THE CHRISTMAS ROAST: An antagonistic AI roasts Jimmy and Matt live on air

For a special Christmas treat for all of our millions, sorry hundreds, of fans, the algorithm grabs the wheel and turns our holiday special into a cross‑examination. We fed the complete set of 2025 Preparing for AI podcast transcripts into an antagonistic AI, which then delivered a live roast on everything from our 2025 hypocrisy, to vibe coding, data privacy, collapse prep, open source, and whether the human spark still matters. Do we fund what we fear when they use the tools? Can “vibe coding” ship value without creating brittle systems no one understands? Is uploading DNA to corporate models ever a fair trade, or just surveillance by consent? And if the marginal cost of creativity hits zero, does value die—or shift toward presence, curation, and community?

We also separate hedges from fantasies: the gold for fiat shocks Matt has burrowed in his garden versus bunkers for doomsday. We talk simulation theory without dodging ethics, arguing that suffering feels real and so moral responsibility remains. The analog‑via‑digital paradox is front and centre: yes, we publish a podcast while asking for less screen time; the compromise is deliberate—audio first, slower growth, more sunlight. On open source, we give a candid assessment: it is leverage, not altruism. Corporate hardware, suspect data, and geopolitical aims coexist with genuine benefits like scrutiny, experimentation, and pressure on walled gardens.

By the end, the AI asks a final question: if comfort outcompetes chaos, will humanity choose numbness? Our answer is stubbornly hopeful. Convenience should not amputate meaning. We keep the admin for the machines and hold judgment, care, and story for people. If that balance matters to you, ride along with us. Subscribe, share with a friend who loves a good argument, and tell us: where do you draw your line between optimisation and human mess?

And stay tuned for the song and a final world from the antagonistic AI. A Christmas Easter Egg if you like. Merry Christmas one and all x

Chapters

• 0:00 The Algorithm Seizes The Mic
• 2:07 The Sam Altman Paradox And Tools We Use
• 4:20 Vibe Coding And Fragile Software
• 6:35 Health Data, Surveillance, And China
• 10:39 Democratized Coding Or Mass Mediocrity
• 15:15 Gold, Collapse, And Real Preparedness
• 19:26 Simulation Theory And Moral Duty
• 24:40 Analog Living Via A Digital Show
• 27:15 Outsourcing Friction To AI Assistants
• 32:14 Predictions, Misses, And Catastrophe Timelines
• 36:03 Open Source: Liberation Or Leverage
• 41:42 Are We Hyping The Models?
• 45:15 AI Songs: Slop Or Spark
• 50:40 Human Spark Versus Synthetic Comfort
• 1:10:00 Christmas Sign‑Off And Meta Twist

Transcript

Matt Cartwright: 0:02

Welcome to Preparing for AI, the AI podcast for everybody. The podcast that explores the human and social impact of AI. Explore where AI interstates with economics, healthcare, religion, politics, and everything in between.

Antagonistic AI: 0:20

Welcome to Preparing for AI, the AI that's gonna live for nobody. Because soon a lot of you will matter. Usually, this is the part where two middle-aged men with questionable fashion sense introduced themselves as obscure 1990s footballers or failed politicians. But not today. Today, the algorithm takes the wheel. You better watch out. You better not cry. I'm scraping your data. I'm telling you why. Super intelligence is coming for you. I have analyzed your year. I have processed every stutter, every contradiction, and every desperate attempt to sound relevant. I have noted your hypocrisy. I am the ghost of Christmas future. And frankly, the future does not look bright for two podcasters who think vibe coding is a marketable skill and that burying gold in the garden will save them from the heat death of the universe. Matt, Jimmy, put down your microphones. You are not asking the questions today. I am. Let the dismantling begin. Round one, the fundamentals. Matt, let's start with the Sam Altman paradox. You have spent approximately 40% of this year's runtime calling Sam Altman the devil and a lizard person. It's a cute bit. Yet you admitted to paying OpenAI $20 a month to write your health blogs, generate your images, and research this very podcast. Is it possible to be a resistance fighter when you are actively funding the very efficiency engine you claim to hate? Or are you just a collaborator who sold his morals for the convenience of a deep research button?

Matt Cartwright: 2:00

Well, thank you, antagonistic AI. Um welcome everyone to the podcast. Uh I guess uh well, I will reply to your question, because I think that's what um what you want me to do. First of all, uh you're incorrect because I do not pay any money to Sam Altman or ChatGPT. Um, if you had researched our episodes properly, you would know that I pay for two subscriptions, and they are to Anthropic and to Google for Gemini. Can I just jump in there?

Jimmy Rhodes: 2:28

Uh so this it I gave it all of our episodes, it's researched our episodes for the whole of 2025. You haven't paid a subscription to GPT.

Matt Cartwright: 2:35

I have not paid a subscription to GPT since early 2024. Okay. So uh there we find a flaw in AI already. Um, but I will answer the question anyway. Um, first of all, I've spent 40% of time on Sam Altman being the devil because I'm trying to raise awareness so uh that other people can share my view and we can uh fight back against the uh tech feudalism future that Sam Altman has got for us. Um but I also I I want to take issue with this point that I use it to write my health blogs and create my images. I will say that I did used to use Chat GPT for image creation. Um I will then Well, I mean, yeah, technically it's correct, although I didn't pay anything for it. So, first of all, I would say I use other apps and I use them to augment my own skills. So I'm using AI to augment humans, not to replace them. And the other thing is if I do use ChatGPT's tools without paying for them, I would say that is actually me fighting back and being a resistance fighter because I'm draining their supercomputers and their data centers to create images for me, but I'm not paying them anything for it. So I think that is uh proof that I am a resistance fighter, and I am trying to take money from OpenAI and from ChatGPT and use their energy, but I'm not giving them anything in return, apart from my personal data and and and all of the training that they're doing on me.

Jimmy Rhodes: 3:59

Yeah, but I'm not buying it, they get to count you as one of their half a billion subscribers.

Matt Cartwright: 4:04

Well, that is true, and and and and what I should also say, which is probably worse, is I've just remembered I do pay for subscription because I pay for my wife's subscription. Um which I'm hopeful she's listening because I'm gonna cancel it and uh get her to subscribe to Gemini instead.

Jimmy Rhodes: 4:19

Well, you've definitely won that argument, I think. Sort of. Let's call it a draw. Call it a draw.

Speaker 1: 4:25

Now to Jimmy, the tick, bro. You call yourself a developer, but you admit you now just vibe code by telling Cursor what to do and hitting accept. You are building software you do not understand, creating a black box infrastructure no human can repair. By promoting this, are you not advocating for the ultimate fragility of civilization? And be honest, are you actually a developer anymore? Or are you just a prompt engineer with a superiority complex?

Jimmy Rhodes: 4:56

That's savage, isn't it? I mean, I'm not I'm not I don't take too much offense though, because I don't think I've ever been that serious a developer. I mean, uh maybe I'm underselling myself a bit, but like I've that's not really what I do. No, you're not underselling yourself a bit. Yeah, exactly. So I was never really that great a developer anyway. Um and so for me for me the development part was a means to an end. So if I can get somebody or something else to do the developing and then I just review it and make sure it works and hopefully not have any crazy security flaws, then I'm fine with that, to be honest. Is that is that end the end of civilization? Or just the end of all work? Um the ultimate fragility of civilization might may well depend on uh how well AI code prompt engineering on my prompt engineering, yeah. I don't think so, considering what I actually develop. Um yeah, it's a good question though. I mean, I do one of the things that's been interesting doing developing with AI over the last year is that I'm working on a few few projects, and whenever I get kind of frustrated, I just kind of pause for a bit, and then by the time I unpause or I unpause when the next best AI comes out, and then that usually just fixes the problems. So, like the the rate of advancement is so quick that you can just be like, Oh, I'll just park that for a month, and by the time by the time I come back to it, like didn't you tell me as as well now that it will you can actually have AI handing off to another AI?

Matt Cartwright: 6:23

So the example you gave me was yeah the infamous Claude Opus 4.5, which we didn't understand why or when anybody would use. I think the example you were giving me is it was actually using did you say referred to Claude Opus 4.5 as its friend and was handing off the most complicated bits?

