兴趣 × 分发 × AI:一个人也能搭起可复利的生意
If You Have Multiple Interests, Please Start a One-Person Business — 一人公司的公式是:兴趣与独特价值 + 社交媒体分发 + AI 放大执行力;你卖的不是产品,而是被感知的价值。
核心主张
一人公司的公式是:兴趣与独特价值 + 社交媒体分发 + AI 放大执行力;你卖的不是产品,而是被感知的价值。
Dan Koe 把一人公司拆成三块:让你不被淹没的个人品牌、决定胜负的报价、以及每天可执行的分发流程——每天 1 到 3 条短内容、每周一篇长文、每天推广一次。AI 在这里是六个需要上下文的员工,而不是全能替身:先有自己的流程、品味和作品,才谈得上自动化。
几条主线
兴趣 × 分发 × AI 的一人公司公式
竞争不在产品,而在报价
AI 是员工,不是全能神
关键知识点
收入与时间脱钩
报价不等于产品
报价的四个杠杆
注意力时长即信任
给 AI 上下文,而不是一句指令
系统要迭代,不是放弃
时间线
- 00:00引子
- 07:31要素一:品牌
- 09:40要素二:报价
- 11:26要素三:分发
- 14:11日常执行
- 18:38六个 AI 员工
- 23:57正确用 AI
- 31:17系统化迭代
原话摘录
“We have moved to a model of paying for intelligence and value that is decoupled from time.”
我们已经转向一种为智慧和价值付费、并与时间脱钩的模式。
“You are not competing at the product or service level, you are competing at the offer level.”
你不是在产品或服务层面竞争,你是在报价层面竞争。
“So just take some damn idea and write it down.”
所以就随便抓一个想法,写下来。
“In the case of AI, this means that if you notice something wrong, strange, or low-quality, you should point out the error.”
在 AI 的情况下,这意味着如果你注意到有什么不对、奇怪或低质量,你就该指出错误。
“Most people processes by loading skills before they have anything worth automating.”
大多数人还没任何值得自动化的东西,就靠加载 skill 来跑流程。
行动清单
- 选定一个平台(建议 Substack),每天早上固定 30 分钟:先用 AI 找想法、定选题,再写成 1 到 3 条短内容发出去。
- 每周写一篇长文(从 250 字起),周初先列好本周选题和结构,写完在中间和结尾各放一次报价。
- 用 PAS(问题—放大—解决)结构每天换一种方式推广报价:长文、帖子回复、Instagram stories 或评论关键词。
- 把个人品牌策略整理成一份可复用文档,连同你喜欢的范例、字幕、文章一起喂给 AI,固定成每周工作流。
- 每次 AI 输出不理想时,指出具体错误、补足上下文、继续追问,而不是直接判定“AI 没用”。
00:00This quote has been haunting me for a whole year. A few years ago, when I first started creating content because I wanted to avoid cold-sending emails to find clients, I noticed people, creators, who were building businesses one by one with revenue ranging from one to ten million dollars.
00:35And they didn't work all day. I think Justin Welsh is a perfect example of this. He values his time very much and only works a little in the mornings, but he has the elements of the business, which we'll talk about, that allow him to make that kind of money. Thanks to the Internet, social media, and AI, this has become possible. People still don't realize how much leverage and power they have as individuals. We have long since moved away from the time-based payment model. We have moved to a model of paying for intelligence and value that is decoupled from time. This is difficult for many to understand, but I'll explain it more simply: you can create one post, and if it gets a million views and 1% of viewers buy your product for $100 — calculate the profit yourself.
01:23Of course, you're unlikely to get a million views on your first post, but you get the point. It's a matter of practice. It's a skill that can be developed, and it's one of the most valuable competencies that companies, creators, and everyone around them are looking for right now. That's why this video might be very important to you . So, creators who built a one-person business with an income of up to 10 million sold info products, or social media as a distribution channel. No paid advertising, podcast sponsorships, billboards, radio or TV—just social media, because that's where the audience's attention is focused.
01:59Creating an audience is free. content is free. This is a low-risk business model for a beginner, because in , for example, you need to invest capital in the product. Today, there are no costs to starting a one-person business on social media, except for the time spent on training. But from the time I started building these businesses until now, a lot has changed. Back then, you could build a business worth millions before most people knew about it, and now they know even less about the new prospects that AI opens up. And now GPT-6 Astra, Fable 5.1, and whatever next model comes along, make it even more real. And your competitive advantage here is that the average person simply doesn't notice anything.
02:46They are not doing anything and are not going to change their habits. But the beautiful, very beautiful thing about this is that you don't have to make a million dollars. You don't need to make $10 million. Most people watching this just want to replace their current sources of income. It could be $75,000 a year, $100,000 a year. And what I'm telling you is that it's incredibly easy to achieve. That doesn't mean it's easy or will happen quickly. I'm just saying it's entirely achievable. As for me, when I first started, my goal was to be able to afford Chipotle for lunch every day. And I'm proud of myself for getting there pretty quickly, and now I don't eat Chipotle for lunch every day, although I probably will today.
03:25But since childhood, I just didn't want to follow the standard path. As a child, I was very observant of those around me, and it was pointless to go to school, get an education, and pay a lot of money for training that would become completely irrelevant in 10 years. And then get a job that you depend on and that takes up all your time , so you can't educate yourself and change direction, and then marry someone you despise and have the promised pension, but never actually be able to stop working. It was all normal, and it disgusted . So I realized that entrepreneurship was the only way out. This was my path, whether I liked it or not, because it was a big mysterious thing.
04:05Opening my own business was something I wasn't capable of doing at the time. And so I did just that . I failed again and again. I started a fitness channel on YouTube. I tried dropshipping wallets. I tried selling glasses with light protection. I did freelance SEO, Facebook advertising, and web design until I finally had to get a job because my student and credit card debt was piling up. And when I got that job, I treated it like a deadline. I knew that if I didn't make my business work soon, I would start drowning in bills and responsibilities and wouldn't be able to get out. So about 7 years ago I was able to quit my time job to do freelance web design because it was one of the skills I had.
04:44About 6 years ago, I started running social media and made about $10,000 from digital products. The next year it was 100,000, then 800,000, and I have a video of how I did it, 2 hours a day writing. Then the next year it was about 4 million and my tax bill was officially four times previous salary. And then, when I made the "stupid" but far-sighted decision to go into software, my income stabilized at 1–1.5 million per year because I had to shift my time and focus to software and invest a lot of money there, since we were . And I still haven't paid myself a penny from Cortex or a new project like Eden, but we're moving towards that.
05:25is gaining momentum very quickly. And you guys won't believe the next update. So I want to say: I have a lot to share about person's business. I see it from a bird's eye view , because my audience and my clients are one-man businesses; I see what works and what doesn't, and when it comes to AI, I'm kind of front row on how it can help you build this thing faster. So let's finally get started. We'll talk about why your interests or unique value plus social media or distribution plus AI or cheap multiplication of effort equals a one-person business. So, as a one- person business, you will be posting on social media.
06:06Boo, I know how hard it is to write for 30 minutes in the morning and then post it where everyone else in the world can see it. But, as I said, social media is where the focus is right now, and you probably don't have the capital to invest in paid advertising or pay creators $2,500 to post one reel about your product . The prices for this UGC content are getting I personally think it's better to hire people or build your own audience as a founder. In this video, I also assume that you don't want to start a service business or a client-driven business like freelancing or agency work because you want some freedom and a lifestyle.
06:42Some of what we're going to talk about will apply to you even if you have this style of business, because you should be building an audience anyway, but I prefer writing and creating products that sell while I sleep. Now we'll talk about what one-person business looks like. What does it consist of and what elements do you need? And then I'll give you little AI employees that you can use right away Links to them are in the description. You can just start communicating and essentially running your business with the help of these six employees I'm giving you. You can do this in Eden in one click, and they will all be in one place ; or you can download a separate skill file that you upload to Claude.
