Fable 5 Is Coming Back: Here's What Actually Changed

Nick, we need a moment of silence because it's finally happened. I have it on good authority the Fable 5 is back. Officially back. I couldn't believe it. Actually, but at this podcast, guys, I tried to type Fable 5 down and it the pen broke. The pen literally That's how powerful this model is, guys. I can't even write the name down without my pens breaking. Nick, how pumped are you for this? We talked We talked about it last week. We have a lot to dive into today's episode, man. We've got a lot happening right now. So pumped, man. I feel like my personal GDP plummeted after that Fable export control thing, and uh we're finally back to 3% interest rates. It's going to be fun. I'm really stoked. So I actually have a bunch of prompts lined up that I'm just going to like throw into Fable. And I've decided to just always have Fable-ready prompts on hand for situations like this. Yeah. Um so the second that that ticker gets back, I'm just throwing them all at the wall. Yeah, well, I think it's going to be interesting bringing it back like like a bit of a test like did we ever exaggerate it? Did we think it was too pop- Like I think when it comes back, we'll be able to see what life was like pre- and post-Opus once again. Do you know what I'm saying? Yeah. this could be pre-Opus and post-Opus. This is how we'll remember things. You know how there's like AD, Anno Domini? There's like pre-Opus like post-Opus, PO. And everybody's been PO'd about Opus, let me tell you. Well, we Well, do we We've got Fable coming back. We have also Sonnet 5 just dropped today, so it'll be yesterday for everybody watching. So this is as crisp and fresh as it gets. Unless Jamie gets out and say that would be amazing, but um I I think that's pretty That's what it's double clicking on. So shall we kick off with Sonnet 5? Um and then we can touch on Fable. We've got a few other controversial things that we have to look into. So Yeah. Yeah, let's do it. All right, let me share my screen. It's so unfortunate that Sonnet 5 dropped and like within 24 hours they figured out their export control thing, huh? Like I can't imagine that was planned cuz it just doesn't seem right. Why would you Why would you do both? Sonnet 5 is such an underperformer relative to Fable. It is. [clears throat] And I'll be honest, like everyone's got different views. I've seen them. We're going to go through X on a minute and see what the kind of the people's voice voices are. But we we can walk through this together on on the podcaster. Generally speaking, though, these benchmarks I I don't really think they mean anything. Like to be honest with you, I I always do my own tests these days. But there's a graph that they showed actually. I think it's Okay, this is probably the graph worth looking at. So we have these difficulty settings. So the X axis is how much it costs per task. And the Y axis is just how effective it is. So let's see what we're doing with Do Nick, I feel like I'm in a grad scheme assessment center. But like what model should we use? Like [laughter] I mean, it's clear it's basically just cheaper Opus, right? Like I think that's basically how they've been trying to sell it to us. Um it's the new quote unquote generation. That's why it's Sonnet 5. But yeah, I mean, just looking at this, it's it's just cheaper Opus. One massive step up from 4.6, but who the hell's using 4.6 anyway? Yeah. Um me having to retire with my 4.6s. But it's like I think that Sonnet 5 seems to be the highest performance that currently exists that is not Fable. That's like But only by a fraction. I mean, we're talking like 84 to 85 difference. A lot of people I've seen have used that are talking about that it's actually quite slow. Mhm. Uh it's very slow. And it's not as accurate. Especially for somebody that like me has just been using Opus 4.9 on fast mode exclusively. Um you know, obviously I could switch I could save some money, but it's also like why don't I just use Opus 4.9 on regular mode and save money? Uh listen, I mean, like cost optimization and stuff is is starting to become a big deal for companies. And you see companies like Coinbase starting to roll back and route to different providers based on cost level and perceived efforts. And I think the sort of like mid-level token management layer is going to grow more important, which is essentially why uh you know, Anthropic and and providers starting to roll out like mid-tier models. But it's kind of like the it's kind of like the iPhone, you know? There's like the best of the best of the best, and then they start rolling out like kind of like mid-tier consumer ones, and they're not remembered for their mid-tier consumer ones cuz those really aren't the ones that do any benchmarks or that are important. Um they're remembered for the best. And so, yeah, in my case, I don't really give that much of a crap because I think we basically reached approximately AGI level on most tasks anyway. What difference does any of this stuff realistically make uh for me at this point? Somebody that just like has the money to spend? Very little. But um yeah, societally, I mean, you know, giving access to Sonnet, which probably runs cheaper on their own inference stack, is valuable cuz uh that means that they'll be able to serve more customers, continue scaling up their their customer base without like, you know, downtime and stuff stuff like that. And um probably make it more palatable for people in lower cost of living countries uh that, you know, currently have to arbitrage like Chinese models and and stuff like that. I Dude, I love a good Chinese model. I love a good Chinese model. I love a good Chinese model in the morning. I actually It's funny. I I think the the way forward is adversarial AI. Like you might have seen a big thing Well, big like but it came out recently on People say it's mixture of agents. I prefer Ministry of Agents because I think that [laughter] sounds It sounds a lot Some guys are like, "Hey, you know it's mixture." I was like, "Dude, I know, but Ministry sounds way cooler." And I think that's really important. But I did that like, you know, you might you have your Fable Fire as like your orchestrator, and then underneath that we've got GLM 5.2. We have Opus 4.8. And we I If you're doing super cost-efficient, this is I think is a strat. Your most powerful model is your orchestrator. Call that Fable Fire at the time of recording. And then as a team, you have your ChatGPT subscription. This is if you want to like optimize costs, right? Your chat Your ChatGPT subscription, GLM 5.2, and then the best DeepSeek-V4 model. Or DeepSeek-V4, too. And and the best DeepSeek Like that would be unbelievably like token efficient. That's like slightly more expensive, but the power of that would be great. Because I think that Coinbase and stuff are doing. Like these companies that I'm talking about that are optimizing their usage. Like their token usage goes up double and then their cost go down by half. I'm sure it's because they're doing exactly what you talked about. Orchestration with the smartest model like Fable, choosing shittier models to to delegate to. Do you think as well like I think it's true that a lot of it is in the harness. Like when I say harness, I mean things around the model. Like it's super intelligent, but it's that kind of like prompting and system and tools around it. Like for example, it if you're going to deploy an AI into somebody's business like you know, you're watching this podcast and great. I've got a number I would like to move up this quarter, this month, whatever it is, right? You have a great model, but you're really crap at asking the right questions. Which I think you find a lot, right? When when you go into business like you just wasn't thinking about it correctly. A good model, you know, even if it's super intelligent, that hasn't been told, hey, you need to question their assumptions. Hey, you need to be asking, should we really be focusing in on this? Or hey, here are like several business principles that they need to be think like that will proactively challenge you that isn't sycophantic. Like