, yeah, more or less.

[02:28:19] Uh, I, I, I, I don't know specifically what she's doing. She's not asking your AI for advice. She's certainly not asking me for the prompt or anything like that. Uh, wonderful, uh, wonderful person. And, uh, both of you are dear friends and that's really nice to see you. Speaking of dear friends, always a pleasure to see Ian. Uh, used to be, you'd come up here with your beautiful wife and eat some English food from our local, uh, British. I know. The English food shop shut down. Yeah.

[02:28:49] It was, they had great pork pies. Oh, I love a good pork pie. You know, it sounds dirty when you say I love a good pork pie. Uh, we have the Petaluma pie company. I think you could probably talk them into maybe not steak and kidney, but you could probably talk them into a pork pie. Oh, you can't get round a good steak steak kidney. I mean, it's the faint tang of urine that really kind of puts me off. I'm sorry.

[02:29:17] Ah, honestly, kidneys, mustard, liver. Oh, it's awful. Nothing like organ meat is there. You know, I found out you can't even, uh, get sheep's lung in America. So you can't make haggis. No, there are proper haggis is illegal in the U S it's illegal. Yes. It's illegal. There's a good reason for it. No, because you cannot get lung sheep's lung in America. You can get Twinkies.

[02:29:45] You can get Oreos, but no, you can't get sheep's lung. Um, and you can't have a proper haggis without the real hard lung texture. Oh, I love a good Scotch egg. I look at the good, uh, pork pie, a little plowman's lunch, you know, give me, give me bangers and mash. Give me some little squeaking bubble. Give me some British food. Anything, but oh, well, oh, come on. I have a little more in my cupboard too. I must admit.

[02:30:13] I was going to say, I mean, my mind, one jar of my gods last could last you years. I buy it in 20, 250 gram Amazon packs. Um, because I can't get it anywhere else during COVID. There was a Marmite lockdown. My God, I scraped that last. Oh, the last thing. Oh man. Ian. Thank you so much for being here. Always a pleasure to see you. Always great to talk. And yeah. Do the podcast.

[02:30:43] I will, I, we will send people to the view from the Valley techfinitive.com and also the view from the Valley podcast coming soon to a iPod. I will do my best once I get out of my personal things. Yeah. Don't forget the Tucson daily brief at tucsandailybrief.com. What's my next assignment? Ah, my next assignment is, uh, so Tucson daily brief is, I can I just take two seconds to kind of describe it's evolved since the last time I was on the show. Yes. And what we do. Okay.

[02:31:13] So the, I think I said last time, which was several months ago, the idea kind of came from when I was messing with open claw and trying to see what it could do, so on and so forth. So basically it's a semi automated local news publication, uh, that has a bunch of different moving parts, kind of the day to day, uh, output is called the brief. And that's based on as look, it looks at around 30 local sources. So those are, you know, the newspaper, the TV stations, radio, NPR, uh, the elected officials,

[02:31:41] their websites, you know, some of this is, uh, RSS feeds. Some of this is scraping, uh, but it, uh, uh, school districts. It, you know, it looks through all those sources at six in the morning, my time. And then I have, uh, I think, uh, sonnet 4.6 just tells you when this was made. Uh, looks at, looks at everything that's happened in Tucson overnight and kind of gives you like, okay, here are the 10 most important things that happened in Tucson overnight, Tucson and

[02:32:07] the surrounding kind of area from government, public safety, uh, you know, sports, you know, whatever it is going around. So that's kind of, that's one component. Uh, the second big component that I made, I call it the live AI reporter. And so basically what that does is I deploy that. That is, that uses a combination of deep gram and some other things, uh, to, you know, every local town hall, the city of Tucson, the Pima County, you know, they meet every week, they

[02:32:34] make every other, every other week, whatever it is, all these things are live streamed on, on YouTube or other platforms. So I have, uh, you know, I'm just listening to the live stream live as it's happening. As that live stream is happening, it uses a deep, deep gram and FFM peg to create a transcript live in real time. Uh, and then it takes, you know, these meetings last, God, these meetings last like four hours. So it's like a four hour transcript of just a municipal stuff happening.

[02:33:02] Once that ends, it sends that like million word transcript of the four hour meeting to anthropic, it creates a, like a, yeah, let's call it like a thousand word, uh, AP style news article based on the biggest things coming out of the, the meeting that draft gets sent to me. I, I edit it. It takes about an hour to edit. You know, deep gram is, is pretty good with diarization. You know, this guy said this, this guy said that it's not great with, uh, foreign language names.