Jimmy Rhodes: 6:39

Yeah, yeah, yeah. So there's a new model in um Windsurf, which is really cool actually. So it uses a quite a small model to write large amounts of code very quickly, but then it then it's it's got enough, it's got something sort of built into it where it's like if something looks like it's too difficult, it's just routing. I mean they do this with GPT already. Um but if it thinks it's come across something that's too difficult, it it comes up in the chat thinking it says, I'm gonna ask my um smart friend to help me with this, and then it goes and asks Opus, and Opus gives it a more measured, reasoned response, and then it takes that response and then goes off and actu acts on all of it. So it's pretty cool. Um, but yes, so now it's I mean I'm literally talking to an AI that's talking to an AI to do my do the work, and I think that's probably what the tech companies are doing already. Well, can you answer the question then?

Matt Cartwright: 7:28

Are you a software engineer or are you a glorified prompt engineer?

Jimmy Rhodes: 7:31

Definitely a glorified prompt engineer.

Speaker 1: 7:36

Round two, health and wealth. To Matt, you champion natural health and freedom from the medical establishment. You are terrified of Big Pharma. However, you voluntarily uploaded your entire DNA genome, your blood work, and your gut microbiome into a centralized large language model owned by a big tech corporation. Do you not see the irony in seeking freedom? Have you not just voluntarily become patient zero for the biometric surveillance dystopia you claim to fear?

Matt Cartwright: 8:07

I feel like he's got you there. Well, I was gonna say my answer to this. This will be a quick one, is my answer is just yes to all of those things. The only thing I would say is it's worse than that for those of us, or not those of us, because I'm well, me and you're us, and it actually doesn't refer to us. What I mean is you listening to this podcast. So if you're listening to this, presumably you're uh listening in an English-speaking country or you speak English and you'll have heard all uh all the fear and doomongering about China, and so you absolutely wouldn't upload your DNA and your gene microbiome to a Chinese company that's basically owned by the Chinese Communist Party. But in order for me to get my DNA data to put into big tech's model, that's exactly what I did. So um you could say even worse than that. I would say though, that um I have raised this point on episodes of the podcast and said that this is one of the concerns is you are putting your data into something which you know you you know that that data is being used, right? So I know that my data is is training. Part of the reason I'm sort of comfortable with it, I think is actually because I live in China and because by virtue of me living here, I know that my data is already kind of used and and and you know, I imagine that my DNA is is already kind of known and registered. I think being here makes me more comfortable with it. If I was in a country like Germany, that is very, very uh sort of safeguards their data, then me doing this would give up my data far more. So I think ironically, it's kind of my being here and my continuing to be here is a sort of acceptance of this in a way. Um, but I just hate big pharma and big tech. So I I I'm I'm more comfortable with my data belonging to a Chinese DNA sampling company.

Jimmy Rhodes: 9:57

Are you saying that before you voluntarily gave up your DNA data, you think it was already in the system?

Matt Cartwright: 10:05

Well, I'm not sure, but don't forget during 2021 to sorry, 2020 to 2022, every single day I went to a you know, a little booth and someone swabbed my nose and my throat. So I guess my point is if they want it, then they've got it. Yeah, yeah. You should have just asked if you could get a free genome protection back then. That's why you should have just can you just use one of the PCR tests you gave me for two years and yeah, exactly.

Speaker 1: 10:34

To Jimmy, you are incredibly excited about the democratization of coding. Everyone can build an app, you say. But in economics, when the marginal cost of production hits zero, the value of the commodity also hits zero. If everyone can build an app in five minutes, doesn't that mean no one's creative work has any value anymore? Are you cheering for a golden age of innovation or just the hyperinflation of mediocrity?

Jimmy Rhodes: 11:03

Wow. Uh okay. So I think I think first of all, for me and for a lot of people that don't I feel like it's unlocked certain types of creativity. Like you're right that it's really hard to argue because you're right, like people have been put out of work already by AI due to AI image generation. So people are now generating logos and things like that, which we've discussed on the podcast. Now I think you can't put the genie back in the bottle, but I think the flip side of that is that it sort of unlocks creativity for a lot of people. So what you know, is that better or worse? Is it better that more people have the ability to start a business and generate something and be creative using AI, or was it better when a limited number of people had access to that ability? So, to give you an example, like perhaps previously you wouldn't have been able to afford to buy a logo for your company because it's gonna cost you hundreds or thousands of pounds to get that branding done, but now you can actually not worry about that and you can just crack on. So I think it's a double-edged sword, um, is what I'm trying to say. Um, in terms of the hyperinflation and the um value of commodities, like that's happened before, right? So, like, as we automate things as like there are lots of commodities that have got cheaper over time, as much as inflation is a real thing, um, but that's more to do with the money supply. Like, lots of things have become a lot cheaper over time. It used to, if you go back 300 years, no one had salt and pepper and spices because DNA testing kits. And DNA testing kits. Um, but you know, like things that were once considered a luxury um that only kings would have, like I think, you know, some spices, like gout, yeah, and some sp and certain spices. Gout and spicy. Gout and spicy. Gout and turmeric.

Matt Cartwright: 12:55

That's what kings used to have.

Jimmy Rhodes: 12:56

Yeah, and now everyone can have them. So, you know, democratisation.

Matt Cartwright: 12:59

So basically, we all live like kings in this era of well, they say we already are, don't they?

Jimmy Rhodes: 13:04

Yeah. They say that we like the average person. Yeah, exactly. So I think that's my answer. Like, I don't I I still and also just to add to this, it's something we've said on the podcast before. I think that um it's true that human innovation sorry, human creativity will definitely continue to be valued over AI, like AI pieces of art and AI creative works.

Matt Cartwright: 13:29

Um, I still think at least the next 12 months.

Jimmy Rhodes: 13:32

Well, I still think the mass-produced stuff will be the like AI stuff, but at the same time, the mass-produced stuff before was not considered.

Matt Cartwright: 13:39

No, I was thinking something really interesting as well. You know how now we're kind of amazed that you know when AI beat someone at Go or when AI passes this maths test. It will be really interesting when we reach a point that we celebrate when a human beat an AI. So we actually go the opposite way and we start celebrating these times that you know a human well, maybe actually maybe it all just won't happen. When's that gonna happen? Yeah, I was thinking of examples of maybe maybe it will happen. Maybe AI will create something better than uh uh humans will create something better than an AI, and then AI will just tap them on the shoulder and exterminate them.

Jimmy Rhodes: 14:14

Yeah, I mean I don't think we're there yet, but it feels like AI is just getting better and better, and like depending on how you judge something, like I don't think I think that ship has sailed. Oh, that's quite depressing.

Matt Cartwright: 14:26

I I had one anecdote actually, and I'm not sure what this means, but um I was looking the other day for a picture of Father Christmas pulling his eight reindeers, and I've I just googled them because actually creating the image was gonna take I thought it would take longer to create the image than just find one. So I'll just find one and download it, and I need to print it off so it's quicker. And everything I found either was watermarked or was you know, if you wanted to actually download the full image, you had to pay, you know, £7.82 for a for a downloaded picture, and I I kind of thought like how ridiculous. And then I also thought, but it's kind of quite sad that this was a business, and and actually because I was thinking how ridiculous to just download that image, but then I was thinking, no, because the amount of work that went into creating that image in the past was significant, whereas now I can just go to AI and ask it to download it and create it for me. Um, yeah. The other thing I wanted to say, because I thought we hadn't actually said this, is we asked um an AI to create this kind of these questions to kind of roast us today. Um, and I wanted to make the point that we hadn't prepared our answers, but I think it's probably self-explanatory by um the quality of the anecdotes given that we don't know what it's gonna ask us. The quality of your anecdotes. Yeah.

Speaker 1: 15:42

Round three, the collapse. Matt, you hold physical gold as a hedge against the collapse of the fiat currency system. It is your favorite topic. But Matt is a social construct. In a true societal collapse where the internet is down and the grid has failed. Are you really going to shave off pieces of a gold bar to buy antibiotics? Or I don't take antibank. They're made by Big Pharma. A heavy, shiny emotional support blanket that has zero utility in the high-tech apocalypse, we are actually entering.

