How To Practice Writing Fiction (And Write Stories With AI)

Okay, quick video here. Today I have this paid newsletter called Commercial Fiction Club, which is where I've been writing all about my fiction endeavors. And I'm documenting what it's like to build a new writing vertical, specifically here fiction to seven figures in revenue from zero. So if you're at all interested in how I built these verticals, but also how to do this in the world of fiction writing, I encourage you to check out Commercial Fiction Club. Anyways, a topic that I've been thinking a lot about recently is how to practice. And last week I wrote this newsletter all about how to isolate certain skills and get better at writing by not thinking of writing as this big umbrella term. You know, oftentimes people say, I want to become a better writer. That is a very difficult thing to do because the word or the words becoming a better writer are a bundled term. You know, if you say, I want to master the art of business writing or fiction writing, these are umbrella terms. And it's not until you unbundle the umbrella term that you realize that in order to become successful at any specific type of writing or even writing in general, it's actually a smaller subset of skills, usually a long list of sub skills that you need to get really good at. And so when you get good at lots of individual sub skills, eventually they all ladder up to the outcome of you being a proficient writer or you being proficient in a certain type of writing. Okay, so I've been thinking about this idea a lot lately, and one of the skills that I've chosen to isolate and really work on as it relates to fiction writing is this idea of writing story seeds. Now if you're not familiar, a story seed is basically the elevator pitch for the story before you go and write it. A faulty belief that I use used to have as a writer is thinking, you know, in order for me to understand what my story is all about, I have to go write 300 pages of it. Or in order for someone to really get the point of my story, I gotta go write 300 pages of it. And what I've learned over the past decade, really pursuing the career of being a full time successful writer, is that this is not true. You can get the, the majority of the value of the thing that you're trying to communicate or write in a very small container. And if you can't summarize the thing that you're trying to say in a small container, then increasing the size of the container doesn't actually improve the quality of it. If anything, it actually makes it worse it makes it harder to understand what you're trying to say. So I have become a really big believer. And you know, this translated into a lot of the nonfiction and even business related writing that I've done over the years where I've become a really big believer in. If you don't know what you're trying to say in the headline, you're not going to know what you're trying to say over 300 pages of a book. And so in fiction, if you don't know what you're trying to say in the story seed in the little elevator pitch, then I promise you, writing 300 pages isn't going to solve the problem. If anything, it's just going to exacerbate the problem because you don't know what you're trying to say in the small container. How are you going to say it in the big container? So if this is a skill that I want to improve, or if this is a skill you want to improve, then the first step is isolating the skill that you want to work. On side note, I'm reading this really interesting book right now called Peak. I think the subtitle is the New Science of Expertise or something. Basically, the gist of the book that I've gotten so far, which is something I've always intuitively understood, I just had never really read a book and heard it articulated this way to me before, is that practice in and of itself is not leads to skill acquisition and certainly not mastery. If that were true, then anyone doing anything over a prolonged period of time would achieve mastery at something. And anyone doing anything, the more that they do it further in their careers, they would get better and better and better and better and better just automatically. And we actually know that's not true. You know, there are lawyers or doctors that peak in skill, and then as they continue to get older, their skill declines, right? Or there are people who are really intuitively good at something, but then, you know, they get good really quickly, they might have a natural ability for it, and then they plateau and they never actually become world class at the thing. So the point of the book is, so why does that happen? And that happens because eventually, you know, irrespective of biology and natural talent and things like that, eventually we all reach plateaus in whatever skill we're trying to develop. And the only way to push beyond those plateaus is not just brute force hours. It's actually what the author refers to as deliberate practice, meaning you have to isolate specific sub skills that you want to improve at. And Then you have to come up with ways of deploying volume and increasing repetition of just that sub skill. So metaphorically, something I've been thinking about a lot is chess. You know, chess is actually very similar to writing in the sense that you don't sit down and go, I want to become a world class writer. I want to become a grandmaster chess player. That is the outcome of lots of sub skills getting mastered. And so when someone is on the path to wanting to get really good at chess, for example, they