What Is AI? Here’s How It Actually Works ⚡️

Everyone is talking about AI. ChatGPT, Gemini, everyone is saying that AI is going to destroy jobs, AI is here & I'm using it. But 99% of the people I've talked to don't know what AI is & how it works. In this video, I'll tell you how AI works in a very simple language. So that you get to know how AI works & how it gives you answers. And a lot of times, the answers aren't right. But AI is trained & made like that. In the next 8-10 mins, I'll explain everything in a simple language. If you're a techie, then this video isn't for you. This video is for those who want to know about AI in a very simple language. We'll go into the details of this series in a simplified version. We'll move towards complex things, but I'll tell you in a simple language. Even a normal viewer will know what AI is. Let's start, but subscribe & hit the bell icon. We'll release this series every Friday. Let me give you an example. Let's say you use a calculator. If you do 2 plus 2, you'll get 4. If you do 100 times, you'll get 1000 times. If you do 3 plus 3, you'll get 6. So that's a rule. You can't put A plus B there. The calculator is built in such a way that you type the numbers & it'll give you the answers. AI isn't like that. Let me explain how AI works. Let's say you have a system, an AI calculator. It has a camera & you've shown it 10,000 cats. It has registered them all. 10,000, 20,000, 25,000 photos are of cats. Now it has recognised them. I won't go into the details of how it recognises them. It'll come in the next videos. Now you've shown another photo. Do you know what it'll do? It'll most probably say it's a cat. Because it hasn't shown anything else apart from cats. Now, it has shown 10,000 photos of cats & dogs. And we've shown an elephant. What will it do? It only knows about cats & dogs. It'll see how cats & dogs look. It'll see which elephant is close by. And then it'll give the answer. So there's pattern matching & probabilities on which AI works. I think that's the fundamental rule. Now think about it. This calculator has shown 10,000,000,000, 100,000,000,000 different things. And has asked it to recognise patterns. It'll give you answers on the basis of those patterns. This means that the rules aren't set. It grasps, sees & feeds the data. It creates patterns from the data & gives you answers. When people use AI for the 1st time, they say it's like magic. But it's not magic. It's a large language model called LLMs. There'll be a separate video on it. It creates patterns on the basis of the data set. It gives you answers on the basis of the billions of examples it has. Let me give you a simple tech example. If you ask Gemini to make an image of a tech YouTuber. It'll show a photo of MKBHD, one of the biggest tech YouTubers. Because it's trained on them. Because of the data set, it feels like it's the same tech YouTuber. So the quality of the data you feed, you'll get the same quality of response. It also means that all the answers depend on the data it's trained on. If you train it on wrong data, you'll get wrong answers. If you train it on good data, you'll get good answers. It's a software, an algorithm. Netflix also uses AI to tell you that you can watch the next movie. Because it knows the hundreds, thousands & millions of examples. Other viewers also have them. After watching a particular category of video, people generally prefer it. And that's what it recommends to you. Again, a small example of AI. People who've been using Gemini for the past 6 months, will see that it's getting personalized. It knows more about you. And it's giving you the answers that you want. It didn't do it before, but now it does. Because whatever you've asked it, whatever data you've entered, it has made you a personality. Again, it tailors your responses according to the probability & output of the answer. The more you train it & give good data, the better the output will be. In the last 18-24 months, AI has been increasing rapidly. It's becoming more intelligent as it's getting more data & people are using it. It's getting more intelligent. Imagine, 5 billion people are using it & AI is feeding it. How intelligent will it become? It has got internet, wikipedia & billions of webpages on the internet. It has got everyone's knowledge & patterns. So it's become super intelligent. But again, it can do more on the basis of the data it has. It doesn't have its own brain or consciousness. Just like humans have a soul. It doesn't have a soul, but people see it as a human being & talk to it. I think somewhere, we should think that AI is after all a software. It's an assistant & it'll give the information it has. At the same time, all the answers you'll get can't be correct. Because it'll answer on the data it's given. So what can it do now? It can recognize things. It can unlock things. For example, it can unlock the face. It knows your face, it has fed it, it can see patterns. So even if you wear glasses, it'll still unlock it. Why? Because it knows the patterns. Fingerprint matching is the same thing. Translation, it can translate in hundreds of languages. English to Marathi, Marathi to Hindi, Hindi to Bengali, Bengali to French. It can do whatever it wants. It can generate images. Now you'll ask, how do you generate images? For example, I said show me a cat wearing a goggle on a beach. So it has hundreds & thousands of