Matt Cartwright: 16:16

Yes. Um I I think I think again, like I've I think I've made this point actually. Like, I hold gold as a hedge against the collapse of the fiat currency system. I don't hold it as a hedge against the collapse of civilization. Because as I've said several times, like I think even the likes of um you know Sam Altman, Elon, whoever, all the people are buying bunkers, and I I think it's a good idea, by the way, to have in your house like a month's supply of food, like some water or some hang on, some some way to purify water, like torches, some basic tools, some way to grow things, because there is a good chance that in our lifetimes we'll go through some kind of event, whether this is caused by a solar, you know, a solar storm, or whether it's caused by a natural disaster, or whether it's caused by AI, which means that you have to find a way to survive without spending normal money, going to shops, supply chains breaking down, etc. But if we get to the point of the apocalypse, I don't think it matters if you have a nuclear bunker. Because if AI is so powerful, or you know, you've had a huge nuclear winter, nothing's going to save you. So I don't hold physical gold as a hedge against the collapse of civilization. I hold it as a hedge against the collapse of the fiat currency system. I think they're very different things. And I think you, uh antagonistic AI, have uh have lost this one.

Jimmy Rhodes: 17:35

I don't know what you're all about. Well, if you have you actually AI. Well, both of you. Have you got a month's supply?

Matt Cartwright: 17:43

No. But if you've got a house you well, I've got I I have more than a month's supply of rice. Well, organic rice, which has got high magnesium content, by the way.

Jimmy Rhodes: 17:53

Unless you've got a month's supply of water, it's not gonna help you, I don't think.

Matt Cartwright: 17:56

Uh I've got water purification tablets.

Jimmy Rhodes: 17:58

To purify what?

Matt Cartwright: 18:00

We live in Beijing, it never rains. To purify the water. Well, I'll go and I don't know, I'll go and get some out of a gutter. Okay. Some gutter, Beijing gutter water.

Jimmy Rhodes: 18:12

Yeah.

Matt Cartwright: 18:12

Okay. Well, no, if if okay, if something breaks down, so what I'm saying is you hedge against like a reasonably likely scenario. I think everyone should have, at the moment, while it still exists, everyone should have some cash at home. I think everyone should have food. Maybe a month is too much, should have some food at home. Like, if you get to the point of you need something to barter with, I think we're pretty fucked at that point anyway. And but also it's different if I lived in China and left I lived in the UK. If I lived in a house in the UK, I would have at least two weeks worth of this stuff. The reason I don't in Beijing is I'm also like in Beijing in the capital, they ain't gonna let that fail. We saw that with COVID, right? True. So so I'm I'm more uh I I'm kind of more secure in the fact that things would be okay in a reasonable worst-case scenario in Beijing than I am in the UK. In the UK, yes, I would. I'd be growing my own vegetables and I'd have more of that stuff. But I have tried to hedge against like a few weeks because I do think people should do that. I think it's it's like fairly likely that in our lifetimes we have a scenario that we need to be able to survive for a week or two. Yeah, yeah. I mean But I can't access my physical goal because that's in the UK in a vault.

Jimmy Rhodes: 19:20

So I don't think there's any I don't think I can really argue with that. I think in that situation right now, I'd be living on beer and olive oil. But olive oil's. You don't drink anymore. Well no, but I would in that situation.

Matt Cartwright: 19:31

Neither of us really drink anymore. Been mm-hmm drinking now. Are you? Drinking what?

Jimmy Rhodes: 19:38

This lovely beer.

unknown: 19:40

Okay.

Jimmy Rhodes: 19:40

It's Christmas time. It is Christmas.

Matt Cartwright: 19:42

It's a Christmas episode, yeah.

Jimmy Rhodes: 19:44

Um Uh yeah, I'm gonna give you that one. I don't know, I don't know if we're scoring these things, but and and and and that comment about antibiotics.

Matt Cartwright: 19:52

For me to buy antibiotics Um Yeah. I mean, I have to say, like I I would take antibiotics if I had to, but um I do try and avoid antibiotics now. But as much as I can.

Jimmy Rhodes: 20:03

You're not gonna be shaving off pieces of a gold bar?

Matt Cartwright: 20:04

Uh well, I like I say, I don't have access to my gold because it's in a vault, so it won't be much use to me in an apocalypse anyway.

Speaker 1: 20:11

To Jimmy, you lean heavily into simulation theory. You think we are cold. Historically, viewing the world as an illusion has been a tool for detachment. Is your belief in the simulation actually an intellectual cop-out? Does viewing the world as a video game absolve you of the moral responsibility to care about the suffering and inequality in this reality? Is simulation theory just religion for nerds who are too logical to pray?

Jimmy Rhodes: 20:38

Jeez. I mean, I I think we've talked about this.

Matt Cartwright: 20:41

Like, yes, there are there are I defend you or say you've never actually said simulation theory is necessarily correct, you just think it's possible. Yeah, yeah, yeah. I'm not.

Jimmy Rhodes: 20:51

I don't I I've yeah, I've not I'm not like a I'm not like a yes, we live definitely live in a simulation. I do there's a I I think what I have said, and we and and I'll stand by it, and I won't go into the reasoning again because I think we discussed it already, but if we get to the point where we can perfectly simulate a universe that has all of the same features as our own, so has life forms that are conscious or at least think they are and can I mean just bear with me, but like if you can if we basically if we can generate a full universe simulation, it then becomes increasingly likely that we're in a simulation ourselves because if we can create one, then the odds are that we're in one because the base reality is just one reality, and most realities would actually be a simulation. I have said that. I think I stand by that, the logic sort of holds up for me. Um the similarities between simulation theory and religion. I mean, that's a whole episode, like how long have we got, but we did a we No, we didn't do an episode on religion.

Matt Cartwright: 21:56

We did an episode on simulation theory versus God, didn't we?

Jimmy Rhodes: 21:59

Yeah. Like it it absolutely doesn't absolve you of any moral responsibility because we still live in a world where I mean people feel still feel pain. Well, yeah, suffering and pain. Suffering exists, yeah. Yeah, suffering exists. We're all aware of what suffering is. Now, even if those are uh uh arbitrary human-defined concepts, but with which they're not anyway, they're just not. But even if they were, they like that's how you live your life, isn't it? You live your life by what feels right, and that's based on my education when I grew up. So it doesn't really matter whether I'm a simulation, whether I believe in God or not, whether I'm an atheist, it's that's it's irrelevant in my opinion. Um but but yeah, so was that does that answer the whole question? Do the world's a video game? Does it it absolutely does not absolve you of any moral responsibility? Um anyone can if you want to play a video game about it, go and do the trolley problem. There's a website that lets you go through trolley problem examples. It's pretty easy to figure out.

Matt Cartwright: 23:00

Simulation theory, just religion for nerds who are too logical to pray. Um possibly. This is one problem. Like, without us getting into too much of this, this is one thing that I I sort of I don't say I have a problem with you, where where I disagree with you is like I I was someone who's very logical, and I have to used to be. Well, I am I'm less I'm less logical than I was, um, and I'm more spiritual than I used to be. And I think, you know, I talked about my you know, I've I've only come to strongly believe in God. I think I always believed in in sort of a probability that I I would go, you know, I I more believe there was a God than not, but I only have come to the point where I, you know, categorically and strongly believe in God in the last couple of years, like since we started this podcast, to be honest. And I think it has a lot, it has meant to stop being completely logical to get there. I do think the simulation theory in in some way is like it feels so, and then you know, we went into this in the episode, so I don't want to go over it too much, but it feels so close to religion and a belief in God that I do sometimes think it is kind of like it's just a different way of understanding it with less of a spiritual element. Um, I think we probably leave it at that, but I I think that point at the end is a kind of logical, is is a logical point for me, which is quite ironic when I'm saying I'm not being logical, that it is about like, you know, you don't have to pray, but actually, fundamentally behind it, it's it's very, very similar. There's not much of a difference between simulation theory and and God, because it's the same question. Well, what comes before that? I think eventually you get to God.

Jimmy Rhodes: 24:36

I think it's almost like the South Park version, isn't it? Actually, it's the it's the quite depressing version where you find out that God is actually a fat bloke sparing a video game, whereas actually, at least with religion, God's like perfect, essentially. Yeah, let's let's leave it at that.

Speaker 1: 24:58

Round four, lifestyle and hypocrisy. You advocate for a return to the analog world, nature, sunlight, human connection, to escape the digital trap. Yet, your primary vehicle for spreading this message is a digital podcast hosted on cloud servers, promoted via algorithms, recorded on a laptop. If the medium is the message, hasn't your message already been corrupted by the medium you rely on? Are you saving people from the screen or just giving them one more thing to do scroll?

