Transcription
Prompts no longer work. I just deleted about 2,000 prompts that I had been collecting for the last 3 years. I meticulously selected each one, checked its functionality, and thought I would use them my whole life. Perhaps this is one of my best decisions. And now you will understand why. In this video, I will demonstrate a new revolutionary technology of smart projects, which today makes 99% of all existing prompts, sold on the internet, or presented as secrets, useless and unnecessary. And today, in just 10 minutes, you might do exactly as I did. Because I will share with you a method that allows you to work with language models as equals. You will work with them like an expert. You will not spend 3-4 hours researching to find information. This is already built-in. This is already in billions of parameters of neural networks. You just need to use it correctly. And for this, you don't need to write such prompt posts that take you several hours, which could be spent on completing the task. Everything is done as simply, quickly, and easily as possible. And now I will demonstrate all of this to you. And now, look at how often this happens. A person turns on ChatGPT, starts a new dialogue, a new chat, and writes a simple request: "Write me a post about fitness." And here, yes, since I have a large number of chats and ChatGPT interacts between them more or less, it knows that it needs to write in a lively, motivating, and slightly personal style, yes, but it still starts to produce something that you wouldn't publish anywhere. If you publish it in a thematic channel or blog, you won't get reactions, you'll get a lot of unfollows, because the first thing it starts with is: "Fitness is not about the body, it's about choice." Well, that's maximally outdated information. Sound familiar? You write a simple request, and it gives you complete nonsense. As if someone who doesn't understand the topic at all decided to put words together and give you a result. Well, you can't post this anywhere, you can't publish this anywhere, and refining it can take more time than writing it from scratch. Well, maybe it's because of such a simple request. Let's complicate this request and write it according to the rules. You have selected, like me, your prompts, collected a huge database of these prompts, and write a more sensible request. You are an experienced copywriter with ten years of experience. You give it context, understanding for whom the post is. And here it is, it adapts it more or less, when you look in the mirror and don't recognize yourself. After childbirth, the body changes. Okay, a cool fact. Agree, it's already better, but it's still some kind of soulless writing that needs to be reworked, improved, corrected, changed. In general, again, this takes a lot of time. It's important to understand that no matter what prompt request we have, it still limits the language model and doesn't let it break free to all its parameters that are already loaded, that it already has. And for this, there is another way. But first, I will show you the final result of prompt requests. Well, let's say you found a super prompt request and write: "You are a top copywriter from an agency with a turnover of 50 million rubles. Your specialization is fitness for women. The audience is this, income is that, so more details. The more details, the better in one sense, but in another sense, we are simply confining our chat within limits. And here, of course, we get an even better result. But how many such iterations do we need to make? How many limits do we need to set? I looked at myself in the mirror and thought: "This is temporary." That is, it's already in the first person, storytelling, AIDA, but a year passed after childbirth, and my stomach still looked pregnant. My arms are heavy, and my chest... ellipses. Well, as I say, it interacts well in chats. It understands more or less. And I have many created elements in projects where it takes this data more or less. But if you start this chat from scratch, you will get a completely different variation, which will be comparable to this first iteration. Well, and here, Marina, everything is the same for me. I want to lose weight, but I have no strength at all. So I spent another 20 minutes refining the so-called ideal prompt request, and the result is still like a carbon copy. I read it, I understand, yes, technically everything is correct and within the limits that we specified in our prompt request. But where are the live emotions? Where is the engaging delivery? Where is that magic that makes you read to the end? And now imagine, you go into a neural network, write one line, not even a sentence, literally one line, and immediately get a result of the highest level, like an employee with a salary of 200-300,000 rubles per month. And at the same time, you get such a result without long prompts, without settings, without dancing with tambourines, without watching a bunch of videos, without taking a bunch of courses, you get a perfect result with a simple task. In GPT, there is this menu on the left, "New Project." Here you can create a project called "Content Machine." That is, I have an instruction that you will then take for yourself, and it is already developed, and it is called "Content Machine." It contains all the methodologies, which it determines independently depending on the niche, and it determines the niche from your request. That is, you need to just transfer this instruction here. And I will copy it now, open it, and paste it here. This is how it looks. Look. Here are additional fields. Well, in general, there is a lot of data here, and the data that it determines and determines where you form it, how you form it. But the main essence here is that we do not confine our neural network within limits, we give it the opportunity to choose. And for example, since we most often use Claude for writing, I will create this project there as well. Here in the menu, there are projects, new project. Well, and I will name it exactly the same, "Content Machine," and I will put this same instruction here. It can be transferred, and I just paste it. This is what changed everything. GPT and Claude projects. With the latest updates, they reveal incredible possibilities to us, but most people continue to use neural networks in the old way, opening a chat, loading prompt requests into it, no matter how long, no matter how precise they are, they still limit the neural network in its capabilities in one way or another, while projects allow you to create and use absolutely all parameters of the neural network. That is, you reach all the neurons of the neural networks. And this is exactly what was intended from the beginning when creating neural networks. Therefore, those who switch to this new revolutionary way of using neural networks right now will get colossal results before everyone else. It seems like science fiction, but now I