aren't just trying to get good at playing entire games at once, right? The same way that a writer might go, oh, I want to get better at writing novels, so I'm just going to focus on writing full novels. Inevitably, what will happen is when you do that, you will reach these plateaus and you won't know how to improve. So how do chess players train? They train by isolating different parts of the game. So you've probably heard the term in chess, there's something called openers, right? What is an opener? An opener is the first handful of moves in the game. There are chess players, there are entire chess books that are dedicated to just chess openers, just the first few moves of the game. That is what it looks like to isolate a specific skill. So when I think about that in terms of writing something, side note, something I've always found that's really helpful for me is thinking metaphorically. So not just thinking about the thing that I'm trying to work on, but thinking about how do people work on things that are similar or tangentially related in other industries? I've always found thinking metaphorically to be really effective. And so with writing, I've been really thinking, okay, so if we use chess as the metaphor, what is the equivalent of an opener in fiction writing? You know, in chess, a lot of books are written and a lot of times players will study what's called like the mid game, you know, or the end game. These are different sections of a conventional chess game. You know, you have your openers, there's probably other words I don't know, but like, generally it's like you have your opener and then you have mid game and then you have end game, right? Just to simplify. So if you want to get better at chess, you don't just keep playing entire games. What you do is you isolate these different parts of the game and then you practice just that part of the game over and over and over again. So a skill that I really want to improve at with fiction writing is writing seeds. Now why is that an important skill? Well, because it's really hard to improve at the skill of how do I tell a cohesive, overarching story. If in order for me to practice that skill, I have to go write 300 pages, that means my feedback loop is 300 pages, if not more. Right? My feedback loop might be a year, might be two years for me to go write that story and then go, all right, so how did all the pieces click together? Right. It's significantly easier to practice. All right, so how do I think in terms of story arcs if all I have to do is just think in terms of the seed of the story? Like, what is the elevator pitch? How would I articulate this to someone else? So let me walk you through the little exercise that I came up with, and then I'm going to show you why doing this work manually is so valuable and so important as it relates to AI because when you take the time to go through and do this work manually, that is what allows you to automate it with AI. If you don't take the time to learn how this works manually yourself, then no amount of AI compute power is going to help you do it. Because the thing that people misunderstand is that the bottleneck to using AI is not your knowledge of AI. The bottleneck is your ability to articulate the thing that you're trying to automate. And everybody wants to automate things that they don't know how to articulate because they don't actually know what they're asking for. Okay, so instead of thinking, how do I get better at writing fiction? How do I become a successful fiction writer? Which is like, too big of an aspiration, right? That's an umbrella term. We have to unbundle it. When we unbundle it, we might isolate a handful of different skills. You know, you have the skill of writing the. The elevator pitch, the story seed. You also build the skill of taking that story seed and flushing it out into an outline. You also have the skill of world building. How do you make this an immersive world? You have character development. How do you isolate individual characters, like character development? Talk about unbundling. You don't just think, how do I make all my characters great? It's actually, well, what are the different archetypes of characters? And then how could I create frameworks for each individual archetype of character so that when I sit down to write and I know I'm creating that archetype of character, I'm cycling through the smaller subset of frameworks in My head. Pacing is another skill. Narrative voice is another skill. Word choice, metaphors, tone, language, that's another skill. Right, so it's not I want to get good at writing fiction. It's that you get good at writing fiction when you are good at all of these different sub skills. Okay, so we're not going to try to improve all of these at once. What we're going to do is we're just going to pick one and then we're going to create an exercise to practice just that one metaphorically. Another sport I've been thinking about a lot is golf. If you wanted to learn golf, you don't just go play nine holes over and over and over again, right? Sure, you go play nine holes or you go play 18 holes. But also you have all these individual clubs. They all do different things and they all are used at different parts of the course. There's a reason why when people are working on their golf game, they go, I'm going to go and just work on putting today. That's the same thing as a chess player going, I'm just going to work on openers today. Which is the same thing as a fiction writer going, I'm just going to work on practicing