Loading summary
A
Grumpy Old Geeks, a weekly talk show hosted by Brian Schulmeister and Jason DeFilippo discussing the finer points of what went wrong on the Internet and who's to blame. Welcome to Grumpy Old geeks. I'm Jason DeFilippo.
B
And I'm Brian Schoellmeister. Happy Canada Day, Jason.
A
Oh, Canada, oh, Canada. I don't know how that song goes, but close enough.
B
That sounds more like a Christmas tree,
A
but, you know, it's kind of otan and baum.
B
Normally that would be appropriate for the weather, but. Oh, my God, it is, it's. It's coming up on 7pm Here it is still 97 degrees outside. It's. It's roasting up here.
A
Let me check. Carrot, let me. Carrot's telling me to off. Oh, no, it's saying. It's saying it's partly cloudy on carrot and it's 77 degrees here in Los Angeles.
B
I miss Los Angeles. We're also under a severe thunderstorm warning, so if, you know, the house explodes behind me, you know why.
A
Okay. Those Canadian tornadoes coming to get you maybe.
B
I definitely felt like tornado weather when I walked outside a bit earlier, but hopefully not. It being Canada day. I do have an update on my Canadian citizenship. Citizenship journey as well.
A
Oh, okay.
B
I took the citizenship test today. I was finally approved enough in the process to go take the citizenship test. I actually studied for. For 10 minutes.
A
And did you pass?
B
I've always been very good at taking tests. 20 out of 20.
A
All right, just 20. It only takes 20 questions to become a Canadian citizen.
B
Well, there's an awful lot of paperwork and tests and other things involved, but that is the only. The citizenship test is a 20 question test. Yes.
A
Oh, that's pretty sad.
B
And you could do it online. I didn't have to go anywhere. It was great.
A
So you had chatgpt.
B
Well, the honor system, you know, we're Canadians. We don't lie.
A
Oh, we'll get to the honor system in a minute. We'll get right to that in a minute.
B
But.
A
Well, congratulations. I'm glad you.
B
I fully expect with Canadian bureaucracy being what it is, about three years from now, I'll be in citizen.
A
All right. And in four years from now, you're going to adopt me so I can move to Canada and get my teeth fixed. It'll be great. Well, we're going to move on here with some interesting news from Brian Brushwood. I love this guy. He's been a friend for years. I was one of the. On one of the Original scam schools. He's doing a daily little news show now where he like plays an old scam school, then talks about an item in the news. And this one caught my eye because I'm like this, this, this hits home, home to me. No skill hacker uses Claude and codex to hit 14 sites. So I'm like no skill hacker. I like that. So a cybersecurity firm analyzed the complete working directory of a novice attacker who used anthropic squad code and OpenAI's agents to breach 14 organizations. According to OA Labs, the hacker relied on vague low skill prompts while the AI handled reconnaissance, identified vulnerabilities, generated exploit code, validated access and harvested data. Well, researchers say it's another example of generative AI dramatically lowering the technical barrier to cybercrime. The attacker made a classic rookie mistake by running the operation through a third party server.
B
So why didn't Claude warn him?
A
Why did not Claude warn him? Why didn't Claude say what? He didn't ask obviously what a VPN was, or how to roll your own or at least bounce your IP and mask your traces and all this stuff. So when the provider noticed malicious activity, it captured the entire working directory, including more than a thousand AI sessions. The attacker's prompts, policy violations, IP address, LinkedIn profile, resume and location in Addis Ababa. So to paraphrase the great Michael Bolton in the masterpiece Office Space, this kid was a no talent AI clown. What an idiot.
B
I mean, you know, no talent clown. It's obviously not going to cover all the bases and it just proves to me just how wildly insane and wild west all of this AI stuff is. All the talk about guardrails and protections and it all just goes out the fucking window randomly for no reason whatsoever. Like it's, it's crazy.
A
Yep, yep. And it reminds me of in, in Hackers when Joey hacked the bank across state lines. In serial killer goes, it's universally stupid, man. That's what this kid did. Universally stupid.
B
Yeah.
A
Yeah. Well, my adventures in no skill coding and using Codex in Opus. This week I. I've been working on building a best in class cms. And I'm going to tell you right now, I posted a video on our discord about this guy. He made a fantastic video about trying to get his, his AI to cut a sandwich in half and it's fucking hilarious. And how just it's impossible to get it to do exactly the thing that you want. Over the past two weeks, I've probably spent about 40 hours dicking around with this thing, trying to get it to do the things that it needs to do to be a first class cms. And then I just gave up. I'm just like, this is a waste of fucking time. It is terribly difficult because when you, when you think about things that need. Well, Brian, you and I have worked in major corporations. You've worked in the music business, I've worked in the movie business, like those movies behind me. We worked with things called lawyers and compliance officers and things like that where a single pixel cannot be out of place. You try and get a fucking AI to get exactly what you need it to do without knowing what the fuck you're doing. It doesn't happen.
B
But all we hear is how all these AI tools are ending the need for programmers. Jason, how could this be? Either they're all shipping slop that barely works and is like MacGyvered together with patches that were made by actual humans, or it's all just a fucking joke, or both.
A
What I think is happening is these kids are really good at building an mvp. A minimum viable product, or a minimally, minimally vibed product or minimally, minimally vibed piece of shit. Because none of this shit is ever going to make it to production or it's their hobbies, it's just hobbies, you know, and if these kids are working inside of a corporation, well, they're going to find out real quick with the, the token minimization act coming along that there's just no way to get this shit done without knowing what the fuck you're doing. And that goddamn Instagram post keeps coming back to haunt me. I looked at my, my comments are so vicious and, and I still laugh at them because I'm like, you kids still don't know what the fuck you're talking about. If you don't know how to code and you're trying to build a program. It's like, it's just mind boggling how short sighted these little shits are and are not actually figuring out that what I'm trying to tell them is actually good advice moving forward if they want to have a, like a career in the business. You have to know how to code. All of the, all of the old school coders that have reached out to me and like, yeah, you're right. You're absolutely right. I'm like, thank you. At least somebody fucking gets it. They're the old guard and people are like, oh well, you know, you guys shouldn't make a video. You should write it on a cave wall because you're that fucking like old man. It's like, come on.
B
See that what's interests me about that post is not, is not the content, it's not what you said, it's not the discussions taking place. And then what's interesting to me about that post is it is wildly outperform anything else that we've ever done. And I'm convinced, I am 100% convinced it's because of the vitriol and rage that's in the comments. That's what gets it promoted, that's what gets it pushed up in the algorithm. That's why it keeps appearing in other people's feeds. That's the reason. And that these companies for that.
A
Yeah, 100% man. 100%. If everybody just agreed with me and patted me on the back and said good on you mate, it would have not gotten anywhere near what it's gotten. And you know, it's just keeps happening. People find it and then they, because it just keeps getting re swizzled up because of the hatred that all the little shits keep doing. I just wish any of that would translate into actual people watching the show that we're making right now.
B
Yeah, that would actually make social networks somewhat useful. But they're not.
A
No.
B
And they don't.
A
No. Well, let's move on to a little more, a little more fun news. Ford says it has hired 350 veteran engineers after discovering that its AI driven quality control systems weren't catching enough problems before parts reached the assembly line. Shocker executives admitted the company leaned too heavily on automated inspection and expected AI to turn engineering requirements into high quality products. Instead, Ford brought back experienced specialists, many former employees and supplier veterans to identify failure points before production.