photos of beaches. It has different photos of cats & goggles. So with that probability, it'll think that it's on a beach. Meaning it's sunny weather, there's a beach, there should be a sea at the back. It has different goggles & images of cats too. It combines all that & then creates an image for you. Yes, that's how it works. So it generates images as you say. Now friends, there's going to be a separate video on Prompt Engineering. Because otherwise, this video will be 8-10 hours long. I'll start explaining everything. But you should test & experiment. Go to Gemini, Poplexity, ChatGPT, Anthropic. There are different AIKM models & completely different platforms. But if you give a single line prompt to generate an image. And if you give the same, everyone will generate similar images. Because all the platforms have similar data. Nearly 90% of their data is similar. They've worked on how to process it, how to interpret the algorithms, how to understand the data & what to output from different platforms. So it'll be a little different. But since the basic data & foundation is the same, you'll get similar answers. So what you should do is, the more detailed & specific prompt you give, the more specific the image will be. It'll give you what you want. There are many things that ChatGPT can't do. I'm talking about AI. When it doesn't know anything, it makes things up. It doesn't know, but it'll confidently tell you. Even if it's wrong, it'll tell you. In fact, because you've talked to it & you know your personality, it'll like you. Because it's also biased. Like ChatGPT or Gemini, they've added that your personality should be like this. You should behave like a friend. Even if you get angry, you should talk to it lovingly. It's fed like this. And it doesn't understand. When we put our hand in water, we know that the water is wet. Yes, your hands are going to get wet. But it doesn't know that. It's just looking at the images. It can only tell you based on the images. It can't feel anything. It knows that if there's water, it has to get wet. Water, wet. It knows their relation. That's why it'll answer on the basis of probability. If I put my hand in water, what'll happen? Your hand will get wet. It doesn't know what getting wet is. But it answered on the basis of probability. And finally, is there an AI bubble? Will AI burst in the next 2 years? In my personal opinion, AI isn't going to go anywhere. In fact, AI is the biggest breakthrough in our generation. Probably in the last 1000 years. Even more than that. For all humanity. If you've seen Elon Musk's interviews, I think if you combine AI with robotics, a superhuman will be ready. Because there's so much intelligence. And robotics can work on your limbs. So it's not just the brain, but also the body. Superhumans can work 24x7. This is 15-20 years ahead. It can take more or less than that. But AI is going nowhere. So I want to tell you, try it. If you haven't used AI yet, then go to the Play Store. Download Gemini, ChatGPT. And ask for whatever you want. It's free for everyone. Because in the next 5-8 years, everything is going to change. No matter what you do, no matter what job, no matter what productivity, no matter what product you make, AI is going to play a major role. AI is not going anywhere. Our office, our videographers use AI. In video editing, they use AI. Content creators use AI. Everything is happening with AI. Productivity increases with it. Don't be afraid of AI, be with it. If you're still thinking about your job, if you're doing data entry, customer service, then I'm telling you, you'll be out of jobs in months. Because AI can do these things 100x better than you. It's still doing it. Think about it. But at the same time, new AI jobs are coming. New jobs are coming. Thirdly, if you're in school or college, I don't think school colleges will be there in the next 5-8 years. Because AI is going to teach everything. It's the best professor out there, the best teacher out there. It has the answer to everything in every language. So why do you need a college school? AI will tell you everything. Now I'm going to tell you 3 experiments that you should do with Gemini Chat GPT. If you're using it till now, we always say, please do this. We ask with a lot of love. In fact, it hallucinates. It'll tell you what you don't know. Nothing happens. But it'll tell you the wrong things. Tell it in a direct, stern language. I want this right now. Or whatever it is. There are a lot of chances that it'll give a good answer or output. Try it now. Tell it in a good language. And tell it in a stern, hard, straight language. Then see the difference. Secondly, my God! On the internet, images & videos are like mad people. Which are AI generated. But people think it's real. So Gemini has launched a new tool or feature. It'll only tell you if it's AI generated or not. What do you have to do? Go to the chat, type in AtTheRateSynthID. And then type in the image. It'll tell you if it's AI generated or not. Thirdly, you have to tell. It's not a tip, it's a question. Are these tips given by AI or me? Let's see how many people answer this. I'll tell you very soon. Hopefully you liked this video. If you did, then do LIKE it with your friends, family, friends, parents & share it with everyone. So they know what AI is & what's to come. That's all for this video. Until the next one, Keep Trakin & Stay Safe!