WEBVTT 00:00:00.000 --> 00:00:01.600 You know, when you look at the creator economy 00:00:01.600 --> 00:00:04.900 right now, it really feels like the only way 00:00:04.900 --> 00:00:07.280 to win is to be the loudest person in the room. 00:00:07.379 --> 00:00:10.140 Oh, for sure. It's the attention economy. Everyone 00:00:10.140 --> 00:00:13.460 is chasing viral fame, trying to be the next 00:00:13.460 --> 00:00:15.880 big thing on TikTok, you know, filming themselves 00:00:15.880 --> 00:00:18.059 in public, just desperate for that attention. 00:00:18.339 --> 00:00:20.600 If you aren't visible, you don't exist. That's 00:00:20.600 --> 00:00:22.679 kind of the rule. That's the common wisdom. Yeah. 00:00:23.280 --> 00:00:25.460 But then you read a document like the one we're 00:00:25.460 --> 00:00:27.399 covering today and you realize there's this. 00:00:27.920 --> 00:00:30.679 This whole other group of people doing the exact 00:00:30.679 --> 00:00:32.979 opposite. The quiet ones. They're anonymous. 00:00:33.539 --> 00:00:36.560 They aren't winning literary awards. They're 00:00:36.560 --> 00:00:40.979 not doing book tours. But they are quietly, you 00:00:40.979 --> 00:00:43.820 know, sitting in the shadows, generating somewhere 00:00:43.820 --> 00:00:47.000 between $800, $1 ,100 a day. Which is just a 00:00:47.000 --> 00:00:49.579 staggering amount of income for a writer. It's 00:00:49.579 --> 00:00:51.700 absolutely wild. Right. And they're not doing 00:00:51.700 --> 00:00:53.939 it by writing the next, you know, great American 00:00:53.939 --> 00:00:56.380 novel. They're doing it with these extremely 00:00:56.380 --> 00:01:00.679 specific niche fiction books, things like Age 00:01:00.679 --> 00:01:03.840 Gap Werewolf Romance. And they're using a workflow 00:01:03.840 --> 00:01:06.500 that looks less like writing and, I mean, more 00:01:06.500 --> 00:01:08.459 like industrial manufacturing. It really is. 00:01:08.599 --> 00:01:11.099 It's the ultimate example of systems engineering 00:01:11.099 --> 00:01:14.120 applied to creativity. It's not art in the traditional 00:01:14.120 --> 00:01:17.730 sense. It's supply chain management. Welcome 00:01:17.730 --> 00:01:21.489 back to the Deep Dive. Today we are unpacking 00:01:21.489 --> 00:01:24.670 a fascinating guide called How to Write and Publish 00:01:24.670 --> 00:01:29.349 an AI Novel on Amazon. By Maxa. And I want to 00:01:29.349 --> 00:01:31.010 be clear right off the top, we're not here to 00:01:31.010 --> 00:01:33.230 debate the ethics of AI art today. That's a whole 00:01:33.230 --> 00:01:35.510 other deep dive. We're here to look at the system. 00:01:36.030 --> 00:01:38.109 Because what really struck me about this source 00:01:38.109 --> 00:01:40.849 is that it doesn't read like a creative writing 00:01:40.849 --> 00:01:44.010 class. It reads like a technical manual for fulfilling 00:01:44.010 --> 00:01:47.290 ultra -specific demand. That is the perfect way 00:01:47.290 --> 00:01:49.269 to frame it. Most people think AI novel means 00:01:49.269 --> 00:01:50.950 you just, you know, push a button and a book 00:01:50.950 --> 00:01:53.049 pops out. Right. But this guy argues that if 00:01:53.049 --> 00:01:55.510 you do that and just get garbage, the workflow 00:01:55.510 --> 00:01:57.370 we're going to look at is a seven step process. 00:01:57.430 --> 00:02:00.670 It involves market research with like data science 00:02:00.670 --> 00:02:03.569 tools, architectural plotting so the AI doesn't 00:02:03.569 --> 00:02:05.730 get confused and a very specific distribution 00:02:05.730 --> 00:02:08.090 strategy. So let's make this out from it. We're 00:02:08.090 --> 00:02:10.900 going to look at the economics first. Why? These 00:02:10.900 --> 00:02:13.599 tiny little micro niches are printing money while 00:02:13.599 --> 00:02:16.039 the big publishers just ignore them. Yeah. Then 00:02:16.039 --> 00:02:17.759 we'll get into what they call the surgical workflow. 00:02:18.400 --> 00:02:21.800 Using perplexity for research, Claude for drafting, 00:02:22.000 --> 00:02:24.680 and a tool with the truly hilarious name for 00:02:24.680 --> 00:02:27.979 the covers. Nano Banana Pro. I still can't quite 00:02:27.979 --> 00:02:29.879 believe it's a real tool, but we'll get there. 