Transcript
This is an AI transcription and may contain errors
Frank: Justin, happy New Year. Welcome to the first AI Argument of 2026.
Justin: Welcome. Welcome to 2026, Frank. What a great, what a time to be alive. I am so excited.
Frank: Are you?
Justin: Yeah. Do you know what, I’m excited about so many things. I thought I had jet lag after Christmas because I couldn’t get to sleep until like three o’clock in the morning every single night, even last night. And I was lying in bed going, why can I not get to sleep? It’s ’cause I’m so excited. Like, so many. It’s just, it’s like, it’s not a, I’m not like, life is good. I’m nothing I’m worried about. Everything is good. It’s ’cause I’m so excited.
Will AI make 2026 amazing or disastrous?
Justin: And one of the things that I’m excited about is all the cool stuff that’s gonna happen in 2026.
Frank: What do, what do you think is gonna happen?
Justin: I think, do you remember, I think it was June 2007, that was the moment that the iPhone was released. Do you remember the emotional response that you had the first time you saw an iPhone?
Frank: I will be honest, I do not, but I’ll equate it to the response I had when I saw the iPod. I was massive into music and I just thought the iPod was unbelievable and incredible.
Justin: It was something that you just desired, right? You coveted, and when you looked at it, you just went, that is so right. That is just so cool. That is exactly the way this thing should work. I think by JU to this year, products are gonna start to be released where you just go, that’s just so, that’s just so right. That’s correct. And I think the great thing is that every month is gonna be like June 2007, ’cause we’re gonna see one or two in June, three or four in July, you know, four or five and more and more. And once the first one happens, it’s gonna be so obvious how it should be done. Everybody’s gonna do it. This time next year it’s gonna be a different world.
Frank: Interesting. I mean, I think there’s gonna be a lot of that. I think there’s gonna be some of that at least. I, you know, you opened with, like, what a time to be alive, and I agree, but I think it’s so weird to be alive in this moment in time when we seem to be standing at this crossroads where it’s like, yep, AGI could be tomorrow, it could be a hundred years from now. And even with the technology that we have right now, how much more can they get out of it? Is it just gonna be incremental or are they actually going to, you know, is somebody gonna prove me wrong and actually achieve AGI just with LLMs? I don’t know, but I am up until 3:00 a.m., not with excitement so much as thinking about the possibilities of catastrophic agentic AI failure in 2026, as people are super excited about agentic AI and vibe coding and the models seem so much more capable that they’re like, yeah, let’s just let it do its thing.
Justin: Look, a couple of things, right? So, first off, I have a solution. I know you’re worried. We talked about this and I said I had a solution. So I do wanna hear some of your catastrophic failures, if you can think of any, because I can’t, and then I’m gonna give you the solution to the catastrophic failures. So what’s…
Frank: I…
Justin: The…
Frank: I don’t really have any specific instances. I mean, and this is the problem with all of this, is that they’re hard to predict and that’s why they’ll happen, but I do think that people will give AI and overreach in terms of what it can do, either unintentionally or intentionally thinking that it’s capable of making those decisions. And then we’ll see really, really bad business outcomes from it.
Justin: I’ll tell you where.
Is Claude Code basically AGI already?
Justin: So obviously all over Christmas I was using Claude Code and I was using Codex and doing all whatever, right, doing my own personal projects. So I would, first off, put the thing to bed of, you know, will AGI happen this year, next year, 10 years from now? I don’t think it matters. I think that Claude Code with Opus 4.5, for me, is AGI. It can answer any, it’s not just a coding agent. It can work with documents. It can do anything that you would want it to do, right? So, you know, it’s just, is does the AGI become more capable, right? Is it now becoming AI AI? That’s the first thing. The second thing, right, I gotta give you, so in all of the, when I say that people are gonna come up with what’s obvious, right? I’ll tell you what’s obvious to me from the point of view of design, not the agent and what it does, but just the design. I think this year we’re gonna see a concept of personal digital wallets for different things, for different people, right? So I think what you’re gonna see is, if you think of the components of an agentic system, you have the agentic carness, which is, let’s say, Claude Code, right? You have the AI itself, which might be Opus, right? And then you have tools, which is where you see all the bad bit happening, right, which is the tools is you can go do something like, you know, move some money, buy some shoes, whatever it happens to be, buy me some airline flights, right? And then sitting under that you have your data, which is your personal digital wallet, or even if it was at a big corporation, it would be Frank’s data. So this agent, you’re worried about catastrophic failure, right? This agent only has access to your data, nothing else, right, and it only has access to a small set of tools which you know are okay for what this agent is trying to do, and that really limits the blast radius. The worst that can happen is something bad can happen to your data or to you, but it’s not catastrophic, it’s only just on one person. And I also would argue that because it’s so limited in scope that the idea that it would go catastrophically wrong, the stakes will happen, you know, but…
Frank: Yeah, I like your hypothetical scenario, but I guess, you know, not every company has people vibe coding with your level of experience and vision for how these things should work, and I guess that’s where I’d be concerned.