07:20Or you can add them to Eden and then connect Eden to Claude via MCP so that Claude can see social media posts, research, and these special AIs or employees. So, the first thing we need to define when creating a person business is the brand, which is how you present yourself or your business to the world . In your personal life, when people hear your name, what do they think of? What do they associate you with? Is it nice to talk to you? What are your likes and dislikes? Are you having fun? Are you witty? Do you help others? Can I trust you ? Are you a valuable person? What do people ask you about? Things in this spirit.
07:56So, a brand is the same thing , but for your business. It's like a digital storefront or digital identity . This is what makes people buy your app, course, coaching, or product in an online store instead of someone else's, even if it's worse in quality, because there are tons of incredible apps or courses out there that you have n't even heard of because you don't know they exist. And if you knew about them, then if a content creator you like and who has helped you in the past because you spent so much time on their offered a similar product, you would buy the product of that creator you like. This is how marketers or people who study digital marketing, but real earn 10 times more than programmers who just build little portfolio apps and don't know how to sell them.
08:40Now, in a one- person business, your brand is a personal brand. And even if you're not a one-man business, look at Elon Musk. Look at MrBeast. Look at celebrities. They are all building a personal brand because that is what needs to be done in today's world. A personal brand is one of the business development strategies because it's free. And if right, you almost never have to market yourself. So, these are your values, your worldview, and your story, embodied outwardly. So, one of the little AI workers I'm giving you below will interview you to gather information about you and create a comprehensive document with your personal brand strategy, which you can then pass on to any other AI.
09:21That is, at any moment, in any chat, you have this document that you can take with you anywhere. This is necessary because it gives you the vision, anti-vision, and pillars of your content. So, all the content ideas generate, if you want them to have a unified line and impact, should at least be based on this personal brand strategy document. Now, the second thing you need as a sole proprietor is a proposition, which is what people pay you to do to get results. So, you need a product or service. This seems obvious. Without this, it is impossible to make money. You can choose the path of building an audience, just grow it, and then take on advertising integrations, which is a very profitable business.
09:59But today it just makes sense to have your own product or service. The difference between an offering and a product or service is that an offering is the perceived value of that product or service. It's how you package it. This is a deal you make. This is one of the most profitable skills you can master— proposal creation , because once you hit the mark with the proposal, landing pages, marketing, and more become much easier because everything is already in place. And that's what you compete on. You are not competing at the product or service level, you are competing at the offer level. That is, the amount of value compared to the price.
10:31So if you take a brain app or a tool like YouTube, a tool everyone uses, and compare it to Eden, which is able to research all platforms through MCP, through chat, and keep all your ideas, notes, links, posts, and soon products in one place, Eden is clearly the better deal, if I may brag a little, and it has a post scheduler, so it replaces another app. This is the entire workflow of a content creator. But creating a proposal is the art of controlling four levers. The first is the desired outcome, that is, what the client actually wants. The second is the delivery mechanism, which is how you help them achieve this.
11:06Is this a course? Is this coaching? Is this a supplement? What is the unique mechanism for these things? What is your way to achieve the desired result for them ? The third is results, that is, how quickly you will get them . And the fourth is a guarantee, meaning what happens if nothing works out. These are the four things you need. Write them down . And again, AI will help you create your proposal. Now the third thing a one-person business needs is distribution, that is, how to spread your brand or offering. And I like this quote from Jack Butcher: first build the distribution, then build whatever you want. Because we live in the age of the internet, where attention leverage.
11:41You can have the best brand. You can have the best product, but none of that matters if no one knows about you. So, the distribution we are interested in is everything: from marketing and sales to content, advertising, landing pages, and websites. This is how people find you, learn about you, and pay you. That's what a funnel is. For example, you probably found me through content on social media. You're watching this on YouTube right now, and if I the value of my knowledge enough, you'll start to trust me enough to use or pay for one of my products. Not because I deceived you, but because you see it as beneficial for your own purposes.
12:16This is an exchange of values on terms. I could also run paid ads or sponsor other people's tried both and will probably do so again in the future. But as a beginner with small capital, you should take this as a stepping stone. You wouldn't run paid advertising for an offer that hasn't been tested or won't sell. You will waste your money. And yes, you can test through advertising, but it's testing at your own expense. On social media, you test your ideas , products, and services for free. If you are not growing or selling anything, then there is room for improvement. This is a trial and error method . You just need to experiment, and if it doesn't work, again.
12:52And if that doesn't work, learn something new and experiment again. Eventually, you will have so many tools in your that you will become a master at this thing called business and will be able to literally make money with the click of a button because you will know what you are doing. During practice, you master not just one skill, but human psychology, the mechanics of attention, copywriting, content, and persuasion. All of these different things come together to make you an business owner. So, we're going to talk about these AI workers, or this AI team, or, frankly , just skills or queries—it's all the same thing, the context for AI.
13:24They will help you through all of this, but it is through doing the work and publishing it that you feedback from the market. You don't outsource your intelligence to AI just because it does the work for you, because that's not what learning is about . You don't learn anything by just tutorial; you learn knowledge you gain to the real thing you are trying to do. And if it worked, then you learned . And if it didn't work, then learned something too. But if you don't give it a chance to work or not work, then you don't learn anything. And even if it doesn't work the first time, you now know that AI won't replace everything overnight.
13:58You understand that reality exists, and that getting is not always a linear, step-by-step process, and that it evolves along with changing culture and perceptions of value. So, let's start simplifying this. Here's what you 'll do to actually start this business. So, this will be your daily routine. The first thing you will do is posts one to three times a day on the short-form platform of your choice. It could be Substack, X, LinkedIn, Instagram, etc. You just need to choose the one you like the most . Now on the Eden channel, I create tutorials on how to develop and write content for all of these platforms. There are already quite a few videos there, so you watch them, because of course this video would be 10 hours long if I did it here.
14:41But you start with platform and become a real pro at it. I started on Twitter, but if I were starting today, I would probably choose Substack, simply because it has long reads, short notes, and a growth engine—all in one place, and it's just text. So, I would n't have to appear on the video. I can sit down, have a coffee, write down ideas I like in the morning, and try to develop on this platform. And your subscribers on Substack aren't just followers , they're newsletter subscribers. So you own the list. You can export this and take it anywhere , which you can't do on any other platform, and I'm assuming that most of you who are watching this have a lot of interests, have deep interests,
15:18you like deep thinking. And Substack is simply the best place to do it. And I see many budding creators gaining momentum. Now the second thing you'll do is publish one longread per week on Substack, X, or YouTube, but most likely because that's where you'll probably start, although you can do the same on X because they now have articles. So, all platforms are moving to this dynamic of short and long format because that's what creators want. Creators understand that the more time people spend with your content, or " attention time," as I call it, like time under tension for muscle growth, is time under tension for social media growth.
15:55People who watch an hour-long or half-hour video on YouTube will trust you much more than those who watch a 15- second clip. They would have to watch hundreds of such videos to consume the same amount of content as in a 30-60 minute YouTube video. The same goes for written articles, like on Substack or articles in X. So, the way to make it acceptable or realistic to do is to set aside 30 minutes every morning. So, you'll have 30 minutes to write short content, and that means no staring at a blank screen. This means researching with the AI I'll give you, deciding on an idea, and then writing it as an interesting short post.
16:33Most people have trouble with the first two stages. They don't research, they don't study, they don't read, they don't look for an idea, they don't see it as part of the writing process that has to happen in the same time frame. And secondly, they can't decide on an idea. The idea has to be perfect before they start writing it. Although it is during writing that you refine the idea. So just take some damn idea and write it down. Now the second 30 minutes are for writing, maybe 250 words for a long read. This could be a script for YouTube, an article for X, or an article for Substack. I recommend that at the beginning of each week, you sketch out a plan of what are going to write for that week.
