El "Truco Secreto" de ChatGPT que Nadie te Cuenta.

Laura Cilleruelo shows how to get better results from ChatGPT using voice commands, contextual prompts, and iterative questions. The trick is treating the AI like a conversation, not a one-shot search.
Hello, everyone. Pay attention to today's video because in the next few minutes, I'm going to give you three tips. Three very simple tricks that are going to completely change how you use artificial intelligence, and above all, the results it gives you. But first, I want you to watch the following video. Excuse me, do you know if there's a bakery near here?
Uh, yes.
Would you mind telling me where it is? No, I wouldn't mind.
Well, why don't you tell me?
I do want to tell you.
Well, then why don't you want to tell me? I do want to. Go ahead.
Well, tell me if you haven't asked me. But I just asked you.
No, you haven't asked me if I wanted to tell you.
Okay, okay, you're right. Let's start over. Where is the bakery? Which bakery? Where's the closest bakery?
But the closest one. Where?
Where's the closest bakery from here? The closest bakery from us now.
Ah, that bakery. It's on fire.
Yes. What we just saw, and that seems so funny, is what happens to us every day at work. With the boss, our colleagues, with our partner, with our children. We want to say something, and the other person doesn't think the same way we do, doesn't process it the same way, and doesn't understand it. And the exact same thing happens to us with artificial intelligence. But in here comes a very important but contrary to what everyone thinks, it doesn't need to be super technical, speak very, very well, write even better, or have a degree in quantum physics. Next, we're going to see three very simple ways to solve this problem. Okay, let's open our ChatGPT screen. Although what we're going to see you can use in any artificial intelligence application you use. It could be ChatGPT, it could be Gemini, it could be Cloud, or even if it's dip C, it doesn't matter. So, as I was saying before, 99% of people do this. Hello. I want you to give me ideas to increase sales for my business. He'll probably answer me differently because he has a memory of me. But if we send him this type of prompt, the first thing he's going to do is give us mediocre answers. Look, because I don't understand much, he throws you specific ideas divided into quick ones to execute. If you give me more details about your business, I'll fine tune it. But here's a first set of ideas. Limited Hook Offer email campaign to inactive users Irresistible Pack in short, here, he's giving me a very mediocre answer that from what I'm seeing in most cases and most of the options won't work for me. And this isn't his fault, it's my prompt's fault, my messages. Because when we start talking to artificial intelligence and we start writing, we use the telegram model that I call it, okay? We're used to Google and also our brain is lazy by nature. It wants to save energy and we don't give it absolutely any context, as you're seeing here. So the best thing to do in these cases, the simplest thing to do, and what will save you time so that you can start using artificial intelligence and not have to learn a lot about prompting, which we will see in later steps, is to use the live model. Notice that here you have these little lines. Notice that next to the microphone you have these little lines here that tell you to use voice mode. And that is what we are going to do next. Hello, Chat, how are you? Hello Laura. Well done. I am delighted to chat with you. How are you? Tell me what you feel like doing today if you have any ideas in mind. Or I can just chat. Chat. Look, I have a client who has a candle company. They have been on the market for six months. Many people still know about them. They have also been creating content on social media for six months, but they are not getting results. I would like to be able to help them and come up with strategies that could be disruptive and help them first of all to get people to know them and then to win customers to increase sales of these candles. Can you help me?
Of course. For a candle brand that is starting out, the first thing is to capture attention and differentiate themselves. Some strategies that could work for them are, for example, collaborating with micro influencers who have a very niche audience so that they can write authentic reviews of the candles. They could also create a limited edition with a very special design or aroma that arouses curiosity. Another idea is to hold workshops or live events.
Excuse me, Chat. Uh, I wanted to tell you that for some of these things you are telling me, they would need to invest in ads and right now they do not have the budget for this. Do you understand? In that case we can focus on organic strategies. A good idea is to take advantage of cross collaborations with other brands that have a similar audience but are not direct competitors. Perfect. And you, and you, in order to contact them, could you help me find these contact people and write the emails so that I can write to them? Sure, I can help you with that without any problem. If you tell me a little more about the type of brand you would like to collaborate with or the profile. Okay. Notice that by doing it this way, since we have a conversation with them, we will naturally and instinctively enter into a lot of context. And I guarantee that it does. I'm sure you know someone that you meet on the street, an acquaintance, a friend from school, from childhood, and they give you a lot of context. Let me tell you some stuff that makes you look at your watch, thinking, hey, when is this going to end? By nature, most of us tend to ramble on and on. And when we have a person on the other end of the spectrum like this one, ChatGPT, whose voice, as you've seen, is incredibly human, I guarantee that the results will be much better. You'll be able to elaborate much more and give many more details. Now we're going to move on to the second trick, and this is going to give you different results than 80% of the people out there. ChatGPT is trained with all the knowledge in the world, but if you've noticed, he usually gives us pretty similar answers if we ask him about the same topic. Well, there's a way to prevent this from