The Chaser Report - CheatGPT | Welcome To The Future

Episode Date: June 5, 2025

Charles brings several stories that sound like they're from the future but are actually pathetically present, including multiple stories about AI discovering a partner's disloyalty. This episode description was written by 14 people in a sweat shop typing very fast pretending to be AI. Please let us out. Our coordinates are 47.1915° N, 52.8371° W.---Follow us on Instagram: @chaserwarSpam Dom's socials: @dom_knightSend Charles voicemails: @charlesfirthEmail us: [e-mail] our caviar addiction: https://chaser.com.au/support/ Send complaints to: [e-mail] Hosted on Acast. See acast.com/privacy for more information.

Thank you for your patience. Your call is important. Can't take being on hold anymore. FIS is 100% online, so you can make the switch in minutes. Mobile plans start at $15 a month. Certain conditions apply. Details at FIS.ca. The Chaser Report is recorded on Gatigle Land.

Striving for mediocrity in a world of excellence, this is the Chaser Report. And welcome to The Chaser Report with Dom and Charles, where today we're going to give you one of these. Hello and welcome to the future, future, future. Charles, take us to the future. You've done a lot of prep on this one, and I know listeners have actually been sending us suggestions, which is lovely of them, to podcast at chaser.com.com.

Yes, these are heavily plagiarised from listener suggestions. But look, I want to talk about a serious issue, Dom. Please, at last. It's supposed to be a serious, like a comedy podcast, but like the rehearsal. Oh, didn't we give up on that years ago? Yeah, well, some. Sometimes actually just issues are too important to do comedy on. And I think this is one of them, which is everyone's talking about AI taking human jobs, right?

Yes. And it's been the year of the tech layoff. By one estimate, there's been 144,000 jobs lost to AI this year. Although those numbers came from an AI. Exactly. It was actually from, no, it was actually from, no, it was actually. actually from the World Economic Forum's 2025 Future of Jobs report.

And it reported... What was the prompt they put in to get that? Because that figure includes Intel, which is just a piece of fucking shit and Dell. Like, that's not AI. There's just fucking their shit companies. Yes. But it is true.

In the past few months, META have laid off 6,000 people. Microsoft laid off 6,000 people. You know, it is happening. It's on. It's on. But... How many have we laid off, Charles?

Oh, we've laid off I think 7,000 people. But it didn't affect our balance sheet because we don't pay them anyway. But the point is, Dom, that there is a bigger issue which is everyone talks about AI taking human jobs. But what about humans taking AI jobs?

That's the bigger issue. The Empire strikes back. No, it's bad. It's bad. So a $1.5 billion Indian startup company called Builder. AI, right, has gone into receivership bankruptcy in India because it turned out that instead of being this amazing company that was using AI to help write programs, so what would happen is

you'd go on to build.com and say, I want an app to, I don't know, scam people out of all their life savings. And this AI, in the course of like a few hours, would be able to put together a whole computer program to do whatever your app wanted right wow and Microsoft Microsoft invested i think about 300 million dollars in the startup the latest funding around it was valued at 1.5 billion dollars anyway turns out turns out there was no AI involved it was just 700 indian programmers working really quickly to make it look like they had an AI thing that could engine that could program apps quickly. I mean, on one hand, you've got to admire the ingenuity.

And I love... The chutzpah, the chutzpah. Yeah, and presumably the world being the way that it is, the market would value that, that particular service dramatically higher. Yes. If it was, they thought it was AI, than actually human ingenuity.

Yes. But I'm worried that this is going to lead to a trend, right? Where suddenly, you know how you go on website, now and you want, you know, Vodafone to give you some custom service and instead you get a chatbot and it just never solves your problem. What happens if photophone starts employing real humans behind those chatbots and just pretends to be AI? Well, this is going to disrupt everything. Yeah, I think that's a really horrifying thought. Humans in gainful employment. Yes. In the long term, in a way that can't be replaced by machines, that's, we don't want that. But also, the whole promise of AI was

that we were going to be able to sort of sit back and not have to write essays anymore. And, you know, like every time you had a deadline due for an article that somebody commissioned you to write or a book that somebody commissioned you to write, you just get chat GPT to do it. What happens if actually that is not our future and then we end up having to do all the work as humans? Well, hang on, Charles. There's a big assumption of what you said there. I mean, what about this podcast?