B
I'll tell you what I, I, I want, I want there to be a law and this will never happen because you know, we can't actually pass laws that are relevant to any technology that comes out within the first 15 years apparently. I would like a law that lets us know when we buy any product, particularly a car. Was AI used in creating this and if so what specifically? Because I do not want to buy a 50,000 ton piece of steel that travels along at 100 miles per hour that was vibe coded.
A
Well, you know, yeah, it could be like a nutrition label for your car, you know, 23% I AI, you know, whatever. I don't think it's going to happen, but.
B
No, yeah, but it'd be nice because I really don't want to buy AI created products. I just don't. Yeah, but except for grumpy old geek shirts.
A
Yeah, look at this.
B
Get those.
A
That's what I'm saying. Right there, right there. Well, the so called gray beard engineers are now mentoring younger staff while helping retrain the company's AI systems. Come on, guys, what are you doing? Ford says the strategy is already paying off, projecting roughly $1 billion in savings this year and earning the top ranking among mainstream brands in, in the latest JD Power initial quality survey. So, Brian, after spending millions trying to replace decades of experience with a chatbot, Ford finally rediscovered the radical innovation known as hiring people who know what the fuck they're doing.
B
Oh, how the tables have turned. And I think we're going to see this a lot more often because all we've been doing for the last almost two years is reporting about companies cutting jobs because of AI and to save money because of AI. And now we're finding out, oh, this major company that has real products hired people back to save money because the AI sucks.
A
Yep. Do you get a free NFT when they fire you for, you know, being fired for AI? I think that would be appropriate. Here's, here's a bored ape to look at while you're sitting, you know, at home waiting for us to call you back to rehire you.
B
It has the attached smart contract termination contract. Yeah.
A
Oh, yeah.
B
Yes. Well, Ford may be rehiring people, but Volkswagen, the auto giant that leads Europe in electric vehicle production, plans to cut 100,000 jobs, or 15% of its global workforce at German plants over the coming years. The plan represents the automaker's largest overhaul in its 89 year history. Refreshingly, they said that AI has nothing to do with this. AI was not even mentioned.
A
Awesome.
B
It's pretty much just Trump's tariffs and Chinese car sales that are, that are killing them right now. So they gotta, they gotta cut back a little bit and save some money and maybe make a better looking electronic car because their electric vehicles are kind of butt ass ugly.
A
Yeah, that new micro van failed miserably.
B
It's, it's so ugly. One of my neighbors has one in orange.
A
Yes, Orange one.
B
Anytime they park that in front of our house, I just go, okay, goddamn popsicles. Back in front of our house again. I hate that car. I hate looking at it. It's so ugly.
A
Now look, I have an or, I have a very bright orange jeep and mine is beautiful to look at. That, that, that micro bus is just hideous.
B
Yeah, that thing would be ugly in any color. But it's particularly bad in orange.
A
Yes, it is. Well, a Growing Reddit community r almost homeless is becoming a lifeline for people facing eviction, unemployment and financial collapse, highlighting how quickly housing insecurity can affect people who once had stable careers and then got fired because their company said that AI was taking their job. Members include software developers, remote workers, and others who have submitted hundreds or even thousands of job applications without success sharing practical advice, emotional support, and survival strategies. Experts say the rise in homelessness is driven less by increases in mental illness or substance abuse and than by housing costs that continue to outpace incomes.
B
We've been talking about this for a long time. We've known that the lower middle classes have not been saving money. There's tons of reports about it. There's, there's no savings and when you have no safety net like in the US and rising healthcare costs, rising gas costs, rising food costs, all a lot of people are one paycheck, maybe two if they're lucky, away from, from not being able to pay their rent, defaulting on their mortgages, being homeless. That's where we're at right now. And if you lose your job and there's no jobs out there right now, you're screwed.
A
Yeah, yeah. No, I mean, like, seriously, when I have to make a decision at the, at the end of every month, it's like, do I pay for my health insurance or do I pay for my car and my car insurance? I always pick the car and the car insurance because the health insurance is only going to put me up after I fall over from a mental collapse for at least two days. But the truck I can sleep in if I have to.
B
Also, if you get a job, it drives you there.
A
Yeah, yeah, that too. That too. The community also pushes back against stereotypes that portray unhoused people as fundamentally different from everyone else, emphasizing how layoffs, medical debt, disability, or simply a run of bad luck can rapidly unravel financial stability, which is what we're seeing over and over again. It is. Wealth inequality continues to widen. Online communities like this are increasingly filling gaps left by strained social safety nets sense. I, I, I'm a member of this group now because it's, I, I went and checked it out for this article and it is like, scary, man. It's fucking scary. A lot of people that are like, you know, high level people who used to be lawyers and just are not being able to make the ends meet. And on that note, I did get my California food handler's license that I talked about last week. I know a lot about food poisoning and how to avoid it, and it only cost me $9.99 but now I can go to into any commercial kitchen in California. So you know, it's just one of those things that I'm just, I'm, I'm, I'm racking up these things in case, just in case. Because, you know, we're podcasters. Come on, this is. There's no guarantee that there will be an audience next week.
B
I know. And even our rage baiting social media posts are not getting us more listeners.
A
So you know, I tell you man, I tell you what the this episode is sponsored by storyblocks. If you've been listening to grumpy old geeks for any length of time, you know we're constantly producing content between the show, clips, social posts and all the other projects we work on. We need assets that are high quality, legally cleared and ready to drop into a project without wasting half the day. That's why we keep coming back to Storyblocks. Storyblocks is a 100% human human made stock media library built for creators. Every asset comes from real filmmakers and artists, not AI generated content. And everything is professionally curated, pre licensed and ready to use. I can grab footage, music, sound effects, templates or images without worrying about licensing. Headaches and unlimited downloads mean I can experiment until I find exactly what works. They've also built tools that speed up the workflow even more, like AI powered search and editing features that help you customize content while still supporting the artists who created the original assets. Head to storyblocks.com gog to access the human made stock media library that's essential to our workflow. For a limited time, they're offering 15% off any annual plan and that discount is only available through our link. Again, that's storyblocks.com gog for 15% off
B
annual plans While Tesla has settled yet another lawsuit that involves a death related to one of its driver assistance systems. As first reported by Bloomberg, the lawsuit, which revolves around the death of 71 year old Joanna Story in Arizona, has been resolved, but terms of the settlement weren't disclosed. Story's family filed a lawsuit against Tesla in 2023 following an incident where the grandmother stepped out of her car to direct traffic around a separate collision that was caused by sun glare. She was then hit and killed by a Model Y that was using full self driving. Notably, this case was the first reported incident of a pedestrian fatality related to Tesla's automated driving technology. Besides the lawsuit, the crash led to a federal investigation from the National Highway Traffic Safety Administration, which examined how full self driving Operates in poor visibility conditions. Not well, as we know.
A
Yep. No lidar. Come on. Illinois has become the latest battleground in the fight over prediction markets after the state passed a law treating platforms like Kalshi as unlicensed sports betting operators. Starting July 1, Illinois would require licenses costing up to $15 million and impose new taxes on sports event contracts. Well, Kelshi sued, arguing its markets are federally regulated financial derivatives overseen exclusively by the cftc. Not gambling. Subjected state rules talked about this ad nauseam. The broader dispute hinges on whether betting on the outcome of a game is really a financial hedge or just sports gambling with a different legal wrapper. More than 40 states have sided against Kalshee while the CFTC is pushing for broader federal authority and accepting public comments through July 27th. Well, so maybe, Brian, your final fantasy league can file for SEC approval.