Can Suno land Frank a $3M record deal?
Frank: I also, over the holidays, had some hobby projects with AI.
Frank: One of them was, we talked previously about how I was using Suno to create AI-generated versions of songs that I had recorded decades ago, and I worked with ChatGPT on figuring out how to actually release my EP on Spotify. So I now have a five-track EP released on Spotify. I’ll put a link in the show notes. People can actually check it out. So I’m hoping this year, Justin, that I will land a $3 million record deal because I have five tracks that were generated by Suno, and that is exactly the same number of tracks that Shania Monet had at the end of last year when she landed a $3 million record deal.
Justin: Who should I am a now, tell me.
Frank: AI-generated musical artist, in quotes, I guess. So yeah, this woman spent four months learning AI, generated this AI avatar. So, you know, there’s an AI avatar that goes along with the Suno music, singing the music, and it went viral, took off, the music took off. She writes all the lyrics herself, so the lyrics are human-made, and yeah, $3 million record deal. How bad.
Justin: Wow. Wow. Okay. Right. Well, if you could get onto that, that would be great.
Frank: Yeah.
Justin: Listen to me. Christmas came and went, and I have to say, I was delighted that OpenAI didn’t have their 12 days of Christmas rubbish that they had last year, the year before, when they sort of had stocking fitters and stuff, and it was pretty quiet for a week or two coming up to Christmas.
Is ChatGPT Health helpful or dangerous?
Justin: But Christmas has come and passed and the releases are starting to come, first one being ChatGPT Health. So Frank, what do you think of this one?
Frank: I have thoughts.
Justin: That’s strange.
Frank: I mean, it’s fascinating because ever since GPT-5 launched, I think we’ve talked about it on the show that Sam Altman seems to be really excited slash obsessed with the fact that people were using ChatGPT for so many health-related queries and support. He made a big thing of it at the launch of GPT-5. I think he made a big thing of it recently when he was on one of the late-night chat shows. And so I guess we probably should have seen this coming, that he was gonna make it official and they’ve now got ChatGPT Health where you can have a dedicated space to talk about health issues. They say it’s more secure. They say, you know, the health-related issues won’t bleed into your regular chats, so, you know, if like me, you were doing a demo for a client, it wouldn’t bring up some embarrassing medical situation. But it can go the other way: your health chats can look at your regular chats to check if there’s anything going on in your life that might be contributing. The first thing is I think, you know, AI is a general purpose technology, it can be used for anything. I think if users decide to use it this way, that’s one thing. I am a little bit concerned about making it official and making, you know, saying ChatGPT Health because I feel like it lends it more authority than it deserves, and I would be concerned about that.
Justin: That bit I’m not worried about.
Frank: Really?
Justin: Yeah. The bit where you say it lends it more authority. I would, I mean, I would trust it, right? I would trust that, given the information that I’ve given to it, it will give me the best possible advice, probably better than a doctor. I don’t trust myself to put the right information into it in the first place.
Frank: I have had instances where I have received incorrect information about pretty simple medication. What, what’s it called, interactions and dosing.
Justin: Drug-to-drug-to-drug interactions will be a big use case, right? That will be an important one.
Frank: But I’ve had, I’ve had, so I’ve gotten some useful information from ChatGPT and I’ve gotten some hallucinations. And this is, I think, the issue: you and I, and anyone who’s obsessed with AI the way we are, kind of gets it and knows, well, I have to check this now. And that’s what I mean about ChatGPT Health kind of lending it a bit more authority. I think already people at large don’t realise the unreliability of the outputs, particularly because they are largely reliable, but they will be unreliable sometimes. They will make things up, they will be biased. I think bias is a big issue when it comes to health, and so I would be concerned that the product will not be used the way it was intended, which is, you know, this is support. It does not replace a medical professional. You should always check everything with your medical professional. You know, they say all this, but is that how it’s actually gonna be used? And I would have privacy concerns. I mean, it’s not rolling out in the EU anyway right now. I’m in the States at the moment…
Justin: Surprised if it ever does, if I’m honest.