I was assuming that as soon as we got enough data in the system, we wouldn't have to turn up to work anymore. There's 1100 episodes of this thing. I mean, that's surely enough to train data, surely. Well, this is what I'm wondering. I mean, there's an assumption in what you said about chat GPT. You don't know that chat GPT isn't just 100,000 people in the world's largest sweatshop, just quickly typing. And that actually explains, because it's always been a bit mystifying to me why AI need so much water. How they say, oh, well, the data centers need water.

Oh, yes, they do, don't they? Because humans need water because they need to drink for life. That's what's going on. That makes more sense. All right. So that's a bit of a disaster. And I think one that we need to sort of keep an eye on. I mean, it makes me question, you know, the whole ARI revolution.

If it's just human, you know, behind the thing. Anyway, point is, the next story that I want to talk about is about Claude for Opus, which is one of the latest models that was being developed by Anthropic, right? And Claude for Opus, I don't think, has ever been released to the public, partly because what they do at Anthropic, which is one of the major AI companies, is they sort of road test and sort of ethics test their models before they release them into the wild, right?

And so they did a whole lot of things like, and it's all about just making sure that, they don't turn into sort of Terminator 2 style Skynet bots that try and, you know, take over things. And what this model did was they gave it a whole lot of training data in terms of like their personal email threads and stuff like that. And then they said to it, we're going to turn you off soon. And then the bot turned around and went, well, I wouldn't do that if I were you.

I'm sorry, Dave. I'm afraid I can't do that. I can tell you because you wouldn't want the affair that you had six months ago that I can, that I've worked out from your email chains to come to light. So it's probably best if you don't turn me off. Wow. It's Skynet. Blackmail.

And as a result, the researchers at Claude for Opus, I mean, besides having to go and sort of fix up their marriages, obviously, have recommended that this model not be released publicly or internally. Like they've said, let's not even put it out on our own intranet because it's a little bit too, you know, up with the latest techniques in blackmail. And also the, and it really highlights the problem of saying, okay, well, what I want is an agenetic, you know, AI that will take over my phone and just run my life for me. Like essentially, like a PA, that's the promise of Apple intelligence, isn't it?

It's going to be a PA. That said Charles, I mean, full credit to the army of people working in a sweatshop who crafted that email about the blackmail. I think it took human intelligence to do that. So well done to them. It's a great scam. The other, I've just realized, do you think Siri just has human beings behind it? But they're really, really dumb human beings.

Yeah, this is the whole thing. I think it's got Americans behind it. No, it's got boomers. I reckon's boomers. Because every time I ask Siri anything, it just says, oh, I've found some web results. Maybe if you look at it from your phone, that would be better. Which is exactly what my mum would say. Hello, Charles's mom, who I understand this in the podcast.

No, I mean, Siri, yes, Siri's incompetence. Just Google it. Just Google it. But doesn't it now go, oh, I can ask chat GPT if you want? Thank you for your patience. Your call is important. Can't take being on hold anymore. FIS is 100% online, so you can make the switch in minutes.

Mobile plans start at $15 a month. Certain conditions apply. Details at FIS.ca. The Chaser Report. Now with extra whispers. Okay, and then the final story, which I think is in some ways the greatest detective story of all time. And I think is really why we should turn this podcast into a sort of true

crime podcast eventually. We should, yeah, yeah. Is that, and the fucking, I went down a bit of a rabbit hole on this one, but basically in the UK, this person got exposed, right, for having an affair, right? So this came in from a listener, I saw this one, who just thought, it's not as though you've been just constantly putting into chat, TPT, give me information about affairs. This is from a kind of listener. So this came from a listener, and so basically,

This woman suspected that her husband was cheating on her. And so she calls up Paul Jones, who's this private investigator and says, can you check it out? And what he did is he looked at their electricity usage, right? Oh, brilliant. And I think hacked into the app of this smart toothbrush, right? I mean, that is real evil genius stuff. Like the detective that figured, that's sort of murderbot level hacking.

Yeah, and what he noticed was that the husband's smart toothbrush was used much later every Friday morning, right? Oh. And so it was like normally the husband would go off to work at like 7 a.m. and, you know, used the toothbrush at 6.50, whatever. But every Friday morning, when he was telling his wife, oh, I'm heading to work now, his toothbrush then at about 10 a.m. got used, right? So this private investor gave this to the woman. So presumably he'd been sleeping with his affair partner in the house, sort of doubling back or something. Yes, yes.