00:02:30.000 --> 00:02:32.020 And then finally, we'll talk about the context 00:02:32.020 --> 00:02:35.039 cage and how to stop the AI from hallucinating 00:02:35.039 --> 00:02:37.860 and, of course, the Amazon SEO game. Sounds like 00:02:37.860 --> 00:02:40.240 a plan. Okay, so let's start with the money. 00:02:40.520 --> 00:02:42.719 The source makes this distinction that I think 00:02:42.719 --> 00:02:45.300 really explains everything. It contrasts the 00:02:45.300 --> 00:02:47.159 traditional publishing model, you know, your 00:02:47.159 --> 00:02:49.979 random house, your penguin, with this new AI 00:02:49.979 --> 00:02:53.319 indie model. And it all just boils down to the 00:02:53.319 --> 00:02:55.560 size of the audience you need to survive. It's 00:02:55.560 --> 00:02:57.340 all about overhead. I mean, if you're a traditional 00:02:57.340 --> 00:02:59.879 publisher, you've got offices, editors, marketing 00:02:59.879 --> 00:03:02.400 teams, legal departments. Right. So for a book 00:03:02.400 --> 00:03:04.759 to even be worth your time, it needs to sell 00:03:04.759 --> 00:03:07.259 tens of thousands of copies. You need mass appeal. 00:03:07.360 --> 00:03:09.580 You need the next Harry Potter. But mass appeal 00:03:09.580 --> 00:03:12.419 is incredibly expensive to market because you're 00:03:12.419 --> 00:03:14.500 competing with Netflix, Fortnite, everything. 00:03:14.860 --> 00:03:17.020 And this model just flips that on its head. Completely. 00:03:17.060 --> 00:03:19.560 This model relies on what the source calls niche 00:03:19.560 --> 00:03:22.520 hunger. You don't need millions of readers. You 00:03:22.520 --> 00:03:25.919 need a very small group of absolutely obsessed 00:03:25.919 --> 00:03:28.500 readers. Obsessed. Yeah. The source lists these 00:03:28.500 --> 00:03:30.699 examples that, I mean, they sound like jokes 00:03:30.699 --> 00:03:34.360 to the average person. Mafia romance with neurodiverse 00:03:34.360 --> 00:03:38.819 protagonists or cozy mystery specifically for 00:03:38.819 --> 00:03:41.740 45 to 60 year olds. I genuinely laughed when 00:03:41.740 --> 00:03:45.620 I read. Age gap werewolf romance. But the argument 00:03:45.620 --> 00:03:48.740 is that these readers are like starved for content. 00:03:48.960 --> 00:03:51.680 They are. I mean, if you love that specific subgenre, 00:03:51.680 --> 00:03:53.520 you can't just walk into a Barnes and Noble and 00:03:53.520 --> 00:03:55.460 find a whole shelf for it. You might find one 00:03:55.460 --> 00:03:57.659 book a year if you're lucky. So when you go on 00:03:57.659 --> 00:04:00.199 Amazon and you find an author who produces exactly 00:04:00.199 --> 00:04:02.919 that flavor, you don't just buy one book. You 00:04:02.919 --> 00:04:05.419 buy their entire back catalog. You subscribe 00:04:05.419 --> 00:04:08.259 to their newsletter. The loyalty is just off 00:04:08.259 --> 00:04:10.680 the charts compared to general fiction. It's 00:04:10.680 --> 00:04:13.939 the long tail theory. Kind of turbocharged. And 00:04:13.939 --> 00:04:15.699 the math in the document is pretty compelling. 00:04:15.960 --> 00:04:18.279 They break down the unit economics. You price 00:04:18.279 --> 00:04:21.259 a book at $2 .99. Which is the impulse buy price. 00:04:21.339 --> 00:04:22.879 It's basically the price of a coffee. No one 00:04:22.879 --> 00:04:25.040 thinks twice about it. Exactly. So in $2 .99, 00:04:25.180 --> 00:04:28.379 you only need to sell about 50 copies a day to 00:04:28.379 --> 00:04:31.279 generate, what, $5 ,000 to $7 ,000 a month? And 00:04:31.279 --> 00:04:33.939 that's just one book. The strategy here isn't 00:04:33.939 --> 00:04:37.040 to write one masterpiece. The source explicitly 00:04:37.040 --> 00:04:40.060 talks about the four book series strategy. If 00:04:40.060 --> 00:04:42.680 you launch four books in a series, they all act 00:04:42.680 --> 00:04:45.040 as a funnel for each other. You can hit that 00:04:45.040 --> 00:04:47.899 50 sales a day mark across the whole ecosystem. 