writing story seeds today. Okay, so this is a really important. I mean, honestly, I've always understood this intuitively, but this really hasn't consciously clicked for me until very recently. So that's also why I wanted to make a video about it and write a newsletter about it, because it's helping me crystallize this for myself. But this is such an important idea to understand when it comes to improvement and acquisition of a skill. So what do we need to do to isolate a skill and get better at it? We need to come up with a simple framework or a simple template that we can use. Now, why is this so helpful? Well, because when you sit down to practice, if you don't have a mechanism or a routine for practicing that sub skill, then you're actually using unnecessary mental bandwidth. Because now instead of focusing on the thing, you also have to use mental bandwidth to think about how you're going to focus on the thing. Okay, so I'll give you a simple example. Let's say you want to get better at putting with golf and you work with a golf instructor. The golf instructor is like, here's the exercise I want you to do every time you go to the putting green to practice. I want you to start, you know, five feet away. I want you to make 10 putts, then move 10ft away. I want you to Make 10 putts. Then 15ft away, I want you to make 10 putTs. Then 20ft away, I want you to Make 10 putTs. And if at any point you miss, you start over again. All right? People do this with shooting free throws. People do this with shooting threes. You know, you do this when you're learning a musical instrument. It's like, I want you to play these 10 scales in a row. None of this should seem novel. What I think is novel is, is that I went to school and got a degree in fiction writing, and nobody explained this to me. The way that everyone has articulated writing, to the best of my knowledge, the way that I've been exposed to it is everyone talks about writing it like it's this big, amorphous, intuitive thing. And imagine if someone was like, I want to get good at playing golf. And you're like, well, you just got to buy a bunch of clubs and show up to the. To the golf course and just rely on your intuition. That's a horrible strategy. And I don't understand why. That's sort of the way we talk about skill acquisition in the world of writing. So that's why I'm fixated on this idea. So what do we have to do? We have to come up with a framework to practice this sub skill. So I am a big believer with writing that you always want to try to minimize the number of variables that you're working on at once. Another framework I come back to often is Picasso's minimalist bull. If you're not familiar, Picasso would do this writing or this drawing exercise where instead of trying to draw something that was really complicated, he would reverse engineer objects that he wanted to draw and try and figure out what are the fewest points on my sketchpad? What are the fewest things that I need to connect together to communicate the thing that I'm trying to draw? Do I need to draw in detail, like, every single hair and expression and nook and cranny for someone to know that this is what a bull looks like? No. If you keep stripping things away, eventually what you end up with is like a sideways hexagon. And then it's like, yeah, I sort of intuit that that's a bull. That was his minimalist bull exercise. And I think about that all the time. So with writing, I'm always asking the same thing, which is, well, instead of trying to figure out, how do I say this in 30,000 words, how do I say it in 10,000? And instead of 10,000, how do I say it in 1,000, and instead of 1,000, how do I say IT IN 100? And instead of 100, how do I say IT IN 10? And so, before we even start thinking about how do I write a fully fleshed out story seed, I actually want to start with one sentence. Everything in writing is always the expanded version of some smaller version of it. So I want to come up with one single sentence. So then I ask myself, okay, so if I was to reduce a story down to a single sentence, what are the fundamental pieces and the pieces that I came up with? And this is also from me studying all sorts of other fiction writers and people who write about fiction writing. I won't claim all of this, but certainly the ways that I'm putting the pieces together are maybe original. So I came up with, you have five pieces. You have the character's desire in the beginning, some sort of weakness that they have, danger or obstacle, the action that they have to take and how they change as a result. And so if you string this together into a single sentence, what you end up with is when character experiences weakness, they run into danger and must action to change. Now, do you need to write 300 pages in order to communicate that about your story? No. If you don't know how to say your story in one sentence, you're not going to know how to say it in 300 pages. Okay, so first we start with the single sentence seed. Then what we want to do is we want to build some pattern recognition for ourselves. So something that I always do when I'm coming up with writing exercises is, is I think about works that I love, and then I use those as examples because chances are I've seen it a million times. I've thought about it for a prolonged period of time. You know, like, it's a very easy example to pull from relative to you trying to come up with something original on the spot. So to build pattern recognition and so that I can drill in this, the five different pieces here and this single