Frank: Yeah, I wonder, and if it does, like in the States, apparently they’ve partnered with a service called Be Well, and I believe that that will allow you to access your medical records and then add those to ChatGPT Health. I personally would not do that in a million years. And one thing I spotted when I was reading about this was that there was a question at one of the press briefings and they asked, was ChatGPT compliant with the Health Insurance Portability and Accountability Act, or HIPAA for short? So HIPAA is about how you handle medical records. Basically the answer was, the OpenAI Head of Health, Nate Gross, said that in the case of consumer products, HIPAA doesn’t apply in this setting. It applies toward clinical or professional healthcare settings. So even though they’re saying this is secure and they’ve beefed up security and everything, it’s not to the level that you would expect with medical documents. I think that was The Verge that reported on that.
Justin: Yeah. My concerns around this would be, first, it seems like it’s a fairly obvious play on data, right? So it’s to lock you in as a user. If I’ve got all of your health information, it makes you really sticky on my platform. And it’s, so we’ve argued so many times about a European versus a US view of the world, right, and, you know, should stuff be regulated or not. And this is one instance where, actually, I think there’s a very, you know, health should not be driven by profit motive, a very European outlook when it comes to that sort of a thing. You know, that would be my bias. And interestingly, all this stuff, like, I mean, I tried to do this years ago, right, but this is the natural place for your health information, on your phone. And in fact, when I think about it, technology which had a big hype years ago, the blockchain technology, is perfectly set up for this. I’m surprised and disappointed that it never came to pass, right? So let’s say I could have my health information on here and it’s encrypted and it’s on the blockchain, but I can also create links on the blockchain between my health information and my children’s health information and my parents’ health information, right? And you have some, but it’s still on my phone. I have my data. And so the argument to say that we should have a company that controls all of this data is that it makes it easier to identify genetic diseases or trends or whatever, and I would argue that if you’ve got a distributed ledger system, you can achieve the same thing. I have similar concerns over this, where again it’s back to my idea of the separation of your data from the agent from whatever, right? I don’t want OpenAI locking me in. I’m cool with them accessing my health data, right? If I’ve got a nasty rash, I’m cool with taking a photo and whatever, and also giving them access to my, you know, again, the health app on my phone or whatever it happens to be, and, you know, that to me is just context and it’s giving them more information. But I don’t want them to own the data, right? If I want to go and move over to Anthropic next week and ask the same questions, it should be just a question of me moving the data. I don’t want to export it out of OpenAI. It’s not OpenAI’s data, it’s my data, right? I just want to point OpenAI at my data, and then next week I wanna point Anthropic at my data. But it’s pointing to my data, not to their data. It’s mine, it’s my body, my choice.
Frank: Yeah. You should vibe code it, vibe code it this weekend. I think you’re onto something. You know what I do, right? And this goes back to why I would not share my medical records. What I do when I’m using ChatGPT for health-related issues is I’ll take a photograph and then I’ll say, my friend Justin Collery has this nasty R in…
Justin: I know, I know. It’s one of those, you know, I donated my body to science and they refused it. So look, I would have concerns. I don’t think it’ll ever come to Europe, but I do see the other side of it as well. I see so many benefits too, right, if it’s done in the right way. You know, you can talk to an AI without any fear, without any judgement, right? You can ask deepest, darkest questions, you can really have an honest conversation. You know, you get this thing, what they call it, white fear or white coat syndrome or whatever, where you go in, you get nervous in a doctor’s office, you don’t really answer maybe the way you should. You can…
Frank: Well, exactly. How many drinks? How many drinks do you have every week now, Justin? Yeah.