will demonstrate it to you. I am launching two requests here. Here I will write a post about a psychologist. And now you will compare and see the difference, why we prefer to use Claude for writing posts and texts for the most part. Here is a version of a catchy post from a coffee shop. And here, it already forms it from such a simple request into a post, which would have taken us at least five sentences in a prompt request. Opened a coffee shop without a franchise, without investors, just with a dream and trembling hands. So you see, it uses metaphors, it doesn't use participles, it writes clearly to the point, it determines a specific style and adapts it for any social network. Here you see, you can specify, and I will just write it a post for a Telegram channel with a call to action for a lead magnet with a price list for a coffee shop and list all the methodologies you used to write the post, what you took as a basis. Describe all of this in detail. So you will see for yourself that it did not invent this out of thin air. And it writes excellently. Below. First, it will write the post itself, and after that, a detailed analysis. Post for a Telegram channel. How much does it cost to open a coffee shop? This question follows us even in the queue for coffee. You feel the difference, right? People think you need a franchise, millions, a cool spot near the metro, and a separate barista god. But for us, it all started with an Excel spreadsheet and the belief that we could handle it. That's why we made a simple and honest price list without fluff, marketing flourishes, and rose-tinted glasses. And here are the advantages: minimal starting budget, how much equipment actually costs, and how much you need, what for. Pay attention to the difference. And our request is as simple as possible. We didn't write to it that you are a copywriter who has gone through many trials, received a Nobel Prize. That's useless. We don't confine it. Instead, we open up opportunities to all the knowledge that is already available in it. And we don't need to delve into it unnecessarily. And look at what it used, precisely that AIDA formula that we formed with a detailed prompt request from the previous example. A project based on its own research, on connecting all neural network parameters, it accurately determined the niche, audience, and selected the right words. No additional prompt requests, no additional instructions. And here you can refine it further in a dialogue mode with this neural network. Let's see what Claude wrote to us here. Of course, it gave us a slightly detailed breakdown, but still. Post about psychologist services. Yesterday, a client told me: "I don't know how you do it, but in an hour I understood what I had been struggling with for 3 years." And she came with a typical request of the year 2025. Tired of being anxious every day, but don't know why. In recent months, more and more people have been coming to me with similar problems. And you know what? It's not that we've become more anxious. It's that we've finally stopped pretending that everything is okay. Well, this sentence, in fact, I don't like it. I think it will already grate on everyone's nerves. This can be fixed, you don't need to write this. Try not to confine your language model, but rather specify it in more detail. And I will just specify from whom. Psychologist services, from a girl, about anxiety. I just added a supplement, what I didn't describe and didn't confine it. That is, as you can see, these are smart projects. And the biggest advantage is that in addition to the instruction, we can add files here. And you know what these files can be? These can be your best examples, which these smart projects will be guided by. These projects are called smart for the reason that the instruction here says not to limit itself, but to find research, data, without spending 10-20 minutes searching for research. And as you can see, it is literally browsing through lots of websites. We are not asking it to do this. This is our prompt request, as simple as possible. It browses data, makes comparisons. It went to different social networks, looked at what everyone is writing. And here is a post about services. Lena, what if I go crazy from this anxiety? Tanya was crying on the phone at 11 PM on Sunday. Monday is an important presentation. My heart is pounding, my hands are shaking, my thoughts are racing. What if I fail? What if I get fired? What if? What if? What if? Familiar? Girls, I get 10 calls like this a day. And you know what? You are not going crazy. Women experience anxiety almost twice as often as men. This is not your weakness, it's a feature of our nervous system. We are biologically more sensitive to stress. Our hormones fluctuate every month. We are pressured by social expectations to be perfect everywhere and in everything. But here's the problem. We are used to enduring. It will pass on its own, everything will heal, you need to be strong, and anxiety accumulates like a snowball. Until one night you wake up thinking, I can't take it anymore. And it's normal to ask for help. In the last 3 months, a mother of two, an IT specialist, a student, a freelancer have contacted me. Each has their own story, but the root is the same. We don't know how to communicate with anxiety. The good news. Anxiety is perfectly amenable to work. In three to five sessions, you can learn, but spending several hours on each session, sitting with a psychologist, traveling, commuting. In current realities, no one wants that. And, well, here is my short lesson. Watch for free on YouTube. Well, as an example. The only thing I didn't like about all of this is this. At least it grates on my nerves, where I see that, girls, I get such requests every day, they write to me like this every day. Well, here you can neutralize this or replace it with some other element, but there are no participles, no unnecessary inappropriate metaphors in this text, but the text turned out great, and I found it really interesting to read. And it reveals the topic step by step, relying on a lot of data, a lot of research that we didn't ask it to conduct, and where we didn't confine it within these limits. Sales copywriter, personal coach. Let's try it right here. Yes, and here, as you understand, in Claude projects, you can also add your example posts, your research, some documents, and your materials will turn out even cooler, even with the simplest requests. Post about coffee. There won't be a post here now. I even know that before it starts writing, it won't. Coffee is. And based on research of coffee trends for 2025 and the psychology of coffee lovers, I will create a catchy post, identifying key insights. And it writes: "You don't drink coffee, coffee drinks you." How do you like the headline? Just a request for a post about coffee. Go to any neural network and write a post about coffee. It will write to you: coffee is this and