B
I. I just love it. It's. It's a bet. You're betting. It's. It's so ridiculous to call it not a bet. Unbelievable. Anyways, as I've mentioned a couple times, I. I have been watching a ton of World Cup. I've tried to catch as many games as I can, so the TV is at least on in the background even if I'm not actively sitting there and watching the game. So I can see it. And I can tell you I have. I have used a pirate stick to. To enjoy some of the US coverage, the Fox coverage, and I also have the Canadian coverage that I subscribe to through TSN Sports Network in Both are inundated. Inundated with ads for gambling.
A
Oh yeah, of course.
B
Every. Every ad is one of the apps for sports bet, betting and gambling. It's unbelievable. It's. It's actually sickening that this is happening. I have been beating the drum about this for at least 10 years. It's gotten worse and worse and worse. It's. It's unbelievable. And it isn't, of course, just the normal people that admit that they're actually taking bets and it's gambling. It's also the poly markets and it's also the cow. It's cowshis. There are ads for all of them. And I don't know if you've seen this one. I know you haven't really been watching much of the World cup, but guess what? One of my favorite assholes is doing a commercial for polymarket. Rick Rubin.
A
Make it sound like blue. I think it's my favorite comic about him.
B
God, this guy, he drives me nuts. If you don't know who Rick Rubin is, he is unfortunately a Very famous music producer who basically does nothing. He just sits in a corner and spouts Rubinisms and somehow keeps getting hired by everybody. And everybody thinks he's a God and he's just insane. But of course, because he has no soul and no humanity and no common decency. Have you no decency, sir? He dug a, he did a commercial for Polymarket. He says, if you could ask one question, what would you ask? Which is also the same thing he told the Green Day when he was producing them.
A
Exactly. And every, every other fucking artist. Now look, I, I read Rick's book on creativity and it is actually a pretty good book. But I, I see what you're getting at though. He's, he's kind of like the Kara Swisher of music production. He's like, tell me what you think it sounds like.
B
I think your bass needs to sound more green. And then he just fucking sits in the corner to big cloud of fucking weed smoke and then he gets a fucking million dollars. Unbelievable. Are you not surprised that that's not the only Internet age technology tool that Rick Rubin has endorsed? He is also a big fan of AI Vibe coding because he, AI Vibe produces.
A
Are you serious?
B
I don't know if he uses AI for his producing. That would actually be a step up. He would actually be doing more work than he actually does now if he did.
A
That's what I'm saying. Just like you know how hard it is to write a good prompt. I'm like, that sounds way his comfort zone. Oh God. Well, you know, get the paycheck. Get the paycheck.
B
You know what? He's sitting in his multi million dollar Malibu mansion and I am not so obvious.
A
I am in a garage and you're in an addict.
B
He is doing the right thing and I am dumb.
A
Well, speaking of dumb people, Brown University economist Roberto Serrano says he may have uncovered the biggest AI cheating scandal in Ivy League history. After an advanced mathematical economics take home exam produced for 40, count them, 40 perfect scores out of 86 students and an eyebrow raising class average of 96.
B
Oh God, I hope everybody that took that Canadian citizenship test at the same time I did, didn't get 20 out of 20. Or else I'm screwed. They're going to come back and tell me I cheated.
A
Yep. When the same students took an in person final, the average crashed to 48. Literally half of the 96.
B
Yeah, and that's probably why you should be having in person tests.
A
Yep. Serrano found answers that closely matched ChatGPT outputs and called the evidence Overwhelming. Overwhelming, sir. The incident adds to growing concerns that AI is eroding critical thinking, turning professors into plagiarism detectives instead of educators. This week on the plagiarism detectives on A E. Princeton has already abandoned a 133 year old honor system tradition because of similar concerns. Brian set Anna to zero.
B
Unbelievable.
A
Oh, no. Brian. Absolutely fucking believable. I'm sorry.
B
That's true.
A
Absolutely believable.
B
Absolutely believable. I mean, you know, back in. Let's be honest, if we could have cheated that easily back in the university days, we would have to.
A
I would have gotten a degree. I quit after English 101. I'm like, fuck this, I got a D. So I'm like, why am I going? I don't need to take English. I'm a photographer, God damn it. I'm an artiste.
B
Artiste. All right, well, let's go. Let's move down under for a second. The Australian Competition and Consumer Commission, the acc. You down with acc, you know me. Has sued Amazon Australia for allegedly rolling out ads but then unfairly changing the cost of avoiding them. The ACC website claims that from 2023 to 2025, Amazon price contracts were unfair in that Amazon could unilaterally make negative changes without any recourse for the customer. So when Amazon prime rolled out ads in 2024, the suit said that the version of the contract enabled Amazon to unfairly force ad free customers onto a more expensive tier if they wanted to stay ad free. Meanwhile, anonymous reports are strongly hinting at a big case against Amazon's marketplace ads from the US Federal Trade Commission. That still in progress suit would if real allege that Amazon made its ad auctions deceptively opaque. What? That's the first I've heard of this. Keeping sellers unaware of how much competition there is for ad space on a given search and tempting them to pay more than a fair price for ads. It is worth noting that Amazon made over $68 billion in ad revenue last year. The five big tech firms, Amazon Meta Alphabet, Microsoft and TikTok, reportedly rake in two thirds of all ad revenue now, according to a Moffett Nathanson report from last year. Also in 2025, Amazon was the largest advertiser in the world by amount spent, according to Ad Age. A little crazy.
A
That is crazy.
B
Staying down under. The social media checks implemented in Australia after the country banned their use of for teens under 16 have shown little evidence of being effective, according to a study by the University of Newcastle published in the British Medical Journal. The study surveyed participants between 12 and 17 years old before and three months after the law was introduced. It specifically looked at the participants use of TikTok X, Facebook, Instagram, YouTube and Snapchat. Based on the information they gathered, more than 85% of teens under 16 continued using those social media apps despite 2/3 of them reporting that they had encountered age checks. Approximately 48, 54 to 68% of responders under 16 just kept on using their accounts. How you ask? Well the most common age check the Australian teens encountered was to self declare their age. A method criticized by authorities in the country as well as anyone with a fucking brain cell to rub together. Among the responders, 24 to 39% encountered self declared age verification while 13 to 27% got through checks by uploading a selfie which they then just drew beards
A
on there from mustache. This is ridiculous. Like Ms. Paint.
B
That said, the study also showed that affected teenagers had found other ways to keep using social media. Around 15 to 19% of the responders said they used fake accounts to access the platforms while 9 to 29% reported going on social media using someone else's account. The smart 11% of the teens said they used private browsers to get around the restrictions. There were a few teens who also reported using a vpn. They are the geniuses.
A
Yeah, seriously.
B
So I mean this just goes to show you where everybody's screaming these aren't effective. Well they're not effective because the social media companies are basically just saying hey you old enough to use this? And check yes or check no. Yeah, I mean what the fuck. Relatedly, in a press release, the Australian government announced that it will double the maximum penalty for any social media companies breaking its minimum age law from 49.5% million to 99 million Australian dollars or more than 68 million US dollars. It's clear big tech are not doing enough to comply with the law. Anthony Albin is the country's Prime Minister said you think these changes reflect the seriousness with which we take any failure by social media companies to comply with our world leading law? So along with the new penalty threshold, the Australian government is granting its esafety commissioner Julia Grant more enforcement power she can demand Social media companies provide evidence of how they're stopping children under 16 years old from starting an account. You see the form? Are you over 16? Yes. No. What more do you want from us?