sentence seed, I go, all right, so what would the one sentence seed for Star wars be? You know, the original Star Wars Episode 4, I've probably seen a thousand times, you know, so I've thought about it a lot. So I fill out each one of these things. Character, desire. Luke Skywalker wants to be a pilot, fight with the Rebellion. He's unknowingly a Jedi. His weakness is he's naive, he's unfocused, lacks confidence, etc. Danger. The Galactic Empire is trying to take over. Action. He needs to use his skills as a fighter and his budding skills as a Jedi. Using the Force in order to help the cause and change. He builds self esteem and basically becomes like part of the rebellion. Right, and so when we reduce it down to a single sentence, what we end up with? What we end up with is when a princess falls into mortal danger, a young man uses his skills as a fighter to save her and defeat the evil forces of a galactic empire. Great. Okay, so once we have this single sentence that's amazing, then we can start to expand it. Okay, Again, this is the whole key, is you want to reduce things down to Picasso's minimalist bull. What are the fewest number of pieces that you need to communicate the idea? And then writing is just the expansion of that seedling of an idea. Right? So then we expand this single sentence into a more fleshed out story seed. And this story seed should really give us all the pieces of the story that we want to use in order to have clarity ourselves over. All right, so what's the whole arc here? Who's the character? Who are we following? What do they care about? What obstacles do they run into? What's the point of this story? And you need a little bit more space than a single sentence, but you don't need 300 pages. You could probably communicate all of this in two or three paragraphs. So after studying a bunch of stories, I came up with an expanded template that I put here. Okay, now what we've done is we've isolated the skill, and we have a one sentence template prompt, and we have an example to reinforce pattern recognition, and we have an expanded template. So notice, if I want to improve the skill of writing story seeds, then if I didn't have this every time I would sit down to practice this skill, I would also have to think about, all right, and so how do I go about practicing the skill again, which wastes mental bandwidth. Instead, now I have the exercise that I created. I have some pattern recognition, and I just go, all right, so now I just have to rep out story seeds. Okay, and the whole key here, this is a very important point when it comes to combining this with AI as much as possible. The rules and constraints that you set for yourself when creating exercises are you want to follow them as closely as possible in the practice that you do. So you don't want to deviate or you want to try not to deviate. And the reason is because bridging the gap here to AI the reason is because what you want is you want perfect matching. You Want to be able to tell AI later? Here's the exercise, here's the framework, here's how I'm thinking about it, and here are my reps following those instructions to a T. And the reason that's so valuable is because. And this is gonna get a little meta here. The reason this is so valuable is because when you create exercises to improve yourself, what you are really doing is you are creating instruction for yourself, which is no different than the instruction you create for AI. It's so much easier to instruct AI on what you want its help with when you've already instructed yourself. The instructions that you create for yourself end up becoming the prompt for AI. And second, the way that you execute the instructions that you've created for yourself, your examples, your reps, become the training data for AI. So the exercise becomes the prompt and the reps become the training data. And this is such an unbelievably powerful idea because when you take the time to do this work yourself on the front end, you basically go, I'm going to invest 10 or 20 hours into doing this myself so that I understand it. I'm going to crystallize my knowledge here. I'm getting even more leverage out of it because I'm crystallizing my learnings and my takeaways by writing it out in a newsletter, which is benefit to me because it crystallizes it for me, but also is a benefit to you. This is a paid newsletter. This also becomes a product, right? So think about the leverage on the leverage of the thing that you're working on. And then on top of that, you do all of that for yourself and then you can automate it forever using AI. The reason that, like, I'm so excited about this and why I wanted to make a video on it is because I just want to show you high level how this works. So basically in this newsletter, what did I do? I created a mechanism to improve myself. I took the time to articulate it, which reinforces my learning. And also I get leverage out of it because I turn it into a product, AKA a newsletter for you. Then I practice the skill and I go, okay, great, I have my exercise, I have my rules. Now I'm going to go practice it myself, I'm going to do it manually and I create a bunch of story seeds myself, following my instructions to a T. I get leverage out of that because I include it in the newsletter, I share it with you. That pattern recognition helps you improve, but also that pattern recognition becomes my training data for AI. So then When I'm done with all of this, and I go, I created the exercise, I created the instructions, I crystallized it. I got a bunch of