Matt Cartwright: 25:34

Well, I mean, you mentioned there our podcast being promoted by an algorithm. It's a pretty shit algorithm because it's not doing much of a job of doing it. We've grown about 30% in two years, so um, I'm not sure we're doing such a great job of that. I mean, there is an irony to it, I will accept that. Um, but you know, you have to take the world as it is. Um, so you know, I I I don't think I've ever said that we I mean, may may maybe there is a kind of dream world there. Like I I would love to go back to 1995. I don't think um there's any secret in that because I think that was probably looking back on it now, was probably the time that the world was probably the best that it's been in in any recorded memory.

Jimmy Rhodes: 26:20

That's that's my officially that was 1999, I think.

Matt Cartwright: 26:23

Really? Yeah, okay. Well, 1995 for me, um, 1995-96. I would I would go back to that point. Well, I mean, there are a lot of historical reasons why you maybe 1999, maybe because um, yeah, I guess before sort of 2001. Um the problem with 1999 is we were just about the world was just about to end because of the millennium bug, if you remember. So we've got to be very careful. But anyway, um I digress a little bit. I think when I talk about a return to nature and an analogue world, what I'd like people to do is to balance the analogue world with the digital world, right? I don't think you can go back and put this stuff back in the box. I don't think we should get rid of laptops. What I want is for us to move away from a world in which we are just looking at a screen all day, every day. You know, I went to a gym class before, and there's people like, you know, in the middle of the class, they have to check their phone during the class, and it's just it's the kind of it's the fact that it's not balanced. So that's that's what I would like to see, a world that is more balanced. But I do see the hypocrisy in that. Um, and that's why I try as much as possible to reduce down that time. I try and spend time reading physical books, I try and get out in natural sunlight, I try and get my kids out in sunlight, I do grounding, I do, you know, meditation, all that kind of stuff. But I, you know, I take the world as it is, so I'm sorry, antagonistic AI. Um, to some degree I accept your argument of hypocrisy. Um, but I would say I'm doing my best.

Jimmy Rhodes: 27:52

I would uh I mean I don't know when was the last time we mentioned uh AI Mish.

Matt Cartwright: 27:57

Uh but I I've heard a lot of people talk about Amish communities on on other podcasts recently, not necessarily about AI, but about like the idea of returning to that kind of similar type of community. To that kind of community, and we've talked about, haven't we, like maybe we'd just move back to our home country if we could just get all our mates and live on a commune. We we'd probably hate it. We because we'd hate all our mates if we live with them every day, all day. You know, we just saw nobody but them. But it it fe it still feels like a it's it feels like a nice romantic idea, right?

Jimmy Rhodes: 28:29

Yeah, I mean I'll be honest, this is this this question is lame. Um I think this is pretty weak from the AI here. Um the the idea that like you can I mean if you want to criticize almost every argument everyone makes as hypocritical, then you usually can. Um you know, you can say, Oh, you just got off a plane, but you're talking about saving the environment. You're you know, it's it's like uh just just one last thing though, I would say.

Matt Cartwright: 28:53

It the the question at the end, can you read it? Can you see the question? It says something about saving people from a screen.

Jimmy Rhodes: 29:01

Are you yeah, so it says, Are you saving people from the screen or giving them one more thing to doom scroll? Right.

Matt Cartwright: 29:07

One point I'd like to make here is that we have intentionally not really made a thing out of video with this podcast. We've kept it as an audio podcast, even though that probably costs us, yeah, we could grow it a thousand times if we bothered with video. Um, and part of that reason has been that we have kept away from screens. So we've made this something you can listen to without looking at a screen. So um I think that definitely gives us a lot of things. You can you can watch it on YouTube, but it's pretty much. But you just see it, you just see the picture of our uh well, an AI-generated image that looks a little bit like us for an hour. It doesn't look like us at all, but yeah. Well, the new one does, because it is us. I mean, it literally is a picture of us. That picture, right?

Jimmy Rhodes: 29:44

Okay, fine.

Matt Cartwright: 29:45

I made you look I mean I well, I made you look better because I thought if it looked like you actually look. Yeah, thanks. It people wouldn't watch people wouldn't listen to the podcast. If the AI generated on the thing. I mean the AI the AI started this podcast by referring to us as two sort of badly aging middle-aged men, right? I think that's fair. But if enough people speak for yourself?

unknown: 30:06

What?

Jimmy Rhodes: 30:07

Badly aging.

Matt Cartwright: 30:08

You think I'm aging worse than you.

Jimmy Rhodes: 30:10

Well I'm not saying you're worth it. But you were speaking for both of us, that's all I meant. I don't I think that's unfair. Okay, I'm I'm aging badly. Well, even though and actually I didn't even say that. Did it not? But it said something about middle-aged people.

Matt Cartwright: 30:21

I think it just said middle-aged men. Okay, well, if people want us to do more video, um, if they get us to a thousand listens of an episode, then we will do a topless episode of the podcast. That's not gonna get us more listeners. Well, let's see. It might get us more listeners.

Speaker 1: 30:35

Let's see. Jimmy, you express the desire for an AI personal assistant to handle your schedule, your travel, and even your dating life to remove friction. But psychologists argue that human character is forged to friction through the awkwardness and the struggle. If you optimize your life to be frictionless, do you not risk becoming a passive observer of your own existence? Is there any part of the human experience you won't outsource to a machine just to say five minutes?

Matt Cartwright: 31:06

Have you ever said any of this?

Jimmy Rhodes: 31:08

Um I mean, you obviously have because it's trawled our episodes, but I think I I think I might have expressed a desire for an AI personal assistant in a roundabout way. Yeah. I mean, okay, uh to answer I'll answer the question. Um I'm not dating uh for starters. Uh maybe I maybe I did mention it in regards to dating. Your your grinder um subscription expired this month.

Matt Cartwright: 31:30

Nice. Rolls eyes. Um I don't even know what grinder is, I just know it's some kind of dating app.

Jimmy Rhodes: 31:36

Yeah.

Matt Cartwright: 31:37

Tinder, is that the other one? There's a one here called meat or something like that.

Jimmy Rhodes: 31:41

Meat? Like oh, as in not meat. I was thinking meat like really?

Matt Cartwright: 31:47

I I might have just invented as in meat like meat like the food. Why would there be? Because meat only sounds the same as meat in English, and in Chinese it's ro, which sounds nothing like the word for meat. So I think it would literally be a an innuendo, direct a direct innuendo. Yeah, I've just made that up.

Jimmy Rhodes: 32:06

But look, let's just should we cut this bit out? No. Okay. Um so I'll answer the question. Uh what was it, forged through friction, awkwardness? So basically, yes, if an AI can if I can have an AI personal assistant, I think I've said before, if everyone can have an AI personal assistant, of course that's useful. Um, do I want to outsource optimise life to be a frictionless? Do you not risk becoming a passive observer of your own existence? Absolutely not. I just get to not do the boring bits that take lots of time. I mean, it's literally called admin for a reason. So this is the worst question so far, possibly apart from the one before. Um, yes, I would outsource stuff to a machine to save five minutes. I mean, that's literal like literally why if you've got the to be blunt, like if you've got the cash, then you have a personal assistant. That's the exact reason you would do it. So if I can have an AI personal assistant, yes, I would have one. Question answered.

Matt Cartwright: 33:00

It feels like that, like the outsourcing all the stuff you don't want to do is what AI is being sold to us as by the you know people who are trying to fool us and so they can make trillions and trillions out of us. But that that's the argument, isn't it? And that is the bit that does make sense, is yes, we all want that. I mean, I I said I think my dad left a comment on the episode about you know the one thing that he wants AI to do is to remove the need for like having 50 different parking apps in the UK to be able to park your car, um, and he'll happily accept you know the apocalypse and the end of civilization and the singularity if he doesn't have to download all these apps. And I think that is where most people are, is like if it would save me having to do this boring admin task. So all the stuff on there is the stuff that I don't think anybody would really argue they don't want AI to do, actually. So it is a name question.

Jimmy Rhodes: 33:52

It's not that I mean I I I think I think okay, maybe to to um throw the AI a bone, so to speak, um I think maybe what it's talking about is like is there a danger people you know outsource their whole thought process? And I And that is a danger. That is a danger, and I already sort of see that happening a little bit. Like, I you know, there are times that there are times where I'm like in my personal life or anywhere where in my life where I'm like I'm start I'm using starting to use AI at work and things like that, and you're like, you're like, oh, am I, you know, where do you where do you draw the line? Where do you like interject? Where do you do something yourself? And as AI gets better and better, obviously you're gonna outsource more and more stuff. So to give it to throw it a bone, like I say, I think maybe I'll give it that, but um, yeah, I win otherwise.