A
Yeah, we're, we, we put it there for them to, you know, come on.
B
Yep. So they are actively investigating potential non compliance with Facebook, Instagram, Snapchat, TikTok and YouTube.
A
There's not much to investigate. It's already there. Right in front of you.
B
Yeah, it's right there. You can take the screenshots of the forms. A study from researchers at New York University and Northeastern University is claiming that at least half of the safety features meant to protect children on social media platforms do not deliver on their promises. The study, published by Heat Initiative and Cybersafety Research center, tested 86 features across Instagram, Snapchat, TikTok and YouTube. In the study, researchers created dummy accounts to mimic children of various ages. In addition to some adults accounts, the study examined three different scenarios. One where a child was using the social media platform naturally. Another where a teen was trying to circumvent a safety feature. Yes, I'm over 16. And a final case where a malicious adult actor would try to bypass the protection features of a separate teen account. According to the analysis, each of the social media platforms had a failure rate of at least 50% when it came to its advertised protective features, like those that prevent adults from messaging children or stop underage accounts from accessing harmful content. Spokespeople for Snap, Meta and YouTube contested the study's findings and statements to the New York Times, which had reported that it was able to replicate all of the study's findings. Okay, so they're doing fuck all.
A
Yeah, exactly. Yeah, I'm sure. Well, you know, they're. It's what you get when you're vibe coding. Yeah, it's like, dear AI, make it so kids can't get in. Okay, how about a checkbox? Perfect, we're done.
B
Well, I think it's even more malicious than that. It's. How do we make sure we have users in the next five years? How do we keep this, this gravy train running? We got to get the kids hooked early, so let's not work very hard at stopping them from getting on.
A
Yeah, what's the bare minimum that we can do? And by the way, what's the ROI versus the actual account or the cost of the fine? Is that, you know, how does that math work? You know, is it going to. Are we going to make more money if these kids sign up in five years and we just pay the fine now, or how does that work? I'm sure that there's somebody in there with a calculator doing that math.
B
Chat cheap. He did that math for someone.
A
AI companies are running to a problem they arguably create themselves. Brian. As the supply of high quality training data shrinks, many firms are paying contractors to generate funds. Fresh Examples for increasingly specialized AI tasks.
B
Hang on a second. How about you go back and pay the people who you stole their content from first?
A
No, no, we don't have to. We don't have to do that. No.
B
All right. Screw those guys. Screw all the people that we already took all the data. The data from. Yeah.
A
You mean every human that had ever lived before two years ago?
B
Yes, pretty much, yep.
A
Well, according to interviews with contractors, many of those workers are quietly using AI chatbots to produce the very data that's supposed to improve other AI chatbots.
B
The Ouroboros is real, Jason.
A
It is feeding itself. It is. Companies have rules against the practice and try to detect it, but workers say it's easy to avoid detection if you know how to clean up. The obvious AI tells.
B
Did you use any AI to generate the data you were feeding us?
A
Yes, no, or no, no, no? Many also say they're driven by low pay, temporary contracts, and the constant fear of losing work over small mistakes. So after spending years vacuum vacuuming up the Internet to train AI, the industry has apparently invented a closed loop ecosystem where underpaid humans use AI to manufacture synthetic data for other AI. So, model collapse in 3, 2, 1.
B
Actually, if you asked, if you asked ChatGPT, it would be model collapse in 3, 6. 1943.
A
Oh, God.
B
Well, California has launched a new portal which tracks AI related job losses in the state. According to the office of California Governor Gavin Newsom, it's meant to serve as an early warning system for widespread job cuts due to artificial intelligence, allowing the government to proactively determine where interventions may be needed the most. The website says Newsom's office worked with the California Employment Development Department as well as with the California Policy Lab at the University of California to conduct research to measure AI related job losses. They used unemployment insurance claims data combined with AI exposure measures to come up with the figures in the tracker. Anybody can see the data for themselves and the tracker will be updated monthly. As Bloomberg notes, the tracker was built during a time when there's increasing pressure on authorities to be more proactive when it comes to AI driven job loss. And for politicians, Keen is being seen as the champion of the masses against AI, especially if you happen to be widely rumored to be running for president.
A
Yeah, no shit. Did they vibe code this new portal, Brian?
B
They don't say how it was built. I'm hoping that they didn't fire people and have to put themselves on their own tracker to get this built. Get this out there. So this will let you see potential AI exposure by different groups such as age, education, gender, industry, race and ethnicity, as well as by region. For instance, people in the 25 to 35 age group seem to be most vulnerable to AI related layoffs and females more so than males. So. Interesting. I will be checking out this website and seeing what data comes up. I think it's, it's a great idea and it's pretty fascinating.
A
Yeah, no, we'll check it out. You know, I like transparency.
B
Damn it.
A
Hyundai factory workers in South Korea have voted to authorize a strike as the company ramps up plans to deploy humanoid robots across its manufacturing operations.
B
I thought they were going to strike because they were protesting their team getting knocked out of the World cup so damn early.
A
Could be, could be. But not this time. Not this time.
B
All right.
A
Union members want a formal role in decisions about automation and AI, arguing that increasingly capable robots pose a real threat to long term job security. They're also demanding hefty performance bonuses pointing to generous AI fueled payouts negotiated by workers at Samsung. Hyundai maintains the robots will handle dangerous and repetitive work rather than replace people outright. But the union is rightfully calling bullshit. The company has already announced plans to introduce Boston Dynamics Atlas robots with ambitions to deploy more than 25, 000 humanoids across its factories. And I'm sure that there is an Atlas Shrugged joke in there somewhere, Brian, but I just couldn't find it.
B
Oh man, if Ayn Rand was alive, she would have the biggest hard on right now about all this stuff. Everybody liked you in college, then they went out into the real world. Your shit don't work. Your shit don't work.
A
That's right. Google has reportedly put limits on Meta's access to its Gemini AI models after Meta requested more computing capacity than Google could provide. The shortage delayed some of Meta's internal AI projects, while other Google customers also experienced capacity constraints, though on a smaller scale. In response, Meta has reportedly told employees to be more efficient with AI tokens. Even after spending billions on GPUs and data centers, the biggest companies still can't buy enough compute to satisfy demand. And Google has already acknowledged the capacity shortages have limited cloud growth and contributed to a growing backlog.
B
Or you could stop shoving AI into all of your goddamn products, especially the ones that absolutely don't need them, because if it's there, people will use it and then use the capacity.
A
Well, this, this, this tale gets a little more tricky, Brian, as we go along. It does, it does, and it gets, it gets kind of funny here. Meta is reportedly preparing to launch a cloud computing Business renting out AI, compute power and hosting AI models in a move that puts it in direct competition with aws, Google Cloud and Microsoft Azure. The strategy follows similar deals from Xai, which has already leased data center capacity to companies like Anthropic Google and Reflection AI. Now Meta has committed nearly $183 billion to AI infrastructure, including enormous facilities in Louisiana and Ohio, despite little evidence that Meta AI or its llama models are generating meaningful standalone revenue because nobody likes them, because they fucking suck. So, Brian, let's put this in perspective. Google cuts off Meta because they use too much Gemini and can't handle the load. Meta has too much capacity because no one even themselves uses Meta AI. So they're going to lease Compute instead of running their own shitty AI because they saw that SpaceX was making some money because Grok is shit. They have capacity and said, yeah, that looks like a better business model. Do you see? Do you see how this fucking works?
