First, Sonnet changed engineering. Then, Opus outclassed Sonnet. And now, Fable 5 and Mythos 5 are outperforming Opus. By now, you've seen the headlines, you understand that this model is an absolute beast. And you also know Anthropic is rugpooling Fable 5 from our subscription plans, ProMax, Team, and Sebast. This is completely unprecedented. Forget the June 23rd subscription. Rugpool everyone was mad about. On June 12th, the government pulled the entire model. Both Fable 5 and Miffals 5 are no longer available. A federal export control order made Anthropic suspend the best two models that anyone has ever seen after the government apparently found a jailbreak. I'm not sure the federal government has the skills, resources, and agentic engineers to confidently find and address these brand new jailbreaking techniques, but apparently they have. The very controversial piece
[00:01:00] here is that Enthropic says these same jailbreaking techniques work on models like GBD 5.5, and it's not specific to Fable 5 or Mythos 5. So, it's very weird that these same tricks work on GBT 5.5, but Fable is the model that got rugpulled. It's not clear what's true yet. What is absolutely clear is that this is temporary and Fable will be re-released soon. So instead of focusing on all the mania, all the hype, all the chaos of this situation, let's refocus ourselves on three actionable ideas you can use for all next generation mythos level models. In this video, I have three observations on Claude Fable 5 specifically for agentic engineers that can help you get the most out of this model. I'll list them on the screen here and at the end I want to share exactly how this model is changing my approach to agentic engineering. Now running a model in isolation is meaningless. I used Cloud Fable 5 to orchestrate itself along its two little brothers Opus and Sonnet. They all ran the exact same five
[00:02:02] specs each in their own agent sandbox. They looped until the work was complete and created a single public URL where we can see their results built end to end. These are full stack applications orchestrated by Fable 5. That's 15 sandboxes in total. This is enough for us to understand what this model really gives you because it's not price per token. Each one of these URLs is live built by the respective agent. Here we have Fable recreating Simon Willis's LLM price index. We have a hacker news clone. We have a scikitlearn model generator. We have a chat room where we're chatting directly with the pi coding agent docs. We had an agent build an agent inside of a full stack application. And then we have this brand new type of application and we can directly chat to any one of the agents inside this chat room. But you'll notice something interesting here. Sonnet, Opus, and Fable were able to do the job. So what is Fable for? In this video, we break down the model that will do what
[00:03:00] Sonnet did. Change engineering forever. Let's talk Claude Fable 5. This is one of the first models that is truly feeling like a payto-play model. But then there's the other side of this coin which is this very very important statement. Fable 5 is not an Opus replacement. It's a tier above Opus. It's a mythos class model with mythos class pricing. So that's the other side of this coin. Now the question is is this actually true? Right away we can address this. In our total of 15 agent sandbox applications, we can see the price differential between Sonnet, Opus, and Fable. Sonnet ran $55 worth of tokens. Opus ran $91 worth of tokens and Fable ran 200 per token. Fable loses this battle. This model is not giving you an improved price per token. And the raw numbers don't look good. Throughout all these applications, Fable shipped this with a million tokens. Opus did it
[00:04:00] with 700K and Sonnet did it with also around 700K. Where is the value does more with fear tokens is not true. This model is not doing more with few tokens. So what is it doing exactly? So this is the first observation I really want to nail home with this model. It's not about price per token. It's not about intelligence per token. What we're really looking for is this price per intelligent agent hour. Time is our most scarce resource. This is what we're getting out of the new fable model. There's a sweet spot here. This leads us to another really important observation that we'll talk about in a moment. This model is only really useful if your missions, if your tasks, if your specs are big enough, are complex enough. We're getting into this really interesting place where we're not spending on tokens anymore. We're spending on agent time. We're spending on useful agent time. The best, easiest measure is in hours. But here's the breakdown. This is where Fable completely wins, right? And it's not really close. You can see here the time taken to generate every single one of our applications and there is a gap here
[00:05:01] right a sizable gap. Fable is completing the work with more tokens more expensively about 20% faster. So this is the key observation when you're using Fable 5. What you're getting here you're not buying tokens you're buying intelligent agent hours. Get out of task top to bottom mindset and get into feature shipped work completed end toend work done. the spec is driving Fable to those next level results where we can truly delegate, get out the loop and just judge the results and iterate from there. This model kind of pushes us further away from prompting back and forth, babysitting the agent, and repeating that cycle. It's more and more becoming about specs, proper delegation, proper looping, proper closed loop structures, making sure your agent can validate all of its work, and then understanding the review process. So, that's the headliner here, right? The harder the mission, the harder the task, the harder the feature you're trying to ship, the more Fable makes sense. Looking at our applications that were generated, the price just keeps going