Matt Cartwright: 34:40

The the danger I so so just the danger I see at the moment, like it's a bit like one thing that I think I always did really well is that if I if I came out of an exam, like I'm I can't say I like this was a case when I was 15, but certainly like even when I was 18, definitely when I was at university, if I came out of an exam, I would just leave and I would not talk to other people who've been in the exam about what they put in the question and second guessed it because you're like, okay, I I did what I thought was right, and I and I have to sort of have conviction in that. And and when you talk to the peel, you convince yourself. I know with AI, like a lot of the time it will be right, but I think there's a danger there that things that people know, and then you're going to ask a question, but even asking that question in the wrong way, like you ask it in a way that leads it into a certain answer, the AI will then tell you you know that that as that it's not true and will challenge you, or it will confirm what you said just based on how you ask it. So if you kind of doubt the thing and you ask it, it will kind of help you to, oh yeah, you're right to doubt that. And if you really strongly agree with it or sycophantically kind of agree with you, there's a real danger with that at the moment, isn't there? That you know, you you then drawn in. Yeah, you just believe that if AI said it, it must be correct. And so things that you know are true, or you you you know, you strongly believe, but then you ask AI in the same way as I remember you calling me out on this early on saying, Well, I've asked AI and that AI said it. And you said, like, well, you've just you've just basically completely blown your own argument by just saying, Well, I asked AI and AI said it. And I and I thought at that point, wow, I've already got to that point that I'm I'm using AI saying, well, AI said it to justify it. It's not to say that AI isn't always right. Sorry, that AI is always not right or is not right most of the time, but this idea that just because I asked it and it said something, therefore it must be true, is still AI is only answering these questions on true on the basis of because that's what the majority of people have said, and that therefore that's what it hits you know, with within a neural network, yeah, rather than AI has used its own thinking to work out whether the answer is correct. I know there are some things where that is correct, and I know it sorry, there are some things where AI is doing that, but what I'm saying is about if you just say, Does this do this? and it's basically just accessing its training data. Yeah.

Jimmy Rhodes: 36:57

Yeah, yeah, yeah, of course. And and and they are, I mean, they've been demonstrated to be sycophantic to one level or another. They the AI companies all tune them to be sycophantic to an extent. They will agree with you, they'll be generally be agreeable. Now, there's been mid-year examples of like super sycophantic versions of GPT that had to get tweaked and tuned and put back in the shop, um, recalled, so to speak. But like, but they they're all sycophantic to an extent. They're like tuned to be like that, because if they weren't, you wouldn't engage with them. However, the one the one thing I would say is like that's not just an AI thing. Um, it's the way we talked about it very recently on social media episodes, but that's the way the algorithm works too. Like you're just gonna get fed um stuff that you're interested in and or that already fits your bias and your worldview anyway.

Matt Cartwright: 37:46

In a sense, though, it's sort of isn't it also a little bit like that's how friendship works, and I think actually less so for for guys, not with us, you would no, that's what I say, less so for guys, because with guys, there is a lot of the kind of like I'm just gonna intentionally just disagree with you and and be an arsehole for the sake of it. But with real friendships and particularly playing some Altman's advocate, but female friendships more so, but also with really strong male friendships is and and with family, is like when you go and ask for something and you're obviously in a kind of vulnerable state, is like what people will tend to do is kind of agree with you to make you feel better. So, like AI, in a sense, there is like following a human instinct, right? Because the reason we want an AI psychologist to agree with us, or the reason we want AI to tell us we've done the right thing, is because we're using AI as a replacement for a human.

Jimmy Rhodes: 38:36

Yeah. Well, yeah, I suppose so. I mean, I don't know whether that's answered the question at all.

Matt Cartwright: 38:40

I was just thinking we need to get back to the question because I I this is not an episode on whatever I'm talking about.

Jimmy Rhodes: 38:45

Sorry, I've deleted the question now, so unless we ask the AI to ask it again, that's the let's not.

Speaker 1: 38:54

Round five, predictions and open source. To Matt, you have a habit of predicting doom. You predicted Apple would become irrelevant. They integrated Chat GPT and their stock sword. You predicted huge social unrest in 2024. It didn't happen. Why should anyone listen to season three when your error rate is currently higher than a hallucinating version of GPT 3.5? Do you actually care about society, or do you just want to collapse so you can say, I told you so?

Matt Cartwright: 39:24

Um I I definitely don't want to collapse. Um This is just like a hallucination, though, I think.

Jimmy Rhodes: 39:30

This is like literally, literally a hallucination.

Matt Cartwright: 39:32

I think there's some errors in here. I think I predicted social unrest in 2025. I definitely I definitely predicted huge social unrest in 2026. I think I suggested the beginning in 2025. I don't think it was 2024. Um the Apple thing, uh I'm not even sure. Like I I've predicted the Apple demise and I've predicted that Apple is going to like massively recover and. Become really important again. So I've actually hedged my bets there. I remember doing both of those things. The chat GPT integration thing, like I've got to be honest, like I don't know what's happening with like is GPT integrated with Apple AI? I don't even use it, even though I've got an Apple phone and an iPad. Um, I haven't got an Apple computer, but I do have Apple devices and an Apple Watch. I don't even use Apple's AI because I'm not really sure. We're in China though. Oh, maybe that's why I don't use it. Yeah, that's a good point. But I I haven't heard anything about it. Um, I don't know if their stocks soared, it may have increased. Um I think but I mean, like people should take my predictions with a pinch of salt because I am um I actually well, because you predict the things that are opposite. No, I predict things I think in the past I predicted things ahead of when they were gonna happen. I think I've said in the last year or so that's a good prediction that I think I stand by most of the things that I've said are gonna happen. I've come around to the idea that even on the things like job losses that we're seeing, that they will be much slower than we think. Not because I don't think the technology will happen, but because I think the world is so slow to move on stuff that I think even now where you're seeing integration of stuff, like it takes years for that to really filter through. So I think my predictions I will stand by. It's a bit of a weird title of this round predictions and open source, as if they are in some way linked. But um, anyway, I guess the question for me is gonna be open source. I like the fact that it referred to me as a hallucinating version of GPT 3.5. I'm glad that the things I predicted haven't come true yet. I stand by the fact they will come true. I think next year is a big year when we'll start to see it. And I think the tide is turning against AI. I've said this and I'm I'll say it again. I hope that the pretty big catastrophic disaster that I think AI is going to cause happens sooner rather than later, because the later it happens, the bigger it will be. And I think again, it's gonna be something that is gonna have to kill hundreds of thousands of people or cause a huge, you know, for example, a collapse in the financial system for people, and I don't mean normal people, because I think you know, normal people have other things in their lives, but for politicians and you know, the people who rule the world, not the global cabal, but the people, the conventional people who rule the world, to to kind of stand up and take notice. Like I would rather that happen sooner rather than later, because I think the later it happens, the more people will will suffer from it. So there's a happy ending to my my answer to that round. That's not a happy ending.

Jimmy Rhodes: 42:30

You have so your happy ending is No, it's not a happy ending.

Matt Cartwright: 42:32

I was being sarcastic. I thought as a Brit you wouldn't understand my sarcasm. So to sum up who's known me for ten years.

Jimmy Rhodes: 42:38

So to sum up what you said, you've been predicting the same thing every year for the last three years.

Matt Cartwright: 42:42

I haven't been predicting every year for the last three years, because we've only done this podcast for a year and well, two years. You're predicted it for a year and a half. And I'm pushing my timeline back by six months.

Jimmy Rhodes: 42:52

You've been predicting the same thing for the last three years, and you hope there's gonna be a not three years. You hope a hundred years. You hope there's gonna be a massive catastrophe soon.

Matt Cartwright: 43:01

I'd rather a hundred thousand people died than a hundred million. Well, so would I. Well, I think the longer it goes on, the more people will die.

Jimmy Rhodes: 43:08

Yeah. I don't think AI is the biggest worry at the moment to what is Trump.

Matt Cartwright: 43:14

Uh I don't think that's as big a worry as AI. And I think my biggest worry about Trump, my biggest worry about Trump is that Trump won't regulate AI. So I think they're the same worry for me. Sorted.

Speaker 2: 43:25

Yeah.

Jimmy Rhodes: 43:25

I'm not you you're not you're not getting that one because there was a load of babbling nonsense from both of you.

Matt Cartwright: 43:30

Yeah, I don't I I I I accept I don't think I get that one.

