A serious blover.
Aren't you serious? You seem to be the most serious bloger of all. Your balcony.
What? What? So keep it down. The main thing is, a serious blogger, the main thing is that a serious blogger from the seventh floor doesn't get anything down.
The point is, the bloger doesn't fly away. Everything else doesn't matter.
I've got it. I'm here. I'm sitting on this balcony, and I feel like I'm gonna get in with him right now because he's on my knee.
Yeah, the balcony for the flower, not for you.
He's not, he's Paris. Paris.
And the Parai. We need to put a song in.
Paris, Paris. Yeah, Tanya, all of this.
I'm all right. I've had a day.
I'll go get you someplace else, Tanya, I'll be sure.
Cuddy?
It's obvious, it's gonna be helpful.
Good luck!
And that's eight hundred million, that's $2 billion, they're attracting funding now. Althman said he'd be porn on the Internet, ChatGPT. We just discussed in France, Italy, the feeling that the police are just driving around with the town blinkers. That sounds very accurate. This is Sanja, Sas, Tanya, as we do, California is driving this, which is a firefighter driving. It's such a distinctive sound. Yes, very distinctive.
It's very loud compared to this. It's so melody.
What a melody. I was walking down the street today, the policeman was driving by. I thought I was deaf. I thought I was deaf. They seem to me to be more important than they are to make everyone swallow.
Look, Sasha, remember how a traffic in Moscow on the Rublev-Uspen Highway this morning, they don't sleep from the blink. This is all the stuff that's getting weird. And California and Paris.
Yeah. So what's the subject you want to start talking about today?
Still, considering what OpenAI did the peddate there a week ago, and that you can make your apps, and the apdite was before that, which is what you can do at the store. And plus they've made a system that can create their agents. I think it's a very-- I want to make it clear to the audience again, yes. That's a little more to go. Illnar, that's the subject. I've been talking to people all week, asking questions. What kind of app are you planning on doing? And, among others, for people who see us, who not only plan to do, but who generally wants to figure out what's going on in artificial intelligence now. We'll have a special release on Wednesday, by the way. I've taken a special episode on the subject just like that, the best action you can do in your life in terms of artificial intelligence. Come to see Wednesday. I'm not going to go to bed. So what the topic was to raise. And we talked about it, and I was talking to my partners last night. From the point of view of the development of annexes. Should we just throw up and make annexes inside ChatGPT? You remember, Illnar, when the GPT showed up, you and I are right there. I just think it came out on the same day. You and I started making an app, started thinking what we can do. First, we were thinking of making a financial application. We even did, yes, there's some stuff. Then we thought we could make some annexes to HR, and it turns out that, uh, in fact, OpenAI had a very raw product that was unproductive. I think a lot of people spent money, and they kind of, like, don't work. And the second story, why would people, what kind of user experience would people have? Because yesterday ' s discussion had given me the feeling that, in fact, the current mobile applications, they had a very different user experience in their own way. I'm not talking about mobile applications, the payroll or mobile applications, I don't know if there's a CRM or a disk manager. I'm talking about applications that provide different formats for some kind of service in some kind of, uh, some very simple stuff.
For example, an annex to see where your plane is flying, whether it's an app to pick, pick, I don't know, some kind of route. Well, I mean, here's the information service. And right here, you think that if I'm in the GPT chat room, I'm gonna make this app, then the client, the client, he won't pay the money. I mean, why would a client pay money if he, if he could use GPT chat and he could get to the app, he's obviously a literate man. And why would he have to separate the annex to find out, for example, where his baggage is at the airport or to find out what the weather is today. He also simply learns the weather through GPT chatting in a much broader format. Here. Or return to his old standard application, to which he's used to. Because I still told me two years ago that the OpenAI would obviously be doing its operating system. And this heading that was online, that OpenAI is doing its operating system, is worth an app. It's more like an operating system than an annex, yes, they do. The audience wants to ask, what would you see the apps now? Because it's a time when, if this topic comes to work, if OpenAI doesn't say it in three months again, it's not a big deal. Yeah, and if this thing works, you can be the people who will be at the dawn, there, these applications, they'll be used by people, and so on. But while it seems like this system, it only needs some very big operators who add some interface to it and it doesn't make sense. Even if you take the Salesforce. Salesforce a few days ago, he had a huge conference. It may not be a Russian-speaking market, but it is still the largest CRM operator in the world of systems, uh, task management. They have fully integrated with various LLM systems, in particular with chat-- with GPT chat, Antropic, and have established their agents system. So now, in the Salesforce, you can turn around and make a bunch of agents. They even said they, they have, they have, they're all through Slack, including the possibility of creating a huge amount, for example, inside agents. And the question is, why would I have an annex Salesforce in the GPT chat?