00:04:48.240 --> 00:04:50.699 The document claims this has an annual potential 00:04:50.699 --> 00:04:55.939 of over $240 ,000. That is a top 1 % income for 00:04:55.939 --> 00:04:58.360 being essentially a Pulp Fiction writer. And 00:04:58.360 --> 00:05:01.660 just look at the cost. book can cost you anywhere 00:05:01.660 --> 00:05:04.420 from a thousand to five thousand dollars just 00:05:04.420 --> 00:05:06.500 to get it ready for print right editing cover 00:05:06.500 --> 00:05:10.680 design all that all of it here your cost is the 00:05:10.680 --> 00:05:14.300 subscription fee for claude and maybe mid -journey 00:05:14.300 --> 00:05:17.300 20 maybe 40 bucks a month the barrier to entry 00:05:17.300 --> 00:05:20.040 has just completely collapsed so it's not about 00:05:20.040 --> 00:05:22.360 writing the next great gatsby at all it's about 00:05:22.360 --> 00:05:25.000 feeding a very specific hunger exactly you're 00:05:25.000 --> 00:05:27.459 moving from mass appeal to obsessed reliability 00:05:27.459 --> 00:05:31.110 okay so if the barrier is zero Why isn't everyone 00:05:31.110 --> 00:05:34.470 doing this? Or I guess, why is most AI fiction 00:05:34.470 --> 00:05:37.670 so bad? Because most people skip the engineering 00:05:37.670 --> 00:05:40.069 part. They just open up ChatGPT and say, write 00:05:40.069 --> 00:05:42.350 me a book about a werewolf. And the AI writes 00:05:42.350 --> 00:05:44.410 something generic and boring and just full of 00:05:44.410 --> 00:05:47.089 plot holes. This workflow, it's all about constraints. 00:05:47.610 --> 00:05:49.490 Let's get into that workflow then. So step one 00:05:49.490 --> 00:05:52.750 is research. And this isn't just, you know, brainstorming 00:05:52.750 --> 00:05:55.610 ideas. They're using data. to find what the source 00:05:55.610 --> 00:05:58.209 calls content gaps yeah they're acting like data 00:05:58.209 --> 00:06:01.589 scientists the guide recommends using perplexity 00:06:01.589 --> 00:06:05.250 specifically in its pro search mode you are not 00:06:05.250 --> 00:06:07.430 looking for what's popular okay if you search 00:06:07.430 --> 00:06:09.790 popular romance books you'll just get crushed 00:06:09.790 --> 00:06:12.410 by the competition you're looking for complaints 00:06:12.410 --> 00:06:16.120 complaints you mean like bad reviews Sort of. 00:06:16.120 --> 00:06:18.660 The prompt strategy is to search Reddit threads, 00:06:18.920 --> 00:06:22.000 Goodreads reviews, forums, anywhere. Readers 00:06:22.000 --> 00:06:24.620 are saying things like, I'm so tired of reading 00:06:24.620 --> 00:06:27.199 X or why are there no books about Y? You're looking 00:06:27.199 --> 00:06:30.620 for frustration. Frustration equals demand. That 00:06:30.620 --> 00:06:32.620 is brilliant. You're literally finding the market 00:06:32.620 --> 00:06:35.779 failure. Precisely. And once you find a potential 00:06:35.779 --> 00:06:37.959 niche, let's just stick with our monster romance 00:06:37.959 --> 00:06:40.839 example, you have to validate it. You go to Amazon, 00:06:41.120 --> 00:06:43.300 find the top books in that weird little category 00:06:43.300 --> 00:06:47.040 and check their bestseller rank. Or BSR. What 00:06:47.040 --> 00:06:49.040 does that tell you? If those books are selling 00:06:49.040 --> 00:06:51.339 50 plus copies a day, you've got a green light. 00:06:51.519 --> 00:06:54.040 If they're selling zero, the niche is dead. Move 00:06:54.040 --> 00:06:56.740 on. OK, so we found a hungry crowd. Now we have 00:06:56.740 --> 00:06:59.199 to actually plot the story. And this is where 00:06:59.199 --> 00:07:01.060 the guide brings in a concept they call reverse 00:07:01.060 --> 00:07:03.360 thinking. This was one of the biggest takeaways 00:07:03.360 --> 00:07:07.259 for me. It's so key. The rule is you must plot 00:07:07.259 --> 00:07:10.319 the ending first before you write a single word 00:07:10.319 --> 00:07:13.279 of chapter one. Why is the ending so critical 00:07:13.279 --> 00:07:15.689 for an AI? because of something called prompt 00:07:15.689 --> 00:07:18.529 drift it's the technical term for what happens 00:07:18.529 --> 00:07:20.910 when an ai writes a long story it starts out 00:07:20.910 --> 00:07:24.089 strong but by say chapter four it forgets the 00:07:24.089 --> 00:07:26.889 main character's motivation by chapter 10 it 00:07:26.889 --> 00:07:28.930 completely forgets a subplot you started in chapter 00:07:28.930 --> 00:07:31.769 two the story just It wanders. It creates those 00:07:31.769 --> 00:07:34.550 infamous hallucinations and plot holes. Massive 00:07:34.550 --> 00:07:37.310 ones. So by deciding the ending for, I mean, 00:07:37.310 --> 00:07:39.889 literally asking the AI to generate 10 satisfying 00:07:39.889 --> 00:07:41.949 endings and then you just pick one, you create 00:07:41.949 --> 00:07:44.350 a destination, you turn that ending into a PDF. 00:07:44.629 --> 00:07:47.430 The source calls it ending reference dot PDF. 00:07:48.170 --> 00:07:50.829 Now, every single time you ask the AI to write 00:07:50.829 --> 00:07:53.790 a chapter, you are referencing that PDF. It acts 00:07:53.790 --> 00:07:55.600 as an anchor for the whole project. It's like 00:07:55.600 --> 00:07:58.160 a GPS. If you don't put in a destination, the 00:07:58.160 --> 00:08:00.660 GPS just drives around aimlessly. But if you 00:08:00.660 --> 00:08:03.740 lock in the destination, every turn is calculated 00:08:03.740 --> 00:08:06.019 to get you there. That's a great analogy. The 00:08:06.019 --> 00:08:09.319 AI is the car, but that PDF is the satellite 00:08:09.319 --> 00:08:13.259 lock. So why is that PDF step so critical for 00:08:13.259 --> 00:08:16.839 the AI? It acts as an anchor so the AI never, 00:08:16.939 --> 00:08:19.180 ever forgets the destination. So we have the 00:08:19.180 --> 00:08:21.629 ending. But we still need the middle, obviously. 00:08:21.769 --> 00:08:24.610 The source calls this next phase the Bible. Yeah, 00:08:24.670 --> 00:08:26.990 this is where you build your constraints. You 00:08:26.990 --> 00:08:29.089 need two more PDFs. The first one is the character 00:08:29.089 --> 00:08:31.529 profiles. Okay. And the source notes something 00:08:31.529 --> 00:08:34.830 really specific here. Ask for detailed physical 00:08:34.830 --> 00:08:37.820 descriptions. If you don't lock down that the 00:08:37.820 --> 00:08:40.600 hero has a scar on his left cheek, the AI will 00:08:40.600 --> 00:08:42.899 forget it or move it to the right cheek or just 00:08:42.899 --> 00:08:45.960 remove it entirely three chapters later. Consistency 00:08:45.960 --> 00:08:48.000 is really the enemy of large language models. 00:08:48.240 --> 00:08:50.379 It really is. So you lock the characters in a 00:08:50.379 --> 00:08:53.399 PDF, then you build the structure, and the source 00:08:53.399 --> 00:08:55.379 swears by the save the cat method. Which is that 00:08:55.379 --> 00:08:57.200 famous Hollywood screenwriting formula, right? 00:08:57.259 --> 00:09:00.320 Exactly. Blake Snyder's beat sheet, it's got 00:09:00.320 --> 00:09:03.580 15 beats. Opening image, theme stated, catalyst. 00:09:04.250 --> 00:09:07.389 Dark Night of the Soul, Finale, all of it. Why 00:09:07.389 --> 00:09:10.409 save the cat specifically? I mean, why not just 00:09:10.409 --> 00:09:13.870 a standard three act structure? Because AI naturally 00:09:13.870 --> 00:09:16.789 wants to resolve conflict. Have you ever noticed 00:09:16.789 --> 00:09:19.889 that if you ask an AI to write a story, everyone 00:09:19.889 --> 00:09:22.429 tends to get along a little too quickly? It tries 00:09:22.429 --> 00:09:25.120 to be helpful and nice. Yes. Oh my god, yes. 00:09:25.340 --> 00:09:27.580 It hates tension. It just wants everyone to go 00:09:27.580 --> 00:09:30.039 have some tea and resolve their differences calmly. 00:09:30.320 --> 00:09:33.460 Save the Cat forces tension. It forces a dark 00:09:33.460 --> 00:09:35.419 night of the soul where the hero has to lose 00:09:35.419 --> 00:09:38.360 everything. By forcing the AI to follow this 00:09:38.360 --> 00:09:40.379 beat sheet, you stop the story from becoming 00:09:40.379 --> 00:09:43.379 this boring sequence of nice events. So you upload 00:09:43.379 --> 00:09:45.720 your ending PDF and your character PDF to Claude, 00:09:45.799 --> 00:09:47.799 and you just say, create a chapter -by -chapter 00:09:47.799 --> 00:09:49.940 outline using the Save the Cat structure. That's 00:09:49.940 --> 00:09:51.980 the prompt. I love the analogy the source uses 00:09:51.980 --> 00:09:54.340 here. It compares this to... To building a blueprint 00:09:54.340 --> 00:09:56.539 before you lay a single brick. Most people just 00:09:56.539 --> 00:09:58.559 want to start writing. But this workflow spends 00:09:58.559 --> 00:10:00.480 a huge amount of time just building the constraints. 