[00:06:00] up. This model is just going to use more and more compute. You can see the cost roughly equivalent. It's just using double opus, double sonnet. And this is the big takeaway, right? On about 80% of these tasks, the sibling models, the brothers of Fable, Opus and Sonnet, did the job at a fraction of the price. put on our harder task here, specifically our multi- aent chat room application here. This is a significantly harder task. We have an entire micro IDE in here and our agents are operating on this inside of the UI full stack application. You can see all the personas in the bottom left. We're of course spinning this up using the pi coding agent under the hood on the server that fable oneshotted through this entire application. We can prompt this directly. I'll run on all give me your perspective on the application. And so all is going to kick off every single agent. You can see they're now starting to think here in the top left. We're chatting to everyone in this chat room and everyone is going to give me the work received. So they're showing their red files and they're reporting. So
[00:07:00] we're in this really really interesting application where we can just chat back and forth with all of our agents in a chat room. And so we just address all of them with at all. We can of course address our agents specifically. But you can see we got a nice perspective from every one of the agents in the chat room. And again, this is one of the harder tasks. We also have our PI documentation support agent. We have our market direction scikitlearn model predictor. We have hacker news clone much simpler and an LLM price index just presenting information again much simpler. You know what we're looking at here is all the fable version. So of course these are going to be the top-notch best versions but we also paid a massive amount for these results. And the big kicker here is that we didn't need to. This is one of the big ideas the kind of uncomfortable ideas we're going to get to in a moment. But one of the key observations, the larger the task, the harder the task, the more fable makes sense. And here's my kind of extreme version of this. If you can cure cancer with a million Fable tokens, $10 per million tokens is nothing, right? It's absolutely nothing because doing this work, curing cancer or whatever
[00:08:02] hard valuable work you're doing is is worth a lot more than $10 per million tokens. $10 per million tokens is nothing. At the same time, if you're centering a div, you're making a donation to Enthropic. Okay, let's just say it like it really is. The price premium here scales directly with the mission. It will create that chat app for you. It will do that, you know, minor front-end backend work for you. It'll create the migration file for you, but you'll be mostly wasting money. So, first takeaway, what we get out of Fable is price per intelligent hour. That's the new thing to compare up from price per token. If you just look at price per token, 2x opus, nobody likes that. That sucks. They're rugpooling Fable from the subscription plan. What we get from Fable is price per intelligent agent hour. Time is the resource that if you prompt contacts harness engineer properly with this model, this is what you get back. This is worth all the money in the world, right? It's time. All right. And so we can see this directly in our receipts and this small
[00:09:02] microcasm benchmark. The results show up pretty quickly. Just 15 items to compare here. 15 full stack applications that Fable orchestrated. Okay. And speaking of orchestration, that brings us to our second big observation for your agentic engineering. Cloud Fable 5 is not an intern. It is an orchestrator. I've been using this model since release and this is one of the standout pieces in terms of raw performance. Cloud Fable 5 is not an intern. It's not a worker. It's pushing toward the most valuable thing an agent can do and that is orchestrate. Every engineer's progression looks like this. Now it is you start with a base agent. You then make it better. You learn how to prompt and context engineer. You then add more agents. You then customize them. You specialize them to outperform any agent without that unique information. And then at the last level, you orchestrate every previous level. Last week we talked about Cloudflare's review
[00:10:02] software factory. They are using multi-agent orchestration. They understand the value of it. And of course, Enthropic understands the value of multi-agent orchestration. If we go to the system card and we just search for multi- aent. Guess what pops up? An entire section on multi- aent. What did they find? They found what everyone finds when they start scaling their work with agents. What do they do here? They have a bunch of benchmarks where they're specifically battle testing multi- aent orchestration with the best model and then with the best model scaled up to three, five, 10 async and then I think they have some versions where they're just running unlimited agents. Spin up as many as you want, right? Async sub agents. And as you can see here, accuracy on the left, latency per task on the right. What you get is exactly what you would expect. If you scale your compute, you scale your impact. aka if you use more agents and you have a great model that can steer them. If you have an orchestrator model like Cloud Mythos, like Fable 5, you get better results and you get them faster. Now, it's not