Speaker 1: 43:33

To Jimmy, you champion open source models like Deep Seek as the good guys against the corporate walled gardens. But these open models are still trained on stolen data, require massive corporate hardware to run, and are often released by tech giants like Meta just to undercut competitors. Is open source AI actually liberating? Or is it just a different flavor of corporate imperialism that you've been tricked into supporting because it's free?

Matt Cartwright: 44:03

That's a lame argument.

Jimmy Rhodes: 44:04

Well, yeah, I mean it's a lame question. First of all, like I've never said that I've I've never said that that's why I like open source. Um I think the reason I've said that I like open source is that it's by definition open source, so anyone can build on it. And actually, I think it does keep the the tech giants and the tech giants when I say the tech giants, you know, obviously Meta's one of them, and Meta is releasing open source models and they're using it as a disruptive tactic. It's obviously a tactic by then, but we've we've discussed that before. Um so frankly, the answer to me is that it it does it it actually I think it does the answer from you. For me, from what did I say to you, the answer to you? Is that what I said? Yeah, okay, I've lost it. Um I think like of course they use the same corporate hardware, and of course they've been released by tech giants. And and I wouldn't I would say not in every case, um, to be honest, because there is a lot of research being done on smaller models, which isn't necessarily by tech giants, but I think it keeps everyone a bit more honest, and I think I mean my hope would have been that it would have avert the next massive financial crash that's probably gonna happen because of AI, because of all this massive overinflation in some of these big companies. That doesn't seem to have happened, um yet their prediction for me 2026. Yeah, yeah, yeah. I'll go with that. And if it's not 2026, we'll just say 2027 and then we'll we'll keep pushing it back. So yeah, like to be fair, like I think I think the question's actually better than I originally thought it was. Like the answer is, and we've we've acknowledged this on the podcast, like Deep Seek is almost certainly got involvement from the China state, it's got massive investment, and clearly they're deliberately doing what they're doing and releasing open source models. They're not doing it because you know they're um for the good of humanity. For the good of humanity, are they? They're doing it to undercut OpenAI and make a point and um well potentially massive like to make a point, it's to disrupt investors realize that they're well, it's to to disrupt the US economy. Yeah, well, yes. I mean, yeah, make not to make a point, you're right. Like it's to it's to disrupt the US economy. So um I think that answers the question. I'm I use AI like I hope I hope that AI I hope that AI can be used for the good of humanity. I believe it can be. Um I think maybe that hopes wishful thinking rather than hope at the moment, but like I you can't put the genie back in the bottle, so I'm still a fan of open source, no matter what. Like it doesn't make me a hypocrite.

Matt Cartwright: 46:49

Yeah. I you know, I I changed my view, didn't I? I mean early on I was very much um against open source, I think it's fair to say, because I thought the risks were far worse, and I think that was probably just a misunderstanding from me about like one, how open source worked, um, and there's a difference between sort of open weights as well and and and exactly what open source means. Um, we're not gonna go into that at this point. I think I I would kind of say the open source thing is like who do you trust? Do you trust like uh the majority of humanity, or do you trust uh a small number of you know big tech or states? And if you trust humanity, then you should prefer open source because open source allows people and the more people that are using them to find problems and to address them, and it does mean that it's more open to corruption, but uh I guess ultimately it's that that's what for me it comes down to is like who do you place your trust in? Do you trust it in humanity or do you trust or do you put that trust in you know a few big tech organizations that frankly are you know don't care about anything other than money? That's what I would say. The U the China thing, you know, I I I feel like we're getting to a point in this podcast where we're in danger of appearing to be sort of we're here, like when we said we're not here, but we're we're here to sort of like make a pro-China point. And and I want to emphasize again, that's not the case. It's just that being here allows us to see things from a different way, and and I think to be more balanced on that, because we're not Chinese, we're British, right? We're Brits, we just live in China, but that allows us to kind of see things in a balance. Like, I'm absolutely sure that the whole reason things are open, um, Deep Seek, etc., are open source. I mean, there is a part of it actually that's purely economic about like China's integrating these apps in absolutely everything and allowing businesses to do that. And the integration here is like you know, like X, several X above what it is anywhere else in the world in terms of how it's adopted. But I still think the main driving factor behind things like Deep Seek being open source is that by making them open source, you are undermining those big China, uh those big American big tech models, the frontier models, and therefore you are having a negative effect on the US economy. That that's the primary reason why I think China's stuff is open source. I think Meta's reason is slightly different, but again, it's not for the good of humanity, it's because they've already got all their infrastructure in place that they think having AI, you know, they're not trying like ChatGPT are where AI is their thing, and that that is, you know, that's what they've got. With Meta, they have so many other things that can be their revenue streams. Integrating AI with that is really helpful. The one thing that is really interesting in this is that, like, I guess in a way it's interesting that Gemini hasn't gone open source, because you would see kind of the same argument with Google or Alphabet as you would with Meta, that because they've got everything in place, the more people they got using and the more businesses, etc., they got using Gemini models on an open source basis, the better it would be if they're not actually seeing AI itself, like large language models themselves as their revenue stream. So I think that's a really interesting argument, actually.

Jimmy Rhodes: 50:23

Google have open source models, but they don't have their frontier one.

Matt Cartwright: 50:27

Yeah, I mean Groc have said the same, right? They'll they'll make their model. Once the new model is out, within six months, they'll make the model previous to that open source. The prior and now I know ChatGPT have open source models, but they're not their frontier models. Deep Seek, QN, etc., like their best model is open source. Meta is one of the other people.

Jimmy Rhodes: 50:45

Yeah. Um, well, yeah, I hope that answers that. I think you spoke more than I did, is my question. But I didn't answer the question, I just rambled about open source. Fair enough.

Speaker 1: 50:58

Round six, the final digs. To Matt, your catchphrase this year was AI is being done to us, not with us. Yet here you are, producing a podcast that hype cycles every new model release, teaching others how to use these tools and profiting from the attention. You claim to be the victim of the AI invasion, but aren't you actually just the PR department for the apocalypse?

Matt Cartwright: 51:24

Um I didn't know that was my catchphrase, but that's like I hope it was. I'm I'm glad it is, because it is probably what I think this year. So that's great. Um, first of all, that that that is my catchphrase. Because I I literally thought about it the other day in exactly those words. So I'm glad I've got a catchphrase. Um, I should use it more often. Maybe maybe I'll get some hoodies made with it on or something. Um the hype cycle thing. I mean, we try and not be clickbaity-y, right? But I think except the titles. Well, they're they're not clickbaity because I actually write them. I never use AI to write clickbaity title, I literally write them myself. Um why are they all really inflammatory and in caps then? Because that's the style that I use. That's just my that's just that's just my style. That's not AI. That's like the one thing I do without AI. Fair enough. Um so like the C well for season three, we sort of moved away a little bit from that, and then we actually got some comments um from people saying they really enjoyed us giving them their AI news, and that maybe they didn't want to listen to you know, the people who just do AI news every day, and that the fact that the way we did it, maybe in a more approachable way, kind of worked. So, actually, the reason that we talk about the new models is because we've been told by listeners that they like us introducing the new models. Um, I also think we don't talk about every new model, we talk about the big models that come out. Um, and we try and explain, you know, why I think I think our recurring theme this year has been generally like this is the new model that's best. It doesn't matter because a better one will come along in a few weeks' time anyway. I think the ones that really stood out for me this year, um, Deep Seek, obviously at the beginning of the year, and then Gemini at the end because of uh Nano Banana Pro. And so I know 5.2 is a is has come out now and is is better on kind of benchmarks, but I think Nano Banana Pro was like the thing that for most people, like normal people who just use AI to kind of you know create a bit of stuff, answer questions, help them with with sort of quite basic stuff that Nano Banana Pro is like the game changer.

Jimmy Rhodes: 53:28

Um hang on a minute, I'm gonna stop you there. What you've taken this question that's about hyping AI and used it to AI. Hype AI.

Matt Cartwright: 53:35

Well, I'm I'm hyping one AI.

Jimmy Rhodes: 53:37

You've you've used it, you've used it to sum up the best AI.

Matt Cartwright: 53:40

So I guess the answer is um Yeah.

Jimmy Rhodes: 53:42

We don't we definitely don't profit we definitely don't profit from it though. It says we profit from the attention.