00:10:00.860 --> 00:10:03.379 That's the key word right there. Constraints. 00:10:03.700 --> 00:10:07.159 AI is creative, sure, but it's chaotic. You have 00:10:07.159 --> 00:10:09.679 to fence it in. So we aren't writing yet. We 00:10:09.679 --> 00:10:11.960 are just building constraints. Right. We're building 00:10:11.960 --> 00:10:15.000 a context cage so the AI stays on track. I love 00:10:15.000 --> 00:10:17.740 that term, context cage. It really implies the 00:10:17.740 --> 00:10:21.720 AI is this wild. animal that needs to be penned 00:10:21.720 --> 00:10:24.220 in so okay the cage is built now we actually 00:10:24.220 --> 00:10:26.440 have to generate the pros now the fun starts 00:10:26.440 --> 00:10:29.980 the source recommends using claude 4 .5 opus 00:10:29.980 --> 00:10:32.960 in project mode let's walk through the actual 00:10:32.960 --> 00:10:34.620 drafting phase because this is where the industrial 00:10:34.620 --> 00:10:36.779 part really kicks in yeah this is the assembly 00:10:36.779 --> 00:10:39.980 line you open a fresh chat you upload your three 00:10:39.980 --> 00:10:42.879 holy grail pdfs and you start with chapter one 00:10:43.340 --> 00:10:46.039 But, and this is really crucial nuance, you don't 00:10:46.039 --> 00:10:48.419 just say write chapter two. You have to reprime 00:10:48.419 --> 00:10:50.879 the pump every single time. How so? Your prompt 00:10:50.879 --> 00:10:53.679 has to be something like write chapter two, aligning 00:10:53.679 --> 00:10:55.940 closely with the save the cat structure and the 00:10:55.940 --> 00:10:58.279 character profiles in the attached files. You 00:10:58.279 --> 00:11:00.039 have to constantly remind it of the constraints. 00:11:00.120 --> 00:11:02.360 If you stop reminding it, it starts to hallucinate. 00:11:02.559 --> 00:11:05.600 And even with all of that, the source mentions 00:11:05.600 --> 00:11:09.210 a very specific technical hurdle. The context 00:11:09.210 --> 00:11:12.049 window limit. This is the bottleneck. Even the 00:11:12.049 --> 00:11:15.409 best AI models, like Claude Opus, have a limit 00:11:15.409 --> 00:11:17.350 on how much text they can hold in their active 00:11:17.350 --> 00:11:20.509 memory. For a novel, you usually hit that wall 00:11:20.509 --> 00:11:24.450 around chapter 15. And the AI starts to get dementia, 00:11:24.710 --> 00:11:26.470 is how they put it. It forgets what happened 00:11:26.470 --> 00:11:28.190 in chapter one. Right. I think we've all seen 00:11:28.190 --> 00:11:30.350 that in long chat threads. The bot starts repeating 00:11:30.350 --> 00:11:32.970 itself or contradicting itself. The source offers 00:11:32.970 --> 00:11:36.120 a workaround for this that feels very... Manual. 00:11:36.360 --> 00:11:38.519 It is manual, but it's effective. It's a hard 00:11:38.519 --> 00:11:41.039 refresh. Once you hit that limit, say chapter 00:11:41.039 --> 00:11:43.700 15, you take everything you've written so far, 00:11:43.840 --> 00:11:46.259 compile it into a new PDF called chapters 1 to 00:11:46.259 --> 00:11:48.879 15, and you start a brand new chat session. Oh, 00:11:48.879 --> 00:11:51.700 wow. And you upload that new PDF as its history. 00:11:51.820 --> 00:11:54.139 You're manually giving it a long -term memory. 