that, there are such and such varieties, there is arabica, there is roasting level. It will write a level that cannot be published or posted anywhere, and we get an incredibly cool post with incredibly cool engagement, which is guaranteed to gather a large number of reactions. >> And here is where the most interesting part begins. By working through a project, we do not limit the neural network, we give it the opportunity to select the context we need from billions of parameters that the neural network has. That is, it accesses the necessary context, the necessary neural connections, which carry out deep analysis and determine the best tools for solving your task. What target audience, what triggers are used, how texts are correctly formed in this particular niche, or how and through what methodology certain tasks are solved, if it is not related to writing. That is, the advantage, I think, is obvious. Another plus is that one project can work for different niches, for different directions at once. So, now we will quickly look at two more projects here. The first is a sales copywriter. And the second, I will create right away, in parallel, is a personal coach. There is also a cool smart project called "Personal Assistant." And you will see it yourself, you will be pleasantly surprised by the information it provides about finances, sports, general development in depth. It searches for research on the topic. The personal teacher, personal coach will come with this instruction. And the sales copywriter will come with this instruction. And for the sales copywriter, I will simply write: "Write a letter to partners about cooperation in an IT project." For the personal coach, I will ask it: "Teach me English." And it will not just start telling me the alphabet, the sequence of learning, it will start conducting research right now. It conducts research not just based on the request, I want you to understand this immediately. Depending on the niche, on the direction of your request, it conducts separate research and selects the best methodologies based on real research. It selects current methods that are trending, that are in demand now, and offers you not just within the limits of the prompt request, where you told it: "Teach me according to these rules, according to this methodology," but it offers you the most effective methodology that works right now. And it diagnoses the situation, determines that you are in the category of professional skills. English is needed for career growth, expanded opportunities, and access to information. From your request, I understand the level, motivation, format, and time constraints. That is, you can discuss this with it additionally here, but in fact, it creates a complete program for us. Look, let's wait a couple of seconds here and see what volume of data it will compile and provide to us. In the meantime, let's see what we have regarding the post. Make it as short as possible for an email newsletter. Here is the post. Here is a specific offer. In principle, we can immediately see what information we need to provide. From a simple request and the simplest request, I will ask it to program all of this into a game. And let's see what we get here. For logic, of course, it's better to use Opus. Well, it works better with logic. Claude Opus 4 is directly used for programming. It works better with logic. It has a larger chat volume. Again, pay attention to my prompt request. Just program this game for me, which you have described yourself, so that I learn through play. And here, teach me English. I didn't write to it: "You are a teacher from Hogwarts, who knows English perfectly, can speak fluently, knows three spells in English: 'Teach me this magic, teach me this English.'" I didn't write all this. These are limitations that often lead the neural network to a dead end. English Quest. Enter your name. Giveg. Start test. We have energy, we have health. Word battle. Okay. Water. Aha. What's next? We have skills, achievements. Words learned: one. Thank you. It has slightly misplaced the words itself, but as I said, for coding, Claude is used here. But the point is that, please, we can publish it via a link and always log in to play. Look, you can easily publish it via the "Publish" button and share this link. Imagine you run a thematic channel about English and say: "Guys, today we are learning these 10 words. Please learn them, you have two lives." Make it so that it caches and doesn't launch a second time, and then say, who has what results. Imagine if you add this to your channels, blogs, chats, social networks, just post it on your feed, share it with friends and say: "Friends, today let's learn something, or let's conduct a cool research." This works incredibly well. And this will take you to a completely new level. And these are just three out of seven smart projects. Imagine you have a team of seven top experts. These are your personal employees who work 24/7, deliver the best results, and don't get tired. Right now, you have an incredible opportunity, literally a few days after the release of this video, to get these seven instructions absolutely for free. Yes, you might have thought now, there will be some special offer. No, follow the link in the description of this video and get these seven instructions for your smart projects absolutely for free. Well, and subscribe to my Telegram channel, where I show how I earn on thematic channels on Telegram, on Dzen, how I run my own websites, and in general, how you can grow very quickly using neural networks. I directly show what I use. And there is a section in my Telegram channel where every week, every month, I show what has been done, how much has been earned from it. And you see what is working right now in the current market realities. Well, and let's realize what we have just seen. First, you can safely forget about tens of hours spent searching for so-called cool prompt requests. We don't need them anymore. Second, you can safely forget about unnecessary customization of our chats, because it only complicates the work process, it only confines everything within limits, and our neural network starts to hallucinate. You've probably encountered this. Now you can also forget about the long search for employees to perform a particular task. Well, now one project is 1,000 specialists in your pocket through a project to which you can provide absolutely any information, and it will rely on it as a supplement. Of course, take advantage of this colossal advantage, because many are still spending time searching for prompts from the internet. Competitors are already using smart projects and are outpacing many by tens of times. Therefore, this is your competitive advantage. As much as I would like to say, share this video. It's really cool. It's better not to tell anyone about it. Just give this video a like. Well, and subscribe, because we might simply never see each other again.