OpenAI has AgentKit, yes, when you're using, uh, small blocks, roughly, a mouse can rigate the agent's behavior, and then build it inside your existing application. Yeah, well, you have a parole, I don't know, school, there, with some educational courses, you have your mobile application, and there's an extra service inside that mobile application. I mean, some kind of chat with some behavior. It's like a story. The second story is when you put your own applications in the GPT. Well, that's how the Convoy case was when you can glue Canva, and she's doing something there. You're talking more about AgentKit now. Do you understand?
I'm talking about the possibility of an application. They just have two things there, too, inside this AgentKit, as far as I understand. First part of this is about making agents, which is where you put in the app, the application interface inside the GPT chat is coming.
Again, again.
I don't have a fee, you know, $20,000.
It's not just about the story. I mean, they seem to be rolling the test version.
I think it's just about.
Elnar, I'm just saying the exact same thing as the interior application in the GPT chat room.
Then, in fact, about the app, yes, if we say that inside, uh, our standard interface where we write messages, there, in the GPT chat, it's answering, we can actually tag some apps, and these are the ones. The apps are gonna do something else. That's how they showed the example of Convoy, yes, there, the Salesforce, too, is probably another application that could be built. Oh, that's obviously gonna be helpful, and in general, they're only gonna be used by major companies. So the same Canva, which already has a large number of users, you just give her another point of entry into your application. So someone has a canva installed there, a web-based account, you can do it. And those people who have already gotten used to working with it just get another point in the entrance. You're inside the GPT, you can tag the app you've always used, and just start it, in fact, through the interface you're already used to there. Because many things are in GPT chat room. Oh, will that be common to small companies? My opinion is not. Just because there's a very small number of users on, you know, some small companies. We'll take something out there, I don't know, local, uh, such a small school, uh, those, those that teach anything that doesn't matter, but how much does she have? Well, there's a thousand, two thousand, five thousand students, yes, there's a lot of them.
But to support this interface, therefore, to put it in the GPT chat, to keep calm, to do something about it, to do something about it, to--
To fix, do something about it. For a small number of users, I don't see much sense in this. It's better to look at AgentKit and, on the contrary, add ChatGPT in its annex. I mean, not your ChatGPT app, but, like, the reverse side of the sale-- look. For some super-city, which, as a result, already has a large user base, add another entry point for the same users. Maybe a plus-minus someone's gonna come up, but I doubt it. I mean, a man who, I don't know, never worked with Canva, I don't think he'll find out that Canva can be called, but only with some presentations, yeah. It's where we showed it, we're talking about it now. And some other applications will appear there somehow, and they may be called. Well, no one will know if you don't use that app. If you use them, you'll probably know. Some integration will work there. Here. And as for a little bit more business, yes, if we're talking-
Yeah, but I want-- now I want to add. Except the agents are still apps. So they, uh, generally plan to do their system inside the installation applications, modify them somehow under their internal infrastructure, under their internal visual. They're probably gonna unload some visual elements inside. There was a discussion that there would be some area before there was some kind of coincine or commercial, and something else. I mean, they're basically trying to make an internal application system. That's what I asked you at the beginning, including not about what's right now, right now, where's it going?
Uh, what mobile applications, uh, can, uh, be useful when you're inside ChatGPT? It's like a checkout. We came in yesterday and tried to buy something on Etsy. I mean, there's a system of different goods, Etsy, and I think the Russians are called analog... What's that called? It's a system where you can buy.
They're--they're their master, something like that.
Yeah, something's good.
Some master, yes.
Well, it's a hendmeid of goods. I mean, it's--
Yeah.