[00:11:01] always faster. Token usage across agents also adds a time cost. It's not always going to get the job done faster. Sometimes it's slower, it costs more, but you still get the accuracy, right? You still get the raw performance when I'm prioritizing what I'm looking for out of my models, out of my agents. It's always this, right? There's a trifecta for every single agent. You're always trading these three things off. This is the trade-off triangle. It is performance, speed, and cost. As a northstar, I'm always sacrificing speed and cost. But then very quickly, depending on if you're running a product agent or if you have a subscription where you can blast through tons and tons of tokens. That equation is going to change very quickly. There are many things you cannot deploy a fable level model or even an opus level model because the economics do not make sense. But back to multi-agent orchestration, the idea is simple and enthropic knows it. Again, everyone using models at scale that has pushed past the progression of agents base, better, more custom orchestrator. You know that if
[00:12:00] you want to push your results, you add agents and you add specialized agents and then you let the orchestrator do whatever it needs to do to get the job done. And so that's what we're seeing here. Enthropic seeing it in the benchmarks. This is not new for them. They found this pattern with Opus and they've been building their models to be better orchestrators. Another way to say that is that they're building their models to be better prompt engineers. So this is the second big takeaway for Fable 5. Okay, this model is not a worker, it's a leader. If you have this model center or make a stupid small change, you're wasting it. You're legitimately wasting it. This model is not a worker, it is a leader. And so, you know, in my head, this model is performing more and more like a principal engineer. For sure, if you're just vibe coding, if you're firing off random ad hoc prompts, you'll never see this capability out of this model. But if you are agentic engineering, if you're writing great prompts, if you're setting up your context, this thing can perform like a principal level engineer. What do principles do? They can do the job. Of course, they know how to do the job, but what they do best is they
[00:13:00] delegate to others. Okay, so Fable 5 is the ultimate orchestration model. If you're thinking about how to get maximum results out of this model, it's in delegation. It's in orchestration. It is in building multi- aent orchestration systems, multi- aent orchestration agent harnesses, so on and so forth. So this is the second really, really important observation. This is the pattern I used to create this right to benchmark this model. I had Fable 5 spin up multiple sandboxes. Five on itself. Five have Fable, five Opus, five Sonnet. They each got their own sandbox. They each ran the exact same five specs. And the interesting part here is in the results. Just to mention this, I think that every model release there is some fear that um everything can just be done with a prompt. I firmly do not believe this. I think engineers will continue to have a place in the world. In fact, we're going to need even more engineers that actually know what's going on. The way I think about this is the floor in the ceiling. So, models like this raise the floor, but they catapult the ceiling up as well. If you have been, you know,
[00:14:01] following channels like mine, if you've been doing the work using agents and trying to build systems that build systems, going to that metaentic engineering level, the ceiling is much higher for you. You can do a lot more with a model like Fable. That's a big observation here. Uh, Fable 5 is an orchestrator. Treat it like an orchestrator. Treat it like not even just a co-worker anymore. Like this is I'm really thinking about Fable as a principal engineer that knows how to given a great spec. I should preface with that. garbage in, garbage out for every system, right? That's just a foundational truth. But if you are writing great long specs with high detail with validation testing and review steps that make the loop very very clear, this model can perform like a principal engineer. And the best principal engineers delegate to scale far beyond themselves. Okay, it's the same thing. Fable 5 only has a million context window. And you know, a great example, just scrolling back up, I had a single Fable 5 instance kick off all 15
[00:15:00] agents. And you know, I don't need to like prove this in any capacity, but let's just pull down the session I was using here. And this is the entire session to spin up the agents, run it, execute it, and then I use it to also build the presentation that we're looking at right now. Check out my context window. 62% 600,000 tokens. If I was not delegating, this work would be impossible. Look at the token usage inputs. Well, more than a million, okay, total combined outputs 100,000 looks like 2 million total. Look at the costs, right? The estimated costs from running all these agents and they're on sandboxes. Just to re-emphasize the point, this is impossible without multi- aent orchestration. And many things are impossible without multi-agent orchestration. That's why we've been talking about it on the channel for months, probably over a year now. I have no idea. I've lost track of time. We've done so many of these videos all the way back to some of the original ideas before Cloud Code was released and then of course Cloud Code's original release of sub agents, right? Going way back there. You know, drop a like if you were with the channel for that long and drop a comment as well. Shout out to you. But um here we are proving it in the future.