you could have like Fable six level intelligence that's super sycophantic. It's like, Nick, that ice cream business in Norway is a great idea. Go and do that right now. Like do you know what I'm you there's like one Norwegian ice cream seller in the in the comments down below that's going to kick off about this. But it's like all that to say like I think that illustrates the point that you you want a a you need to ask the right questions. So I think the harness, the skills, the way that you prompt it still is important and we don't see that on these benchmarks. And we're going to pull up some some people have done a Sonic five in a sec. But Nick, what are your what are your thoughts on that, man? Well, dude, the entire time you were thinking I was like, why do we make a J harness for coding but not for business? I mean like I know that there was sort of like a quote unquote advancement recently. Like there was a new drop that I think tried to solve this, but could you imagine a harness that's just optimized for solving business problems, not necessarily coding problems? Don't get me wrong, there's a fair amount of overlap in the like Venn diagram of, you know, effectiveness at coding and then effectiveness at business, and I think they're they're they're reasonably overlapped, but they're not fully overlapped. So, optimizing for business tasks, you could probably actually crank GDP with one of these things now. You probably could. Yeah. We we we probably could actually. I mean, you could build like a Jack and Nick bot that's like trained on all the stuff we've done. I mean, I actually think that would be fun. Like, if you had to systematize a way of thinking, and the objective of the harness was basically try and make the error rate as low as possible. Like, that such that if they were to follow it, they're most likely to win. Like, it not winning would be unlikely. I I do think there probably is some You could probably codify cuz I think a lot of decisions you make sometimes like sometimes they're based on intuition, which is so fascinating. I got sent the video from Steve Jobs, and I I says before we jump into this on a five, they are they are super related. And it was him doing the presentation of the iPhone, and he was basically explaining like, um, what's the pro And he had the BlackBerry. Now, these things have keys on it. He was like, "Keys are great, but what if we think of something new that doesn't need a key?" And what was so interesting about that? And it was like, dude, there were gasps when he released the iPhone. Well, iPhone 1's like this [ __ ] this big, right? And about that thick. It was [ __ ] like this thick. And [snorts] um, there were like gasps in the audience when he released it. And it was But he was a innovator. Like, he cuz he said like, he was famous for this once. He said, "If you did the customer research, everyone said they don't want a screen. They want their keyboards." He said, "They don't know what they want. Keyboards are crap. They're not They're not where it's at." But he literally, if you just followed this rubric of like listen to what your customers say, we wouldn't have the iPhone. He said, "Yeah, basically keyboards are [ __ ] We We need like a full screen." And he did it, and it was so freaking fascinating. So, it's like But is this aspect of innovating and thinking differently and just going with your gut and having this like intangible taste for style. And I said that to say like if we if we're building up this mega business harness, some of it I think could be a little scientific, but sometimes a business is like I just feel that this is wrong. Or do you know what I mean? It's just kind of like intuition lever that is tough to get into the models. Yeah, it's funny cuz everybody's like the customer's always right, customer's always right, customer's always right. But um at least in maker school like one of the main failure modes is when people just listen to their customers too much. And so basically once a week on the frequent community call I got to come jump in and say like, "Hey guys, stop listening to your customers. They have no idea what they want." Uh this is especially prevalent on jobs platforms like freelancer.com or Fiverr or Upwork or whatever. Somebody will like just pump in their demand to chat GPT. Chat GPT will generate the shittiest scope possible. And then they'll be like, "Do it this way." And then you're like, "This sucks. Like I can't do it that way." Or like with cold email. Um you know, a lot of people that have hired me in the past for cold email are like, "Okay, Nick, we want to like target this market with this offer. We want to say it in this way." And I'm like, "Okay, um why are you hiring me?" And they're like, "What do you mean?" And I'm like, "Well, you already have everything you need in order to do the campaign. Have you tried doing the campaign?" And they're like, "Yeah, but every time we tried doing it it doesn't work." And I'm like, "Oh, right. So you're hiring me cuz I have the expert knowledge and skill set. So let me do my job. I'll do my job, deliver you those results. And we're going to do it completely differently than you think. So it's it's one of those same sort of situations I think. Like yeah, you can only listen to your customers so well. You'll only ever see incremental improvement, sometimes in the wrong direction. Steve Jobs is like, "Nah. You know, I'm uh Napoleon Bonaparte, the Napoleon Bonaparte of phones, and I'll do it my way." And then, you know, he did and he crushed. He did crush. And that was so fascinating as well, isn't it? I think it's such a good point that you make, man. You've got to know it's like you have to this is like J- it's like the whole Jeff Bezos thing, right? Which is like um I think even, you know, people made this point that like if you hit home run, you don't just 10x, 100x, you can 1,000x. So, you've got to have those bits. I think like business is a game of risk, and sometimes it's worth taking calculated risks. Like if anyone's seen the show Dragon's Den or Shark Tank, sometimes you'll have people that appear on it. And I've just got the Sonnet 5 thing here in the background actually as well. Sometimes you'll appear on it like, "I spent I remortgaged my house. I spent $500,000 on a banana peel machine that is going to change the way we bananas and I it's going to be the best like the there's some aspect of like the delusion. So, it's like you you need people around it's it's it's a balance and it's like I think it's tough to quantify what that is, but you're looking for but yeah but it's okay but at the same time you need to take big bets. Speaking of those big bets, I'm going to pull it right now. Sonnet 5 test we can have a look a little look about what it's doing and what it looks like, but I like I'm kind of really enjoying this idea of actually having a bit of um a harness. I think that'd be a really fascinating thing to do. Yeah, me too. Like you think about it, that's the the layer at which you and I can probably meaningfully contribute. I'm not going to be able to build my own model at this point, but I'll be able to build the harness by which the model works. Yeah, cuz I don't think we need to build the model actually. I think Right. It's already smart enough. It's already there. They already got like a lot going into that, but I I think it's like the post-processing stuff. So, this is a comparison. So, Sonnet 5 on the left, 4.8 Opus on the right. Let's have a look at that. It's a fun spot to begin and have a good look. Yeah, it's cool that they're um you know, they're like judging it based off of the design more so than they're judging it based off of like the benches cuz the benches only show you a small story, I would say. It's more about like the vibe, the feeling. I would agree. Yeah, I think design's probably the best way I mean that I always think like websites are a really good way to test what they're like and how they're going. Yeah. But I mean this This cool, too. Look, it's like literally making animations. I think the Opus ones are better. They just cost way more. Oh, did the sun Did you see the price of the sun