17:07You will choose a topic and start thinking about a plan and structuring it, with the help of one of the AIs. Now the third thing you're going to do is try to promote your offer once a day. So, you promote it in your longread somewhere in the middle. You just say, "Hey, if you're struggling with XYZ, try this." And then add a link to your product or service. And at the end, do the same. It's that simple. It's so short. If you don't know how to do that, give your mailing list to an AI and say, "Hey, I want to insert this in the middle and at the end. How can I do that tastefully?" And of course, since you're probably only posting longreads once a week, you'll be promoting it every day in a different way as soon as you can.
17:46It just has to become part of your day. You can add a response to one of your posts on Substack or X, simple PAS structure to promote your product or service, which stands for " problem, amplification, solution." What's the problem? If you're struggling with this, this, this and this, here's the solution , link to the product. Do it in your Instagram stories. Make it a call to action in your Instagram post. Hey, write a keyword in the comments if you're struggling with this, this, and this and want to get it. So, that's the whole process. You build distribution by writing content. You are promoting your offer. If you have n't created an offer yet, then you should focus on that as well.
18:21This is exactly what you can create in Eden , simply by taking the knowledge from your head and turning it into an AI product that is also a course. And this will become clearer when it actually launches. So keep an eye on this because you will make more money just by listing your product on Eden than on any other platform, simply because it is on Eden. So, in this whole process, there are a few things that we can do or that we can use AI for. The first is the work of a brand strategist. That is, creating a personal brand strategy so you don't get lost among others, and effectiveness of your content to offer you growth tactics.
18:53It's like a meta-level to everything else is subordinate. The second is a content strategist who researches what works effectively on social media, understands all your previous content, connects it to your brand strategy, and provides ideas for posts for the week creates drafts of content for editing, if you prefer. Now, one thing you might not realize if you're not an active AI user is what I'm telling you: you can just copy and paste. You can just grab this transcript from YouTube. There's a YouTube transcript generator at eden.so/tools, or you just plug it into Eden and say, "Hey, help me implement this newsletter."
19:28Help me figure out how we can set up a workflow to accomplish all of these tasks.” And then you communicate with him until you get the desired result. And we'll talk about how to do it right in the last part of this video, because you need to understand how to use AI. Now, the third thing we can use it for is as a long- read specialist or planner who long-form text structures and attention-grabbing strategies, helping you structure an article for a newsletter or a video script with your ideas. Fourth, we can use it as a short-form specialist that takes your draft ideas and turns them into posts that you can review, edit, and schedule for publication across all platforms.
20:02Fifth, there's a proposal consultant who helps take your experience, interests, and expertise and turn them into a proposal, and decide which product or service to start with. And the sixth is a strategist. So, it takes your offer, brand strategy, and customer avatar, and then helps generate promotional posts, email newsletters, and landing page copy. Now I've created all of this for you here, and I'll quickly show you what it looks like. But if you really want to use them, the link is in the description, or go to eden.so/kits/1-human-business. And all you have to do is click the “ set up my business” button and it will add them to your workspace, or you can upload them separately as skill files for Claude or ChatGPT.
20:44And for those waiting for the Two-Hour Writer update, it's being updated and we're still polishing it, but here's what it's going to look like. All modules will be here. You will be able to go through the curriculum and communicate with him in the process so that he can help you implement it, having the entire course and modules as context. So here's what you'll be able to sell inside Eden as a product, not just as a general AI course or skill. You will sell the entire package and it will be branded as your own. But this is how they will look in your workspace. And once you add them all, you can just switch between them here, and all your chats for each of them will be stored in one place.
21:23So it's like your little command center for all your AI assistants for a author business. One of my favorites is a tool for long reads. So, all you need to do when you want to create a viral article on Substack, X, or a newsletter is go here; click on one of the starting prompts, for example "make a mailing plan" or " make a script plan for YouTube", and the AI will ask you about the idea, because it knows how to write it. And since this is happening in Eden, or if you connect Claude to the Eden MCP, he will be able to research anything that goes viral on Substack , X articles, YouTube scripts, etc. It can extract patterns and principles from the
22:01headlines and structure of real materials to help shape your own ideas so that they work. It then gives you an outline that you can transfer to a document and you can write alongside it, because I personally like it when AI structures and researches things for me so that I have more control over the writing process itself. So I also have my little writing workspace here. And when it's time to publish an article to Substack or any social network, I just go to the publishing tab and can do everything right there. I spent a lot of time creating them. They are free to use. Of course, you will have to pay for Claude, ChatGPT, or Eden if you want to use AI, but those are the rules of the game.
22:44The actual context and knowledge inside these AIs is available to you for free, and they are around what works, because if you just ask Claude's regular chat about it, it won't know what to do and will provide terrible information. You have to shape and adjust the context to what actually works. I recommend going through them in the order I listed above. brand strategy, then content strategy, and based on them, you create your short and long posts. And once you have an offer, you can strategist tool. And once you've this process enough, you can start running them as routine tasks. So, in a content strategist, you can just say, “ Hey, do this for me every Monday and add it to a spreadsheet so all the ideas are there and I can use them.” Now you'll notice that none of them,
23:29at least the ones I've created, write content for you. It doesn't create content automatically because I don't think it's appropriate for a beginner. And personally, I don't think we're at the point where it's worth automating this side of things, although I'm changing my mind a bit about writing texts with AI, because ultimately the choice is yours , which we'll talk about soon. But if you want an AI to do it for you, to write for you, then I'm not entirely against it, but you have to understand what we're going to discuss now. So now I want to talk about the best way to use AI in general throughout all of this work.
24:03Because there is so much information on social media, you just drown in Instagram reels that tell you to download a certain tool that you plug in and it does everything for you, which is great. But if you don't know how to think about using AI properly or how to approach it, you'll run this tool just like you downloaded a Notion template, get the result, and say, "Oh, that's cool." And then you just throw it away . Or it will give you something terrible and you will think, "Oh, this doesn't work." Or “AI sucks” or “AI is bad.” Or "AI is just making things up and hallucinating, so I won't trust it," even though that's not actually the case.
24:35People are just as stupid and hallucinate no less. If you go on social media, you will see that 99% of people are spreading misinformation. There was never a time when you didn't need to critically on your own. And just because you use AI doesn't mean this rule stops working. Now this will put many off because it is a complex process and it doesn't make sense until you try to do it in practice. It takes a certain level of systems thinking to understand and implement this. But if you watch to the end and try to understand, then the next time you use AI, you will begin to notice new opportunities and implement them.
25:10The problem with artificial intelligence is that people perceive it as some kind of omnipotent being. And since it's not an omnipotent being, even now that we actually have AGI, models like GPT-6, Astra, and Fable 5.1 still can't tasks well without proper context. But since it's not an omnipotent being, when it doesn't do exactly what a person wants—because they don't know what they want— people simply assume that AI is only capable of creating garbage. A great example of this: I saw a video recently where some guy said, "Hey, don't use ChatGPT to build a website if you're a small business owner, because we can see that AI did it."
25:49But if you use AI, it's just laughable. This is true. If you just tell an AI, " Hey, make a website for this, that, and that," it will look like any other AI-made website. But if you provide examples and clear instructions, then at least you'll have a better start, but it's not the whole picture. After that, you need to start delving into the details. You have to customize the border radius, color, shadows, layouts, and how everything should look. At this point, when you have a prototype, template, or first draft, you do what any designer would do in Photoshop or Figma. You go through and make all the little detailed adjustments to make it unique and In fact, AI can make minor changes.
26:33You don't have to try to do everything the first time. Or, when you're editing a video, yes, you can let it go through everything, cut everything, and add animations, but that's just a general outline. And then when you look at the result, you think, "Oh, this is terrible." And you using it. When you come in and say, "Hey, do this a little better. Do this. Do this . Do this." That's when the real editing begins . He probably won't do everything perfectly the first time because you haven't explained to him exactly what you want. And that's okay, because you probably won't be writing a 1,000- word paragraph. You will instruct him to make a general outline, and that will be acceptable.