Not a mass story, but...
Well, there's different things. Yeah, there, yeah, yeah, yeah. And Etsy is different. She's in Europe, too, and she's in Europe, and she's in a big system. They've done this integration, we've been meaning to buy it. And the point is, you're like, why? I'm buying it, I'm having a guest account. I say, can I use my Etsy account? I can't. I'll go in and get the goods. Then who am I gonna claim to be? Is it like if I buy a plane ticket to Expedia, I call Delta, there, or United, I say, or whatever, Air France, and say, "and you'll change my ticket?" They say, I'm from the people you bought. I mean, if you bought a ticket from, uh, Expedia, you'd like to go to Expedia. Yeah, you don't... And this is where what's going to happen? I'm gonna be the service to whom-- I'm gonna be at ChatGPT, so call? Well, obviously not. I mean, the function--
Sam, I'm sorry.
Yeah. So, and therefore from the point of view, and therefore from the point of view of the annex, what is the annex? The app, the right thing they're doing. Why would I use that mobile application? So what value does it give me? People might be able to develop some kind of cool thing there. I'm here to go through some of my training projects, even in the sense of, uh, there, so I don't know, my students could get some kind of information on the ChatGPT base. But there's a question here, too, why do you do it inside?
And here, Sasha, the system with the agents is much closer. AgentKit they showed when you were inside-
Yeah.
Your mobile application can be built by an agent. Well, that's quite simple.
You still need a programmer.
Yeah, but that's different. Yeah, but these are different things, right, Ilnar? One thing, the existence of a store inside ChatGPT is me, well, this piece. It's clear that the construct they've done is gonna tell you now, and she's very interesting, just to keep them out of the way, they're totally different in logic, by their own.
Yes! Because, look, just using an agent, they're making it convenient now to use the interface separately. You used to use ChatGPT on API. Okay, now, I know that this has made it easier now. But the fact and the interface have made it so that they can create such agents on their side very easily. Just the fact that it's in the mobile application inside, uh, what are the mobile applications? I mean, if we move on to the fact that I'm ChatGPT now is my main application for the day, I've been using a cell phone much more than before because ChatGPT showed up, and I'm in the middle of it. ChatGPT is a base-- ninety percent, I guess my time is now. Well, if you don't talk about the meetings I'm going to have. Well, I don't know how if I went to Google, there, or I would have read the mail. I'm doing a lot of things in ChatGPT right now. I mean, in theory, I can imagine that there are some apps, but what? And wouldn't that do it by ChatGPT, would it? I mean, the OpenAI will turn itself around. That's a request, I guess. Just before we're at agents, I mean, Ilnar, we'll move, yeah, so the un- and the audience will also take us off, looking at the sco-- as you can see, this function, as you see it alive, as long as you see it, as you can see it alive, as long as you can see it, Do you think that such a system can turn around? Because basically it's like me when yesterday, uh, I read that information that Salesforce launched the creation of your agents, so you're like, uh, it's some kind of micro-functional on the base, there, Salesforce, right? And it's some kind of micro-shot, it's not a integration into the vast space of these mass-produced operators. There's a ChatGPT with eight hundred million users. Obviously, these-- these users will be getting more and more, and more.
Here.
I can tell you as the man who uses Canva every day. I'm much more comfortable opening Canva and doing it there. And if I need some kind of fraimvor, right? That's how the idea that you're not gonna be a full presentation in GPT chat, is that you can do some kind of base, right? Oh, I'd rather use AI Mode in Canva than GPT chat, but I'm at least so comfortable and somehow better. I mean, I don't really understand the point.
Canva, Canva, Canva, yeah, by itself. But I mean, not AI, it's still a question.
It's like if you use it, like you say I'm in the GPT chat room, I want to use the environment to develop, to write the code. I'm going to Cursor to write, for example, a code in Cursor. But I can't connect my coolest GPT Pro chat. Is it possible, Ilnar, to connect or not to Cursor?
We should watch. API seems to have come.
I don't think you can.
But I don't use Cursor.
