Transcription
Did ChatGPT make you a millionaire? My answer is that I believe that with the help of ChatGPT, you can indeed become a millionaire faster than without it. Well, in our youth, there was no ChatGPT, we had to figure things out ourselves. Yes, yes, yes. Well, something like that. What do I mean? I mean that ChatGPT is like having a best friend, a professor from some Cambridge, who simultaneously teaches a course on logic, rational thinking, and finance. And you can come and consult. That is, he won't tell you unequivocally: "Taras, do it this way." Or like, for you to make money, put everything on red at the casino. But what I mean is that it's like a cool advisor with whom you can really consult, but the choice will still be yours. In general, a correctly formulated prompt to a neural network can save tens of hours, bring order to your thoughts, show blind spots, weaknesses, or, conversely, strong points. Right? This is Evgeny Taras, and today we will tell you how to use ChatGPT, which is on the level of a Cambridge professor, not just as a search engine, but as a real partner who will help you easily achieve success solely through correct prompts and questions to it. Prompting, in short. Prompting, in short. In this video, you will learn about prompts that you never guessed before. They will definitely change your life. Will they really? Definitely. If they don't change it, they can come to you and ask for their money back. The guy said it, the guy changed his mind, or what? No, we didn't take the money, for fairness. We'll have to organize a cooperative. Well, shall we go? Let's. Taras, how do you prompt a neural network? One of the most important things that I do most often is role-oriented prompts. I set a clear role for ChatGPT at the beginning of the prompt. That is, you are my project manager, you are a marketing expert, you are a higher mathematics teacher, you are a business strategist, a marketing specialist, and so on. And do roles really help? They really help. And this helps to understand in what context to give advice at all, with whom I am actually communicating at the current moment. Despite the fact that it is, as it were, the same neural network, we perceive it as the same entity. Nevertheless, we start at the moment when the neural network accepts this role, communicating with it as a specific expert. And this, well, actually, helps to maintain context and respond to you expertly. And it also helps you understand, uh, what question to ask next, for example, after the first answer. Well, okay, I won't troll it. A minute of entertaining psychology. Actually, when you assign a role to a neural network, then the identity or role is at the top of the pyramid, and it sets everything below it in a cone shape. That is, the system of beliefs, context, rules, and so on and so forth. And a person works exactly the same way. That is, when we switch to the role of, say, a wonderful family man, a father of children, then we sort of pull up the contexts and memories associated with it. Our behavior changes, and everything else. When we are at work, say, you are in the role of an engineer, right, you have different contexts, your voice even changes sometimes. And this is noticeable, for example, from the outside. Some people don't notice it, but most people change. And the neural network works approximately the same way. This is because it pulls up the lower layers like a pyramid, and based on them, you narrow down the context of the search a bit. Therefore, the neural network doesn't wander around from here, from here, from here, from there, but it becomes more focused, and as a result, you get a better answer. Therefore, 100 100 100% no magic, pure science. If you don't do this, there's a high probability that the neural network will drift in context, give examples of random associative connections that are actually irrelevant. When you don't assign a role, you get too broad an answer, and the neural network doesn't understand how it's talking to you, and can actually give a simply abstract answer. For example, if you ask it in the context of marketing, it won't understand and will respond in the context of general business or some strategic stories. It might go completely off track. That is, it won't understand that it needs to talk about, well, hypothetically, attracting customers to your coffee shop, but instead will suggest using some kind of informational approach to generating some kind of customers in CRM or something else. These are completely unrelated things, so this is important. Write in the comments for what area of life or activity you most often use a neural network and do you assign it a role. While preparing for this video, I asked the neural network itself to analyze my prompts and my questions to it in order to identify some key approaches that I use. Because many things arise simply intuitively, based on previous experience or a large number of experiments. And one of them is, first and foremost, to use the neural network as a mirror. That is, I don't try to turn the neural network into a smart search engine, or use it as a source of truth, like, go, find me this or that, or do this or that for me. Although that happens too. But often I understand the potential of the intelligence behind the neural network. And therefore, I often ask the neural network to reflect back to me some of my false beliefs or my cognitive biases. That is why, for example, I often ask it to highlight cognitive biases in my question, because the very structure of constructing a question can often include an assumption about something. That is, what does this mean? That is, with a question, I might be looking for an answer that confirms my knowledge in something. For example, why, you know, are cats to blame for all my troubles? Well, the question itself implies that the neural network has already started looking for an answer. And why, you know, yes, indeed, all problems are because cats are shitting in our beautiful entrances? And by doing this, you are already setting limitations, setting a context. That is, it won't go looking to argue with you that there will be no alternative culprits besides cats. Yes, yes, yes. And even if we look at this at a higher level, it's looking for culprits