Internet will NEVER be the Same Again! | Rise of AI Videos | Dhruv Rathee

Hello, friends! You may not have even dreamed that technology will reach this level. Watch this video. Can you guess which place is shown in this drone footage? Which country is it shot in? Whatever your guess may be, it is wrong. Because this video is neither of a real place nor has it been taken by any drone. This is a video made by Artificial Intelligence. All these videos on the screen all of them are fake. This man wearing a cap or this man reading a book, these people don't exist in reali

ty. They have been generated by the computer. Now it is almost impossible to tell the difference between what is real and what is computer generated. And now, to make videos, you don't need a camera. You can make such videos with only a keyboard, in a matter of minutes. All of this has become possible today, because of Sora AI. "A new AI platform called Sora." "You can't hide from these changes." "No country, no economy." Sora AI is an Artificial Intelligence software, developed by OpenAI. The s

ame company which shook the world 1.5 years ago when they launched ChatGPT. In ChatGPT, you have a text-to-text conversation with AI. You write something in text and send it to AI and the answer is generated in text. After this, OpenAI launched the DALL-E software which was text-to-photo. You wrote a text prompt and the type of photo you wanted to generate and you got an AI-generated photo. And now this Sora AI is a text-to-video. Now you simply have to write in simple English about the kind of

video you want to generate. And it will do that for you. Like, if you say, "I want a video of a young person who is reading a book while sitting on a cloud." Write this prompt and in response, Sora AI will make a video like this in a few minutes. The videos I have showed you, text prompts were written to generate each of them. Some of them were generated by a long text prompt, while others had a short prompt. But Sora AI can not only generate text-to-video, but also photo-to-video. This is a ph

oto of a Shiba Inu dog. If you input this photo into Sora AI, this video is generated. Not only that, it is also possible to do video-to-video. Give it the original video here, where a car is driving on the road. And if you tell it to change the environment around it. Like, you want it to look like the 1920s. While keeping the car red coloured. So this video is generated from the original video. Do you understand what this technology means? Entire industries can turn upside down. Animators, Vide

o Game Designers, Graphic Designers, what is the need for them now? Stock photo platforms, stock video platforms will be completely destroyed. Why do you need to film for weeks for making movies, when in a few minutes AI can give you footage for the movie? We will talk about how this will affect the world later in the video, but before that, I would like to tell you about the level of this technology before the arrival of Sora AI. Exactly one year ago, in March 2023, this video went viral on th

e internet. Will Smith eating spaghetti. This video was made using AI software of another company, but a year ago, this was the peak of this technology. After this, some other text-to-video AI tools were released to the public, like Pika Labs and Runway. Before the arrival of Sora AI, Runway was an industry leader in this field, but look at the quality of AI-generated videos generated by it. There is a vast difference. After Sora's release, actor Will Smith was so impressed with this technology

that he made a video of himself eating spaghetti as a joke. To show the level of technology before and after the launch of Sora. Sora can now understand how our world works, how motion works according to the laws of physics. This is the reason why, when the puppies are moving in this video, it knows how the snow on their heads should move. It is now able to understand emotions too. Watch this video. When this cartoon sees this candle burning, you can see the excitement play on its face. Sora is

able to understand even the smallest details like how elephants walk, the exact motion of their feet, this is why these videos look so real to us. You might be wondering about how Sora is able to do all of it. So, its basic functionality is actually very similar to ChatGPT's. Where on one hand, ChatGPT uses tokens on the other hand, Sora uses patches, its equivalent. Whatever you ask ChatGPT, it divides the text in tokens and it generates its answer through tokens. Similarly, Sora AI tries to un

derstand photos and videos, by dividing the visual data into visual patches. And even while creating its output, it uses these visual patches to make the video. I explained this process in detail in my ChatGPT course, about what exactly are tokens and how they work. If you have taken the course, then refer to the explanation in Chapter 2.3, the visual patches can be understood as an equivalent of tokens. And if you have not taken the course yet, then do check it out, because this technology is c

hanging the world. It is disrupting industries. If you don't learn about this, if you don't use it in your daily life, you will be left behind. The first two chapters of the course are quite theoretical, in which I explain the concepts behind it. What is the technology behind ChatGPT? How do tokens work? What is temperature? What is the RLHF method? But the useful things begin in Chapter 3, These are the practical lessons. In Chapter 3, you will learn how to use it in your studies. For example,