of the Wait a second. Sun, it was more for that one. Look at that. Yeah, that's weird. Okay, this is not a one-to-one. 1,000 tokens. This is 8,000. And I do think I mean it's tough to see, right? I mean that kind of does look okay-ish. But I mean it's like worse or comparable performance for three to four x the cost. I think this thing even though it's cheaper, it just takes 10 times more effort to get round there. That's kind of the problem. Interesting. I wonder though, like with one-to-one comparisons like this cuz these are probably presumably one-shot comparisons. I do wonder if like some of these one-shots are just going to be pretty lucky. Like I think probably the way to do it is you do 10 shots and then you compare the 10 shots median average. Do you know what I mean? Like 10 terminals printing hello world, 10 car animations, or whatever. Otherwise, you'll probably select like a really good one from one model and a really bad one from another model. The delta will look really massive. Um yeah. Dude, that's actually a really good point. I think this is this kind of corroborates a lot of the feedback that I've seen and I've used it myself. Like this one for example, if you look at this this table here, which is intelligence versus cost. cool. That's cool. Cost per intelligence index task. I like this. I like this. And then you know, this is kind of log-marith- uh I don't know how to say log-marithic. Well, like is this like a log scale? Sure. So it's like two [laughter] cents to three cents here, but then over here it's like literal dollar jumps. So I think actually this would be somewhere in like Nantucket if you just followed off. But this is cool. This way MiniMax is good. Yeah, MiniMax is a kind of broke the line actually. MiniMax is a really cool one. GLM 5.2 Max. Look at that. That's really cool in terms of like how smart it is for its task. This this is a good ratio. A lot of people just want what the hell is the best one, right? This one right here. it's kind of like not the TLDR of is not better performance for more cost. I think it's kind of essentially what's on it on the fibers. My take on this though is like you should just select the models there that are like 70% or so and like those are the ones you should give to the Fable orchestrator. Like you know GLM 5.2 was there. I think Kimmy K 2.7 code was there. I think you know Opus 4.7 and Opus 4.8 were there. Like those are the models that you should be delegating the tasks to because realistically 70% of Fable on audit is probably enough for most of the economic work that you you and I are doing. So like you don't actually need to go Fable for everything if you didn't want to and if you wanted arbitrage cost. And that's particularly relevant because one of the main constraints with the new Fable re-release is you'll only be able to use your plan for up to 50% usage and then afterwards it's going to roll over to usage pricing like like API pricing essentially. Um so I'm fine with that. You know you're fine with that. But it's like the average person that's going to be using these models going to be fine with that? Probably not cuz I mean Fable's getting really expensive these days. I could easily see a token budget in like the the thousands for somebody that does not make a thousand. I do that's so true. Actually I was watching an interesting video discussing topic like why is it the case that whilst GLM 5.2 so epic people have not adopted it at scale. And there's a couple layers to this but like two layers I think that worth discussing is like one is that on an enterprise corporate basis they're very um behind. And I think there's two or three big brands that they trust and like OpenAI, Claude maybe Gemini just mixed in based on who they are and they follow that. But actually I think it's like the benefits you get from cheaper tokens only really manifest when you hit a certain work threshold. Like when you're spending five, six, seven, ten thousand dollars a month on tokens it's like actually yeah a 20-50% reduction would be meaningful to me. And then it's like let's opt out. Oh you're a a month on it. Maybe then you want to bring it down. But if you're average Joe Pickle, you know, he doesn't need to jump onto Deep Seek to ask about his accounts once a week. You should just Yeah, that's true. The best model. Well, it's kind of like um when you work with a business and the business is I don't know under 100,000 bucks a month, under about a million bucks a year, the most effective usage of your time is probably going to be to increase revenue. Cuz if you can increase revenue by even like 10%, well, then you can make them an extra $10,000 a month, right? Um a mistake that a lot of people do when they come into small businesses is instead they try and decrease expenses. But it's like what's the expenditure of like a $100,000 a month company? Like, you know, maybe it's 50k. So, you'd actually have to decrease their total expenses by 20% to get the same benefit that you could get just increasing their revenue by 10%, which is kind of kind of stupid. Um but when when you start working with the real big enterprises, like, you know, $10 million a month, $100 million a month, like billions, um well, then like your actual usual first step is is costs optimization. And that's because if you can boost their margins up by like even 1%, you can usually make a ton of money, right? But it's actually harder to like distribute more. So, I think about it kind of in that way. Like if you're a really big company, these cost optimizations probably make sense. Just like any other cost optimization. Um but if you're not, you know, it's probably not that big of a deal. Un- unless you're just like a total newbie freelancer beginner with like 20 bucks in the bank account, then like yeah, get your zed.ai subscription and and fire it Um you know what I mean? It was like on a personal finance level that advice everyone gave it. Dude, you need to um stop buying those lattes. And if you were to not buy one latte a day, you invest that £5 compounding over 30 years, you got a million It's not a million dollars, but like the whole idea is like when you're 60, you could have Yeah, it was it was just kind of like I was like, dude, you're not going to latte your way to a million dollars, okay? You're not like drink the latte like have the latte like you're not going to like save like you know dancing at the margins like it's what we we say 5 pounds a day then what the hell does that do, right? So it's like it's the same kind of thing I think. The what one interesting thing I think it's worth looking into are the strings coming along with this Fable release actually because it's not as I think that first of all they've changed the model which a lot of people aren't actually physically aware of. They've kind of tampered off on some of these cybersecurity concerns but one thing that we may be seeing coming very soon is digital ID for language models. Right. Yeah, that's going to happen for sure. I mean I I think it was always going to happen but this is accelerated timelines with like the US administration just jumping in and being like no a Fable to non-Americans. Obviously Anthropic played hardball with them and they were like well if we can't release Fable to non-Americans we're just not going to have Fable at all. And I think that's really like what what happened here. Uh so kudos to them. I'm I'm thankful to them for that considering it's Canada today. Let's go baby. Get Fable in the hands of us. Um you know Canadians but yeah like we're all going to have to do ID verification and I bet you like people in the developing world probably not going to get anywhere near as much preferential access. That'll be a big like wealth divider as well. Mhm. Well dude one of the tailwinds the fact that I think it was a Chinese model did a distillation attack on Anthropic, right? Didn't they do that? M5.2? That was the 100% distillation. Same thing with Gemini. Like 2.7 code. Dude, what's that Scooby-Doo mean where they pull the hat off and it's like behind JLo? Who's this? It's like oh it's 4.8 again. Um Apparently dude, apparently this is