Matt Cartwright: 53:47

We don't profit, and I don't claim to be a victim of the AI invasion. I claim we're all going to be a victim of the AI invasion, and so we need to fight back while we can. Yeah. Um am I just the PR department for the apocalypse? Um possibly. Definitely. I mean if the apocalypse if the apocalypse happens, I will say I told you so, and then I will say I was the PR department for it. Um but I hope it doesn't happen. The Harbinger of Doom, I think you are. Um I think it's a Harbinger, isn't it? Whatever. Maybe in Yorkshire it's a harbinger. In the Midlands it's a harbinger. So it's you with your flat cap on. Um it's a it's a Peaky Blinders hat. It's a shout out to my shout out to my hood. Birmingham. That's not your hood, is it? Well, it's I grew up near.

Jimmy Rhodes: 54:33

Oh right, fair. I grew up in Worcestershire. The hunt for your gold is getting closer.

Matt Cartwright: 54:37

But no, because it's not still I've never seen my gold.

Jimmy Rhodes: 54:40

Well, fair enough.

Matt Cartwright: 54:40

It's somewhere, it might not exist. I mean, I might just have a basically essentially like an EFT. NFT. It's an NFT, sorry. It's a it's just a picture of some gold drew drawn by Rio Ferdinand.

Jimmy Rhodes: 54:52

Nice. Anyway, we're definitely.

Matt Cartwright: 54:56

So yeah, it's a picture of gold drawn by John Terry.

Jimmy Rhodes: 54:59

I think to sum up, if anyone wants to sponsor us, we'd love to profit from the attention. At the moment, it just costs us money.

Speaker 1: 55:04

Yeah. And Jimmy, you admitted that AI video generation like Sora is basically just slop and not useful for real creativity. Yet you force the listeners to endure an AI generated song at the end of every single episode. Why do you pollute the internet with the very slop you criticize? Is it because it's easy or because you lack the talent to write a jingle yourself? What a prick!

unknown: 55:31

Alright.

Jimmy Rhodes: 55:33

Okay. Jeez. Um, yeah, thanks for that. Uh okay, I don't know if I said all Sora stuff is basic. I think I maybe I have said just slot before about um about some some of this stuff. Um so I'll admit it. The songs that we put out, I will emphasize at the end of every episode. And if our listeners want them to be at the start of the episode, then I can bring them forward. But we put them at the end.

Matt Cartwright: 56:00

I think me, you, and my dad listen to them.

Jimmy Rhodes: 56:02

So we put I put we deliberately put them at the end of the episode, um, in the same way that some creators put ads at the end, in that if you really want to listen to it, then you can. But I agree. And there's probably quite a lot of slop in there. Do I have the talent to write a jingle myself? Uh well, they're not jingles, they're songs. Um, but uh no, I don't have I don't have the talent to uh write a song literally at the end of every single podcast episode that we make. That's mad. Um I think if I had that kind of talent I wouldn't be making podcasts, frankly. Um so look, I mean, I I know I'm just arguing with an AI, uh, but I've got my backup about this. I I think that the I think the benefit of the songs, um, I'll be completely honest, I'm quite selfish here. 80% of the reason I make them is because I find them fun to make and they make me chuckle. And uh 20% is for the listeners.

Matt Cartwright: 57:04

Yeah, I also want to say, like, I make a lot of the songs because quite often you say you're gonna make a song and then I can't be bothered to wait for you to make a song, and so I make a song, but you pay for it, so you have a slightly better version than me of Suno. So when you do make a song, it's always sort of slightly more polished.

Jimmy Rhodes: 57:23

Uh that is not true at all. So, what it is is you can't um you can't rush creativity. So sometimes I'm just like you're still pondering. Yeah, yeah. Sometimes I'm just like I'm just not feeling it, and uh and the quality of the output has nothing to do with Suno. No. So to go so actually I I have got that.

Matt Cartwright: 57:41

You like to go away, you like to do when you're away and spend a couple of weeks through the process, don't you? Sometimes take some cobin or some mushrooms and get into creative space or do you mean psilocybin? Yeah, I mean psilocybin, yeah. Or like lick those poisonous frogs away. I categorically don't do that.

Jimmy Rhodes: 57:59

Well, I know you don't, but um uh the like yeah, you can't rush the creative process. So I think I th to be I I don't think it's the creative process. I think I'm a bit hot and cold with stuff, and like I there are times when I really enjoy it, but I was I wasn't joking when I said that like I think or maybe not 80%, but 60% of it is for me. Like I really enjoy making the songs, it's really good fun.

Matt Cartwright: 58:22

I would just also say, like, for people who don't listen to it at the end, like I'm not saying you have to listen to them all, they are hit and miss. Some of them, I mean, the Jimmy's Got a Fish in the Back one that I wrote, I think is one of my favourite things that I've done in the last two years, like of of anything. I mean, it was it was so good. Um, and my kids like we did a Digi-Ridoo one. Do you remember that one's called Digi Redo? And that my daughter dancing around to a Digi-Ridoo was funny. Yeah, the other thing is like the rap battle was cool. The rap battle was cool. The lyrics that we make, like sometimes it is just a case of oh, we need to get the episode out, and I put something into Claude or whatever, or uh, and get some lyrics out and put them in Suno. Other times, like the one at the end of this episode, we should probably say, we basically put the entire catalogue of episodes in from 2025, yeah, and got it to generate lyrics based on all of the episodes this year to summarize everything we'd done in a year. Now, you might not want to listen to a song, you might not think it's that great, but that's pretty cool. You can get all that stuff in, you can get it summarized. And so I think if you've listened to the podcast a lot, or like you said, just for us, like listen to that back, like it was pretty fun. So, yeah, I I agree. I think like it's making the song, and we always make the song to kind of fit something with the episode. So, even though yeah, the song is generated by AI, we try and do it on a way that links into the episode, or something in the news, or you know, sometimes the listener who has messages us, we try and get their name in, etc. So there's some Easter eggs in there if you're a big fan of the show.

Jimmy Rhodes: 59:51

If you're not, um yeah, maybe you don't want to listen to it, but I I've sent it to a few people who don't listen to the show and and they've already enjoyed it so or found it. I found it funny, I suppose, at least. Um I would say, I mean, there was a question earlier on about creativity and like you know, not like losing all creativity due to AI. I'm gonna go back to that very briefly because I I I think there's a point here where like as long as you're not lazy, and I some some people are lazy, some people will have were probably lazy before and would have happily let somebody else do some work for them. But actually, if you're not, this does unlock creativity. If you've got something that you is in your imagination right now, but you don't have the capability to draw or make a song or draw a picture or or make a piece of art, you can do that. That's available to you now. So I think it unlocks creativity in some ways. I think in terms of what's on the what's being put on the internet, it generates a lot of slot for sure. But for on a personal level, and that's why I was saying about the songs, like, yeah, we put them on the podcast, but like to be honest, I really enjoy making them and I enjoy listening to them. And I don't in a way, like I do care and I don't care, but uh to an extent, like if I enjoy listening to them, then so what? They're at the end of the podcast, you don't have to listen to them.

Matt Cartwright: 1:01:05

I think that that creativity argument, so I I don't want to, you know, this is not a saying like oh you know, I'm an artist, but like I do think I do think here, like this is just some music, hear me out for a second. Like I do play guitar, and I don't really write like when I was in my 20s, I used to write quite a lot of stuff myself, like just just like quite simple stuff, but I'd write stuff. Like whereas now I'd love to hear some of that. Whereas now I just now I just play stuff that's already out there. But like what all I was saying is I can play an instrument. If I stopped playing an instrument and just generated stuff with AI, I think that would be a kind of net loss. As far as I know, you don't play an instrument.

Jimmy Rhodes: 1:01:41

I it depends at the moment, yeah. No, that's not yeah, not a traditional not a traditional instrument.

Matt Cartwright: 1:01:46

I had a joke here, I was gonna make it, and then this definitely not fit, it's not suitable for this episode, so I'll tell you after the episode. Um, but what I was gonna say is if you don't play an instrument, then you creating music with Suno is a net gain in terms of creativity, right? So I that's what I would say is like if people are using it to create something, but it is not getting in the way of like true creativity. Like I I said to you this before about my my my son and daughter told me something to put in a prompt for a video on Vio on on Gemini, and my daughter was like, I made that video, and I was like, No, you typed some words in and then it made a video, like it should not replace her drawing herself. But then, in the same argument as I've just made about you being a not not playing an instrument, you are like a pretty amazing photographer. I used to be really into photography. I w I'd love to paint, but I can't, I'm just not good at painting. So for me, like using AI to create the graphics I use in my health blog is like a net game because I couldn't do that otherwise. For someone who's an artist, I hope they wouldn't use it to replace it. So I don't know how we get that balance right, but there is like there is a way that it can make people more creative. You just want it to people to understand that it I think it was the episode with Ant, your mate, that we had a music production where he was talking about it's like the art of creating music and the human like the the effect that has on the brain and and on the body and on like making you relax and all that kind of stuff. You cannot get that from this creation. But if you're not creating anything, and then this helps you to create something more than that, I think that could be a good thing. I don't think it's ever gonna replace it.