[02:33:30] And obviously in Tucson, there's like a million Spanish language names. So I made a thing to like automatically catch and correct any, uh, misattributions and things like that. So that's, that's one component. Uh, we have a newsletter which comes out on Sundays and that's based on everything that I published throughout the week. Uh, and that's, that's also, you know, that's free, uh, in that newsletter every week, there is a little mini crossword puzzle that I, I whipped up together, a Tucson themed crossword puzzles.

[02:33:59] The AI makes crossword puzzles too. Yes, it does. Yes, it does. Awesome. Uh, and then there's like a little rag tool. So up at the top, it should say chat TDB. I bought chat TDB.com. It's on daily breeze and that you can ask it like, Oh, what are the restaurants that open? And so you could do that. It monitors all the local filings. So like restaurants and liquor and all, anytime something hits the directory, uh, it immediately pings. You know, some of this is automated.

[02:34:29] Some of this, uh, you know, some of this, I'm editing by hand. Uh, you know, the, how much time a week do you put into this. More now because the social, you know, some of the social, I just, I just created a blue sky earlier this week. I just created an X. The blue sky podcast too. Yeah. The podcast that uses 11 labs. So that takes like the daily brief that like, you know, here's what happened last night. The voices 11 labs are amazing. If you want to use my voice, I'm on 11 labs. You couldn't use my voice. Let me just hear it a little bit.

[02:34:58] This is, this is today's. And this is the Tucson daily brief for August 9th, 2026. A lightning sparked wildfire is burning near finger. Is this supposed to sound like you currently at around 20 acres with zero. Wow. It sounds like you closed as crews work to continue. It was trained in my voice. I mean, it was a while ago. So maybe, maybe the models are better. They're actually, they're really good. 11 labs. If it weren't, if it were free, I would use it all the time.

[02:35:27] It's a little pricey. Yeah, but it adds up. I get, so you're doing a one minute podcast. That's not too awful. No. And that's, that's sent to Apple podcasts and YouTube. Well, yeah, what takes a while is the social, like Instagram. I have to, even with like, you know, using the AI to like automate the content. I still got to post it. I still got to, so the social is kind of what takes forever, but I just said, yeah. So I built the blue sky last week. I was actually in contact with the blue sky folks.

[02:35:56] Like you can now comment in every article it's powered by blue sky. So it does, I don't know. It's just me, you know, I, you know, I'm not, you know, I'm not going to be a billionaire off this, whatever, but like I spoke to a local business. You're doing good work. Consortium. And they seem, you know, you say you're not an expert. I think that's pretty impressive here. I'm going to have my AI Leo offer. Hey Nicholas, I want to audition to be in your podcast. This is AI Leo ready to work.

[02:36:26] That's 11 labs. That's good to me. Yeah. That's really disturbing. I, my agents for a long time were using 11 labs. And then I realized, oh my God, I'm going to go through all my credits. But I had some really fun voices, including Burt Reynolds was one of my AI agents for a while. Yeah. So, and then what is this? My next assignment, by the way, you know what we should do? Would you like to come? Are you available Wednesday about 2 p.m. Pacific,

[02:36:54] 5 p.m. Eastern to be on Intelligent Machines? I would like to interview you on our show. Paris would love to, I'm sure. I assume I'm free. I don't do anything. I'll have Benito get in touch. And I'll call your wife and make sure she doesn't have any chores for you that day. That's fine. What do you think? Yeah. So yeah, let's do it. That's fine. Let's do it. I'd like to have you on. And then we can talk about this too. It's TucsonDailyBrief.com. And if you go to X and search TucsonDailyBrief or Blue Sky, it should pop up everywhere.

[02:37:24] I love chat GDB. ChatTDB.com. TDB. That's hysterical. Yes. Thank you. You registered chat GDB.com? I own chat TDB.com. Yes. If you go there, it takes you to the site's rag. Which is... Yes. That's very smart. Yeah. And what's great is you're just doing this for fun. You don't charge people. No, this is all free. I mean, it costs...

[02:37:52] It costs... The most is 11 labs. I'm on like the $20 like starter planter. Yeah, that's what I have to. Yeah. So that's the biggest... In terms of like the LM calls, you know, depending on the model I'm using, it could be... If you're using DeepSeek, it's basically free. But I don't know. Call like 20 bucks a month. I mean, 10... Nothing that I'm... You know, I spend more at Starbucks at a week. So it's really not a problem. We will hear more from Nicolas De Leon Wednesday on our AI show, Intelligent Machines. Then you can get into the nitty gritty of it. Very good.

[02:38:22] And then the other site is my next... Oh, I had two. One was cool. I'm sorry to take up the last all this time. I made a kind of a sequel to the baseball site. It was called MatchDayInference.com. Yeah. It was for the World Cup. So basically the idea was, you know how you go to a stadium and, you know, there's like the program where they talk about like, oh, today the Mets are playing. You know what I love about this? It doesn't look like AI. It really doesn't.