interesting. Anthropic updated its privacy policy on effective from July the 8th covering free pro max plan saying they may ask for identity or age verification via persona in certain unspecified circumstances. In other words, whenever we want. Wow, okay. That's really interesting. So, you may just need to say who you are. That's I mean, is that fine? Yeah, I guess that's okay, right? I mean, ideally you wouldn't need to, but they already know who you are anyway, right? They've got your email. So, the only problem is if they were to ban you, it's like you can't re-verify. Yeah, like the you know, I feel up until now you've been you'd had so many loopholes. Like it's like oh like I don't want to pay $40 a month for whatever [ __ ] you know, TV subscription. I'll just like go on the Pirate Bay and do it. Um or like oh, you know, I'll just like get use my friend's account. Like we're dealing with we're going to deal with super intelligence here. You're not going to be able to like have those workarounds. Like if the government says you can't use this model, it's going to be very, very, very difficult for you to use this model. There aren't going to be like easily accessible like black markets you can just jump onto. Uh shit's going to be really, really hard. Well, that's the thing that scares me. Yeah, pushes people. Mhm. But like, you know, people are apprehensive about Chinese AI. They're apprehensive about where's my data going? But it's like it's like we talked about before like well, where's your data going right now? Why is it more noble to go to an American company than the CCP? Presumably cuz they've like, you know, we're we're you I and most of the listeners here are more aligned with that like political sphere. And you know, like I I agree with that political sphere in general and I would rather if there were were like two possible outcomes that like the American hegemony win, for sure. But um yeah, I think it's naive to think that they're not also going to be just like using, abusing, and ripping your data. It's like a different form of control, right? It's maybe not as in your face, but it's control nonetheless. Well, well, this is the alignment problem, isn't it? Like them releasing increasingly powerful models that serve the greater good defined by what we think the greatest good is with certain controls by the US government. And it's like I don't know how you can keep the genie in the bottle realistically because I I wonder if it will be viewed as a gun. So like if you think about like we we license firearms in the states, right? Like you need to get an ID, you need to do a sanity check because you have a weapon that can cause harm. You can injure, murder, whatever it is somebody with this weapon. I don't see why you wouldn't see AI as the same thing. Like you can engineer a bio weapon potentially. I mean it's a lot larger scale, but you can already you know, you can walk into Walmart. They Dude, Walmart's insanely huge. I didn't realize that. I walked in one of them 2 years ago. I was like, "Holy cow." Dude, I had like bread in one aisle, guns in the other aisle. I was like, "What the hell is going on over here?" I was like, "You know, you can already Walmart to White your way to some dangerous materials. So I I can I can envisage a world that the models are so powerful. We need to know who used what. But then it's like the Yeah, this is why again I think local is the way. And I think I think it's great that there's a counter pressure from Asia that they need a counter balance. We can't have a world where the intelligence is all clustered and centered around the United States of America. We We need some kind of counter pressure from somewhere else in the world to keep them honest and drive not only to drive innovation, but it provides more optionality. If the US government decides, "No, no, we're not happy with that. We're taking it away today." Yeah, yeah. I I agree with you um mostly on the market competitiveness aspect where if there was not another person slowly but surely creeping up behind you in terms of algorithmic progress, you know, I don't think American AI would progress anywhere near as quickly. Um I'm reminded of a situation I had with my own phone plan a few days ago where I like wanted to switch my phone plans. And the thing that um you got to understand in Canada is it's basically a monopoly. There's like two or three major phone companies and uh you're [ __ ] either way. You're damned if If pick one, you're damned if you pick the other. It's just they're so bad. So I like was like, "Okay, what's the fastest and easiest way to sign up?" So I went to like one of them and I did my little sign up, did it in like 30 seconds, and then it's like, "Error, we can't verify or whatever, so call a rep." I'm like, "Okay." So I pick up the phone, I call a rep, it's 25 minutes. Keep in mind I did my thing in 1.5 minutes just to get to the rep. Then the rep's like, "Okay, like you want to sign up? Okay, let's just redo it all." That takes 40 minutes. The whole thing's an hour. Halfway through, well, not halfway through, at the end because she [ __ ] up. I was like, "Okay, like just cancel all this. I'm going to go with the other guy." Hung up. Did the exact same thing with the other leader. No service, you know, it was just horrific. So the reason I'm bringing this up is because it's like when you don't have multiple competitors in a space, nobody is intrinsically incentivized to be good. And so they hire the worst support reps, they have the worst training, they have the worst like resource allocation. The companies just suck. They're just terrible companies, and I would like short all of them tomorrow. And I am, as we speak, actively considering doing that. And like they're just obviously going down the tubes. But um Same thing with AI companies, right? Like you do need to have multiple people, and China is innovating really hard in algorithmic progress because they're the lagging leader. They can just distill the main guys and then perform algorithmic upgrades that allow them to achieve similar levels of quality with significantly less hardware. If you didn't have that, then Anthropic and these other AI companies, like you know, OpenAI and stuff like that, they would be nowhere near as incentivized to actually deliver a good quality service. And thank [ __ ] for that. I I think it's that lightning rod of anxiety that hits Sam Altman, that hits Amadeo. I've got Amadeus. The CEO Dario Amadeus. Dario Amadeus. And you know, it's like, "Holy crap, we're behind. We've got a We have to like get this model out. Like we have to get out." So I think it's a good driver for performance, 100%. Um it's going to be exciting. I can't wait to get my hands on Fable. Um I'm going to be fabling my way to I'm going to be fabling all week, all night long, all day long. And uh we'll we'll see what the bill comes out as next week. We'll see how high we can drive this one with only a couple of days. Jack, I wanted to congratulate you on winning these school games yet again. I feel like you've won the school games every school games. So, whatever you're doing, clearly working, brother. Um are you excited for that? Yeah, well I'm going to be uh I'm not going to be in LA for this one. I'm going to be in Montenegro. Whoa. Whoa. Whoa. Whoa. You're not going to the You've won so many of them that you don't even care anymore. You're like uh I've got a I've got a hologram of me that's going to be literally beamed into one of the chairs. And I'm going to be He's got He's got digital coffee, dude. He's going to be drinking digital coffee. He's having a great time. Luke Skywalker over there. Well, well, dude, what was funny, Matt, is like so I I went from Dubai to LA. So, I think it was roughly an 8-hour flight, but then you have to go from Dubai to London. You chill out in London. Then from London to LA. And it was cool. I mean, but guys, it's a 12-hour time difference. And like literally I I I literally looked at my watch and I didn't need to change it. Like it was perfectly 12 hours difference. Something like that. I didn't I didn't have to change my watch. And I walked off and it