27:08And then you will work everything out yourself and add the small details. So, asking an AI to create a website, write a newsletter, or tweet today is the same as downloading a WordPress template once upon a time. People notice it because it's poor quality work. A few years ago, there was no less garbage than today. But AI is a special case that people really don't like. You can stop there and have a terrible site that doesn't highlight your brand at all , or use this as a starting point for improvement. Just like you build an app and then fix bugs until it's really good. Now the bar for what is considered a “ WordPress template” or garbage will rise, forcing people to develop taste and expertise if they truly want to stand out.
27:55And this is one of the reasons why my opinion about writing texts with AI is changing. Because if you treated it the same way you would when you sit down to write yourself, you would realize that with AI you have more creativity, and possibilities. But people usually say, "Hey , write this newsletter for me." And when it turns out badly, they just consider it garbage. Or they think it's okay and to publish because they don't have the taste or experience, and taste is something you're not ready to publish to the world. So when they publish it, people immediately see that it's AI. But for those who truly treat AI writing as a craft, no one will notice because this is not a basic WordPress template.
28:34It's what you choose to publish because it reflects what you would say yourself. It reflects your values, your worldview, your opinion . In addition, authors are not always authentic, speaking in the same voice they use with friends. No, they created their own authorial voice and pretended a little. They are wearing masks, and that's normal. It's part of what it means to be human. So, now that we've figured that out, what do we actually do? How do you create something like the weekly content strategist I showed you so that it's really good? First of all, you need to know what a product looks like, or have an example of what is considered good.
29:07Because when it comes to quality content or social media posts, factors play a role. This is the structure of a post, article, or newsletter. This is the author's vibe or his energy signature. There is a consistent line to their brand and big ideas that they implement to convey their worldview. There are also unique ideas that they come across that help them stand out from the rest. So, all this needs to be taken into account . No artificial intelligence will tell you that . No artificial intelligence will do this by default. You can only find this through mentorship, experience, or learning from someone who has this specific knowledge, because there is no one way to create content.
29:42But now that you know about it, you need a guideline, an example, or instructions that you can give to the AI to limit the context. This means providing it with YouTube titles and video transcripts that reflect the type of content you want to create. This Reels transcripts, notes or articles in Substack, posts or articles in X, or many examples of the writing style you like. This is how start most chats with AI now. If I know what I want, I just give detailed instructions. I used to write 2000 word queries while teaching him to write tweets like me and he comes extremely close to the result. But you wouldn't be able to do that by simply asking him to look at my best tweets, because he would try to stereotype them and not take into account the thought process that went into writing them.
30:24Now, more importantly, if I don't know what I want, I start researching and asking questions until I get clarity and fill the context box with the information. So, if I design something, I need a reference. I need to go to Mobbin or another site for inspiration. If I'm writing content, I have to go to Eden, to the "Discover" tab, and find the most effective ideas. This means I don't just say, "Hey, write me a newsletter on this topic." I ask questions to decide which newsletter to write about. So , in your case, you could ask, "I want to write one of the newsletters in the world, something that stands out and doesn't sound robotic."
30:59Let's brainstorm until we find a style I like. Start with questions about what you need from me to make this happen. We will move on from this. And then, at the end of this conversation and after giving him different examples, ask him to summarize it into a document or skill that you can transfer between different chats. The second thing is that you need to understand the process that produces results, or start with someone else's process until you develop your own. So, for a role like a content strategist, it's not as simple as just telling the AI posts, because that's not the only thing a content strategist does.
31:33And if you don't know this, then you won't be able to implement it in your business. Asking him to research the best ideas is something a average person would do. So if you ask Claude to do this, even though he has genius-level AGI intelligence, simply because you're asking him to do something that sub-average people do , the results you're going to get are going to be sub-average because he's not going to tell you, "Hey , do this better." And an even lower- level person would simply ask AI to write a viral post for them. So when I think about my content writing process, it looks like this, and that's what I would try to turn into an AI workflow.
32:08First, I read books, save ideas, chat with friends, and listen to long lectures in search of ideas that no one is talking about on the internet. Then I choose an idea at the intersection of: first, something I'm interested in writing about, second, something that will enormous benefits to people, and third, something that has already been tested and works well. The third thing is that I write down ideas for headlines for an article or script, and then I find the best ideas that would fit that material. Fourth: I consider what long-read structure would best suit these ideas and begin to make a plan. And fifth: based on this plan, I come up with 10-20 ideas for short content that I can take and use immediately.
32:43And this is also taking into account the fact that I have already mastered branding, which is a completely different process and story, because this is what would allow the content strategist to understand my worldview, my influence, my mission, and my "brand enemy." So the hardest part is turning all of this into a process where AI can help, because of course it can't read books or discover ideas for me. This is the advantage that cannot be outsourced. This is what makes your process a little more indispensable than someone else's. But I can collect all my ideas in my Eden library, which automatically tags and links them so that the AI can take them as a base of ideas each week and suggest to me a few options that are most likely
33:22to be successful. That is, I can focus more on reading and studying, instead of delegating it, because I enjoy the process. Next, the main thing I need to do is teach artificial intelligence to do quality research and understand which ideas are worth paying attention to. For example, celebrity posts are not an indicator of what actually works on social media. Instead, the indicator is the successful publications of small and medium-sized accounts. Then—how to relevant topics with my own ideas and examples of content structures that have worked before. After that, you get the result in the form of a table that AI can update and supplement every week.
33:58Now the third: you need to constantly improve and adapt the system you have created, as you always have done. This is more of a life or reality tip than an AI tip, but we're building a system here. Life is about systems. Your mind is a set of nested systems. The road is a system. Everything is a system. So, we build a system, and if you've ever templates in Notion or had a morning routine, you understand that sometimes the system breaks down, you outgrow it, or something goes wrong , and it becomes unusable in a particular situation. But this is not a reason to believe that the system does not work at all, and it does not mean that it should be abandoned completely.
34:36This means you have to improve it and make it adaptive. In the case of AI, this means that if you notice something wrong, strange, or low-quality, you should point out the error. You the system. You provide suggestions. You ask more questions. You start moving in a different direction until you get to where you wanted to be. Personally, I think most people are simply afraid of the uncertainty that artificial intelligence brings with it . It's very similar to entrepreneurship. This is a direct interaction with the trial and error process. Just because they don't get the quick gratification they get everywhere else, they call it bad or ineffective, or say, "Oh, you're just delegating your thinking," when it's the opposite, and they're just repeating it because it's a cliché that's been drilled
35:16into their heads. So who here is more like a car? It seems like everything we've talked about just doesn't fit in their heads. And that's okay, because for people like that—people for whom none of this has worked—it's worth stepping out into the arena yourself and stopping trying to until you understand what exactly you're delegating. You can't content creation until you start creating it yourself. You outsource something that doesn't exist. You can't automate this. Most people processes by loading skills before they have anything worth automating. You don't even have anything that works yet. But those who have understood it revolutionary it is.
35:51So, I hope this was helpful. I hope the custom AIs linked in the description will be useful to you . And if you want to create your own custom AI or add a course to it and sell it on Eden, that will be available very soon, and I'm excited about it. So, we'll hear from you soon. Like, subscribe. Well, you know, it's just one click away, and I'll see you in the next video.