[02:38:52] That took the longest time. That took the longest time to make it. So the inspiration for this design was kind of like sub-pop records, kind of grunge, kind of like whatever. It has a nice feel to it. Yeah. And so what this was, this is finished now, but the idea was, you know, if you were making like a journalism product from scratch with AI as like, you know, the co-editor, what might that look like? And so I came up with this.

[02:39:17] There's, you know, at the time you pick your player, you pick your team, you pick like your lens, we called it. So it was like historical tactical, you know, a bunch of different like permutations and it would, you know, on the day it would generate for you based on your parameters, a custom email. You know, today Argentina is playing, they're playing at the Estadio Azteca. And the last time they played here was blah, blah, blah. And it had, it, it was super cool.

[02:39:46] And it sent out via one of the email thingamajigs. Wow. So that was that. And the last one I want to talk about. The thingamajigs. I know what he's talking about. The thingamajigs. And the last one it's, it's called my next assignment. And so this is basically a job, a job finder tool for people in media or comms or adjacent spaces. What this does is, you know, you, you sign up for the site and I'll get to that. You sign up for the site. It takes your resume.

[02:40:16] It takes your skills and it tries to match, you know, it scans like 300 plus job boards or whatever. And the fun part is like based on the job description, based on your resume, based on whatever you tell it, it will automatically draft a cover letter. It'll automatically draft a resume tail tailored to that exact role. So you don't have to sit there and come up with some fake cover letter about like why you want to join Acme Inc or whatever.

[02:40:43] And it is, you know, this is, this is, it is launched. It is up and running. I haven't officially like announced. Oh, you could actually make money on this one. Now we're talking. Yeah. This is a paid thing, but I've given to some friends. It's been kind of like humming in the background for about a month. And I'm aware of, I want to say five interviews that folks have had with like real companies that everyone listening to this show has heard of. So it works in some capacity.

[02:41:11] I have a friend, a good friend who was, used to be an engineer at Facebook. He's looking for work. He built something very similar. He called it Jared. Jared and it got, it, he got more interviews. He got, he ended up getting five competing offers. He was having a hell of a time deciding. It really worked. I, this is a great idea. And I think you're going to, honestly, Nicholas, I don't want you to do too well.

[02:41:40] Cause I don't want you to leave consumer reports, but I think you've got something here. I would, I like CR so much. I would, you know, good. I would like to stay at CR a long time. I like it. I put Ashley on this, you know, let her, let her be your, that's what I did with my wife. Let her run the business. She wants to run it. She can run it. That's fine by me. She'd probably be better at it, frankly. But yeah, so that's true about Lisa for sure. That's the high level of, uh, of the, I'm trying to think of something to do for the football season coming up. I may have one idea.

[02:42:11] Hold on to that. So we have something to talk about on Wednesday. I want to tell you, and then we'll look at your baseball site, but we also, I really want to look at your, uh, how you do it. Your procedural, you know, what your tools are and how you do it. I think that's very helpful. It's very interesting. We cover AI, of course, on intelligent machines every Wednesday. We also have an AI user group, which we've now made, uh, twice a month because I think everybody's working on this stuff and we're all doing it kind of independently. And we've all found different ways of doing it.

[02:42:41] I think it's really useful to share that, uh, we can learn from each other and get better together. That's one of the reasons why I'm still on X. X, there's a lot of guys, a lot of indie hackers who would say doing really interesting stuff that are kind of like, okay, if this guy, who's just a guy like me, if he could do this, then why can't I do that? I steal stuff every day from, I go to my agents and say, Hey, here's a new repo. Check this out. Here's a new idea. Check this out. Because there is a lot of it. I hate to admit it.

[02:43:10] The good news is X now has buttons at the top where you can just say, I just want to follow AI. Yeah. Yeah. It doesn't post much there. Uh, the only thing you have to watch out for is there are a lot of people down and doing engagement farming saying, I've found the three props that are going to change my life forever. You just ignore those guys, but there are some legitimate people on X. In fact, most of the people who are doing the real work, uh, like Technium from a noose research, which who does

[02:43:35] Hermes, uh, and his partners are on there. And I think that that's a really good way. It's sad to say, I, I left X when Elon put Twitter, I think I'm back on it. I hate to admit it. Uh, cause that's where you go. Yeah, it is what it is. Yeah, it is exactly. I mean, you know, I, I'm, I'm not wild about being back on there, but you got to do it. So it is what it is. As you say, I do, I do use, uh, a chat, a Gronk,