if you want to go study abroad, you don't need to waste money on a consultant. Use ChatGPT. For example, you can use it to practice for a job interview. Chapter 4 is focused on personal life. How can you use it in food, diet planning, health and fitness, and travel planning? Chapter 5 is my favourite. It is focused on business owners and entrepreneurs. For example, you can do market research and analysis for your business through ChatGPT. You can use it for Customer Feedback Management. Chapte

r 6 talks about ChatGPT's limitations. And now, there is a Chapter 7 for you. It's a Bonus Content of an hour-long workshop that I hosted in Dubai's 1 Billion Summit. In this workshop, I taught the attendees how to use ChatGPT's full potential. But not only this, in April 2024, I will add another major update to this course. There will be new chapters and new lessons because this technology is changing so fast and if you buy this course once, all future updates are free for you. So if you are i

nterested in joining, there is a special discount for you in this video. You will get 40% off. if you are on of the first 400 people to use the coupon code SORA40 You can get the link from the description below or you can scan this QR code. Now if let's get back to Sora, at the first glance, you will definitely feel that this technology has reached perfection. But if you look at it closely, you will see that it is not 100% perfect yet. No doubt this is a great technology, it is absolutely unbeat

able, but not yet 100% perfect. Why am I saying this? Because in the videos I showed you, look at them carefully. The man who is reading the book, look at how the wind is moving the pages of this book in the wrong direction. The astronaut that's walking, look at how his walking speed increases suddenly. Open AI openly acknowledges the weaknesses of Sora AI on its website. These AI-generated videos sometimes fail to apply the physics of the real world. Like this basketball, which is passing thr

ough the hoop. Or the three hands seen in this video. Or this video of a grandma celebrating her birthday, it may seem like everything is fine first, but look at it carefully. How are the people reacting? Look at it more carefully. Did you find the mistake? Then look more carefully. The way people are waving their hands in the background. No normal person waves their hand like this. So, today, these are the minor details based on which it is possible to detect if a video is AI-generated or a vid

eo taken in real life. But the speed with which this technology is improving, I don't think that one year later, it would be possible to detect this anymore. And the people in OpenAI know this, the day these videos become so real that no human eye will be able to differentiate between these, that day, huge problems will be created. Fake news would be spread so easily. Perhaps, this is the reason why Sora AI is not released to the public yet. It is being tested by Red Teamers for security concern

s. This means that it is in the testing phase. Red Teamers refers to those people from Cyber Security who behave like attackers. Their job is to find the system's weaknesses. But even when it would be released to the public, I don't think it would be possible to manipulate photos and videos of any real person because that will always be a major security risk of people using it to make fake videos of any politician and that could cause election interference. But this is a thing that is becoming

a big problem even before Sora is released. Watch this video. This is a video of Virat Kohli in which he is promoting a betting app. "*** is a type of game which can be called an investment-worthy app. Or I can call it an investment-worthy game. In which you can spend a little money to get maximum winnings." Could you recognise that this wasn't a real video? This is a 100% AI-generated video. Some fraud companies have already started exploiting this. Celebrities' faces are used to run these ads

on social media. People are lured with money, and told to download their app. The same thing happened to Sachin Tendulkar. "My daughter is playing this game now, everyone is talking about it, ***. She plays games on the ***** app, and earns ₹180,000 every day." Another completely fake video, but did you see that the technology has reached the level where it is possible to use someone else's voice. For now, there is only one way to avoid such videos, watching these videos carefully. You will alw

ays find small mistakes. Somewhere, the lip movement will be strange. The quality of such videos will be low. If you try to listen to the voice carefully, you will find that the voice is not like the original. It seems like an artificially synthesised voice. Listen to Sachin Tendulkar's voice again. "My daughter is playing this game now, and everyone is talking about it." You can differentiate by noticing the small details, but most importantly, look at the video content. What is being promoted

in the video? What are you being told to do? "This application is absolutely free, any iPhone owner can download it." "The plane flies at 2X, 3X, 4X, 5X, etc." If a celebrity tells you to download any app where you can make 10 times the money you spend on it, this should alert you. Whether it is an AI video or a real video, you should stay away from such things. Apart from this, a technological solution can also be created for these problems. An AI whose job is to identify AI-generated videos.