in general. I, on the contrary, fight against such things, so I usually prompt the neural network like this: "Tell me what you see in my question, help me overcome my limitations in this prompt." That is, I usually, at the end of some of my questions or prompts, I add a whole block related to thinking strategies, where I ask the neural network to highlight what I am hiding from myself, what I don't see, and what cognitive biases the neural network sees in my questions. Give likes if cats can't be blamed for anything at all. Do cats call you? And do you get special notifications from cats when they are offended? That's an answer, so to speak, not of a boy, but of a man. What is the essence of such an approach globally? Actually, in an era when intelligence is becoming like electricity, that is, it is becoming accessible to everyone. And even if you are not very smart, you can go and consult a neural network. Your task, your task is not to accumulate knowledge or anything else. Your task is to correctly structure the process of your own thinking. That is, your task is to see the architecture of how you think about something, because knowledge becomes obsolete very quickly. Plus, you have access to Wikipedia, plus you have access to this, that, the fifth, the tenth. Therefore, the accumulation of knowledge is not so important now, although that is also for practice, but rather the structuring of knowledge, like the architecture of knowledge, how you think about something. In general, my dears, if you thought that the times of encyclopedists were over, they, thank God, are returning. Therefore, with every question to the neural network, it is very important for you to keep one very simple thing in mind: How can I become smarter? Smarter not in the sense of accumulating more knowledge, but smarter in terms of how I think about any problems, what my approach is, how, if I encounter something, how should I think about it correctly. Because as soon as I understand how to think about it, how to ask questions, and so on and so forth, because we have outsourced intelligence, you just need to write the question, and the answer will materialize. And then, it's a matter of turning that answer into action, so to speak, implementing it into reality. Let's continue with your approach. Uh, I remember you talked about role-oriented prompts. Is there any other smart word for, so to speak, other prompts? No, there are no more smart words. There will be simple, good, but effective words. A task with a clear structure and goals. This is what you need to remember always when you communicate with a neural network. So, ChatGPT itself tells me that I do this often and advise you to do the same. What is a clear structure? Clear structure. Say what needs to be done, how to use it. I plan the results of this prompt and the desired style. We are talking, of course, about solving your personal tasks, right? For example, you need to prepare a script for a video. I have no money to pay for the apartment. This is our target audience, by the way. With what question? I have no money to pay for the apartment. This is not a task, this is context. And context. I have no money to pay for the apartment. Task: find money. Wait, wait. No, that's not all. Find money. Within what timeframe? When is your deadline? Uh-huh. My deadline is by the end of the month. Find, find, find money by the end of the month. Excellent. Excellent. So, role, uh, well, let's say, personal finance advisor. You'll set the role. 80,000 rubles. Do we write this in the context or in the task? Well, write it in the task, it's normal. Personal finance advisor. Did you write anything else? Context, task, role. So, how do you plan to use it? That is, what do you need? Let's write, give me a clear plan or several plans, so that within the specified timeframe I achieve the result. Well, we'll probably use the O3 Pro model. No, let's use O3, otherwise it will be too long. Well, ideally, like a style, but okay, like a plan of action, there's a style. Well, in what form do you need it? That is, can you answer? In what question? How is it convenient for you? I would write: "Let's get down to business." Well, okay, let's go. Bang. Yes. Bang. What are we doing? Looking for money for apartment rent. Subscribe to the channel. So Evgeny won't have problems with apartment rent. What will he advise? Well, let's continue for now. While he, you know, what else do you have there? Next point. If you have already set a prompt and received some result, then it's good to make clarifying or refined prompts. For example, you see a result that has some goals. You are satisfied with the goal, but you want to get two more options. You say: "Okay, what if the goal is twice as bad or twice as good, what will the plan of action be then?" You can thus challenge the neural network to generate different results, different brands, in order to get, so to speak, the full range of solutions that will lead you to results. You can, well, ask to break down several options for something too, that is, when five people do a task, 10 people do a task, you do a task, and so on. This also helps to form alternative options, based on resources, approach, and so on. Okay, well, look, the neural network has written us a plan based on the previous prompt. Excellent. Reduce commitments. Negotiations with the landlord about installment payments or deferral of part of the payment. Uh-huh. -1 euro plus time saved. Okay, let's cancel subscriptions to gyms and services. Subscriptions. Well, reasonable. Plus 30-50 euros you'll get. Not bad. I don't care about your personal state. Let's liquidate assets. Cool. Selling electronics. OLX FB Marketpl. In short, let's go to Avito and sell the laptop, while my good pawnshop is instant. Offline part-time jobs, you can get a job as a courier. They say, 1 thousand, 1.5 thousand per month, they say, covers the whole apartment easily. You can get a job at Yandex Divvey. You can get a job now at Yandex Yandex Food. Yandex Food online freelance. Wait, wait. You can also work for food delivery. Excellent. Online freelance, while you're delivering, you can work in freelance, you can consult on legal issues with the help of a neural network. I did that in my student years, but not with the