happening. So let's go back to our CHAT GPT screen and do the following. Instead of telling him that he's an expert in any field and that we want him to give us, imagine I'm going to copy this prompt. I just told him that he's an expert in vegan nutrition and has been working on it for over 20 years. Let me make a brief aside. It's important because amidst all this knowledge, we're giving them a role. I'm sure you've heard of it. We're giving them a profession, and this helps us give them a compass so that amidst all that knowledge, they know what to discard and what to choose. Okay? After all, you're an expert in vegan nutrition. You've been doing this for over 20 years. Right now I'm making a change in my diet and I want to practice veganism from now on, but I have no idea what to buy or how to cook it. If any vegans are watching me, please, this is just an example. Don't kill me. I have no idea about vegan food, so if I say something outrageous, I apologize. Okay? Having said that, what happens? It's very likely that if we leave this pump here, these instructions, it would give us a pretty good result. It would just need a little more. And we're going to give it the instruction to give us that differentiation. Look, in addition to your knowledge, do a search on forums like Reddit or platforms like X about ingredients, recipes and best practices for eating a vegan diet. Look, I'm telling him to go to platforms where real human people interact with him in addition to his knowledge. This helps him stay up to date, stay fresh, and not limit himself to what he already knows. From here, I'm telling him to make a shopping list for me and then a list of step by step recipes. So by doing this, I guarantee that the answers ChatGPT will give you will be much better. And now let's test it. We hit enter and let's wait. Look, he's already telling us that he's searching the Internet, he's gone to forums and now he's given us our basic versatile vegan shopping list. He's given us a list of proteins, cereals, milk and alternatives, vegetables, subtle extra condiments like hummus, olives, tahini, etc. It also tells us the sources from which the information has been extracted and then gives us a list of step by step recipes. I gave you the example of veganism, but you can adapt it to any topic that interests you. But use this trick. In addition to their knowledge, tell them to search on external platforms to give it freshness and differentiation. And now let's move on to the third trick, which for me is the best. But first, if you like this type of content and don't want to have to wait until it's once a week, I talk about this type of thing and the latest news in my newsletter, Claridad D Aria, a newsletter that I'm going to send you not just once a week, but every day at 7pm so I'm going to leave you the link so you can subscribe in the description. And with that said, what's promised is due. Let's move on to the third trick and as I was saying, the one that for me is the most useful. Look, we're going back to our GPT chat screen and we're going to put exactly this I need in your goal. Imagine that. I tell you I need to lose 10kg in the next six months. I'm a total beginner, I have no idea about nutrition and I have no health problems. This is important because if we have any type of pathology, we should never resort to chatgpt. We go to a specialist, to a doctor, that's what they're there for. But for this example, we're going to use it and it's I need to lose 10kg in the next six months. I'm a total beginner I have no idea about nutrition and I have no health problems. Before giving me any answers, ask me the questions. You need to give me exactly what I'm looking for. This is a total game changer. In other words, I don't have to structure a prong and rack my brains spending time so that the artificial intelligence gives me an answer that I like and that works for me. But instead, in this way, I'm going to make it ask me the questions so that I can give it the answers it needs. Look, it's here. Perfect. To help you lose 10kg in six months in a healthy, realistic and sustainable way. I need to know your age, height, current weight, if you have any physical limitations, what your daily routine is like, if you sleep well, what you usually eat in a typical day. So here we would combine all these questions and this trick. With the first trick we saw, we could pick up and give it a voice command and have a conversation with it while we answer each of the questions. Or even if we want to be able to see the questions one by one, we can give it dictation, this little microphone that we have here that says dictate and we could start talking with GPT chat, answering all these questions. Okay, so far we have seen three tricks that you really should put into practice because they will completely change your relationship with artificial intelligence. We have seen that we can talk to it by voice to be able to give it more details and more context and for it to answer us better. As a second trick, we have seen that we can ask it to use external platforms so that it stays more up to date and gives us better answers. And thirdly, we have seen how to make artificial intelligence ask us questions to have all the information it needs to give us its best answer. So I want you to put all this into practice, but I also don't just want you to like this video, I want you to share it. I want you to tell me in the comments what tricks you're going to use or what trick you've used and what results it gave you. Also, pay close attention to next week's video because we're going to see something that's going to help you take advantage of ChatGPT's lies. I don't want to spoil it for you, so I won't tell you anymore. See you in the newsletter tomorrow or next Thursday on this channel with a new video.
More from Laura

El Error con Opus 5.5 y ChatGPT 6 que Comete el 95% (Qué Modelo Usar en Cada Caso)

Claude Ahora Recuerda Todo lo que le Cuentas (pero hay algo que NUNCA Guardará)

La verdad sobre las IAs chinas: ¿Gratis o estafa? #ia #tecnologia #noticias #kimik3 #deepseek

La Verdad de las IAs Chinas Que Nadie Te Cuenta: Gratis vs Realidad (Kimi K3, DeepSeek, Qwen y GLM)

Hablo a ChatGPT y Trabaja Solo en Mi Ordenador: 3 Casos Reales (y Lo Que Nadie Te Cuenta)