00:11:54.360 --> 00:11:55.879 You're giving it a summary of the past so it 00:11:55.879 --> 00:11:57.440 can continue the future. It's like you're clearing 00:11:57.440 --> 00:11:59.840 its cache. Yeah. And the speed difference. I 00:11:59.840 --> 00:12:01.720 mean, the comparison between the free plan and 00:12:01.720 --> 00:12:04.610 the pro plan was startling. The source says on 00:12:04.610 --> 00:12:07.029 the pro plan, doing all this, you can finish 00:12:07.029 --> 00:12:10.070 a full novel draft in a single afternoon. A single 00:12:10.070 --> 00:12:12.889 afternoon. That is, I mean, honestly, it's hard 00:12:12.889 --> 00:12:14.909 to wrap my head around that kind of volume. It's 00:12:14.909 --> 00:12:17.509 industrial scale. It is. But I have to be vulnerable 00:12:17.509 --> 00:12:20.190 here for a second. I still wrestle with prompt 00:12:20.190 --> 00:12:22.669 drift myself, even with all these PDFs and context 00:12:22.669 --> 00:12:25.720 cages. Sometimes the AI just... It goes off the 00:12:25.720 --> 00:12:28.220 rails. It takes a lot of active management. You 00:12:28.220 --> 00:12:30.200 aren't just watching Netflix while it writes. 00:12:30.320 --> 00:12:33.299 You are reading every output, checking for drift, 00:12:33.480 --> 00:12:36.279 regenerating scenes. It's not writing, but it 00:12:36.279 --> 00:12:39.139 is intense editing. It sounds like memory management 00:12:39.139 --> 00:12:41.519 is the real skill here. Yes. You're managing 00:12:41.519 --> 00:12:44.440 the AI's short -term memory to maintain a long 00:12:44.440 --> 00:12:46.220 -term narrative. We're going to take a quick 00:12:46.220 --> 00:12:48.240 break, but when we come back, we need to talk 00:12:48.240 --> 00:12:51.080 about the final polish, how to make sure this 00:12:51.080 --> 00:12:52.759 doesn't sound like a robot wrote it, and the 00:12:52.759 --> 00:12:55.779 tool called Nano Banana Pro that solves the biggest 00:12:55.779 --> 00:13:02.500 problem with AI art. Stick around. Welcome back. 00:13:02.539 --> 00:13:05.740 We're deep diving into the workflow of high -volume 00:13:05.740 --> 00:13:08.899 AI authors. We've researched a niche using data. 00:13:09.259 --> 00:13:11.659 We've plotted the ending first to avoid drift. 00:13:11.919 --> 00:13:15.039 And we've generated a draft using a context cage 00:13:15.039 --> 00:13:17.559 of PDFs. Right. But now we have a raw manuscript. 00:13:18.220 --> 00:13:20.720 And let's be honest, raw AI text is usually pretty 00:13:20.720 --> 00:13:23.820 dry. It's competent, but it's soulless. It tends 00:13:23.820 --> 00:13:26.960 to overuse certain words like shiver, tapestry, 00:13:27.120 --> 00:13:29.659 and delve. It really, really loves the word delve. 00:13:29.799 --> 00:13:31.720 So how do we fix that? The source outlines a 00:13:31.720 --> 00:13:33.940 two -phase editing process. Yeah. And phase one 00:13:33.940 --> 00:13:36.279 is actually using AI against itself. Right. You 00:13:36.279 --> 00:13:37.840 don't start by reading it yourself. You upload 00:13:37.840 --> 00:13:39.840 the whole draft back into Cloud and you ask it 00:13:39.840 --> 00:13:42.360 to look for logic errors. You say, identify continuity 00:13:42.360 --> 00:13:44.820 errors. Did I say the door was locked and then 00:13:44.820 --> 00:13:46.860 she walked through it? Did I change the car from 00:13:46.860 --> 00:13:48.960 a Ford to a s***? Chevy. And AI is good at that. 00:13:49.259 --> 00:13:51.440 Incredibly good. It spots those logical breaks 00:13:51.440 --> 00:13:53.919 because it just treats them as data points. But 00:13:53.919 --> 00:13:57.399 phase two. Phase two has to be human. Phase two 00:13:57.399 --> 00:14:00.500 is the human Polish. This is where you actually 00:14:00.500 --> 00:14:02.960 earn your money. You have to look for tone, voice, 00:14:03.080 --> 00:14:05.679 and emotional rhythm. And the source suggests 00:14:05.679 --> 00:14:09.299 a simple, brutal test. Read the first chapter 00:14:09.299 --> 00:14:11.759 and the last chapter out loud. Out loud. Yep. 00:14:12.000 --> 00:14:15.440 Your ear catches things, your eye misses. If 00:14:15.440 --> 00:14:17.860 you trip over the words or if the dialogue sounds 00:14:17.860 --> 00:14:20.000 like a robot trying to act human, the reader's 00:14:20.000 --> 00:14:22.519 going to hate it. You have to make sure the emotional 00:14:22.519 --> 00:14:25.539 beats are actually landing. If the dark night 00:14:25.539 --> 00:14:27.120 of the soul doesn't make you feel something, 00:14:27.179 --> 00:14:29.700 you have to rewrite it manually. Okay, let's 00:14:29.700 --> 00:14:31.919 move on to the cover. We've all seen AI art. 00:14:31.960 --> 00:14:34.120 It can be incredible, but it has a notorious 00:14:34.120 --> 00:14:37.000 weakness. Hands are getting better. But text. 00:14:37.159 --> 00:14:39.799 Oh, text usually looks like alien hieroglyphics. 