B
I do, I do. Do they? I don't think they do. I'm pretty sure they don't see it.
A
So, yeah, turn off that shitty fucking Instagram and Facebook AI that nobody fucking wants it.
B
Nobody wants. I've only hit the button by mistake and then I get the stupid messenger thing with my AI fucking report to the question I never asked because I hit a button by mistake. Thank you.
A
Yeah, and then I get the notification that says Brian Schulmeister has decided to use Facebook AI or Meta AI. Do you want to too? Because you know, peer pressure. And I'm like, fuck, no. Brian probably mistakenly hit the button because I know he's not that fucking stupid. And you guys use dark pattern designs, so you probably put it where it used to say, okay, but fuck those guys. It's ridiculous the fact that Meta is using Gemini Gemini AI that much that they're getting shut off. Why are they using Gemini AI that much when they have llama?
B
Because they're.
A
The fuck are they doing?
B
They know not to use their own model because it's horrible.
A
It's horrible.
B
Yes.
A
I heard there's a sale on Grok down the street. Get your tokens, get your tokens. Come on.
B
I, I, I'm, I think there's the. I, I saw that Meta placed a big order for generators from Amazon so they can start up their own data centers. Since, you know, Grok is running on coal, they're going to. Meta's going to run on gas. This is just going to be great.
A
Soylent Green is next, Brian. Soylent Green is next.
B
I mean, basically, we're heading towards the matrix. We're all going to be plugged into fucking batteries to power the AI that nobody's using.
A
Yep. Well, Oracle is telling investors that the AI infrastructure boom comes with an equally massive list of ways that it could go sideways. In its latest annual filing, the company warned that delays, power shortages, GPU shortages, supply chain disruptions, environmental regulations, cybersecurity issues, customer defaults, legal liability, and wait for it, weak AI adoption could all undermine its multi billion dollar data center expansion.
B
I just, I asked ChatGPT to count up how many, how many things were in that list and it said 792.
A
So perfect.
B
Yeah, perfect.
A
You should have tried Grok. It had been one bajillion and boobies.
B
That's what Grok would respond with.
A
And space Nazis. Oracle spent nearly $56 billion on capex last year and plans to spend as much as 95 billion. Doll investors are clearly getting nervous. Brian, I looked this up myself. Oracle shares have cratered from a June 1 high of $248.15 to when I checked it this afternoon, $144.30 as AI concerns over spending increase.
B
You and I have lived through a few tech crashes. The initial dot com crash. We've been through all of this before. This AI crash is going to be massive. It's just.
A
Why do you think I'm getting my food service license? Shit. Done. The AI data center boom has created an unexpected growth industry, Brian. Organized cargo theft. As billions pour into massive facilities packed with high end servers, networking gear, cooling systems and mountains of copper wiring, thieves are targeting shipments with increased precision. Authorities recently recovered two stolen trailers carrying $1.3 million in data center supplies, including $300,000 worth of copper wire after they disappeared from Alabama and Florida in what investigators believe was a coordinated operation.
B
Trucks are pulling up at the back of Grok and Meta right now. You need any of them? Data center supplies is there. Like back in college when the van would come down and they would open up the doors like, anybody need some speakers?
A
Speakers? I was going to say they're selling speakers out of the back of the van.
B
I bought a pair of those. They were pretty good actually. I'm sure they were stolen, but.
A
Yeah, of course, of course. Another theft made off with nearly $5 million in copper and electronics. The Department of Homeland Security estimates cargo theft already costs about $35 billion annually, while cargo net says supply chain crime jumped 60% last year. So. Brian, not all thieves are in the C suite.
B
You know, it's good to see the Little guy getting in on the action.
A
I'm telling you. I'm telling you. God, I need to go get my weapons license. Forget this chef. I need to go get like an M16 and some, some, some grenades and a bandolier and let's go. What you're going to see now is it's going to be like the old days with the stage coaches. You're going to have the guy up top riding the shotgun as they're cooking along like the i5 going, I'm going to the data center. You better not pull us over there.
B
Nice.
A
Let's get to some good news here. Okay, first, a little bit of good news before I end on a little
B
bit more bad news.
A
NASA is preparing to attempt its first ever robotic rescue of an orbiting space telescope with a mission to extend the life of the Neil Garrels Swift Observatory. Swift, launched in 2004, studies the universe across visible ultralight X ray and gamma ray wavelengths. But increasing atmospheric drag caused by recent solar maximum activity has accelerated its orbital decay. NASA awarded a $30 million contract to Catalyst with a K Space, whose Link Servicing spacecraft is scheduled to launch aboard a Pegasus XL rocket. I don't know if you've seen the Pegasus XL rockets, Brian. They're really fucking cool. They're horizontally launched from a plane and then they zoom up. They're really cool. They're really cool. Check them out. If successful, the robotic vehicle will rendezvous with Swift, capture it, and gradually boost it into a higher, more stable orbit over the coming months. Now, beyond preserving a valuable scientific observatory, NASA views the mission as a proving ground for satellite servicing technology that could eventually be used to extend the operational lives of other spacecraft, including the Hubble, which is pretty cool.
B
That's pretty good. And then I heard just, just a few minutes ago that Elon came on and said, we're going to build mobile satellite servicing technologies for all of our, all of our, all of our positioning satellites that keep crashing to the ground and burning up in the atmosphere. We're going to put one over every single one of the mobile Tesla stations that we're also going to build. You're never going to build any of it.
A
So there's going to be giga space pit stops where the, where the satellite, the SpaceX satellites can pull in on their rounds and get like, there's going to be like space guys with suits, like, you know, changing the tires like an F1 on the SpaceX Sate, and then zip tip, done, go off on your way.
B
Power them all with nothing but coal. That's how it's going to work.
A
And ketamine work. Yes. Space ketamine. Delicious. Well, a new study from the European Southern Observatory proposes capping the number of satellites in Earth orbit at 100,000. To preserve astronomy, researchers modeled the effects of current and planned satellite constellations and found that beyond the. Beyond that level, telescope observations suffer unacceptable losses from bright streaks crossing their images. Today, roughly 14,000 satellites are in orbit, which I thought was. When I heard that, I'm like, that's not. I thought there were a lot more than that.
B
Yeah, me too.
A
So 14,000. And it's already causing problems with 14,000.
B
I mean, it wasn't SpaceX's plan to put just another 30,000 in the next couple years.
A
Oh, Brian, it gets much, it gets much worse than that. Proposals could push that to 1.7 million. SpaceX has discussed a long term vision of up to 1 million satellites where startup Reflect Orbital wants to deploy 50,000 mirror equipped satellites to provide artificial sunlight after dark. What could go wrong?
B
What's wrong with these people? All these fucking idiots. Like they've read all the same sci fi books that you and I read, but they got the exact wrong lesson out of it.
A
I know that stuff was bad.
B
Bad things happened. Not good things.
A
They were cautionary tales, not a fucking blueprint for your dystopian future, you dumb God.