[00:16:00] Multi-agent orchestration is how you get outsized results. This was true back when you were prompting just a few cloud 3.5 sonnet models and it was true with cloud 4.5 and it's even more true here with Cloud Fable 5. So multi-agent orchestration super powerful. That's the second big observation. If you want to push this model to the limits, treat it like a principal engineer. Give it a large plan. Give it serious work to work through like spinning up 15 full stack applications in their own sandbox and then let it rip. Let it absolutely rip. And we can open up the other examples here, right? We also have Opus. There's our LLM price index. This is a real full stack application. I I just want to like express that it's hard to work through all of this content of what these models are capable of now. but Fable and then we can pull up Opus, right? Opus 4.8 and let's get Sonnet. Let's get this model comparison. And so this application is letting us quickly compare. This is directly inspired by Simon Willis's uh llmmpprices.com. He's had this site up for a while. You know, I told the agent, look at the site, clone it. This is one of the simplest things these models can do,
[00:17:01] right? You can make a clone with any one of these models, any one of them. And they all did a good job on this specific benchmark task. You can see Fable, Opus, Sonnet. You can kind of compare all three. They all do a nice job. And so this is kind of fun for the channel. It's fun that we're talking to you using a application built with our agentic engineering workflow. And what I think is kind of cool about this is that if you're using multiple browsers and you open them up and you go to the URL, multiple people can use the application at the same time. So if you have a friend or a co-worker, open it up and say, Hey, you can talk to this agent in real time. That's kind of interesting. And so I just wanted to like just show some of the different types of applications that this model can build and build well. And I think the point here is these are real world type applications. It's not just hello world. It's not just centering a div. We have CI, we have CD. We have real servers running. We have real applications running. We have persistent data, data stores. These agents built something real. And so if you're a practitioner, if you're building agents,
[00:18:00] if you need any other piece of evidence that the stuff is really working and you can use it in production, this is it. These are real production-ready full-stack applications with persistent data stores and continuous deployment. This is very real. Real applications, every one of them. All right. And so let's get into third and final observation. And this is the uncomfortable part. This is going to go against a lot of what you have seen. Because if you're watching YouTube, if you're on X, you're going to see people talking about how Fable 5 is god-tier model. It's the best model that they've ever seen. Nobody has ever released anything like this. And Dan is going to come and say a very different thing here. And that is that Claude Fable 5 is the first state of the art model that you probably don't need. Think about that. Fable 5 is the first state-of-the-art model that most engineers don't need. Why? Because Sonnet outperforms Fable on price per token. Sonnet and Opus both do the job
[00:18:50] on many hard tasks. But Fable is a bit of a gamble if you're not writing big enough specs. The key thing here is this model requires a great spec. It requires a long spec. You need to be writing specs that have 100 to 200 lines in the spec or maybe a few hundred lines in the spec. You need a detailed spec that includes testing steps and that includes validation steps as well. So if you're not already doing this if you don't have the skill set to do this if you're not already in the habit of writing clear detailed long specs you're not going to unlock anything unique with Fable that you can't get with Opus. You're just going to spend more money. And so this is one of the things that's becoming very clear as we move toward this level. For me, Fable is unlocking some interesting things. But for most engineers, your default should be Sonnet and then move to Opus for harder tasks and for orchestration
[00:20:00] tasks and then use Fable when your spec is really demanding something truly great and you're on a subscription. If you're on a subscription, Fable is just better right? You're already paying for it, you know whatever. But for API users, Fable is a very specific tool. In the same way that Opus is a specific tool, Fable is even more specific. So that's the third observation. This is the first state of the art model that most engineers don't need. Why is this happening? I think what we have here is there's a non-linearity in how models improve. The improvements become incredibly expensive as you're pushing the frontier. And what's happening is very interesting from a market perspective. Anthropic is making it commercially viable to run Mythos class models. But what's interesting here is the sibling models, Sonnet and Opus, have gotten so good that Fable is not as necessary. It's more like a big deal for specific things. But overall, these sibling models are doing the