was like 2:00 a.m. my time, 2:00 p.m. LA. And I found that bro, for like a 7-day trip, the 18-hour journey was kind of like it was good. I had a really wonderful time. And I really enjoyed it. But I think like probably twice in in 3 months. I'll probably go again end of the uh maybe next year. But I think twice in 3 months is um a lot of traveling. Especially since I just came to uh Hungary. But like this is the last school games though. Like there may not be another school games. So, that's it. I mean, you you say next time. Like when will the next time happen? Dude, I don't know. Yeah, cuz I think they're stopping the school games now, right? They're being replaced by the whole thing. They're just doing the They're doing something else. They've created a platinum thing, which is um for those who don't know, it's like to help creators go from like zero to 100k. Which is a big company. You probably seen Hamza talk about it and a few other people talk about it online. That's like the new thing. Um, but don't they get stuck in school with that bro or something? I don't know. That's tough it's tough to say. I don't fully you know. I got to be honest. I haven't really been looking at it as much as I probably should be but No, I I I'm a bit behind on it actually. I'm looking at it right now. But Jack, um, so you're just not going to go to this like the number one school games winner of all time. You're just not going to go. That's just how it is, huh? Man, he's just he's too busy. All I'm in the middle Well, you know what? I really like that. I really like that for you because, um, I've made a lot of decisions surrounding travel. Uh, cuz I've traveled a lot in the last year as well. Like a tremendous amount as you know. And I've had to make a lot of hard calls regarding travel in the last like three or four in the next three or four months as well. Initially, I feel bad about it but then I'm like, "Dude, every time I go travel it's not just the time that I'm traveling that like kind of sets me back. It's like the time that I'm traveling times two. You know, it's like the month that I go to Europe doesn't just set me back a month. It sets me back like two months. It's like the the week that I go down to XYZ location doesn't just set me back one week. It sets me two weeks. When you're doing this for like transcontinental travel I I don't know if I'm just like old and fragile and weak or something. But like you get really [ __ ] up, man. Not not having like your sleep or whatever is one thing. Also, your lifestyle is completely different. For people like you and I, we have like setups that obviously enable us to be very vastly more productive with agents and whatnot. Um, going out of that like perfect setup really hurts. We're media creators so we need like pretty intense media stuff, right? Like we need microphones, we need like lights and stuff like that. It'll like massively decrease the quality of our content. And um, you know, just when you're somewhere else like LA do you want to just sit in your room and work all day? Probably not, right? Like you're sure you're going to go out. You're going to do cool things, fun things. And then you're going to ride out your life stat at the expense of your career stat. And I think you can actually do all of them simultaneously and that's the ideal. You just can't do it when you're [ __ ] jet-setting around and stuff like that. So, it warms my heart to know I'm just not a loser and it's you, too. No, dude, you're so right with that. It's like you've got to pick your moves on the board, like when you're doing it. Like for example, I arrived to Hungary and those who know me very well know that Jack likes brunch. Brunch is I love brunch, dude. I had my eggs, my coffee, it was beautiful. And the cool thing about Hungary is that Dex has come with me to brunch every day. Like I had my brunch, Dex gets his brunch. It was 3 hours from landing to brunch, which is great. But like even this week, dude, I was like, "Dude, I don't think I've been that productive." And I'm like, "Well, someone's like, "Well, you did just move your world." And I was like, "Yeah, that's true." But it's also like you've got to find a gym, you've got to find like where you're getting food from and the Labrousse completely different. It takes three times longer here than it does in Dubai. And it's like there's like I think what I've figured out is like there's a travel tax, number one, on even if it's a beautiful travel. Like there's a time zone difference, there's a time cost, but it's also like there's a time cost in creating a routine. Like um getting into property, getting your Wi-Fi set up, buying the right property, trying to get a new I mean I've got it's taking a week, dude, for Amazon stuff to come here. I know some of you said like, "Oh my gosh, that's crazy, Jack." But just like there there are there is a transition cost, for sure, man. And you've got to build that new routine. And I think if you're constantly moving, that there's just a transitionary tax to that. So, I I I I definitely definitely get that. One of my one of my buddies who traveled a lot said that he thought that 3 months realistically is the minimum that you can do consistently. I I actually prefer six. I think if I was going to do, which I am planning to do, like half and half, I think six is a good amount of time. And ideally, it's in the same city. Like sixes and six months in place A, six months in place B. Yes. Yes. Yes, I think that would probably be my my perfect as somebody up in um Canada where where the temperature tends to swing wildly, it'd be like, okay, place in in the winter and then a place in the summer. A lot of people do that. They're called snowbirds, a colloquialism up here, cuz they like go down to like, you know, the South Florida or whatever in the winter. Um but I'm I'm now realizing why rich people tend to have multiple properties, but it's like it's not it's not like a tremendous number that they bounce around. It's usually like one or two. They have like one main thing and like another main thing, cuz they just they have their setups at both of them. Like everything's really convenient for them. Everything's awesome. Sure, of course there's a transitionary period of like 12 hours when you make it from place A to place B, but yeah, it's like you know, renting from place to place to place to place. It it seems cool and obviously like you're going to want to do it just to get that out of your system. But no, man. I mean, I've been doing it for long enough now that I just know it's I'm allergic to productivity when I'm doing that. You know what I mean? You're ho-posure maxing. You're like, how many how many different bedrooms can I sleep in in the next 9 days? Pretty much. But it's like it it's it's a little bit of that chaos is beautiful, but too much of that chaos it starts to derail like your focus a little bit. So it's like but it's like you need that in good balance. Like equally I think a life where you're doing the same thing every week in the same locations equally is bad. That's actually very dangerous, I think. Uh and this is an unpopular opinion. I I tell this story all the time. I was living in Leeds. I went to my barber shop and he had this balloon with a five on it. I said, "That's really cool. What What's that balloon about?" He's like, "Our jacket's our fifth birthday." I said, "Your fifth birthday?" I said, "Well, I came here when you first opened." He's like, "Yeah." I was like so I He said, "Dude, Jack, you've been coming here every week for 5 years." I was like, "I haven't." He said, "No, no, you really have." I was like, "Holy shit." And I was like, "I I have to leave Leeds." I was like, "I I just have to move." That was the moment? Was that actually the moment or was that a lie there? I had I planned on doing it, but I think when when your routine's similar it it you just like even cuz we had like a 5-year mortgage and it was like oh your mortgage is up. I was like really? That's been like that's got 5 years on that. I was like no no no it's up right now and I was like [ __ ] it. I should honestly do this. I was like I think there's just