Yeah, API, but I can't turn it on, like I already have a plausible version of GPT. I can't, I can't, like, my signature, and my signature includes, for example, the use of extended files of the code. What's the code for me? He's not giving me the ability to do the code as Cursor gives. And as much as you're here, and, as a result, you or OpenAI have to produce some product inside, it has to be inside and use your memory, use your expanded possibilities. This system. Like before, for example, people liked to use it, I don't know, Perplexity. I always say, why? Why would I use Perplexity if Perplexity doesn't give me access to the GPT read? What system does it give me, I'm sorry, about the GPT version? Well, like API about the version. So what's the cost? Every request is tens of dollars. No, no, no, no, no, no, like a system.
It's clear that the details, but the commentaries haven't given us a whole panic. Codex can be connected to Visual Studio code, and we'll get a plus-minus story. So you have a code inside Visual Studio, inside the IDE-kills, it's gonna work. But, I see, it's details. What you're saying, yes, we still have a specific model to do so while we're inside the company's ecosystem, right? There's the coolest GPT chat inside the product in GPT chat, it's gonna be hard for him to get something to work out.
I'm just thinking, we're in a common way for everyone. I'm thinking about people now. Here's the eight hundred million. Not from a unique perspective, there are a few million staff who use or, there, a dozen, maybe a couple of dozen, but from an eight hundred million people who are. And the OpenAI chat is gonna roll on them and there'll be billions.
How will these people sit down every day and sit in this system inside, right? That's what I'm thinking. But there's something about the agent, Ilnar, to add to the people's understanding again. We told you last time, but maybe it's important because they've now pulled a convenient function, yes, for development.
I'm just saying, if there's a business, yeah, there's another basis here, not that you've come up with some agent, but rather you already have a functioning business and an app, there's a mobile or something. Something you're communicating with your clients, so you can add it in. Yeah, that's what you're saying. And before that, it was also possible for API. So you just, you know, you need to program a little more, you just need to put a little more logic on. Now, it's, in fact, there, a mouse moving different objects, connecting them with their shooters and, there, intuitionally understandable interface, you can generate for yourself, well, a little more advanced behavior of your model. I mean, maybe she'll be there with your files, get the information from there.
From there, the information is delivered, um, maybe with some calendar, your calendar will be synchronized or something to do. Yeah, and it's just enough to do it now. And then your programmer who made you a mobile application or the company that created it, it's quite easy to build it. So you're not really gonna have to hire a separate person, I don't know, date-saientist, who's gonna be making an agent, generating something there and so on, and then with your staff. To connect him to this operation there to add that he's editing it there all the time, something else. So it's all gonna be a mouse. And that's what you mean, if you're in there, I don't know, there's some kind of schooling, right? And you have a lot of clients you want to add a new function, you know, you're actually using agents. The only problem is how ChatGPT agents use API ChatGPT. Here, we put up prices, make it possible to make it available from different regions of the country, ow, from different regions of the world, yes, there, and so on. Here. If this is all arranged, an additional function could be built in an annex in a fairly light manner.
You'll have a tip that will be at a pretty high level, so be careful to answer your users.
Well, great. D- move on. I want to, uh, show you right away, some of the things that concern, uh, the market, which is the artificial intelligence market. Now, you're gonna pay a little attention again. Walmart announced integration with GPT chat. The stock grew up here, didn't it? The market is still on. So, HAI Ilona Mask announced that they were twenty, $2 billion, they're attracting funding now. And then some of these moneys are half--- to give NVIDIA's investment by getting their share. So, everybody started to change their shares and all started to grow up, and then they're getting high. And from the most interesting news about the cost to understand the money breakthrough, I'm gonna show you what I'm doing right now, right? It was Meta who bought a man from Mira Murati. Uh, I think it's like the media that writes, it's all considered that the package is more than $1 and a billion. The package is five or six years old. The man gets, uh, Meta. They wanted to buy this company Thinkings Machine all the time, but they didn't sell it. And the question is, why sell it? They'll sell them this company at STMostech. Here, they'll just sell it more expensive. In parallel, Reflection AI, with an estimate, attracted two billion dollars. Uh, I think two, too, yeah, they're attracted? Yeah. Two, two billion dollars. They also have 60 engineers working super cool from the OpenAI, from DeepMind, shorter, from all of it, right? This is DeepMind's exits.