[00:21:00] work in these benchmarks that I ran personally. So, I just want to offer a different perspective there that I think is useful for you to know. Like, look, the model is incredible. Of course it is. But if you need it as an API user, it's very specific. If you're on a subscription, yeah, sure, blast away, have fun. But the defaults should be Sonnet for most things and Opus for harder tasks. And so that is an interesting framing. And here's, I love thinking about this stuff in terms of model stacks. What I mean by model stacks is that you're using multiple models in your approach to agentic engineering. There is a state-of-the-art model. There is a workhorse. And there's a lightweight model. These are the three tier model stacks. Anthropic has given us this, right? We have Fable for the orchestration, very demanding heavy lifting tasks. Opus for complex
[00:22:00] work, for complex work, for research, for really intense tasks. And then we have Sonnet for your everyday workhorses. And the other thing that's very interesting here is that the key constraints of agentic engineering are planning and reviewing. Not writing code. Not making the change. Not running the test. Planning and reviewing. And here's a very interesting thing about this model. One of the things that Fable does really well is it adheres to specs much better. In my benchmarks, my hard subjective empirical look at this, I gave all of these models the exact same specs. And Fable really hews to the spec in a way that Opus and Sonnet don't. It really pushes down on the spec. And what that means is if you have a really great spec, you're going to see Fable do the work exactly as described. And what does that mean? Planning on the front end is incredibly important. Planning on the front end is incredibly important. But here's the interesting thing. Reviewing at the end is much less. The agent can validate its own work and then you can do a quick review or you can even
[00:23:01] let the agent review its own work better. And so what Fable is doing is helping you shift the constraints to planning and away from reviewing. So that's a really interesting observation as well. And it tells you that the very best use of Fable is if you have a really great spec, if you have a long spec that is detailed with lots of testing and review built into it and you want it just done. You want the work done and you want it done really right. That's when Fable makes sense. And so that just re-emphasizes the point there. But I love this model stack concept. To me this is becoming more and more how I think about agentic engineering. Not just what model am I using? But what models am I using? How are they working together? What is the orchestrator? What is the workhorse? What is the lightweight? What are they each doing? And as they go to market and market forces take over, the market will decide who those models are and the prices will continue to fall. And so what you're going to see is that the model stack will be at the same performance, but it will get cheaper and cheaper over time. And that's actually happening at a rate that no one
[00:24:00] would have predicted. So these model stacks are incredibly powerful. And I hope this is useful for you to think about it in this way. And the model stacks tell us one very clear thing. If you are specifically trying to use all three or even thinking about model stacks, you are operating at a very interesting new position as an engineer. And I think it's an incredible time to be an agentic engineer because you can think about all these things. When you're thinking about the trifecta and you're thinking about performance, speed, and cost and which model to use and when to use it, you're an engineer that understands the playing field in a different way than most. And that's a powerful place to be. So think about your model stacks. Fable orchestrates. Opus does complex work. Sonnet does the workload. That is essentially the kind of the ideal model stack at the moment if you're using cloud as your primary model provider. And I want to close this out. I want to share two things. These are the things that are actually interesting to me from a practical standpoint on
[00:25:00] how I'm using Fable. I've been using this model since the day it dropped. You know, I want to close these thoughts by just being really practical here. And this kind of wraps up the three big observations. You know, the third one is the uncomfortable truth. You probably don't need Fable for most things. But if you really do, it is an incredible tool. And I do think Fable is going to be used as a standard orchestrator in many of the best agentic engineering setups. It will be the standard orchestrator. It will just have a time when Sonnet just replaces it and then the sibling model comes in. That's going to be the story of every new model release as we move forward. And so here's kind of really practical how I'm using Fable. Number one, I use Fable for very large very complex multi-agent orchestration workflows. I use it specifically when I have a big mission. When I wrote that big detailed spec and I need Fable to spin up a whole bunch of sandboxes. that's when I reach for Fable. That's when it makes sense. And this is the setup I was talking about where I can kick off 15 sandboxes which is a really impressive feat. I wasn't really, I had faith in the model, but I was surprised by some of the results,