something about living the same day week on repeat that your your brain's like well there's no new there's no new novel information. There's no new novel routes, novel experiences. So there's no point on saving this information because it's just I'll just save my brain power is what it thinks. So you need some level of disruption in routine I think. I found for me personally every 5 years a bit of a sort of shake-up is a is a good thing. But otherwise I just think you just end up in Groundhog Day to be honest with you. Yeah. Yeah, that's really fascinating way of putting it. I read a post by Steve Jobs forever ago um one of his writings and he said something along the lines of you can never plan to meet the most important people in your life. And I thought about that in the context of routines and I was like yeah, you know what? All of the most important people in my life I never planned to meet. Like I it wasn't you know daily consistent YouTube videos that like you know meant me those people. It was usually like events. It was usually like responding to that one weird DM. It was usually like you know getting on a call with some rando. Like all of the most important people in my life have been through serendipity. So you need a sufficient amount of serendipity to inspire magic in your life, but you know you can't obviously serendipity max either. I hate getting older cuz I'm just like moderation really is the key isn't it? [ __ ] Thanos was correct this whole time. I told you this man. He's been right this whole time. But it's like you have to increase your service area to luck and do those additional things. But it's like Yeah, I've thought a lot about this. I don't know if I talked about it last week, but I kind of had a good think about um you know we talked about metrics and like optimizing different metrics. I was like you know AI [clears throat] business, health, relationships, all this sort of stuff. And I played out several scenarios kind of like Doctor Strange Infinity War. I was like it's really it was really good to forecast that out and think about where they all go. But you're right, bro. I I think like the the TLDR of it was you need good balance. So, what I've landed with it, I think is 6 months in one location, 6 6 months in another with little strategic tasty excursions here and there. So, yeah. A tasty excursion will be going to this Montenegro thing next month. Cool. But I think going to LA I kind of looked at that and I was like that's a lot of travel in 3 months. And I've done it we were like traveling once a month. Like 2 years ago, but I was like probably like right now I'm going to set this plan down, especially since Fable 5's come out, dude. Especially since Fable 5 I could do I said I said cancel the plans, dude. Fable 5 just dropped, dude. Cancel the entire calendar. Tell the mom that 4:00 is canceled. 4:00? 4:30? Absolutely, dude. He's gone. That's funny, man. All right, cool, dude. So, you're enjoying Hungary, then? It's a fun place? Dude, yeah. I don't know who's been to Budapest for my It's Dude, it's really freaking It's actually my favorite European city right now. Damn, that's high praise coming from you. You've been to a few times. it above the UK. I'm putting it above Germany. I know I'm like ice flinging you know, sentence by sentence our audience here. But I I'm telling you like Hungary is Hungary's really good, dude. Yeah, it's really good. Huh. Okay. Did you check it out in your travels over here, man? No, I never have. I'd like to go. I have a couple of really good Hungarian friends. I seem to get along really well with like Hungarian people and Romanian people. Uh which is so odd because like if I look through like my top 10 list of friends, like a lot of them are Hungarian or Romanian. You're somewhere there, Jack, too. Don't worry. But uh no, I've I wanted to go. I'm sure there's some sort of cultural influence there. I don't know the architecture does something to your brain that just makes you cooler and I like you more. It's It's the cool factor. Dude, actually they give you they put you in a WhatsApp group when you come off the plane. It's really cool that it's really freaking cool. Dude, one thing we have to talk about is [snorts] you said Claude science just came out, man. Do you want to do you want to touch on that a little bit? Yeah, I mean it was really interesting and we're seeing this with open AI as well as they're trying to like capture the scientific market, you know, about a year ago people were like can AI even make original discoveries? Can it even like move science forward? And I think most of that talk is now quieted down because it's just definitively crushed various Ardo's problems and mathematical conjectures that have been standing for for quite a while. Um but you know, AI is is heavily accelerating science right now. Like if you guys have ever wondered why we have all these like crazy new inventions around like health and longevity, peptides and so on and so forth. It's it's because people are throwing a tremendous amount of AI compute aka AI inference to like solve problems, to like create proteins, to take the current state of affairs, look at every other paper that's ever been written about something and then be like hmm okay, is there a possible mechanism of action there and then just like slowly answer it forward. And so what Claude science is is it's a platform that is basically meant to all in one box accelerate the process of research for people that work in you know, like research laboratories, for people that work in academia and so on and so forth. They basically want to minimize the friction involved in implementing AI into a workflow by basically creating pre-created skills that do things like you know, scan, connect databases, iterate on your current research question, improve your hypotheses and so on and so forth. And yeah, I mean it's looking pretty promising. They had built-in visualization artifacts, built-in skills for like biochemistry, built-in skills for like protein modeling. It's not a far cry to say that like the world with Claude science is probably like measurable percentage points closer to like the curing of various cancers than the world without Claude science. So very excited to see how that continues and although I don't think it'll get as much limelight as like Claude Fable will. Um it'll probably have a pretty noticeable impact on your life. Dude, I I agree. I can't wait for this to I've been saying I think that they're going to solve aging within the next 20 years. Well, 100%. But, it's also like solving things like cancers. It's like the pharmaceutical like the lobbying of the pharmaceutical industry is going to be unbelievable. Ooh, yeah. Yeah. And the drug industry. Cuz like imagine like I did the the gravy train's over. Like we we fixed it. We solved it. You know what I mean? Yeah, we're done with it. We we solved it and I thank you for your help. Actually, put it in your water. So, now you're you live forever by default. Yeah. I did I'm so excited for that. I'm so excited for that. I I also wonder if like the it will ramp up exponentially. And I'd imagine it would do, right? I'd imagine like 80% would get solved right at that last year. Just as we have the intelligence ramp up like they would start with a trickle. Like, "Hey, we solved this thing. Alzheimer's, out of here." Um you know, Parkinson's, boom, it's done. And then you know, we hear like one one story a year, then it's two, then it's four, but hopefully you know, like guys, we're technically Superman now. That's cool. I actually had this interesting question about one of the amazing things that Elon's doing cuz I love is helping people, you know, who can't move to you know, control cursors with their eyes. And the idea is that they could be able to walk again, which I think is the coolest thing ever. And actually, I think at some point we will get this cyber punk world, bro. There's a question I'm going to ask you. Let's say, okay, Elon and the team and what we get to the point where we have a technology where like if say someone like doesn't have sight in one eye, for example. And so, what we can