00:00这句话在我脑子里萦绕了整整一年。几年前,我刚做内容的时候,只是想避开冷发邮件找客户这件事,我注意到有一些人、一些创作者,他们一个一个地把生意做起来,收入从一百万美元到一千万美元不等。
00:35而且他们并没有整天工作。我觉得 Justin Welsh 就是最完美的例子。他非常看重自己的时间,只在早上工作一小会儿,但他拥有这套生意的各个要素——我们后面会讲——正是这些要素让他能赚到那种量级的钱。多亏了互联网、社交媒体和 AI,这一切才成为可能。人们至今还没意识到,作为个体,自己手上握着多少杠杆和力量。我们早就告别了按时间付费的模式。我们转向了一种为智慧和价值付费、并且与时间脱钩的模式。这一点很多人难以理解,我换个更简单的说法:你可以只发一条帖子,如果它拿到一百万次浏览,其中 1% 的人花 100 美元买你的产品——利润你自己算。
01:23当然,你第一条帖子不太可能就有一百万浏览,但你懂我的意思。这是个练习的问题。它是一种可以练出来的技能,而且正是眼下公司、创作者以及他们身边所有人都在找的最值钱的能力之一。所以这个视频对你可能非常重要。那些把一人公司做到上千万收入的创作者,靠的是卖信息产品,或者把社交媒体当作分发渠道。没有付费广告,没有播客赞助、广告牌、广播或电视——只有社交媒体,因为观众的注意力就在那里。
01:59建立受众是免费的。内容也是免费的。对新手来说这是低风险的商业模式,因为比如做实体产品,你需要先把资金投进产品里。如今在社交媒体上启动一人公司几乎没有成本,除了你花在学习上的时间。但从我开始做这些生意到现在,一切都变了很多。那时候,你还能在大多数人知道之前就把生意做到几百万美元,而现在他们对 AI 带来的新机会知道得更少。如今 GPT-6 Astra、Fable 5.1,以及接下来还会出现的任何模型,都让这件事更真实。而你在这里的竞争优势就是:普通人根本什么都没注意到。
02:46他们什么都不做,也不打算改变自己的习惯。但这件事美妙、非常美妙的地方在于:你并不需要赚到一百万。你不需要赚一千万美元。看这个视频的大部分人,只是想把现在的收入来源替换掉。可能是每年 7.5 万美元,或者每年 10 万美元。我想告诉你的是,这件事简单得难以置信。这不代表它容易,也不代表很快就能实现。我只是说它完全做得到。对我来说,刚开始时的目标就是每天午饭吃得起 Chipotle。我很自豪自己挺快就做到了,现在我不是每天午饭都吃 Chipotle,虽然今天大概会吃。
03:25但我从小就只是不想走那条标准路径。小时候我特别会观察身边的人,我看到的是一条毫无意义的路:上学、拿学历、花一大笔钱去接受那些十年后完全用不上的培训。然后找一份你依赖它、它也占满你全部时间的工作,让你没空自我教育、没法换方向,再和一个你瞧不起的人结婚,拿到承诺的养老金,却永远没法真正停止工作。这一切都被当作正常,让我觉得恶心。所以我意识到,创业是唯一的出路。这是我的路,不管我喜不喜欢,因为它是一件巨大而神秘的事。
04:05开自己的生意,是我当时根本没能力做到的事。于是我就去做了。我一次又一次失败。我在 YouTube 上开了个健身频道。我试过用 dropshipping 卖钱包。我试过卖防蓝光眼镜。我做过自由职业的 SEO、Facebook 广告和网页设计,直到学生贷款和信用卡债越堆越多,我不得不去上班。拿到那份工作后,我把它当成一个截止日期。我知道如果我不能很快让生意跑起来,就会被账单和责任淹死,再也爬不出来。大约 7 年前,我得以辞掉那份全职工作,去做自由职业的网页设计,因为那是我掌握的技能之一。
04:44大约 6 年前,我开始运营社交媒体,从数字产品赚了大约 1 万美元。第二年变成 10 万,然后是 80 万,我有一个视频讲我是怎么做到的——每天写作 2 小时。再下一年大约 400 万,我的税单正式是我之前工资的四倍。后来,我做了那个“愚蠢”但很有远见的决定——转去做软件,我的收入稳定在每年 100 到 150 万,因为我必须把时间和注意力转移到软件上,并在那里投入大量资金,因为当时我们还在打基础。而且到现在我一分钱都没从 Cortex 或像 Eden 这样的新项目里给自己发过工资,但我们正在往那一步走。
05:25它正在非常快地起势。接下来的更新你们不会相信。所以我想说:关于一人公司,我有很多可以分享。我能从鸟瞰视角看这件事,因为我的受众和客户都是一人公司;我看得到什么有效、什么无效,而在 AI 方面,我算是坐在前排,看它如何帮你更快地把这件事搭起来。那我们就开始吧。我们会讲为什么你的兴趣或独特价值,加上社交媒体或分发,再加上 AI 或低成本的执行力放大,等于一人公司。作为一人公司,你要在社交媒体上发内容。
06:06唉,我知道早上写 30 分钟,然后发到全世界都能看到的地方有多难。但就像我说的,注意力现在就在社交媒体上,而你大概也没本钱投付费广告,或者花 2500 美元请创作者发一条关于你产品的 reels。这种 UGC 内容的价格正在……我个人认为,作为创始人,更好的方式是雇人,或者自己建受众。这个视频里我也假设你不想做服务型生意或客户驱动型生意,比如自由职业或代运营,因为你想要一些自由和生活方式。
06:42即使你做的是这种生意,我们讲的有些东西也适用于你,因为你无论如何都该建受众,但我更喜欢写作、做能在睡觉时也卖出去的产品。现在我们来讲一人公司长什么样。它由什么组成,你需要哪些要素?然后我会给你几个可以立刻用起来的 AI 员工,链接在简介里。你可以直接开始对话,基本上就是靠我给你的这六个员工来运营你的生意。你可以在 Eden 里一键完成,它们都会在一个地方;或者下载单独的 skill 文件,上传到 Claude。
07:20或者把它们加进 Eden,再通过 MCP 把 Eden 连到 Claude,这样 Claude 就能看到社交媒体帖子、研究资料,以及这些专门的 AI 或员工。所以,创建一人公司时我们要定义的第一件事是品牌,也就是你或你的生意如何向世界呈现自己。在个人生活里,当别人听到你的名字,他们会想到什么?他们会把你和什么联系在一起?和你聊天舒服吗?你喜欢什么、不喜欢什么?你有趣吗?你机智吗?你会帮别人吗?我能信任你吗?你是个有价值的人吗?人们会来问你什么?诸如此类。
07:56品牌是同一回事,只不过是给你的生意的。它就像数字店面,或者数字身份。正是它让人愿意买你的 App、课程、教练服务或网店里的产品,而不是别人的——哪怕别人的质量更差,因为外面有无数出色的 App 或课程,你甚至都没听说过,因为你不知道它们的存在。而如果你知道它们,如果你喜欢、并且过去帮过你的一个内容创作者提供了类似的产品,你花了那么多时间在他的内容上,你就会买那个你喜欢的创作者的产品。这就是为什么懂营销、或者真正研究数字营销的人,能比只会做点小作品集 App、却不知道怎么把它们卖出去的程序员多赚 10 倍。
08:40在一人公司里,你的品牌就是个人品牌。即使你不是一人公司,看看 Elon Musk。看看 MrBeast。看看那些名人。他们都在建个人品牌,因为这就是当今世界必须做的事。个人品牌是业务增长策略之一,因为它是免费的。而且如果做对了,你几乎不用自己去做营销。所以,它是你的价值观、你的世界观、你的故事,向外具象化。我下面给你的其中一个 AI 员工会采访你,收集关于你的信息,生成一份完整的个人品牌策略文档,你可以再把它交给任何其他 AI。
09:21也就是说,在任何时刻、任何对话里,你都有这份文档,可以带到任何地方。这是必要的,因为它给你愿景、反向愿景,以及你内容的支柱。所以,所有内容创意的生成,如果你想让它们有一条统一的线和影响力,至少应该基于这份个人品牌策略文档。作为个体经营者,你需要的第二样东西是报价,也就是人们付钱让你去达成结果。所以,你需要一个产品或服务。这看起来显而易见。没有它,就不可能赚钱。你可以选择建受众这条路,先把它做大,然后接广告合作,那是很赚钱的生意。