B
I mean, this is all, it's all really horrible. But I mean, I had, I had to have a laugh at. SpaceX has discussed a long term vision of up to 1 million satellites. Well, I too have discussed a long term vision of up to 1 million satellites. I have about as much chance of ever getting a million satel there as SpaceX does. You can discuss it all you want. It's never going to happen.
A
Yeah, yeah, and here's the thing. Reflect Orbital, they, they literally want to put mirrors in space to put to point sunlight down at.
B
Have they not heard of lamps?
A
You can see a lamp. Your lamp might be pointed one way, but you can see it from everywhere.
B
It provides light. Why did I just Captain Kirk?
A
That.
B
I don't know.
A
I. I don't know, Jim.
B
Idiots.
A
I swear to God, I'm a doctor, not a podcaster. Both companies are awaiting FCC decisions. As astronomers warn the night sky itself is becoming an increasingly limited scientific resource. Let's move on to some more fun things.
B
Okay.
A
I watched a new Netflix series, I will find you with Sam Worthington and Brit Lauer this week.
B
Do they have a very special set of skills?
A
No. Well, I can guarantee you there's one special set of skills they didn't have. Screenwriting. It was the biggest new Netflix show of the year, which made me think, okay, I'll give it a shot. Yeah. It was an absolute steaming pile of dog shit.
B
Okay.
A
The problem was there was nothing else on to watch that week, so we watched all of the episodes. And I mean, granted, it makes the Night Agent look like Shakespeare, but, I mean, it was. It's terrible. These new Netflix series are awful. They're just awful. The. The screenwriting is so bad. The plot holes are just ginormous. It goes back to that thing where, you know, Netflix is putting out all those things where he's like, you have to explain the plot 17 times because everybody's dicking around on their phone and wanking it it to pornhub while they're trying to watch the show. So they're not going to get it unless you talk down to them like they're a potato.
B
That's the problem. Yeah.
A
Yeah. And that is exactly the problem with that show. But it is still the biggest new Netflix show of the year, which makes me sad. Makes me sad because it's not good. I did give it a thumbs down. I couldn't give it a double thumbs down. You can give a one thumbs up, a two thumbs up, but only one thumbs down. I wanted a big goat seed that I could press. That's what I want.
B
Two anuses down.
A
Exactly. Exactly. But Sugar is back for season two, which I'm happy about. I started to watch the first episode last night, and I'm digging it. I gotta finish it because I fell asleep because I've been sick. But it's good and it's nice to have. Have that back. I. I enjoyed that. I really enjoyed season one and season two. Looks pretty good. I've also been watching Bodkin, which is a 2004. I think it was a BBC series before it became a Netflix series. Brian, you gotta watch this. You'll love it. It's about a. It's about a true crime podcaster and it goes to Ireland to try and do a thing, but they've got the really hot, like, Irish girl journalist who's like, hates everybody's just like, everything. They got the researcher who's just like the comic relief, and they got the American guy who's just a dipshit, and it's just so good. And, you know, I'm an Irishop file. I've been to Ireland many times. I've been on Irish music cruises. So for me, it hits home but it is about podcasting the Irish and it is fucking excellent. You'll love it.
B
I will have to check that one out. Sounds pretty good.
A
Yeah, yeah, yeah. Put that on your. On your drunk plane reviews when you fly to LA in next week or whenever. The Traders New Zealand season three wrapped as well this week. Here's the thing. New Zealanders. I know, I know. We have a few Kiwis listening to the show. Get on your local Traders affiliates and tell them that they need more spectacle. Every other season that I've been through that is an English language season. When it's the end of the show, the last episode, there's at least a little spectacle. You know, they just didn't even really try. It was. It was kind of sad. It was a great season. I love the ending. But there's. They. They need to kind of have something a little razzmatazz. Razzmatazz. That's what they need. They need some razzmatazzi. But all in all, it was a great season. So I'm happy about that. I'm happy it's back.
B
Shiny ring. Shiny golden ring. Something precious.
A
It is New Zealand.
B
Come on.
A
Wet is down there. Get the fucking weta guys on it. You guys couldn't even make a flame turn red for sake. So come on. Yeah, get on it. But Australia starts in a few weeks, so that'll be good. Both of these have two new hosts because the New Zealand host had a new host. She was okay, not great. The new Australia host looks interesting. We'll see. Because, you know, I've talked about the Traders because I've been a fan since before everybody else was a fan, long, long ago. They're in 40 countries now. There are 40 active like productions of the Traders going on right now. That is insane. Ridiculous. And most of them are foreign language, which is awesome. But you know, it's really hard to watch that kind of show with, you know, subtitles. It's one of those things where I'm like, I wish it was dubbed, but then it still wouldn't be as good. So I just stick to the English versions. We've got. Here's the English versions. We've got US, Canada, UK, Australia and New Zealand. So there's five. Five out of the 40 are at least in English language. So check those out. But I love that show so much. It's just. It's fun. It's. That's my guilty pleasure. Brian, you watch World Cup, I watch the Traders.
B
Fair enough.
A
Do you have any desire to see the Odyssey? That's coming out on the 17th.
B
I. Controversial opinion. I don't give a fuck about any of Christopher Nolan's movies anymore. I stopped caring a long time ago. I don't care. I understand he's a master craftsman and everything. Nope. No interest in seeing this. No interest in seeing anything that Christopher Nolan does anymore. None.
A
When did you. When did you go off the Christopher Nolan train? Inception. Oh, yeah, that was. That was an easy one to come off.
B
Yep, yep. I was done. I was. I'm out. I never care again. Like I. I just don't care.
A
Yeah, for me it was Batman 3. Sadly, I thought that was not great.
B
That that didn't destroy it for me. But Inception was. Was my exit point. Did he do Interstellar too?
A
He may have been. I don't know. I gave up. Interstellar was a steaming pile of shit. But he did Tenant, right?
B
Yeah, he did Tenant. Oh, no, he did do an Interstellar. So there you go. I hated Interstellar. I hated Tenet. I hated Inception. I don't want to see anything by this guy anymore.
A
Yeah, I watched the new trailer that came out today and I was intrigued.
B
I was.
A
I was somewhat intrigued because I do like Matt Damon. I'm a fan of Matt Damon.
B
I love him in, In Captain America or whatever. Matt Damon.
A
Yeah. Yeah. See, I never watched that.
B
Team America.
A
Yeah, no, I'm just. I'm a big Bourne fan, so I thought he was great in that.
B
He was great and boring. I liked a lot. I liked the first couple of those movies before it just got a little repetitive.
A
Yeah, well, the first movie was the best. The second movie, the new director came in with the shaky cam and then it was just Blair Witch to me and I couldn't. Couldn't watch it because it was. And shaky for no reason whatsoever. No reason whatsoever. But anyway, I will watch it on streaming when it hits my 75 inch TV in my bedroom. Right, that's it. I'm not going to go to imax. It's sold out everywhere on imax. Yeah, so. And I'm sure it'll be an event, but I'm poor. I can't go to an IMAX movie, period.
B
Yeah, they're expensive.
A
It's ridiculously expensive. But what I will stay home and watch this week as well as Silo because it's coming out now. Brian. I put a link in here called Silo. How did we lose this World? It's a new website that they came out with. I was excited when I saw this because this is the shit that we used to do in the 90s.
B
Yep.
A
Look at it. Check it out. This is what we used to do. This was our bread and butter. Sites like this immersive websites about the IP that got you sucked into it.
B
Yep.
A
They did a really shitty job. Yeah, I'm sorry. They did a really shitty job.