do is give you a bionic eye. And this eye is actually like it's just like IMAX, right, bro? Like it can freaking like look through walls except for lead. It can like look really deep. It doesn't fire lasers. Oh, maybe it does fire laser beams. I don't know. Maybe that's like an upgrade you can get. I like if it was like a freaking like a Superman eye, would you get the upgrade? Would you be like, "I I want that. I want to like I want to get like a bionic upgrade. How Would you be in the market for that? Of course. I mean, I would probably bore my own eyes out to get those cyber eyes, Jack. I would be perfectly fine. And then I would be like, well, got the rusty spoon, got the cyborg eye, let's do it, you know? Or at least one of them, right? No, man. Like, think it do. Dude, we're I mean, technology already enhances us in so many ways. We're already really cyborgs. The fact that I could have a thought and then execute a couple of electrical impulses, twiddle my fingers, and then that thought can be transmitted over a long trans-Pacific distance to you, and then you can read that thought, do the same [ __ ] and send it back. We're [ __ ] telepaths. Now, why do we even have that loop to begin with, dude? Just hook that [ __ ] straight up in my cortex, and let's go. Look at this inefficiency, Nick. Look at this. Look at how much you've wasted. I mean, this should just be instant. 100%, dude. And we could just talk all day, Jack, which I know you want to do. We just like, boom. Like, imagine like it's like this is like an You got actually scientists and he's like, what's the rate of like You're limited by an exchange factor, right? So, typing is like slow. Speaking is four times faster. But if we can communicate mentally, like how much faster could this entire conversation be? Mhm. Bandwidth of thought. I mean, like you are constrained by various um biological structures, like Broca's area and whatnot, isn't really producing like the muscular contractions for speech much faster than we are executing on the speech. And then the neural signals that turn our concepts and translate them into various muscular movements, like those, you know, cortically those aren't really occurring much faster. I think speech is probably pretty close. Like, if we're at 200 words per minute, for instance, uh with speech, speaking relatively quickly, probably like 250 to 300, like, you know, neuronically, I don't [ __ ] know. But like, it's probably something like that. yeah. You're not going to get to 800. But I I start sharing just like images rather than like words. That'll be pretty cool. Oh, dude, yeah. Imagine that. Cuz I get this feedback a lot that I speak very quickly. Yeah. I need I I actually had it written down on a sheet of paper the word slow on it. It's like slow it down. But it But I I find that I I'm sure you find too, bro. It's very easy to just yap. Not yap, but like just communicate. Cuz there's a lot to talk about. Actually happens a lot when people blend words together. It's like um your mind's like racing a bit faster than you're actually talking physically. But I I think the ability to broadcast messages, I'm down for. And we're not we're not too far away from that either. I think it's going to be awesome. Dude, shall we Shall we look at some of our comments? Check out some of the stack the stack guys. Let's do it, man. Do you want to um screen share? Should I? Uh I don't have it up. Let me see if I can pull it up. I do I do have a recommendation for TV show, by the way, as well with pulling this up. There's a dude, there's a show come out our 21 subscribers will love this. It's called Devil's Plan. It's on Netflix. It's a Korean TV show. And they basically, bro, lock 12 people in a a room. Kind of like a hotel. And they have these like chips as currency. And like one person's got to make it to the end and they win this big prize for. But it's kind of like you play games every day. And if you win the games, you get prizes. If you lose, you can get eliminated. There's like a prison system. Like if you're the bottom two and stuff like that. And they have to like ally, manipulate, backstab, strategize, form alliances, break those alliances. And it's And they've got different games. Some are solo, some are team-based. It's freaking rude dude, you would love it. It's really really fascinating to watch. Fascinating. I mean, like your last recommendation to me was Obsession. And um that was nutty. Jack just sent me a text and he's like, "Bro, you got to watch Obsession. Go in totally blind. Don't look up the trailer." And I was just like So then I did, and it was like Dude, I dude, on that point, how far in the movie were you when you were like what the what the actual [ __ ] is on right now? When I was like, "No, no, no, no, no. We're having such a good time." Dude, the bit the bit This won't be a spoiler for everybody, but the bit when like he You're not quite sure if it's worked or not, this wish. And she's kind of just like at the car door, and you're like, "Mm, this is a bit a bit weird." And then you're like, "Mm, yeah." You're kind of going through this And then there was a whole cat thing. Remember the cat thing? Oh, we [ __ ] I don't want to talk about the cat thing. Jesus. Yeah, I remember the cat thing, man. Wow. That was so Dude, that Okay, you want to know our why? Dude, that was a million-dollar budget. I think that's done over 200 mil now, actually. Yeah, that is What's that? A 200X easy? You hear the um director was just like So somebody complained about how much money they made on the movie because it was like one of those like more art ones and didn't do super well. Uh they weren't expecting it to do super well, but then it like obviously made a ridiculous amount of money, and I think they got paid like six grand for it. And um that was circulating for quite a while. Yeah, yeah, yeah. They were like, "I can't believe I only got paid this much." But then people were like, "Well, like listen, you don't really know when movies actually take off and when they don't. You You really got lucky with this one. You could segue this into a big career move. But um now that you bitched about your pay, nobody will ever hire you." And it was like, "Oh, man." That was actually You know, whoever it was actually did a pretty good job. But you did kind of [ __ ] about the money, so Oh, but that's a tough one, right? 373 million on a budget of 750K. 373 now? Oh, dude, that's like a 500X. All right, you see it You see a screen? I'm going to zoom it a little bit. I do indeed. Let's have a look. Cool says it's community member 18. I've been following both of you for quite some time and never saw anything regarding this podcast. Highly suggest you both promote it on the front end of both of your individual channels' videos. I think for whatever reason the algo truly isn't pushing this very much. Not a bad idea. Why don't we do this more? Well, we're trying to keep it the the competitive advantage as tight as possible, guys. We don't want any more than 19 people watching us, So, we're trying to really suppress it as much as we can. It It's It's a good thing. We We actually tried that, didn't we, bro? We cross-posted it, but I'm not sure what it Maybe we need to try that again. Test that again. Some different ideas. Yeah, it was it was so bad when we cross-posted it. It was like it actually underperformed our regular podcast. And I was just like, "Jesus fuck." I'm not crazy. I wonder I wonder why. I wonder why that happened. I don't know why that happened. Fable. We We said bad things about Fable in Tropic Thunder. think when When we land on a highly topical thing, we can we can make that bad. Yeah. I'm sure that's going to be beneficial. Yeah. Emily Jean 304 says, "I'd love to see you guys discussing try some of the European models like Mistral since there's such a rising concern of model monopoly." Damn, what about those European models like Mistral? Um They're just so far I love I agree with the sentiment. They're just so far behind. That's the only problem. They're so far behind. I would love to discuss it. I would like I would like free players on the My My discussion of that