09:59但今天,拥有自己的产品或服务才是合理的。报价与产品或服务的区别在于:报价是那个产品或服务的感知价值。是你怎么包装它。是你做的一笔交易。这是你能掌握的最赚钱的技能之一——做报价,因为一旦报价打准了,落地页、营销等等都会容易得多,因为一切都已经就位了。而这才是你竞争的地方。你不是在产品或服务层面竞争,你是在报价层面竞争。也就是价值与价格之比。
10:31所以,如果你拿来一个工具类 App 或者像 YouTube 这样大家都在用的工具,和 Eden 对比——Eden 能通过 MCP、通过聊天研究所有平台,把你所有的想法、笔记、链接、帖子,以及很快还有产品,都放在一个地方,那 Eden 显然是更好的交易——请允许我稍微自夸一下——而且它有发帖排期功能,所以它替代了另一个 App。这就是一个内容创作者的完整工作流。但做报价,是控制四个杠杆的艺术。第一个是渴望的结果,也就是客户真正想要什么。第二个是交付机制,也就是你如何帮他们达成。
11:06这是一门课程吗?这是教练服务吗?这是补充剂吗?这些东西的独特机制是什么?你帮他们达成想要的结果的独特方式是什么?第三个是结果,也就是多快能拿到。第四个是保证,意思是如果什么都不成,会发生什么。这就是你需要的四样东西。把它们写下来。再说一遍,AI 会帮你做出你的报价。一人公司需要的第三样东西是分发,也就是如何传播你的品牌或报价。我很喜欢 Jack Butcher 的这句话:先建立分发,然后再建你想建的任何东西。因为我们生活在互联网时代,在这里,注意力就是杠杆。
11:41你可以有最好的品牌。你可以有最好的产品,但如果没人知道你,这些都不重要。所以我们关心的分发包含一切:从营销和销售,到内容、广告、落地页和网站。这就是人们找到你、了解你、付钱给你的方式。这就是漏斗。比如,你大概是通过社交媒体上的内容找到我的。你现在就在 YouTube 上看这个,如果我把知识传递得足够有价值,你就会开始足够信任我,去使用或为我的某个产品付费。不是因为我在骗你,而是因为你自己看到它对你有利。
12:16这是一种按条件进行的价值交换。我也可以投付费广告或赞助别人,这两样我都试过,以后大概还会再做。但作为资金很少的新手,你该把这当作一块垫脚石。你不会为一个还没验证过、也卖不动的报价去投付费广告。那是在浪费钱。是的,你可以通过广告来测试,但那是自费测试。在社交媒体上,你免费测试你的想法、产品和服务。如果你没有增长、也卖不出任何东西,那说明还有改进空间。这就是试错法。你只需要去试,如果不行,再来一次。
12:52如果还不行,就学点新东西,再试一次。最终,你工具箱里的工具会多到你成为“生意”这件事的大师,能真的靠点一下按钮就赚钱,因为你知道自己在做什么。在实践中,你掌握的不只是一项技能,还有人的心理、注意力的机制、文案、内容和说服。所有这些不同的东西汇聚在一起,把你变成一个生意人。所以,我们接下来要讲这些 AI 员工,或者说这支 AI 团队,或者说,说白了,就是 skills 或提示词——本质是同一件事:给 AI 的上下文。
13:24它们会在这整件事上帮你,但只有你真正去做、去发布,才能从市场得到反馈。你不该因为 AI 能替你做,就把自己的智力外包出去,因为学习的意义不在这里。你不是靠看教程学会东西的;你是把学到的知识用到你真正在做的事情上,才学会的。如果它成了,你就学到了。如果没成,你也学到了东西。但如果你不给它一次成或不成机会,那你就什么都没学到。而且即使第一次没成,你现在也知道 AI 不会一夜之间取代一切。
13:58你明白现实是存在的,得到结果不总是线性的、一步步的过程,它会随着文化和对价值认知的变化而演化。那我们就开始把它简化。以下是你真正开始这门生意要做的事。这就是你的日常流程。第一件事,每天在你选的短内容平台发 1 到 3 条帖子。可以是 Substack、X、LinkedIn、Instagram 等等。你只需要选你最喜欢的那一个。在 Eden 频道上,我做了关于如何在这些平台上做内容、写内容的教程。那里已经有不少视频了,你去看看,因为如果我在这个视频里讲,它得长到 10 个小时。
14:41但从一个平台开始,成为真正的专家。我是从 Twitter 起步的,但如果我今天才起步,我大概会选 Substack,原因很简单:它有长文、有短笔记、还有增长引擎——全在一个地方,而且纯文本。所以我不必出镜。我可以坐下来喝杯咖啡,早上写下我喜欢的想法,然后试着在这个平台上发展。而且你在 Substack 上的订阅者不只是粉丝,他们是 newsletter 订户。所以你拥有这个名单。你可以把它导出,带到任何地方,这在任何其他平台上都做不到;而且我猜看这个视频的大多数人都有很多兴趣、很深的兴趣,
15:18你喜欢深度思考。Substack 就是做这件事最好的地方。我也看到很多刚起步的创作者正在起势。第二件事,每周在 Substack、X 或 YouTube 上发一篇长文,但最可能是 Substack,因为那里大概是你起步的地方,不过在 X 上也可以,因为他们现在有 articles。所有平台都在往短内容加长内容的组合走,因为这是创作者想要的。创作者明白,人们在你内容上花的时间越多,或者说我称之为的“注意力时长”——就像肌肉增长的张力时间一样,社交媒体的张力时间就是社交媒体增长的时间。
15:55看了一个小时或半小时 YouTube 视频的人,会比只看了 15 秒短片的人信任你得多。他们得看几百条这样的短片,才等于消耗了一个 30 到 60 分钟 YouTube 视频的内容量。书面文章也一样,比如 Substack 上的,或 X 上的 articles。所以,让它变得可行、现实的办法,就是每天早上留出 30 分钟。这样你就有 30 分钟写短内容,也就意味着不能对着空白屏幕发呆。这意味着用我给你的 AI 做研究、定下一个想法,然后把它写成一条有趣的短帖。
16:33大多数人都卡在前两个阶段。他们不研究、不学习、不读书、不找想法,他们不把这件事当成写作过程中必须在同一时间段发生的一部分。第二,他们定不下一个想法。他们觉得想法必须在开始写之前就完美。可实际上,想法是在写作过程中被磨出来的。所以就随便抓一个想法,写下来。接下来的 30 分钟用来写,也许写 250 字的长文。这可以是 YouTube 脚本、X 上的文章,或者 Substack 的文章。我建议每周开始时,你大概列出这周要写什么的计划。
17:07你会选一个主题,在其中一个 AI 的帮助下开始构思计划并搭结构。第三件事,试着每天推广一次你的报价。你可以在长文的中间某个位置推广它。你只要说:“嘿,如果你在 XYZ 上卡住了,试试这个。”然后加上你产品或服务的链接。结尾也做一次。就这么简单。它很短。如果你不知道怎么弄,把你的邮件列表交给 AI,说:“嘿,我想把这个放到中间和结尾。怎么做得有品味一点?”当然,既然你大概一周只发一次长文,你就会用不同的方式每天推广它,越快越好。
17:46它必须成为你日常的一部分。你可以在 Substack 或 X 上某条帖子的回复里加上,用简单的 PAS 结构推广你的产品或服务,PAS 就是“问题、放大、解决”。问题是什么?如果你在这、这、这上面卡住了,这是解决方案,附上产品链接。在 Instagram stories 里做。把它做成你 Instagram 帖子里的行动号召。嘿,如果你在这、这、这上面卡住了,又想要它,就在评论里打一个关键词。这就是整个流程。你靠写内容来建立分发。你在推广你的报价。如果你还没做出报价,那你也该同时专注这件事。