The company was set up for a year, the evaluation was eight billion, and two hearts were involved. What am I saying? Oh, and the information. OpenAI has, uh, set up its own route that they'll be attracting trillion dollars over the next five years because they need to spend more than trillion dollars, they're planning to spend. Uh, remember, yeah, the beginning of the year, when Trampa had them standing, they made this whole deal, Oracle, Arm, SoftBank, OpenAI, five hundred billion. All of them, Pfff! 500 billion. Where will they take them?" So now is the trillion. It's just OpenAI. It's just OpenAI. And, uh, what do I call it? There's a huge break, uh, between the cost, between the start-ups, the $30 or $40 million in small phases, and these big-level launches like OpenAI. And I don't call Meta there or Google or Microsoft because it's still companies working on the market there for 30 years, right? And here, well, there, or someone twenty, there, years. Oh, uh, when there's a gap like that... Before, this gap was like America and Russia, for example, yes? Well, in Russia, I don't know if anyone was given $100,000 on the start-up, all like, "Oh! Happy." So, $30 million in America was given here. Well, it's like, "Well, okay." And now the break-up is so big that it makes sense, it's totally pointless to invest in, uh, this little venchour. Although the likelihood of falling, the likelihood of these companies is still narrowing. The market is captured by giants of artificial intelligence. Well, what was predicted in the past six months, huh? It's a big deal. And now the money goes to infrastructure. And of course, the prices are getting hurt. Well, that's the sums, the amounts are affected by the government, the economy, the investors and all the others. And this week, by the way, there was a lot of news.
I just think they're not really important to people there. Intel's making a new chip. So there's another race of Intel and AMD from the perspective of NVIDIA. China just, uh, they've announced, so they've announced here that, uh, new hygawat factories are building. China has announced that we'll build three times more and now we'll finance less. Well, here's the American race, the USA, the USA and China on the money, of course, it's a hit, including the cost of the starttaps. Napom-- who knows, there's a Polymarket startup. I don't know if you ever traded it or not, but you didn't... trade it, and the stakes were or not. We're bringing it periodically. Yeah, they're attracting two billion dollars now. They're now attracting two billion dollars in the eight. By lure, huh? So, and with it, the company connected to-- this is the operator of the New York Stock Exchange ICE Intercontinental Exchange. Of course, the numbers are being chased in a crazy way. Although the NASDAQ market is jumping this week in parallel, the last six days are jumping, given that Trump has stated that China is introducing new tariffs. China said, "We're introducing new tariffs, rare earth metals, all." NASDAQ is downloading. I've got it in my own right there every day. You better not open it. You better not open it. I wanted to give this part to you to know in terms of money, because there's a lot of discussion that there's a bubble, so they're not gonna have money like that. They have money, lots of them, much more than they were before, much more than they were a year ago. They're gonna improve their systems, not worse, but improve. Because someone says, "There's a thousand people fired there or something." Well, they fired somewhere, and someplace very serious, they're really moving.
Let's get to the subject. Sam Althman two days ago, now I think this is a very serious subject, said that, so they're worried about limiting the system to people who are mentally ill or have some kind of something. Health problems, mental health disorders, mental state, yes, man. And plus restrictions, different for children. They therefore impose very many restrictions. And he signed it up. But we're doing the peddate, there's a peddate soon where you can get the verification that you're grown, and you're getting adult content, adult content, including erotic and sex. So the whole Internet just got down because Sam Althman said it was gonna be a porn on the Internet, on GPT chat, what did he say today that you were fighting for porn? We're just showing that we're doing company for everyone and we're not going to limit anyone. So we'll limit the children to mental health in some way, I don't know how they'll determine it. Oh, so for-- yeah, and we're not gonna limit it.