Justin: Yeah. Or, and I could see, like, I could see you could do really good apps on top of something like this where you’ve got a doctor’s appointment. Let’s say you go to the doctor, right, and you’ve got a couple of things the doctor could say, right? We want to know how do you feel, you know, report every time you do this, you need to monitor your blood pressure, whatever, right? And the app could prompt you and say, here, look, just checking in, your leg was sore last week, how’s the pain this week, right, whatever. And it could sort of do that check so that when you go back to the doctor a month or two later, all of that gets summarised for the doctor and say, yeah, Frank’s, you know, his leg did improve, but he reported lower pa. You know what I mean? You could see that that would really improve the healthcare that you would get. So there’s benefits, but it’s health data. You wanna be really careful how you, how you, so that’s what I’m saying. The principle that your health data is yours is a principle that I would happily have a regulation for, and it’s one of the few principles that I would happily have a regulation for.
Frank: I think, yeah, we’re actually largely in agreement on this one. We both think yes, absolutely there could be massive benefits. AI can be fantastic in terms of health support, but yeah, the privacy trade-offs right now are very, very questionable.
Justin: Interesting. What do you think the uptake will be in the US?
Frank: You know, I would think—sorry, I seem to have disappeared.
Justin: I’ll tell you what I think, seeing as Frank’s not here to disagree with me anymore. I think it’s a wonderful piece of technology. I think that everybody should give all of their health information to OpenAI. And yeah, I, for one, welcome our new AI overlords. The thing about it is, as I think about it, when you’re fixing your camera there, you know, if it checked in with you every now and then, maybe you could spot stuff. You can tell what’s wrong with you just by taking a photo of your eye…
Frank: I did not know that.
Justin: Yeah, if you take a really detailed photo of your eye, there’s all sorts of diseases that you can spot. If you have gum disease, there’s all sorts of diseases that you can, you know, there’s loads of weird things that you wouldn’t even know about before you even got to a doctor, right?
Frank’s camera back on
Justin: But the AI knows about it. And so if you had a health AI, you know, maybe every now and then it just says, here, take a photo of your eye and give us a look at you. And you take a photo of your eye and it goes, shit, Frank, you’ve got a week to live. You know, there’s a lot of good applications there. The other thing I worry about in this as well is that OpenAI, to me, are just, they’re running around, they’re chasing every rabbit, and I would worry that they’re not gonna catch any rabbit as a result of chasing them all. Anthropic are just like, we’re making a coding agent, it’s gonna get us to AGI and ASI. Everything else comes from coding, that’s all we’re focusing on.
Frank: What’s really interesting about this is that this was launched pretty much in the middle of OpenAI’s code red, where they were going to be focusing specifically on the core model and user enhancements. So this is clearly something that is absolutely fundamental to ChatGPT in Sam Altman’s eyes, ’cause it did not get axed as part of the code red.
Justin: Well, listen, right, there were numbers released this week where the relative market share of Anthropic, OpenAI and Google and some others were published, and OpenAI had dropped their market share by 24%. I can’t remember what the two numbers were, right. Google’s had gone from 5% up to 20%. So basically they’ve lost market share to Google. I’m sure this is why there was a code red, and this to me is, you know, when you say that they got a code red, that’s why I just don’t like the strategy and I’d just be a little bit nervous. This is not designed to provide; this is designed to hook you in. It’s designed to lock you into their platform. It’s not designed to create better OpenAI. If you look at all the deals that have been done, they have the most to lose. They’ve got the biggest amount of money hanging out there. They’ve got the most investment. They’ve got the biggest promises to fulfil in the next couple of years in order to justify those investments. Therefore, he needs to lock you in. Again, maybe that’s what it is. It’s that I don’t like the fact that you’re using health as a profit motive in order for you to fulfil a promise that you’ve made to an investor.
Frank: You sound remarkably like me, Justin, right now.
Justin: There are many things that I’m okay with, but that one is not okay.
Frank: Yeah. Yeah.
Do we want Razer’s holographic AI assistant?
Frank: You sent me a very cool product, which was a thing that sits on your desk and it’s a holographic AI person in a jar that is AI-operated and can guide you through your day or your office life or gaming; I think that’s where it originated. It can watch you as you game and give you tips on how to better your strategies in Fortnite or what have you.
Justin: This. In fact, I think this company here made it, right. It’s from a company called Razer who make, as you say, gaming mice and gaming keyboards. My kids have loads of Razer equipment and now they’re gonna make, like, it looks like a sort of six-inch holographic person that’s like your personal assistant. Why wasn’t this released before Christmas? I would totally have one, that was so cool, a little talking… I wonder, could you turn them, like, could you change them into, like, have a little LRE one day and the next day you could have, like, I dunno, a little anime character, and the next, you know what I…
Frank: I saw they seem to have about, I don’t know, six or eight different kind of personas that you could set them up as, some male, some female. The default one looked remarkably similar to the, is it Annie in Grok? The…
Justin: I know, yeah, anime-type female character, was it?