00:14:39.820 --> 00:14:42.559 It's a mess. It's like a soup of letters. A disaster. 00:14:42.940 --> 00:14:44.840 But this guide recommends a tool I had never 00:14:44.840 --> 00:14:48.000 even heard of. Nano Banana Pro. I know, right? 00:14:48.139 --> 00:14:51.480 It sounds like a mobile game for toddlers. Nano 00:14:51.480 --> 00:14:53.639 Banana Pro. It's actually a specific tool you 00:14:53.639 --> 00:14:55.899 can access through the Gemini ecosystem. And 00:14:55.899 --> 00:14:58.220 the source recommends it over mid -journey or 00:14:58.220 --> 00:15:01.460 daily. Purely for text rendering. It can actually 00:15:01.460 --> 00:15:03.779 spell the title correctly. It can spell. If you're 00:15:03.779 --> 00:15:06.000 selling a book called The Werewolf Secret, you 00:15:06.000 --> 00:15:08.519 cannot have the cover say The Werewolf Scrit. 00:15:09.019 --> 00:15:11.320 Midjourney really struggles with that. Nano Banana 00:15:11.320 --> 00:15:14.480 Pro handles typography inside the image generation 00:15:14.480 --> 00:15:17.720 much better. And it handles the aspect ratio, 00:15:17.919 --> 00:15:20.440 the book cover dimensions? Yes. The guide is 00:15:20.440 --> 00:15:25.500 specific. 2560 by 1600. pixels the claim is that 00:15:25.500 --> 00:15:27.580 you can generate a professional looking cover 00:15:27.580 --> 00:15:31.340 with legible title text in about 15 minutes and 00:15:31.340 --> 00:15:34.000 critically it's free that really is a moment 00:15:34.000 --> 00:15:36.820 of wonder for me i remember when getting a cover 00:15:36.820 --> 00:15:39.940 design cost 500 and took three weeks of back 00:15:39.940 --> 00:15:42.379 and forth emails with a designer and now it's 00:15:42.379 --> 00:15:45.659 15 minutes and zero dollars it's just wild it's 00:15:45.659 --> 00:15:47.899 the democratization of the entire supply chain 00:15:47.899 --> 00:15:50.919 it unlocks the ability for anyone to look professional 00:15:51.440 --> 00:15:54.539 So why use Gemini or Nano Banana specifically 00:15:54.539 --> 00:15:56.919 for the cover? It's the only one that doesn't 00:15:56.919 --> 00:15:59.419 mess up the text on the book title. OK, so we 00:15:59.419 --> 00:16:01.740 have the book. We have the cover. Now we have 00:16:01.740 --> 00:16:03.820 to sell it. And the source says this is where 00:16:03.820 --> 00:16:08.419 most authors fail. Great book plus bad SEO equals 00:16:08.419 --> 00:16:11.419 no sales. This is where we have to change how 00:16:11.419 --> 00:16:13.879 we think about Amazon. We tend to think of it 00:16:13.879 --> 00:16:15.840 as a bookstore. You know, we imagine brising 00:16:15.840 --> 00:16:19.190 shelves. But the source is very clear. Amazon 00:16:19.190 --> 00:16:22.169 is a search engine. It's Google. But for products, 00:16:22.389 --> 00:16:24.809 if you aren't optimizing for the algorithm, you 00:16:24.809 --> 00:16:27.669 are invisible. So how did these authors optimize? 00:16:27.950 --> 00:16:31.649 They use AI again. The workflow uses ChatGPT 00:16:31.649 --> 00:16:34.210 specifically in its thinking mode to analyze 00:16:34.210 --> 00:16:36.879 the market. You feed it your novel summary and 00:16:36.879 --> 00:16:39.139 ask it to analyze successful book descriptions 00:16:39.139 --> 00:16:41.639 in your specific niche. And then ask it to generate 00:16:41.639 --> 00:16:45.600 titles. SEO optimized titles and subtitles. You're 00:16:45.600 --> 00:16:47.259 not trying to be clever with your title. Right. 00:16:47.440 --> 00:16:49.379 You know, The Moon's Glow is a poetic title. 00:16:49.519 --> 00:16:52.320 Alpha Wolf's Forbidden Mate and Age Gap Romance 00:16:52.320 --> 00:16:55.279 is an SEO title. You want the keywords that people 00:16:55.279 --> 00:16:57.159 are actually typing into the search bar to be 00:16:57.159 --> 00:16:59.659 right there in your title and subtitle. And then 00:16:59.659 --> 00:17:01.519 there's the pricing strategy. The source is very, 00:17:01.580 --> 00:17:04.730 very specific about the sweet spot. The golden 00:17:04.730 --> 00:17:08.289 zone. It's $2 .99 to $9 .99. This is an Amazon 00:17:08.289 --> 00:17:11.390 rule. If you price your book within that range, 00:17:11.589 --> 00:17:14.990 Amazon gives you a 70 % royalty rate. If you 00:17:14.990 --> 00:17:19.789 go below $2 .99, your royalty drops to 35%. And 00:17:19.789 --> 00:17:23.430 if you go above $9 .99, it drops to 35%. Wow, 00:17:23.549 --> 00:17:25.799 that's a massive cliff. It really forces your 00:17:25.799 --> 00:17:28.500 hand. It forces the entire market into that pocket. 