OpenAI has said that the videos generated by SoraAI, would have a watermark on them. With that watermark, you can always identify whether a video is AI-generated or not. Personally, I use AI-generated photos or videos often in my videos. In such cases, in the top right corner of the video, I always write Representational. You must have seen this written in the corner of my videos often. This is so that you can understand that the photo or video on the screen is not real, it is AI-generated. But

an even bigger issue is the jobs. What will be the impact of this AI on industries and economies globally? Even in films, filmmakers will not have to carry big cameras and film in different places. In a few years, entire films can be made using AI. Last year, the director of the Avengers films, Joe Russo, famously said that it will take only 2 years for AI to reach the level where it can start producing entire films. Similarly, it will have a huge impact on the gaming industry. The video gam

es that took years to make, needed teams of more than 100 people. How easily the same tasks can be done with this. According to this NBC article, filmmaker Tyler Perry was going to spend $800 million to expand his studio. But as soon as Sora was released, he stopped his studio expansion. After all, what is the need for a studio if AI makes videos for us? This means that many jobs across different industries, are at risk of being eliminated. The cinematographer who needed many days to make 2-min

ute long stock footage, and earned money by selling them, or the VFX artist who made the NPCs for major gaming companies. In the short term, the jobs related to film and video production, advertising, and marketing will be affected the most. But at the same time, there is a positive side to it. The good thing will be that we will see democratisation in all these fields. These fields will become more accessible to the common people. If your dream was to become a film producer, but you didn't hav

e the money or the equipment, now you will get the opportunity to show your creativity through Sora AI without these resources. Now even an IT employee will get the opportunity to become a good filmmaker and a musician will get the opportunity to make music videos without spending a lot. Looking at such technologies, it is understandable to be a little scared of them. Do you know, friends, that 600 years ago, when the printing press was invented, many religious leaders were opposing it. Some p

eople used to say that the books are actually given to us by the devil. That books should be banned. In 1501, the Pope, Pope Alexander VI said that without the permission of the Church, no onw was allowed to print books. One thing is clear here, whether you like this technology or not, it cannot be ignored. All people and all governments have to learn to adapt to this new world. Many governments around the world have already started taking proactive steps. Like, in Singapore's Parliament, rece

ntly one of their MPs gave this speech on Sora AI a few days ago. The politician's name is Tan Wu Meng. He said that the Singapore government has now launched a new subsidy to adapt to the new AI developments. This subsidy is for people above 40 years of age so that they can pursue a full-time diploma once again. Because their government has understood that today, if you study something in a college or university, whatever you are being taught today, will it still be relevant 20 years later af

ter technology transforms so drastically? So it is better to tell people to educate themselves again after 40 years. To be able to adapt to the new developments, by upskilling themselves. Interestingly, the government of Singapore had started talking about AI in 2019. And in the latest budget for 2024, they have pledged to invest $1 billion in AI. "In the National AI Strategy 2.0, I will invest more than $1 Billion, over the next 5 years, into AI computing, talent, and industry development." So

, this will be my suggestion to you. The sooner you learn to adapt to this changing world, the more benefits will you reap. But I would like to say one more thing about AI, since 'AI' has become a very trendy word nowadays, many companies are exploiting it and there are many useless AIs present in the market today. When I made my first video on AI, at the beginning of that video, I showed you many things that could be done through AI. Like changing your face, changing your voice. "And if I want,

I can also change my voice with the help of AI." But after taking my course, many of you gave me the feedback asking me why I made a course only on ChatGPT. Why didn't I talk about the other AIs in the course? This is why. Most of the other AIs are just for vanity. They have no practical use in life. Like during the face swap, when I changed my face to Tom Cruise's and Hrithik Roshan's I used a tool called Akool Face Swap. It was an expensive tool, I had spent a lot of money just to get the o

utput of a few seconds of footage. The result was not that good either, because you could clearly look at me and say that the face looks manipulated, it doesn't look real. And there were no other features in it. That's why such AI tools are called Gimmicks. They are used only for show-off. But these won't help you in your studies, nor will it improve your efficiency, and nor will it help you with your business. That's why I focused the most on ChatGPT. Because ChatGPT is actually life-changing.

ChatGPT can be used practically in many aspects of our lives. That's why the entire course focuses on it. And recently, many Indian AI companies have popped up, claiming to have made the first Indian AI, that is later found to be using ChatGPT in the base, at the back end. Apart from ChatGPT, Midjourney, DALL-E and Stable Diffusion are the three text-to-photo AI software that are actually useful. And I think, in the near future, Sora AI will be as revolutionary as ChatGPT was. So if you want me

to make a video giving the overview of all these AI tools, then you can let me know in the comments below. And for now, the link to download the course will be in the description below, with the coupon code. If you liked this video, you can watch my previous video on AI. It talks about Q* AI. In it, I explained what is Artificial General Intelligence. Click here to watch it. Thank you very much!