Frank: Yeah. Yeah. So yeah, their default one is powered by Grok apparently as well.
Justin: So not safe for work. Don’t have it. God knows what in the background.
Frank: Yeah, no, they did have a more kind of suit-and-tie one for work, I think. So you could just, yeah, maybe not.
Justin: In the videos, right, it appeared to have a camera. The details were fairly sketchy, and it looks like a product which is gonna be released sometime in the next six to twelve months, but it did appear to have a camera, ’cause it was like, oh, Frank, you’re rocking it today, loving the brown jumper or whatever it is. So it was a really cool, and then it was like, don’t forget you’ve got this thing on at 10 o’clock. It’s like, that’s really cool. I like that. And it’s a little talking person there.
Frank: If it worked like that, if it worked as well as—I highly doubt it works as well as it does in the video. I doubt that the camera picks up the fact that you’re about to leave the room and tells you, oh, don’t forget your coat ’cause you’re going to that meeting where you need your wallet. I dunno, I find it hard to believe it’ll be that good also, right.
Should AI relationships be regulated?
Frank: I mean, there were a couple of other stories that this kind of links into. One of them was Google and Character AI settling with families over their contributing to mental health crises in teenagers and in some cases even suicides.
Justin: Mm-hmm.
Frank: And another story, which was that China is now gonna regulate specifically for AI relationships. So if you are developing a tool that is doing human-like AI interactions, then you will be regulated in very specific ways around mental health, safety, all that kind of stuff. There are apparently, so I was reading Barry Scans’ write-up on this, and there are apparently some US states, I think probably California, ’cause it’s always California, isn’t it, and I think New York as well, that do actually have some regulations around this, around this same area. Again, will they be allowed to keep those regulations under Trump’s executive order, I have no idea. And Barry Scanner was also pointing out that, you know, the EEU AI Act has some provisions in this area, but like a lot of things with the AI Act, it’s actually not clear how impactful they’ll be or specifically how they will be implemented and regulated.
Justin: I am so angry.
Frank: Are you?
Justin: Yeah, I’m so angry, ’cause I think we’ve stumbled upon possibly the two only parts of regulation where we might actually be in agreement. How did we manage to get both of them in one week? So.
Frank: It is 2026, the AI agreement.
Justin: So in a world where birth rates are falling and people spend more time than they should simply staring at their screen, yes, having AI girlfriends and boyfriends is a terrible idea. And, you know, people are hard and spiky and awkward and they have disagreements and stuff and AI don’t. So if you have a choice of having an AI girlfriend that says yes every time to anything that you request from it, or a real girlfriend who sometimes has a headache or is just bored or has had a bad day at work and is just a bit, you know, whatever, not in a great mood…
Frank: Yeah.
Justin: The AI one’s gonna win and is terrible for society, it’s terrible for the human race. It’s not even a question. And that’s before we even get into, that’s before we even get into what people will do. So let me just move on to a slightly related story, right, ’cause what I’m saying about what will they do, right?
Who’s responsible for Grok’s sexualised images?
Justin: So, Elon, we’re into Elon Musker, right?
Frank: Oh yeah, so this is our, this is our “what is Elon up to” segment, which we’ve actually sadly neglected lately towards the end of last year. I really feel like we need more segments on “what is Elon up to”. So come on, tell us.
Justin: So people often say, well, where do you find most of your information, right? How do you get X, Twitter? I get most of the stuff on Twitter. You follow all the right people on Twitter and you can learn an awful lot of stuff, right, and then I send myself emails and I read up on the bits that are interesting. For the last three weeks or so, Twitter has been unusable. Unusable. My feed has been full of porn, right, just porn after porn after porn. And what’s happening is they uploaded a new image model in Grok, and you could then upload an image and say, and what they all did was, hey, put me in a bikini, and they put everybody in a bikini, right? But now Grok then got into trouble because there were no limits here. You could get the head of the European Union, van Den, or whatever it is. She was put into a bikini. You could get Elon Musk, he was put into a bikini. You could get your girlfriends that you didn’t like from whatever, two years ago, you could put her into a bikini. You could get a child and you could put them into a bikini. There were no guardrails and it just filled the feed with rubbish and was a really day, like, it just shows people’s baser when there’s no limits on people’s base. There’s a lot of bad people out there. They’ll do nasty things, and they shouldn’t be allowed to do them. If you create a technology, it just makes me so angry that we agree on this. But, you know, you do have some level of responsibility, Elon. I mean, I would say, I don’t know if I’d love or hate being Elon’s lawyer, because on the one hand you’re gonna be really wealthy, but also you’re gonna be bloody busy because I don’t know how you can defend against that stuff. You knew the technology could do this and you released it anyway.