00:17:28.579 --> 00:17:31.779 So almost every indie author sits right at $2 00:17:31.779 --> 00:17:34.900 .99 to maximize their volume while keeping that 00:17:34.900 --> 00:17:38.250 70 % cut. It's fascinating how the platform dictates 00:17:38.250 --> 00:17:41.109 the economics, which then dictates the content. 00:17:41.390 --> 00:17:43.390 Completely. You format it with Kindle Create, 00:17:43.609 --> 00:17:45.529 which is a free tool from Amazon. You upload 00:17:45.529 --> 00:17:48.289 it to KDP. And within one to three days, you're 00:17:48.289 --> 00:17:50.750 live on the biggest bookstore on earth. So Amazon 00:17:50.750 --> 00:17:52.730 is treated more like a search engine than a bookstore. 00:17:53.230 --> 00:17:56.130 Precisely. You are optimizing for keywords, not 00:17:56.130 --> 00:17:58.470 for browsing. It's incredible to see it all laid 00:17:58.470 --> 00:18:00.750 out like this. When you zoom out and you look 00:18:00.750 --> 00:18:02.930 at the research, the context cages, the nano 00:18:02.930 --> 00:18:06.900 banana covers, it's... not really about writing 00:18:06.900 --> 00:18:09.640 in the romantic sense of the word, is it? No, 00:18:09.759 --> 00:18:12.279 not at all. If I had to summarize this whole 00:18:12.279 --> 00:18:14.799 deep dive, I'd say this. It is not about art. 00:18:14.880 --> 00:18:17.400 It's about pipeline management. It's a seven 00:18:17.400 --> 00:18:20.559 -step industrial workflow. You have research 00:18:20.559 --> 00:18:23.539 with perplexity to find the demand. You have 00:18:23.539 --> 00:18:26.480 structure with Save the Cat to tame the chaos. 00:18:26.819 --> 00:18:29.740 You have context management with those PDF cages 00:18:29.740 --> 00:18:33.380 in Claude. And you have distribution via Amazon 00:18:33.380 --> 00:18:36.200 SEO. It's just systems engineering applied to 00:18:36.200 --> 00:18:40.460 imagination. Systems engineering applied to imagination. 00:18:41.240 --> 00:18:43.279 That's a really powerful way to put it. The source 00:18:43.279 --> 00:18:45.299 wraps up with a bottom line that I think is important. 00:18:45.440 --> 00:18:48.160 It says, this isn't magic. It works for people 00:18:48.160 --> 00:18:51.859 who think in systems and value volume over perfection. 00:18:52.160 --> 00:18:53.859 It's for the person who's willing to put in the 00:18:53.859 --> 00:18:56.000 reps. That's the truth of it. The barrier to 00:18:56.000 --> 00:18:58.140 entry is low. Anyone can buy a subscription. 00:18:58.500 --> 00:19:01.599 But the barrier to success is consistency. You 00:19:01.599 --> 00:19:04.240 have to treat it like a job, not a hobby. As 00:19:04.240 --> 00:19:05.900 we wrap up, I just want to leave you with a thought. 00:19:06.039 --> 00:19:08.700 We talked about niche hunger today. Those readers 00:19:08.700 --> 00:19:10.660 who are just desperate for something very specific 00:19:10.660 --> 00:19:12.900 that they can't find. My question to you is, 00:19:13.000 --> 00:19:16.680 what is the micro niche that you know better 00:19:16.680 --> 00:19:20.539 than anyone else? What's the weird specific topic 00:19:20.539 --> 00:19:22.880 that you and your friends complain there isn't 00:19:22.880 --> 00:19:25.839 enough content about? because chances are there 00:19:25.839 --> 00:19:27.420 are a few thousand other people thinking the 00:19:27.420 --> 00:19:29.160 same thing, just waiting for someone to build 00:19:29.160 --> 00:19:31.599 the system to feed them. That's the million -dollar 00:19:31.599 --> 00:19:34.519 question, or at least the $240 ,000 question. 00:19:34.619 --> 00:19:36.940 Thanks for listening to The Deep Dive. We'll 00:19:36.940 --> 00:19:37.579 see you next time.