B
It looks nice. And then you can tell. I. I would have loved to have gotten this job. It would have been. It must have been so fun to build. But. Yeah, it just.
A
Yeah, yeah, it's light. It's only like two layers deep.
B
Yeah.
A
You know, back in the day, we. I mean, we'd be working on this for like nine months or a year, and it would be like one of the deepest. It'd be like a fucking encyclopedia of Silo. This just like, you get one like schematic drawing of a PEZ dispenser, and that's just like. And. And basically clips and stills from the show and it's like. Like you guys didn't do any digging into the backstory, really.
B
There's no budget for this stuff, Jason. And you know. You know that there's a reason we're both sitting here podcasting and not building cool websites for musicians and IP and movies is because there's zero budget. Zero. Anyways, I'm super excited that Silo is back next week. It is the one show that I will break up my World cup viewing. It'll be the only thing I've watched, except for soccer, since the World cup started.
A
Yeah. Okay, well. And I think they're only releasing one episode, so you should be able to get back to your game pretty quick.
B
Yeah, well, the other thing I haven't been doing much is reading because lots of soccer. But I think I'm going to be quite upset the next time I go back to my iPad and fire up that Kindle app. Amazon is going all in on bringing artificial intelligence to your reading experience, adding several new smart features to its famed Kindle E Readers. Initially announced in June 2026, the conglomerate frames its AI add ons as making it easier to stay immersed in your books. We call that reading.
A
Reading, yeah.
B
By offering spoiler free recaps and AI assistants capable of bringing context to your reading experience. Again, we call that reading
A
Stay the fuck out of my books.
B
When combined with their Kindle's previous smart features, which allowed users to do everything from lookup definitions to transl. Foreign languages, the rollout is indicative of a publishing landscape searching for new ways to incorporate emerging technologies into your reading experience, whether you asked for it or not. Unfortunately, not every reader will have access to Kindle's AI infusion. Oh, please, please do tell me how I can not have this. Unfortunately, I need to have an old Kindle.
A
I just got rid of all my old Kindles because you didn't come and get one last time. I had one for you.
B
Yeah. So it's rolling all this crap out to newer Kindle devices and American iOS users. So I have to switch my iOS to Canada and see if that staves this off for a little while. However it's asked, this book AI chatbot will only be available on the US version of its iOS Kindle app for now. Kindles will be receiving the Ask this Book feature later this year, at which point I will throw this book out the fucking window.
A
It'll be like the end of Battlestar Galactica, your fucking iPad scattered into the wall.
B
Yeah, great. This is just what nobody wants. Nobody needs people that read like to read, not have AI explain shit to them about the book that they are reading.
A
We live on the theater of the mind, Brian. Yes, we do. Well, I got some good news for you then. Amazon's grip on the digital reading ecosystem just got a little weaker. Rakuten's Kobo E Readers now integrate directly with StoryGraph, letting readers automatically sync their progress, complet books, audiobooks, reading streaks, and detailed statistics without relying on Amazon's Goodreads. It's the first E reader to support storygraph natively addressing one of Goodreads biggest competitive advantages, seamless device integration. Founded in 2019, Goodreads is anything but seamless. No shit. No shit. So, founded in 2019 is a bootstrapped side project. Storygraph has grown to more than 5 million users by offering richer analytics, personalized recommendations, and community features like reading challenges and book clubs. The partnership also gives Storygraph exposure to Kobo's roughly 12 million users across 190 countries as online reading communities continue to grow. I didn't even believe there was 190 countries. I thought there were 182. But, you know, I'm old.
B
I don't know.
A
Maybe the kids have new countries.
B
Maybe. All of that sounds great until storygraph realizes that they need some sort of PR boost and they start to integrate AI into their bullshit. So, I don't know.
A
I think. I think storygraph is one of those things that we would have built in a weekend, and it just kind of kept growing. But the Rakuten E readers, the Kobo E readers, I've heard nothing but good things about those things.
B
Yeah, me too.
A
Nothing about them. Because the thing is, now the problem is I read most of my books from Kindle fucking Unlimited, so.
B
Yeah, that's how they get you.
A
Yeah, that's the thing. It's like I can't afford to buy a new E reader and all new books. It's like. But you know what I do have down the street is the little outside library. You know, the little local libraries. Yeah, we have five of those in my neighborhood. And they're great. I always can find books in there and always put my books back. We have a very, very vibrant little library community in my neighborhood. I don't know if we've got them
B
all over the place here. There's one just around the corner. I drop off whenever I do have physical books that. That don't muster enough warrant to be put into my personal collection, I put my stuff in there. The other thing you can do, Jason, and I'm always reminded about this by many of our listeners, anytime I bitch about trying to get out of Amazon's ecosystem, you can check out digital books from the library with a library card. So you can always do that as well, Jason. You can get your books for free that way.
A
I think I might do that because we have a really good library and it's right up the street. So I should go get a library card. I know.
B
Go get a library card and then you can. You can do the. I don't know what app specifically that they use, but you can reserve and download digital books.
A
Yeah, no, yeah, it's just a basically straight app that you can do that. I think I might have to go do that. You should do that too, mister.
B
I actually have.
A
AI. Okay. There you go.
B
I have. So the Toronto Public Library is very well funded. It's an excellent library system. And I've got the app and got my card and I'm good to go.
A
There we go. You can get the next dungeon crawler, Carl.
B
Not until I finish about five other books first.
A
You gotta cleanse the pallet.
B
Yes.
A
Cleanse that palette.
B
Oh, by the way, I was in a physical bookstore the other day. I took. I took my son because he wanted to get some new books because he
A
wanted to see a dinosaur.
B
He loves bookstores, man. He's so great. Thrilled. So we went to a bookstore. They're called Indigo here at the big chain. And I was going past the sci fi section, massive dungeon crawler. Carl. Display like that book is everywhere right now.
A
It's on fire. Yes, it is. He's a little too young for that one, though. I'm guessing.
B
Oh, no, no, he can't read that. No way. No, no, no.
A
It's even a little. I feel like I'm too young for it.
B
I felt like I needed a shower after a couple chapters.
A
Yeah, yeah, yeah, definitely. Well, here's a fun one. Meta is introducing new usage limits for some features on its AI smart glasses, adding what amounts to a soft paywall starting soon. The on device Conversation Focus feature, which enhances the voice of the person you're speaking with in noisy environments, will be limited to three hours per month for free users.
B
Wow.
A
A $19.99 per month MetaOne Premium subscription raises that limit to 15 hours. Here's the.
B
How much glasses cost?
A
Like 3.99 or 3.50. The new ones are. The new ones are 299. The non ray Ban ones or. But they've got a couple. Couple different ones. Here's the thing, Brian. The change has drawn criticism because Conversation Focus appears to run entirely on the Glass's own hardware without requiring an Internet connection or Meta's servers. Fuck you guys. They've got no leg to stand on with that. They got no leg to stand on with that.
B
That's a load of crap.
A
I'm calling this I Shitification. Very nice. Yes, thank you. And you know what does not require a subscription, Brian? My app, Track A Lot. I designed it with no subscription, but I thank you to everybody who has bought Track A Lot since the past episode or two. Since I started talking about it, 38 people have bought the app so far. So I made back my initial outlay for the Apple dev program. Brian. I made back my $99.
B
Hey. Now if you charge those 38 people a subscription, you'll be rolling in it.