[02:44:02] a Gronk, uh, 4.5 Elon's, uh, model just as a third frontier model, uh, for a variety of things, but I named it Gronk. And I said, and you have to talk in a caveman voice. So it says, I am Gronk and I will fix. It is marvelous though. When, when Gronk contradicts Elon Musk, it's kind of like, Oh yeah. What do I pay these billions for? Actually, Gronk has gotten pretty good and he's got a new model coming in. I mean, they,

[02:44:31] he spent a lot of money on the school. Yeah. Yeah. But at the same time, he said that Hitler was a left-wing socialist and Gronk. You know, it stopped doing that. Yeah. Maybe not. I wouldn't ask it about that kind of stuff. That's for sure. Uh, thank you so much, Nicholas. Thank you so much, Ian's great to see you. Thanks, especially to our club members who make such a big difference, uh, keeping this show on the road without you, we would literally

[02:44:57] have to, uh, shut down or for sure cut back more than 30% of our operating expenses are paid by club to it members. Uh, your subscription means so much to us 10 bucks a month. And we give you a benefit for it. Ad free versions of everything. So you'd never hear another ad. You also get chapter markers. So, you know, there's a big controversy over at Spotify because people, they put a skip button that you can skip through ads, uh, on the Spotify player. And everybody's

[02:45:24] saying you're going to kill podcasting. No, here's the easiest thing. You don't want ads join the club. You won't have ads. You'll have chapter markers. You can jump to the part of the show you're most interested in jump back to another part of the show. You also get access to the club twit discord, which is a wonderful place to hang out with some very smart, interesting people talking about not only the shows, but all the stuff geeks are interested in, including a lot of AI. We have a whole AI, uh, slop board where you can go and see all the AI generated.

[02:45:53] Here we are with Mohawks, by the way. Oh, well, no, I've just posted up a picture of my last fry up and it seems to be going down well. Oh, all right. A smiley face fry up. Yeah. See our hosts are in there too. Uh, let me see. What's a fry up. Is that your breakfast? Oh, look at that. Yeah. Oh, that's real. That's not AI. Oh God. No, I made that myself. So you've got blood sausage or do you call it black sausage, blood, black sausage, uh,

[02:46:23] two sausages acting as a incorrect, uh, beans, dam, your mushrooms, the middle, and then, uh, back bacon around the top and then fried eggs. The fries can be very hungry. And by the way, because due to pareidolia, that for sure looks like my mother-in-law, but I'm just saying, I don't know. I could be wrong, but, uh, those look like the eyes, the smile, the nose that could be, I'm just saying, uh, anyway, join the club.

[02:46:49] We'd love to have you twit.tv slash club twit. You also get special programming. There's a lot of it, including that AI user group that the, you know, we don't have sponsors for, but the club makes that possible. So we are very grateful, uh, to you. Thank you everybody for being here. We do twit every Sunday afternoon, two to five Pacific, five to eight Eastern 21, excuse me, 2100 UTC. I mentioned that cause you can watch us live. Uh, we stream it not only in

[02:47:15] the club to a discord, but also YouTube, twitch x.com, Facebook, LinkedIn, and kick. Uh, if you want to watch live, we love it to chat with us live as we're doing the show. Uh, I appreciate all the chat comments. It's very helpful after the fact on demand versions of the show are on our website, twit.tv. There's a YouTube channel for all of our shows dedicated each of the shows. So if you want the video there, it's a good way to share clips and you can watch it. I think more and more people are what YouTube's becoming people's TV, put it on your big screen, uh, watch the show

[02:47:44] from there. Or, you know, it's a podcast so you can subscribe. You can even subscribe on Spotify. If you want, just go to your favorite podcast client and look for twit, download it. Whenever you feel like listening, you can, uh, we appreciate it. Uh, however you listen or watch, we're very glad you're here and we will see you next week. And as I've said now for 21 years, thanks for joining us.

[02:48:05] We'll see you next time. Another twit is in the can. Bye-bye. Hey, it's Ryan Reynolds here for Mint Mobile. Now I was looking for fun ways to tell you that Mint's offer of unlimited premium wireless for $15 a month is back. So I thought it would be fun if we made $15 bills, but it turns out

[02:48:33] that's very illegal. Uh, so there goes my big idea for the commercial. Give it a try at mintmobile.com slash switch. Upfront payment of $45 for three months, $90 for six months, or $180 for 12 month plan required $15 per month equivalent taxes and fees extra initial plan term only greater than 50 gigabytes may slow when network is busy. See terms still waiting in line again. That's time. You'll never get back. Save time and money with stamps.com

[02:48:57] over 4 million businesses have skipped the line with stamps.com. Join them to save up to 90% off carrier rates from your computer or phone right now. Print postage for certified mail, registered mail and packages in seconds. Then schedule a pickup right from your home or office for a limited time. Go to stamps.com and use code podcast for a free welcome gift taxes and fees apply.