reps in. Now I can take all of that and I can throw it into Claude, or I can throw it into ChatGPT or Gemini. And now I have all the training data and all the instruction to have help creating more of the thing that I just crystallized for myself. So here I wrote out this long form prompt. If you want the prompt, come subscribe to Commercial Fiction Club, because this is where I'm sharing all this stuff. But I just want to show you how this works. I take all the stuff that I crystallized, I assemble a prompt, I take all my reps, I add it to the prompt as pattern recognition and as training data, and then I run the prompt. And here, now I have everything that I need to come up with story seeds for new concepts very, very quickly, which means I get leverage out of these reps and out of the work that I put in forever. Now, if you play this out, what does this look like when you do this for each individual skill? When you do this for each individual skill, you understand how each skill works. You understand how they connect, and you also build prompts, AKA exercises, for each skill, and your reps create training data for each skill, which ultimately means any part of writing a novel. In this example, you can always turn around and go, hey, AI, I want your help doing this one thing. And I know exactly what I need to ask you for, because I've done it myself. And I also know what makes for a good output, because I gave you the training data on the thing that I was trying to create, because I created that for myself, and it allows you to move so much faster. Okay? So here, notice now I go, I put this prompt into Claude, and Claude's like, what's your idea? So let's make something up. I go, I have this idea for techno crime story about Girl who is an assassin in a futuristic society, and she is trying to rob the major bank of the city to steal enough money to save her mother who is dying from a horrible disease. Like, I should give it way more context, but I'm just gonna use a simple example, right? So I go, I want to write a story about this. Now, how much time does it take you to sit there and really flush out all the different options of this story? It'll probably take you a while, right? And there's nothing wrong with that. You should, I'm a firm believer in you should take Time and block time to do things manually so that you understand them and see what you come up with first, you know. But if I have just like the very beginnings of a story and I don't really know how I want to flush it out yet, and I want a brainstorming partner and the first thing I want to do is create a story seed, then why wouldn't I use this prompt and go to AI and go, here's just high level kind of what I'm thinking. Help me come up with a story seed. And then let's take it from there. So look what it does. It takes my exact framework. Those five pieces. Character, Desire, Weakness, Danger, Action, and change. It writes out a one sentence seed. When Kira, a lethal assassin who has long believed she exists only to destroy, discovers her estranged mother is dying from a disease only the ultra wealthy can cure, she must pull off an impossible heist against the very mega corporation that made her Cool, cool idea. And learn whether she's capable of saving a life instead of just taking them. That's not a bad story. Right? And then we expand it. How does it expand that one sentence seed? By using the framework that I gave and looking at the pattern recognition of the reps that I did first, because that is the training data. And here it writes out this whole expanded seed. That's awesome. And then AI also leaves you with some questions to consider, which is cool. So now like think. Think about what we've done. I spent probably eight to 10 hours thinking through this exercise, writing it out in the form of a newsletter and then writing a handful of story seeds following that exercise so that I could get reps in and build my own pattern recognition and training data. 8 to 10 hours of upfront work. I also learned a ton in the process. I built a bunch of skills, but 8 to 10 hours that now translate into a prompt and training data that I can leverage forever for this one individual sub skill. And if I do this over and over and over again for all the different sub skills, let's say Hypothetically, there are 10 major sub skills when it comes to fiction writing that a writer needs to master. Okay, let's use round numbers. 10 hours per skill. 10 hours. That means 100 hours of practice. Not a lifetime. 100 hours. That's not very much. 100 hours. And you understand all the sub skills. You have exercises for all of them. You have your own reps, AKA training data for all of them. And you have then prompts that you can use for all of them at any time in the future. That is insane leverage. Like, we can't even wrap our heads around how much more we can do as creatives today, with that in mind than we could 10 years ago, or 20 years ago, or 50 years ago, or 100 years ago. So anyways, I was really excited about this. I've been thinking a lot about how to practice. I've been thinking a lot about how to isolate skills as a fiction writer, and I've also been thinking a lot about this idea of the benefits of doing things on your own first, and viewing the work that you do not just as practice for the sake of practice, but the compounding benefit of when you practice and you take the time to crystallize it, that becomes your prompt, that becomes your training data for the thing that you're trying to automate at some point in the future.