A
I. I can track shitification. That's right. I am working on adding one tap tracking to the Apple Watch now. So you'll be able to tap tap your watch to track, which should be done pretty soon. Now that Fable 5 is up because I didn't write a single fucking line of code in this thing. But here's the difference. I'm learning Swift, because when things go wrong in my app, I'm learning how to fix them myself. Go figure. Little shits. No idea how long it's going to take to get that through Apple approvals, but because the last time it took about two weeks. But thank you everybody who checked out the app. I've got four or five star reviews and one review so far, so I'm excited. I'm happy that it actually is. Is landing and it's landing for people who have the same problem that I had was because I had a sick dog and I needed to track that because my dog was sick last week and I used it every day and it was perfect. It did exactly what it was supposed to do. So I'm very happy.
B
Brian, you need to rename the app and, you know, hone in on this prime, prime demographic of yours. Maybe call it eats its own puke a lot or something. I don't know.
A
I need to niche down, Brian. I need to niche down. Well, we're at that point of the show where we like to thank the people who make this show possible. Over at Patre. You, Brian and you.
B
Oh,
A
we don't get to thank ourselves, Brian. We're Gen X. We are thankless. We are the thankless generation.
B
Yeah, we didn't do. And nobody gives a about us.
A
Yeah, yeah, we are just. We are. I don't even know what the fuck we are anymore. It's late. You're drinking water. I bet that's vodka. Fuck you.
B
I wish.
A
Over at Patreon, we've got two new subscribers. Kim and Joe, thank you very much for joining along. And we've got also more cowbell. M.G. rose, Melissa, Torin, Stewart, Azam, Chuck, Wuma, Shane and Norma. Thank you all for your continued support at Patreon.
B
All right, thank you, thank you, thank you. And over at PayPal, Jason, I spent about five minutes internally panicking about how to pronounce this name. And then I realized it's probably just. Just M has cats.
A
Yeah, M has cats. I was like, m has cats.
B
And it's just M has cats. So thank you, Ms. Cats. And John with a $40 donation. Thank you very much.
A
Thank you. Over the tip. J. Yeah, this isn't the 13th warrior.
B
You don't have to. I'm sorry, this chat. GPT's pronunciations that it kicked out for me for all these.
A
Go it. Over at the tip chair, we've got Jennifer, Thomas, Adam and Matthew. Thank you all so much.
B
Thank you.
A
And we got some merch. William from Baltimore, Todd from Dripping Springs, and Steven from Columbus, Texas. So thank you all very much for buying some stuff. Thank you for supporting the show. We've got no new subscribers on the YouTube which you can sign up for now because that's a thing. But. But if you want to help support the show and I implore you to do so because it's the only way we stay on the air, go to patreon.com gog you can sign up for as little as $3 a month. If you sign up for the whole year, you get a discount. You can give as much as you want or as little as you want, as long as it's over $3 a month. We've got a tip jar at GOG show donate. We've got the PayPal. We've got the YouTube. You can sign up as a membership on YouTube. There are many levels there and you do get the show a little early on YouTube as well. And ad free. So. So whatever you want to do to help us out, keep this show going so we can say fuck and come up with great lines like I should ification every week. Please help us out.
B
Yeah, what he said.
A
How'd that go, Brian? I was trying there.
B
Now I got the. I got the US about to play Bosnia Herzegovia in about 14 minutes. So until next time, let's wrap this up. I'm Brian Schulmeister.
A
And I'm Jason DeFilippo. Thanks for listening to grumpy old geeks. Get all the links and goodies from Today's episode at GOG Show 753. We'll want to keep the grumpiness alive. Toss a few bucks our way at God Show Donate. Every penny helps keep the show on the air. Love the show. Share it. There's a share button in your podcast player. Use it to spread the grumpiness to friends, foes and everyone in between. And we'll love you for it. Swing by GOG show to join our discord and chat with us and other show fans. Got thoughts? Feedback, Cool links? Hit us up at GOG show contact. And hey, don't forget to leave a five star review at GOG Show Review. It's been a long time since we got a review, Brian. Hook us up. We will read it on the show. And we've got merch. Look at this merch. Great merch. It's actually a very comfy merch. I like it very much. Brian's got his merch on too. Look at that on the YouTube. You can see him. Yeah, it's the same shirt as he wore last week. Let's get you one with something on the front, Brian. I'm gonna really help when it's on the back.
B
You designed it.
A
Hey, well, you bought it anyway. Get that grumpy. Get that grumpy merch at Shop GOG show and stay grumpy.
Date: July 2, 2026
Hosts: Jason DeFillippo & Brian Schulmeister
Special Guest: Dave Bittner (occasionally referenced)
This episode dives deep into the latest tech industry disasters, especially focusing on the chaos caused by artificial intelligence (AI) in everything from hacking and job losses to automating entire industries. The show lives up to its walk-of-shame energy for Big Tech and picks apart the week’s headlines with characteristic grumpiness, cynicism, and sharp humor.
“To paraphrase the great Michael Bolton in the masterpiece Office Space, this kid was a no talent AI clown. What an idiot.” – Jason [03:39]
“Try and get a fucking AI to get exactly what you need it to do without knowing what the fuck you’re doing. It doesn’t happen.” – Jason [05:18]
“It is wildly outperform[ing] anything else… it’s because of the vitriol and rage that’s in the comments.” – Brian [07:15]
“I really don’t want to buy AI-created products. I just don’t.” – Brian [09:39]
“When you have no safety net like in the US … a lot of people are one paycheck, maybe two if they’re lucky, away from… being homeless.” – Brian [12:54]
“The incident adds to growing concerns that AI is eroding critical thinking, turning professors into plagiarism detectives instead of educators.” – Jason [22:09]
“If we could have cheated that easily back in university days, we would have too.” – Brian [22:47]
“The most common age check … was to self declare their age. A method criticized by authorities in the country as well as anyone with a fucking brain cell to rub together.” – Brian [24:36]
“The Ouroboros is real, Jason. It is feeding itself.” – Brian [30:29]
“Meta has too much [AI] capacity because no one even themselves uses Meta AI. So they’re going to lease Compute instead…” – Jason [36:00]
On AI coding nonsense:
“What I think is happening is these kids are really good at building an mvp… [but] none of this shit is ever going to make it to production…” – Jason [06:00]
On social media virality:
“It’s because of the vitriol and rage that’s in the comments. That’s what gets it promoted, that’s what gets it pushed up in the algorithm.” – Brian [07:15]
On homelessness fears:
“The truck I can sleep in if I have to.” – Jason [13:49]
On AI limitations in industry:
“Ford finally rediscovered the radical innovation known as hiring people who know what the fuck they’re doing.” – Jason [10:24]
On synthetic AI training data:
“After spending years vacuuming up the Internet to train AI, the industry has apparently invented a closed loop ecosystem where underpaid humans use AI to manufacture synthetic data for other AI.” – Jason [30:43]
This episode offers an unvarnished, skeptical perspective on the tech industry’s latest blunders, particularly around AI, through incisive wit and blistering commentary. Whether you’re an industry veteran or just an interested observer, the discussion covers current pitfalls, regulatory battles, and the sharp edge where tech hype meets real-world failure.
Selected Quote For Emphasis:
“Try and get a fucking AI to get exactly what you need it to do without knowing what the fuck you’re doing. It doesn’t happen.” – Jason [05:18]
Visit the episode page at GOG.show/753 for full show notes, links, and community discussion.