[00:26:00] specifically just the complexity of what was delivered. You know, the multi-agent chatroom application is one of the first applications of its kind. It's very novel. It's very interesting. You can spin it up yourself. It's very very fun. And I'm going to do an entire dedicated video on just this application on just this template. I think this is a template for the future. And I want to share how I'm using it, how I'm building it, how you can build it with or without Fable. And you can build this with Opus. Just to be clear, you can build this with Opus. And I'm going to show you exactly how to do that. But, I use Fable for very large complex multi-agent orchestration workflows. That's what it's for. It's not for anything small. It's not for anything simple. Commit it. Just commit it to memory. Fable for big complex multi-agent orchestration. And two, I use Fable for the most demanding spec I have. The most demanding spec I've ever written. The most demanding spec I have. And you know, I mentioned this at the beginning. I'm sharing this with my learners. It's going to be on the sub-stack. If you're not subscribed to the substack, go subscribe. And I've been sharing my
[00:27:00] approach to prompting Claude Code as we go from prompt to ship. And so, you know, that's where I use Fable. Big complex multi-agent orchestration workflows and the most demanding spec I have. And I will share how this is playing out because it's changing my approach to agentic engineering. And so here's the update. So going forward, for every software project I start. There is a single prompt that I'm going to use. And what this is is I call this ZTE. It's a north star prompt, okay? And it is the following. It starts with planning. It goes through building. It tests. It reviews. It documents. It deploys to production. And what this means is I have a prompt. I actually just have a prompt for this. And it's essentially a master spec. And every single one of those phases I just listed is detailed in the spec. And it loops back. And at the end of every loop, I decide what to do next. It has a review step. And if review doesn't pass, it loops back to fix. But if review passes, the work is done. And so this is my attempt at Zero Touch Engineering. I want to ship to production from a single prompt.
[00:28:00] And I know this is ambitious. I know this is not trivial. And I know this is something that is going to have to develop over time. But the goal, and this is the 2026 goal, the channel goal, if you will, is ship to production from a single prompt. And I'm not, I don't think I'm crazy. I think this is the north star. We're pushing as far as we can in this direction. I'm pretty sure I'm not the only one with this goal. I'm sure many of you have this goal as well. And this is where Fable and future Mythos models come in for me. Every time this bar increases, every time we get a better orchestrator, we get one step closer to ZTE, Zero Touch Engineering. And that is my challenge for you today. Where are you on this spectrum? Are you at level one? Are you still babysitting your agents? Are you at level two? Are you getting most work done without having to babysit? Are you at level three? Are you able to delegate tasks end to end? Are you at level four, which is you can delegate features end to end? And that's where I want to put you, right? That's where I want to put you. I want to put you in this position where you can delegate features end to end. That is the goal. That's where we're going with the channel. We are going to
[00:29:00] try to push to level five, which is ship to production from a single prompt and I'll be the test case for that. And I know I have great company because I know some of you watching this channel and some of you that are working through our materials are at level four. And we are going to try to get you to level five with us. And that is the north star. This is where all of this is going. This is why we care about Fable. This is why we care about orchestration and sub-agents and all of these ideas. Everything is in service of ZTE and of shipping from a single prompt all the way through to production. And so if that sounds like your jam, I would love to have you on the journey. Drop a comment below. Let me know where you are on the scale and let me know what you're building. I love hearing from you and it really means a lot that you're here and watching and listening and working on this stuff. And I know it's not easy to do. It requires a lot of practice. It requires a lot of experience, but it is the greatest skill set in the world right now to have.
[00:30:00] So I hope this is useful. I've said what I've come to say. Let me leave you with this quote. It's actually an Anthropic quote that I use every day. It's something I reference regularly, which is if it's achievable, it should probably be achieved by AI or AI assisted tools. This is my filter for everything in my life from engineering to just like, is this worth doing? Can I get this done with an agentic workflow? Can I get this done with agents? And if so, I do it that way. And I think that's really the power of this, and that's the principle of the channel. Alright, that's what I got for you today. If you haven't already, make sure you subscribe. Oh wait, one more thing. I know I know I know it's been a long time coming but my Claude Code mastery course is now at 50% off. To be exact, it's currently 5.97 and it's going to go to 9.97 in a few days. So, you know, just very straightforward. If you're interested in the content, there's a link in the description, you can use the code CC50 at checkout and you can get 50 percent off the mastery course. So that should be a pretty good deal.
[00:31:00] I put a lot of good stuff in that course. I'm continuing to update it. We have a community as well in the course platform. And so I'd love to have you there, you know, we're all working on this stuff together and I would love to have you as part of that community. Alright, that's it for today. See you in the next one.