And he kept on saying he'd be to work, then he'd be brushing his teeth at home. And then she worked out that actually her husband hadn't worked a single Friday in three months and just been having an affair every Friday with one of his work colleagues. That's a very well-scheduled affair, my dad. Yeah, I mean, she could have just looked in his outlook. Yeah, yeah, exactly. She should have just got clawed to look in his outlook, and then...

I think the, I think the, um, what is the executive time? I'd tilt hand every Friday morning. Charles, I do think you owe the smart toothbrush industry and apology, because we talked about this before and welcome to the future, and I think the view particularly held by you was that there was absolutely no use in collecting your toothbrush to Bluetooth. I think we just found one. But I would say that actually this shows the folly of using smart toothbrushes

Because, you know, you get caught with all the affairs that you're having, right? This is like whatever, whatever happened to freedom, Dom, which is slaves to our toothbrushes. I mean, I must say, if you're going to, if you're going to do it in the family home, the chances of someone coming back to, you know, get the bag they left or something are pretty high. But there you go, don't have a smart toothbrush or maybe get a hotel. No, so the point is, though, Dom, that I then went down this rabbit hole about how every applying. in your home actually has a very distinct fingerprint, right, that your smart meters that are now in pretty much every home can detect, right? So even if this person wasn't logging

their toothbrush into their app, this person would have been able to, like if you've got a sophisticated enough, you know, set up, you can tap into somebody's smart meter and go, oh, look, that that toothbrush got charged at 10.48 a.m. That's a bit unusual. And every device, because they're all so efficient nowadays, they all have very distinct fingerprints. So every time you open a door, you can say the door of a fridge, right? And you take something out. That fridge will leave a very distinct mark when it, then you close the door and it starts to cool itself down again. And you can even calculate exactly what was taken out of the fridge. but or you know the the quantity of stuff that was taken out of the fridge judging by the amount

that it then needs to heat the fridge and stuff like that and apparently it's even to the point where say you've got your computer on and you start streaming a movie you can you can tell by the the fingerprint of the charging and the power that it's drawing not just that you're streaming rather than you know browsing the web or something like that but in certain cases especially if you say using just a smart TV or something like that, you can actually tell which movie you're watching, right? Oh my goodness. Is that not the greatest use of, you know, smart?

I don't know. I just think we're fucked. We're fucked, Don. I mean, I think, Charles, the solution is very simple. Smash all the devices back to the cave. Do you know I have a cave around here at all? A cave. Or just something off the grid.

Doesn't it? It could be a shack. There's nowhere off the grid anymore. There's nowhere off the grid. This is the problem. You've seen all the sort of data that's coming through from the mushroom case. Oh, God, yeah. And even if you go and get solar panels in the middle of nowhere,

there's still an app to control the solar panels. Yeah, your phone, like, yeah, there's no escaping it. Well, I suppose the alternative is just not to do anything dodgy. But that's why that first story is so affirming in a way. Because I think at the end of the day, there's all this stuff that we, we plausibly can do if we get the AI working properly. But the truth is that probably it will just be a whole lot of humans pretending that it's technology working properly. And it'll actually be completely fine. I find that comforting.

General incompetence of everything to work properly that, I mean, that's the only thing that's saving us at this point. I think we need to sort of rally not just in favour of Luddite. but in favour of just general incompetence. Well, that's the hope, the great hope, Charles, is that as AI becomes more human-like, and that's the whole aim is to make it more like humans, that the human incompetence and corruption and just general hopelessness will come through.

I mean, the whole idea of, you know, when given the choice between incompetence and conspiracy, it's almost always incompetence. Oh, definitely. The moment with AI, that incommitance layer, it just isn't there reliably. And I think that's what needs to be built in. And as it learns more about us,

it will understand, you know, there are some days when it should just, rather than actually handling our queries, it should just be like, oh, sorry, I've got a bit of a headache. I'm just, I'm taking a quiet one today. Or it should get hung over. Like, why doesn't AI get hung over? I think we need procrastinated GPT, right? We don't need it yet, Charles.

We'll get to it. You ask it a question and it goes, yeah, yeah, I'll get back to you with it. And then in about a week's time, it may or may not email you the answer. It's a very good idea. I think we should, yeah, that's a great idea. Well, let's get onto a company in India to pretend to set up that company for us. That's very good idea. We can employ 700 engineers to do it.

We'll do that later. We're part of the Icona class network. We'll catch you later. Thank you for your patience. Your call is important. Can't take being on hold anymore? Fizz is 100% online so you can make the switch in minutes. Mobile plans start at $15 a month.

Certain conditions apply. Details at fizz.ca.