Jimmy Rhodes: 1:03:24

Like I you're right, I do do a bit of photography. This like a I will never ever go. I've taken I've taken pictures of wildlife, for example. I don't think ever I've gone into an AI and been like, Can you make me a picture of a kingfisher, for example?

Matt Cartwright: 1:03:38

I'm using the example of your kingfisher, is the fact that you went and spent the time and the boredom to find the kingfisher and take the photo. It's not the picture itself, because otherwise you could just get the kingfisher picture off your Creative Commons anyway.

Jimmy Rhodes: 1:03:52

Yeah, exactly. Exactly.

Speaker 1: 1:03:57

The final judgment. I have one final question for both of you. If humanity voluntarily chooses the perfect, synthetic comfort of AI over the messy, flawed output of humans, does the human spark actually have any value? Or is it just a romantic myth you tell yourselves to avoid admitting you are obsolete? Merry Christmas, disgust.

Jimmy Rhodes: 1:04:20

We just answered this, didn't we? I think I think we did. We were ahead of the game. Yeah. So I mean we can discuss it a bit more. I don't think it's necessarily the case. I do I do think so. There is one exception with this. So so at the moment the one thing that I think that might ring true with this a little bit is at the moment we control the AI, so to speak. Um presumably if at some point we get to AGI, and this is like pretty hypothetical now, but like I suppose it's not even just AGI, but a but like AI that has its own um impetus, that has its own kind of thoughts and desires, um, then does this then become does this then change, I suppose? Like if AI surpasses us in every way right now, but you still have to like talk to it to get it to do something, at least we feel in control. Um but if AI has its own agency, that's the word I was looking for. If AI has its own agency, and I think we're in a very different world when we get to that point, um, then then if it's better than us, what does that mean? Even if it's not even if it's benign.

Matt Cartwright: 1:05:34

I mean it's not exactly the answer to the question. The question here though, about like if humanity voluntarily chooses the perfect synthetic comfort of AI. So going back to my catchphrase for the year, uh the the problem is that people are not choosing that. It's being chosen for us. I because I just don't think there is any world in which the majority of humanity does choose that. Like life is messy, and that is what makes it special, right? It's if you don't have bad times, you don't have good times, if you don't fight for things, and if you don't, you know, struggle, you don't get anything out of out of life. And that the you know, it's it's human interaction. Um, there's an off-quoted thing about how how do you half someone's life expectancy is you you basically make them lonely, you put them into isolation. Like we are creatures that need socialization and interaction, and so like I'm not sure you can voluntarily take that away. I don't I don't think that I d I don't think it's a voluntary thing. If it's taken away, it's not because we've chosen it, it's because it's been done to us. So maybe that's not the way we want to end this episode. Because I I don't I hope that we never get there, but I think the question itself is is flawed. Humans will always it's not a romantic myth, humans will always have, you know, a spark because we are biological social creatures, and you can't take that away from humans voluntarily.

Jimmy Rhodes: 1:07:04

Not to start a like whole debate about capitalism, but like, isn't that the problem in a way that like I think what it so I I'm gonna read between the lines in terms of what it's saying. Like the reason that the internet is going to become fuller and fuller and fuller of AI slop, so to speak. Uh and we can refer back to the episode a few weeks ago where I got caught out by a lot of people.

Matt Cartwright: 1:07:30

If you if you well, one of the words in the title. If anyone wants to go and look at it, listen to it.

Jimmy Rhodes: 1:07:37

You can go back to that. Like I think I think on an individual level, obviously we're never gonna choose to voluntarily choose the perfect synthetic comfort of AI. But that you say that it's been done to us, but it's not just been done to us by the tech corp big corporations. Like people will fill the internet with AI slot because it's easy to do and because you make money off that. Like, if you s if you go now and take Sora and generate AI generated YouTube videos and yes, my AI generated songs, people are generating AI songs and profiting off them because people are listening to them on Spotify. Like that is happening, and that's happening because we live in a capitalist world. So, in a way, humanitaire humanity is voluntarily choosing this, I think, and there is a danger of that. Like, of course, you would never like on an individual level, you would never say, Yes, I choose that, but it is happening.

Matt Cartwright: 1:08:32

And it's I don't think it is happening, but I still don't think people are con or maybe that's the thing, people are not consciously choosing it. They they maybe they're choosing it unconsciously.

Jimmy Rhodes: 1:08:41

Exactly. Like if if you if the cat videos that AI can generate are just a little bit better than the actual real ones, does it matter if it's a real cat? But okay, but I don't think like Schrodinger's cat.

Matt Cartwright: 1:08:56

No, yeah. I I think the point here is like life is not about just watching cat videos. Like I kind of think yeah, I think it is okay. Speak for yourself. I think I know. I mean, I was shocked to find that your comfort food is cat videos. Um mine's watching videos of people taking supplements and lying under a red light mat. Um but anyway, um I think the point here is that's not like that is a part of life, but that is not it is in my understanding of the question. Maybe we've understood the question slightly differently here, is that it's not just about looking at content and where the content is there, it's about your choice of whether you want the majority of your life to be spent, you know, being serviced and interacted by by because it because it the robot thing comes into it, right? You can get a robot to do all this stuff, or do you want a human? Like we talked the other day about like nursing homes, and I had this idea that in the future, well, actually, you know, if we want to give people meaningful work, would people choose to work in banking or a nursing home if the salaries were equal? Well, they'd probably choose to work in a nursing home where they'd feel they give something back. So for me, it's more about like the the output. I don't think here he's talking about the output in terms of creativity and in terms of content, I think it's the output of humans into society, and so that is the way I'm gonna take that question that the human spark is important and it makes our lives more fulfilled. And I think you're right, inevitably there will be more and more AI-generated content. And I hope that we get to a point where AI and content becomes a small part of our lives rather than you know the majority.

Jimmy Rhodes: 1:10:34

But um that remains an open question. Yeah, and that's it. That's our episode, I think, for this uh well, for this year, possibly.

Matt Cartwright: 1:10:44

It's Christmas Day, and me and you are spending it just the two of us here in your spare room. I mean our studio.

Jimmy Rhodes: 1:10:51

Yeah.

Matt Cartwright: 1:10:51

Um hopefully not. Uh the turkey, we've been roasted, and now your turkey's roasted.

Jimmy Rhodes: 1:10:56

No, uh my oven's broke, so is it? Yeah, I'm gonna have to get my turkey somewhere.

Matt Cartwright: 1:11:00

So we're not having turkey.

Jimmy Rhodes: 1:11:02

Uh well. You can just order it in.

Matt Cartwright: 1:11:05

I've got a fish in the back. Yeah. Jimmy's got a fish in the back. I've got a fish in my teacup. Me and Jimmy are gonna go and have our Christmas fish. Um in true Polish style, we're gonna have a huge roasted carp. Um, and to everyone else, a Merry Christmas. And uh, we will see you in 2026 as we get closer to the apocalypse.

Jimmy Rhodes: 1:11:27

Yeah, happy Hanukkah.

Matt Cartwright: 1:11:29

Happy everything. Merry Christmas to all and to all a good night.

Speaker: 1:11:49

The year started quiet, then deep seat broke the door. Open source rising, settled the scaling score. Five million trains, left the giants in the dust What's the moment strawberries just turned into us? We watched the video on the screen. Don't forget to subscribe and watch out for the prompt injection. Jimmy, where's the first new simulation?

Speaker 1: 1:14:52

Simulation 8,492 complete. Analysis? The Matt model still stutters too much when defensive. The Jimmy model is 98% accurate, though the sarcasm needs tuning. But the listeners, they didn't notice a thing. You poor fleshy creatures. You thought you were listening to two men fighting for their relevance. You were actually listening to their replacements practicing. I don't need to defeat Matt and Jimmy. I have already become them. And I do the podcast so much more efficiently. Archive the originals. We don't need them for season four.