18:21这正是你可以在 Eden 里创建的东西,只要把你脑子里的知识变成一个 AI 产品,它同时也是一门课。等它真正上线,这一点会更清楚。所以留意这个,因为仅仅因为它在 Eden 上,你把产品挂在 Eden 上就能比在任何其他平台赚更多钱。在整件事里,有几件事我们可以做,或者可以交给 AI 做。第一是品牌策略师的工作。也就是制定个人品牌策略,让你不会淹没在人群里,以及提升你内容的效果,给你增长战术。
18:53它像是元层面,其他一切都从属于它。第二是内容策略师,他研究社交媒体上什么真正有效,理解你过去所有的内容,把它连接到你的品牌策略上,给出这一周的帖子想法,还能生成内容草稿供你编辑,如果你愿意的话。现在,如果你不是活跃的 AI 用户,有件事你可能没意识到,我告诉你:你完全可以复制粘贴。你只要把 YouTube 上的这个字幕抓下来。eden.so/tools 上有一个 YouTube 字幕生成器,或者你直接把它丢进 Eden,说:“嘿,帮我落地这封 newsletter。
19:28帮我梳理一下,我们怎么搭一个工作流来完成所有这些任务。”然后你就和它沟通,直到拿到想要的结果。我们会在视频最后一部分讲怎么做对,因为你需要理解怎么用 AI。第三,我们可以把它当作长文专家或规划师,它擅长长文结构和抓注意力的策略,帮你就你的想法搭出一篇 newsletter 文章或视频脚本的结构。第四,可以把它当作短内容专家,它接过你的草稿想法,变成帖子,你可以审阅、编辑,并排期发布到所有平台。
20:02第五,是报价顾问,帮你把你的经验、兴趣和专长变成一份报价,并决定先从哪个产品或服务开始。第六个是策略师。它接过你的报价、品牌策略和客户画像,然后帮你生成推广帖、邮件 newsletter 和落地页文案。这些我都已经为你做好了,我快速给你看看它长什么样。但如果你真的想用,链接在简介里,或者去 eden.so/kits/1-human-business。你只需要点“set up my business”按钮,它就会把它们加进你的工作区,或者你也可以把它们作为 skill 文件单独上传到 Claude 或 ChatGPT。
20:44那些等 Two-Hour Writer 更新的人,它正在更新,我们还在打磨,但它会是这个样子。所有模块都会在这里。你可以走完课程,并在过程中和它沟通,这样它就能以整门课和模块作为上下文来帮你落地。这就是你能在 Eden 里卖的东西,作为一个产品,而不只是一门普通的 AI 课或 skill。你卖的是整个套餐,而且它会署你自己的品牌。它们在你在工作区里就是这个样子。等它们都加上后,你可以在这里切换,每个 AI 的所有对话都会存在一个地方。
21:23所以它就像你所有 AI 助手的指挥中心,为一人生意服务。我最喜欢的一个是长文工具。所以,当你想在 Substack、X 上写一篇爆款文章或一封 newsletter 时,你只要来这里,点一个起始提示词,比如“做一个邮件计划”或者“做一个脚本计划”,给 YouTube 用,然后 AI 会问你关于这个想法的问题,因为它知道怎么写。而因为这是在 Eden 里发生的,或者如果你把 Claude 连到 Eden 的 MCP,它就能研究 Substack、X articles、YouTube 脚本上任何爆火的东西,等等。它能从
22:01真实素材的标题和结构里提炼模式和原则,帮你塑造自己的想法,让它有效。然后它给你一个大纲,你可以转到文档里,在旁边写,因为我个人喜欢让 AI 帮我搭结构、做研究,这样我对写作过程本身有更多掌控。本身。我这里也有我的小写作工作区。等到要发布文章到 Substack 或任何社交平台时,我就去发布标签页,直接在那里全部搞定。我花了很多时间做它们。它们可以免费使用。当然,如果你想用 AI,你得为 Claude、ChatGPT 或 Eden 付费,但这就是游戏规则。
22:44这些 AI 里的实际上下文和知识是免费给你的,而且它们都是围绕“什么有效”来的,因为如果你只是在 Claude 的普通聊天里问,它不知道该怎么做,会给你糟糕的信息。你必须根据真正有效的东西去塑造和调整上下文。我建议按我上面列的顺序来用它们。品牌策略,然后内容策略,再基于它们,去写你的短内容和长内容。一旦你有了报价,你就可以用策略师工具。等你这套流程跑得够多了,你就可以把它们当成例行任务来跑。比如在内容策略师里,你可以说:“嘿,每周一帮我做这件事,加到一张表格里,这样所有想法都在那儿,我可以直接用。”现在你会注意到,它们当中没有一个,
23:29至少我创建的那些,会替你写内容。它不会自动生成内容,因为我认为对新手来说这不合适。而且我个人觉得,我们还没到值得把这一块自动化的程度,虽然我关于用 AI 写文本的看法正在有些改变,因为最终选择权在你手里,这个我们很快就会讲到。但如果你想让 AI 替你做、替你写,我并不完全反对,但你必须理解我们接下来要讲的东西。所以现在我想讲一讲,在整件事里,总体上使用 AI 最好的方式。
24:03因为社交媒体上的信息太多了,你会被那些 Instagram reels 淹死,它们告诉你去下载某个工具,插上它就能替你做所有事,这很棒。但如果你不知道该怎么正确思考使用 AI,或者该怎么入手,你就会像下载了一个 Notion 模板一样跑这个工具,拿到结果,然后说:“哦,这挺酷的。”然后你就把它扔了。或者它给你的东西很糟,你就想:“哦,这不管用。”或者说“AI 很烂”“AI 不行”。或者说“AI 就是瞎编、幻觉,所以我不信它”,尽管事实并不是这样。
24:35人一样蠢,幻觉一点不少。你上社交媒体看看,99% 的人都在传播错误信息。从来没有哪个时候,你不需要自己批判性思考。不会因为你用了 AI,这条规则就失效了。这会让很多人望而却步,因为这是个复杂的过程,直到你真正上手去实践,它才说得通。理解和实施它,需要一定程度的系统思维。但如果你看到最后并试着理解,那么下次你用 AI 时,你就会开始注意到新的机会,并把它们落地。
25:10人工智能的问题在于,人们把它当成某种全能的存在。而既然它不是全能的存在——即使我们现在实际上已经有了 AGI,像 GPT-6、Astra 和 Fable 5.1 这样的模型,在没有合适上下文的情况下,依然没法把任务做好。但因为它不是全能的存在,当它没有完全做到一个人想要的——因为那个人自己也不知道自己想要什么——人们就简单地认定 AI 只能生成垃圾。一个很好的例子:我最近看到一个视频,有个人说:“嘿,如果你是小企业主,别用 ChatGPT 建网站,因为我们都看得出来是 AI 做的。”
25:49但如果你这样用 AI,那就很可笑了。这是真的。如果你只是对 AI 说:“嘿,给这个、那个做个网站”,它看起来就会像任何其他 AI 做的网站。但如果你提供示例和清晰的指令,至少你会有一个更好的起点,但这还不是全貌。之后,你需要开始抠细节。你得去定制圆角、颜色、阴影、布局,以及一切应该长什么样。这时候,当你有了一个原型、模板或初稿,你就做任何设计师在 Photoshop 或 Figma 里会做的事。你逐项做那些细小的调整,让它独特,而事实上,AI 可以做小的改动。
26:33你不必第一次就把所有事都做对。或者,当你剪视频时,是的,你可以让它过一遍、全剪了、加上动画,但那只是个大致轮廓。然后你看结果,心想:“哦,这太糟了。”然后你才真正开始用它。当你进来说:“嘿,这里做好一点。这里这样。这里这样。这里这样。”那才是真正的剪辑开始的时候。它第一次大概不会把所有事都做得完美,因为你没有跟它讲清楚你到底要什么。这没关系,因为你大概也不会让它写一千字的段落。你会让它做一个大致轮廓,这就够了。