help of a neural network, but just by retyping textbooks. Excellent. You can also rent out your car. Great. So, wait, then you won't be able to work at Yandex Food? No, you'll walk with your legs, or on a scooter, or on a scooter, right? Look, you rent out your car, you get a scooter, it's all normal. Social microcredit. I'm afraid there's no such thing in Russia. In general, dear friends, if you were here where we are, then you would have access to a social microcredit with an almost zero interest rate for 2,500 euros. That's how it is. So, saving on what you have, selling something extra, working as a courier, and online freelancing. In total, I read Only Funs. I think it's okay. It's also a good plan. In general, look. Well, overall, the plan is not bad. If we continue to increase the context, what you can do well, what you can't do well, for example. Well, generally, look, another 1,810 euros is more than 850. You can also invest in vaccination. You'll earn 200,000, and spend 80,000. Is there anything to spend it on? There is. You need to eat, right? Forget it. What are we getting at? In general, not the worst advice, actually. And if you provide more context, uh, in this case, for example, what I can do, what I can't do, what I want, what I don't want, then the plan can be, in principle, excellent. The only thing is, it does deprive you of illusions. Let's try my thing, right, what are my limitations. Although, maybe YouTube will be unbanned in Russia, by the way, if you do this. What I didn't ask that I should have, what I, what I, pam-pam-pam. This, guys, is my approach. Uh, a minute of reflection, where we allow ChatGPT, so to speak, to formulate for us some things that we should have asked. Okay. And then it writes: "Limitation of the initial request. No list of assets." See? That is, to sell, to rent out the car, it assumed, and therefore, you know, well, how it usually happens? People ask such a question, like we did at the beginning, and then they say: "Tell me, sell my car, what kind of nonsense is this, my friend, you didn't specify the list of resources yourself?" Well, again, seeing the offer to sell the car, you might say: "Okay, I don't have a car, but I have this, this, this." Yes. But most people at this moment start to get annoyed and say: "What nonsense is it writing to me?" And look further. Location and legal environment. That is, it was necessary to specify the city where we are located, the time resource in available hours, health, driver's license, skills. Oh, this is what we've already started to assume, or, uh, education, risk profile, readiness to borrow, to pledge property is not indicated. What should have been asked? What assets can be quickly monetized to rank them? Look, these are just three of my questions, and in response, it sends me excellent, excellent questions. What to think about so that I write to it again, and it helps me to detail it further? That is, uh, the neural network in this case acts as a kind of mirror, acts as a kind of advisor, coach, and psychotherapist simultaneously, where it teaches you to prompt yourself. The third point, it specifically writes what to add to the prompt for a more accurate answer. For example, location, timeframe, amount, assets, income, skills. Look, it told me everything about myself. It told me everything about myself. It completely exposed me. Risks, limitations, creditworthiness. Well, in general, you'll look at all this. And it even highlighted the identified cognitive biases. And then you can delve into each of these points. You can deepen your initial prompt. You can, for example, latch onto cognitive biases, go, like, become smarter. It even highlighted what it's important for it to learn from this question. Well, look, what I like. It answers me: "Systematic inventory of assets, place of response, fire suppression." That is, it slows me down a bit and acts like a rational guy, saying: "Okay, wait, let's break everything down clearly and so on." Based on this, you can build a real financial plan. That is, if we take all this into account, then it can really give great advice. Cool. And have you tried writing to it: "Hey, bro, just make money for me"? The result will be bad. Let's try. "Can you, bro, just make money for me?" asks Evgeny the Neural Network. And it thinks, it thought quite quickly: "I cannot make money for you, but I can show you how to launch a mechanism with one click that will do it without your manual labor." This is interesting. Even I became interested in what it is proposing. Okay, logical step. Actions require time, capital, skills. They cannot be fully delegated. And let it be for me. I want passive income. We are talking about a system. And in general, you can go into philosophy and talk about whether passive income exists at all. Because in the conditions of modern economic turbulence, even forming a passive investment portfolio still requires attention. Bot arbitrage. Uh, 1-2 weeks for development. Agreed. Media algorithms, and content generation, traffic accumulation. Well, listen, to the point. It's really to the point. I still like the product, platform, digital product, GPT plugin, idea, one-time assembly, publication, capital market, peer-to-peer lending, finally. Peer-to-peer is questionable, but okay. Minimal stack: automation, legal framework, API, keys, risk control, concrete plan. Choose a model, generate 10 templates. Okay, drive traffic, automate advertising, paid listings. Look, uh, you know, we can conduct an experiment. If you want us to go through this plan and actually do everything it told us, and see if we can earn 190 euros a day, write in the comments if you want us to actually do such an experiment for 15 days that will lead us to 190 euros a day. In general, GPT is not a button, "make it beautiful." How you communicate with it, ask questions, write prompts, that's what determines the answer. If it's a chaotic request, you'll get a chaotic, useless answer. And a precise question and strong communication with ChatGPT is what can change your life for the better. Evgeny, yes, in general, I would say that the most effective prompt in this case is when you don't try to get an answer immediately, but try to change your thinking to ask the right question. Everyone, bye.