is related to it It just It sucks so bad. It's such a bad model. They're so behind it. I mean, are they even in the race? But we've seen with Grok and stuff that people can pull through and like make large large jumps, but I don't know, man. Mistral just does not seem to me to be competitive in the very least. Yeah, unfortunately, that's the case. Derek's going to angry at us. Everyone wants to say they're better than Fable. Don't buy the fake hype cuz it's not going to expect it coming back because of your lies. You're a liar, Jack. None of your fake talk will bring Fable back until they bring it back. You're a liar. Liar. Decimation. Bugger says, "One of the biggest factors of a creator's virality is controversy. Taking a position on a debate is subject and defending it. It's risky cuz you find opposition, but also loyal supporters. Understandably, you may want to restrain from doing that if you want to keep your public image untainted to sell some product, but some creators do and even make that their main attraction. They get down in the mud and give a big old spectacle. Sometimes this work great for them, but it's also capable of going bad. It's a gamble. My two cents. Thank you. I agree with you on all fronts. 100 freaking percent. But I I I found a good example of that. I can tell you there's a guy called Alex Fen, and he did it on Open Claw. Hm. He He He pushed all his chips on the table on it. And you know, Nick, you did the opposite. You went You were kind of like I we we were just joking offline actually guys that um Nick is probably cost Open Claw set multi seven figures 100%. Does that matter? Is this Is this the guy Is this the Open Claw guy or something? go to his most popular video, bro, uh you will see Oh, 1.1 mil. Yeah, that was 5 months ago, dude. That feels like a year ago, right? Isn't that crazy? Who's using Open Claw? Oh, [ __ ] I kind of wanted to talk about Open Claw today, too. Isn't that funny? Check this out. crazy. Imagine getting acquired by Open AI, getting unlimited AI tokens, and still dropping this slop abomination. Oh, what happened? On the new Open Claw dropped an app, right? And people are just [ __ ] on how terrible the design is. And like Can you believe they dropped this? It's pretty It's pretty funny. They're trying to get it in the Yeah, that's um I need to have a look at that. It's pretty funny. Does it do anything new? Is it just like No. It's sad watching this drop have instant 50K downloads taking into account this is created with unlimited tokens, but you can see how little effort was put in. I mean, like I agree. Look at these pictures, man. These are trash. Dude, you can't vibe code these apps. You can't. Like I know it's crazy. Unless you've got like You need a really high-level designer to build this stuff out, like genuinely. I don't know if it has a hand. I haven't checked it out, but like there was a company that I trust, and I looked at their app, and I was like, this is 100 You I can I can smell a vibe code app, Nick, from 6 I I just I know if it's been built I've just seen it too much. I've seen too many of them. I just know it. And it's like it's this instant ooh moment, right? We're like, oh, okay. Has this been set up properly type of vibe going on? Yeah, I wonder if that'll always remain possible. Like, I wonder if our taste will increase or our discriminatory taste will increase the same rate that AI technology increases. It's it's I do have to put a jump on this real quick cuz it's really interesting. Cuz Rory Sutherland is making this point. The you know there's this like thing of like politician speak and we must do better. Again, at one point they all do this cuz pointing is rude and they and they they've all optimized in a micro to be the best they can at politicians. They don't answer questions. And they're making the point that actually if only one person did that, they would be the most effective politician. But the problem is that everybody in the class follows the exact same blueprint system. They'll walk a certain way, they dress a certain way, they talk in a certain cadence, and it's become formula like a like formulaic. So, you can identify it. And therefore, it is no longer the most effective way of doing it because it's become a formula. And I think it's the same thing with clothes designs. Like, if you weren't seeing it every day and one person designed it, you'd be like, oh, this is really cool. But people are too You have haters, of course. They just call everything AI slop cuz they think it's cool. But it's you identify because you see it so much. But I think if you hadn't been in that volume, you'd probably say, actually, it's pretty good. Yeah. Yeah, probably. No, I mean, like it's it's all based off the environment, right? Like an environment of products, that's what that's what makes a better product. It's like some company decides to zig when others zag, like Steve Jobs and the iPhone. Or the open claw. Yep. This is funny. You mentioned on the last pod that Hormozi would release some entrepreneurs competing for a 100K video, and then he did like 8 hours later. So, that's really funny. You you gospel Jack. People need to listen to you more often. can we we see in the future in this podcast. We got a pretty funny one from Vlad. He says, I spend half the amount of my brain could be credited to embed with Nick says compared to Jack. Dude, it's like prompt A. By the way, you know what prompt B and like half prompt C. Here's the prompt like D. Like, who's paying for these tokens, Jack? I love these comparison ones. We should get a lot more comparisons about know I actually don't know what this one I I don't follow what this one means, bro. Honestly. Looks like you're spending your brain compute back. You're saying that it takes uh harder to it's like longer to understand what you're saying explanation-wise or something. Your explanations. Like you're saying he's spending half his brain compute credits to embed what Nick says. Compared to Jack. So he's spending double the tokens. I see what he's saying. Like he has to spend more brain brain compute to to follow along. Okay, that's interesting. Or maybe you're just you're speaking higher-level stuff. Maybe Nick said here Cocomelon and you're just like these actuarial statistic. Well, dude, I I'd someone say to me, "How many prompts do you How many tokens do you spend saying, 'Hey, dude?'" I was like, "Not nearly enough. Not nearly enough of them, dude. Not nearly enough." to be dude-maxing. Hit him John Virgin Islands. Okay. We'll discuss that in detail on the next podcast episode. Yeah. Okay, one one more. One more from Sakana Fugu and then we can Sakana Fugu seems to compare with Perplexity and Orchestration. My question is GLM 5.2 allowing more usage limits per 5-hour window than Claude Opus because the monthly cost to rent the model is similar. Also, why doesn't GLM 5.2 have a free subscription or free chat mode like how Anthropic and OpenAI have it deployed in GPT and Claude? It's difficult for me to compare both models cuz the 20 bucks subscription on Claude is not based on prompts that are allowed, but GLM 5.2 is. And they state around 80 prompts for the $16 monthly light plan per 5-hour window. Yeah, all of this is annoying. I agree. It's annoying um companies get to make up their own pricing schemes though, and you know, despite the fact that this sounds pretty sub-optimal, they do crush, you know, Anthropic models on API spend compute. So, maybe the numbers are different if you're just using plans, but for for me anyway, GLM 5.2 is much more expensive cuz I'm just I'm blowing through my usage now anyway. Very well said, man. Just goes to show the level of like mental gymnastics that need to be done in assessing how I get that 2% performance increase, right? Beautiful neck, that's everything. What do you want to finish on? Any comments to send us all off? Just wanted to thank everybody for their non-stop support of the Stacked Podcast. Tell all your friends to get your dog subscribed, your mom, your dad, the whole crew, really. Um, thank you all for watching. Couldn't have said it better. Boom, peace. Catch you next week, guys.