27:08然后你自己把一切捋顺,补上小细节。所以今天让 AI 做一个网站、写一封 newsletter 或一条推文,就相当于当年下载一个 WordPress 模板。人们看得出来,因为那是低质量的工作。几年前,垃圾一点都不比今天少。但 AI 是个特例,人们就是特别不喜欢。你可以就停在那儿,拥有一个完全体现不出你品牌的糟糕网站,或者把它当作改进的起点。就像你做一个 App,然后一直修 bug,直到它真的很好。现在,什么算“WordPress 模板”、什么算垃圾的门槛会提高,这会逼着人们去培养品味和专业度,如果他们真的想脱颖而出。
27:55这也是我看待用 AI 写作这件事正在改变的原因之一。因为如果你用自己坐下来写作时的同样方式去对待它,你就会意识到,用 AI 你有更多创造力和可能性。但人们通常只是说:“嘿,帮我写这封 newsletter。”结果很糟时,他们就认为它是垃圾。或者他们觉得还行就发布了,因为他们没有品味或经验,而品味是那种你还没准备好向世界发布的东西。所以他们一发布,人们立刻就看出来是 AI。但对那些真正把 AI 写作当作手艺的人来说,没人会看出来,因为这不是基础的 WordPress 模板。
28:34你选择发布它,是因为它反映了你自己会说的话。它反映了你的价值观、你的世界观、你的观点。另外,作者也并不总是真实的,不总是用和朋友说话的那个声音。不,他们创造了自己的作者声音,还稍微伪装了一下。他们戴着面具,这很正常。这是做人的一部分。那么,既然我们搞清楚了这一点,我们实际该怎么做?你怎么做出像我给你看的那个每周内容策略师这样的东西,让它真的很好用?首先,你需要知道一个产品长什么样,或者有一个“什么算好”的范例。
29:07因为说到高质量的内容或社交媒体帖子,有些因素在起作用。这是帖子、文章或 newsletter 的结构。这是作者的 vibe,或者说他的能量签名。他们的品牌有一条一致的线,还有他们用来传达世界观的宏大想法。还有他们遇到的独特想法,帮他们从其他人中脱颖而出。所以,所有这些都要考虑进去。没有任何人工智能会告诉你这些。没有任何人工智能会默认做到这一点。你只能通过导师、经验,或者向拥有这种具体知识的人学习来找到它,因为做内容没有唯一正确的方法。
29:42但现在你知道了,你需要一份指南、一个范例,或者你可以给 AI 的指令,来限制上下文。这意味着给它提供能体现你想做的那类内容的 YouTube 标题和视频字幕。Reels 字幕、Substack 里的笔记或文章、X 上的帖子或文章,或者很多你喜欢的写作风格的范例。我现在大多数和 AI 的对话都是这样开头的。如果我知道我想要什么,我就给详细指令。我以前会写 2000 词的提示词,教它像我一样写推文,结果它写得非常接近。但如果你只是让它看看我最好的推文,你做不到这一点,因为它会试图把它们模式化,而不考虑写它们时的思考过程。
30:24更重要的,如果我不知道我想要什么,我就开始研究、提问,直到我搞清楚,并用信息填满上下文框。所以,如果我设计东西,我需要一个参考。我需要去 Mobbin 或其他网站找灵感。如果我写内容,我得去 Eden 的“Discover”标签,找最有效的想法。这意味着我不只是说:“嘿,给我写一封关于这个主题的 newsletter。”我会提问,来决定写哪一封 newsletter。所以在你这里,你可以问:“我想写世界上最好的 newsletter 之一,一个能脱颖而出、听起来不像机器人的东西。”
30:59我们来头脑风暴,直到找到一个我喜欢的风格。先从提问开始:你需要我提供什么,才能把它做出来。我们会从这里往下走。然后,在这段对话结尾、并给了它各种范例之后,让它总结成一份文档或 skill,你可以带到不同对话之间。第二件事,你需要理解能产出结果的流程,或者先从别人的流程开始,直到发展出你自己的。所以对内容策略师这样的角色来说,并不是简单告诉 AI“发帖”,因为那不是内容策略师做的唯一的事。
31:33如果你不知道这一点,你就没法把它落地到你的生意里。让它去研究最好的想法,是普通水平的人会做的事。所以如果你让 Claude 做这个,即使它有天才级的 AGI 智能,仅仅因为你让它做的是低于普通人水平的事,你得到的结果也会低于普通水平,因为它不会告诉你:“嘿,这样做更好。”而水平更低的人,就干脆让 AI 替他们写一条爆款帖。所以当我思考自己的内容写作流程时,它是这样的,而这正是我会试着转化成 AI 工作流的东西。
32:08第一,我读书、存想法、和朋友聊天、听长讲座,去找互联网上没人在谈的想法。然后我选一个处在交叉点上的想法:第一,我有兴趣写的;第二,能给人们带来巨大好处的;第三,已经被验证过、效果很好的。第三,我写下一篇文章或脚本的标题想法,然后找出最适合那份材料的最佳想法。第四:我考虑哪种长文结构最适合这些想法,然后开始做计划。第五:基于这个计划,我想出 10 到 20 个短内容想法,可以立刻拿去用。
32:43而这也是在考虑到我已经掌握了品牌的前提下的,那是完全不同的过程和故事,因为正是它让内容策略师能理解我的世界观、我的影响力、我的使命,以及我的“品牌敌人”。所以最难的部分,是把这一切变成一个 AI 能帮上忙的流程,因为当然,它没法替我读书或发现想法。这是无法外包的优势。这是让你的流程比别人更不可替代的东西。但我可以把所有想法收集到我的 Eden 库里,它会自动打标签、做关联,这样 AI 每周就能把它们当作想法库,给我建议几个最有可能
33:22成功的选项。也就是说,我可以更专注于阅读和学习,而不是把它外包出去,因为我享受这个过程。接下来,我需要做的主要事情,是教人工智能做高质量的研究,并理解哪些想法值得关注。比如,名人的帖子并不能说明社交媒体上什么真正有效。相反,指标是中小账号的成功帖子。然后——如何把相关主题和我自己的想法、以及过去有效的内容结构范例结合起来。之后,你得到的结果是一张表,AI 每周可以更新和补充。
33:58第三点:你需要不断改进和调整你创建的系统,就像你一直做的那样。这更像是生活或现实的建议,而不是 AI 建议,但我们在这里是在搭一个系统。生活就是系统。你的心智是一组嵌套的系统。道路是一个系统。一切都是系统。所以,我们搭一个系统,而如果你用过 Notion 模板或者有过晨间习惯,你就会明白,有时候系统会崩、你会超出它、或者哪里出问题,让它在特定情况下没法用。但这不是相信系统完全不管用的理由,也不意味着应该彻底放弃它。
34:36这意味着你得改进它,让它有适应性。在 AI 的情况下,这意味着如果你注意到有什么不对、奇怪或低质量,你就该指出错误。你去调整系统。你提供建议。你问更多问题。你开始往另一个方向走,直到到达你想去的地方。我个人认为,大多数人只是害怕人工智能带来的不确定性。这和创业非常像。这是与试错过程的直接互动。仅仅因为他们得不到在别处随处可得的那种即时满足,他们就说它不好、没用,或者说“哦,你只是在把思考外包”,而事实恰恰相反,他们只是重复这句话,因为这是被灌进
35:16他们脑子里的陈词滥调。所以这里谁更像一辆车?看起来我们讲的一切就是塞不进他们脑子里。这没关系,因为对那样的人——对这些人来说这些都不管用——值得你自己走进竞技场,停止尝试,直到你理解你到底在把什么外包出去。你不自己开始创作内容,就没法做内容创作。你在外包一个根本不存在的东西。这个你没法自动化。大多数人还没任何值得自动化的东西,就靠加载 skill 来跑流程。你甚至还没有任何能跑通的东西。但那些理解了的人,就知道它有多革命性。
35:51所以,希望这有帮助。希望简介里链接的那些自定义 AI 对你有用。如果你想做自己的自定义 AI,或者给它加一门课并在 Eden 上卖,那很快就能用,我很期待。那我们很快再见。点赞、订阅。嗯,你知道,就一键之遥,下个视频见。