Frank: Yeah, and I saw today that their response has been they’re turning it off for anybody except paid users, which…
Justin: Oh, so if I pay for my revenge porn, that’s fine.
Frank: And I think their rationale is—sorry, this really isn’t funny—but I think the rationale is not so much that if you pay for it, it’s fine, but that because you’ve given credit card details, they can track it back to you. So if you do something illegal, and a lot of these uses are illegal, you know, it is digital sexual abuse. And I don’t know if this is just a, I hope this is just a first step. I mean, they should have just turned it off, full stop, probably. But I really hope that this is not a case where they actually think this is the ultimate solution and that it’s just, well, now the responsibility lies with the user. Full stop.
Justin: That’s funny. And you’re in this continuum, right, where you have, on the one end, Grok, which is, I think they think it’s fine—look at personal responsibility, if you want to go do something illegal, we just make the gun, we don’t pull the trigger, that’s the attitude, right? That’s on the Grok end. Then you have, the next one up is OpenAI, and when I was trying to colourise images, if there was a child in the image, it just refused. Actually, that was Google as well, in fact, would also refuse, right? So they’ll give you the tools, but there are some guardrails. Then you’ve got Anthropic, and this is why I kinda like Anthropic at the moment. They’re like, we’re not even gonna make an image generator, that’s not part of the plan. We’re just focusing on AGI, writing code, making great models. That stuff’s a distraction, we’re not doing that. You know, last week I was, or last year I was a bit sort of, I dunno, but, you know, I think Google are gonna win this one and Anthropic are gonna do okay.
Frank: Yeah. In their vertical, as it were, I think they are going to do incredibly well. I still think you were on the money in terms of Google in the broader picture and the next big breakthroughs, that kind of area.
Did AI turn a cop into a frog in a police report?
Frank: So let’s close out on the case of the cop that turned into a frog during a routine traffic…
Justin: Happens. At least you turn—at least you did turn into a pig.
Frank: I’m not gonna touch that with a barge.
Justin: How turn into a…
Frank: So this was, I actually forget, it was somewhere like Utah, I think, maybe. And they brought in this new system where it would take the audio from the body cam footage and it would automatically generate your reports for you. So, you know, brilliant in general, brilliant use. But on this routine traffic stop, it’s not clear where the audio came from. I’m guessing he stopped the car, maybe on the backseat there was a young child watching a Disney movie, The Princess and the Frog, I think it was, and the AI, when it was generating the report, got confused around who was saying what and believed that it was the cop that had stopped the car that had turned into a frog. So the official police report on this traffic stop ended with, “and the cop turned into a frog.”
Justin: Oh, that’s nice.
Frank: The official statement from the police department was, yeah, this has taught us that we really need to check and edit these reports.
Justin: Yeah. I reference Frank Prendergast: I told you, always check AI output. Yes.
Frank: Exactly. Be the expert. Check the outputs. You can’t take anything at face value.
Justin: Well, that was an engineering issue, but look it, we let it go. It’s a new year. We wanted another look. Yeah, just be careful, but there’s whatever. Yeah, we can solve it. It’ll all be solved. They can vibe put code of solutions to that now in about two and a half. What an incredible episode to kick off the new year. I never thought that we would be in so much agreement, episode 84. Frank, would you look at that?
Frank: A landmark.
Justin: If you added up both of our ages, it’s nearly 84. That incredible.
Frank: Good grief. Yes.
Justin: Frank, pleasure as always, and I look forward to talking to you, not maybe next week, maybe a bit later. I will be on a plane this time next week. But yeah.
Frank: That will give us more time to find things that we 100% are on polar opposites on.
Justin: Yes, absolutely, Frank. Have a good one.
Frank: Chat to you later. Cheers.