But, by the way, you know, there's an interesting thing that in many countries sophths, as well as platforms are huge, so there's e-commerce different platforms or some kind of forum. There are countries that say: the forum should be responsible for what users do, write by users, for example, in the forum or how they use it, or how they use the site. America has a law of ninety-fifth year, and literally last week, there was a new case, in which the federal court decided to go like a forest. There's a law of ninety-fifth year. According to the Act, the e-commerce site is not responsible for the actions of third parties. So if third parties came to this platform and laid down, sell some illegal goods, the area of responsibility itself is not. Accordingly, OpenAI, I think it's the same mechanism. They say we make a platform, you can use it differently, but it's your kind of topic, how you used it. Someone fooled you, didn't cheat, you made something up, you didn't come up with it, did you?
And if they're localizing in different countries, it's not gonna work in Europe, for example.
It's not gonna work, yeah, it's not gonna work. Look, Europe is actually interested. Google, of course, on the one hand, they're fighting it, on the other hand, I don't understand how they fight. For example, Google in Great Britain is occupying ninety per cent of the market. To remember, yes, in Russia, not only Yandex, there's Google in Russia, about 45 tyre fifty percent. So almost half the Google market.
I mean, almost half the Google market. It's not just Yandex, it's British 90% Google. I'm gonna go to the Russians. If someone in the commentaries can correct if maybe 40, 40, 50, I don't know now. The British has added Google now, or, a week ago, it added Google to the Special Forces where they have the right of Google to make any demands. Say, "We think you're incorrect, giving out there, or we think you're too much of your service moving." What Google said to them, "Okay, we'll follow your restrictions, but we'll just-- we're not gonna blow up the peddies, for example. Well, you'll be a little late for the peddates.
Logic is simple: if you have restrictions in the law, please limit yourself. We're just not gonna turn this around in your country. I'm, like, in France, ChatGPT-42, I'm not available. Look, I'm a totally American user. I have all the American variants. I mean, I've got a full-scope check on the American documents. I've got 40, I don't have any sales services inside. I'm always saying, "You're in France. I'm sorry, we're not gonna give you this service." I've never encountered this in America. I didn't even know what it was. He's determined that I'm in France and writing me in my usual chat room. Look, Ilnar doesn't write to me like that, not... Well, you might know that, but I'm just not. He writes to me in the chat, straight answering me in the chat, says, "I won't give you that information because you're in France." I thought there, I don't know, the window was coming up early, but I got a little garlic in Sora two. I'm like, wait, I want to create a video at Sora. I'm on a mission, right? I need to create a video. No, you can't.
Their cream.
Sora's been rolling into America, and you just, uh, need to show your device that you're in America, you're gonna have it accessible and--
Well, that's weird.
France has already been.
Yeah, yeah, yeah.
answering or not answering, that's right. This is likely to be followed by some local restrictions that are already likely to be imposed from the localization in France.
So, from the point of view of the law, you're right that these new systems... And our users, too, write off, what do you think of the laws? We need to limit it, less limit it. What do you do? Because, of course, the market is smacking on all sides now. I mean, they're in, so California's law is now passed again, and, uh, it's pretty serious. The fact that using someone else's, uh, images of people and someone's voice in artificial intelligence through artificial intelligence without consent. It's up to criminal responsibility. And they took it seriously, didn't they? And yet, it is to keep the balance between not limiting and limiting. So, as long as you can see in the last year of the rolling system, you're doing what you want. Some micro restrictions in their countries are being made, and the world says, "You don't want to just pour them out." So you don't want to use the OpenAI in Europe? Well, please, lock it up. Although they block it, if Samalts meets, I think, with all the right-- with any government in Europe, he's completely open, I think any man will meet him. Who's not gonna meet Samaltov now? From the perspective of the world government, I think it's super-access. And although, by the way, there was a case. Microsoft is the first time that... No, Google. Google is a precedent. They restricted, uh, use, uh, Israel access to their cloud AI servers because they used them to cause damage to people. They said you're, like, causing damage to people on the ground. Well, I see they're talking about Palestine, and we're gonna cut you off. Attack, well, that's a totally extra-political story. Like, look-- well, it looks like an internal rule of its own and a separate restriction of its own. So companies are putting their own restrictions on themselves. You want us to cut your piece down here, you're not gonna have it. What's going to happen to that? Uh, how, uh, how's it going? That's very interesting. Very interesting. Also write what you think in terms of the use of such things by the Government and States. Because the United States, for example, has full access to all Anthropic, OpenAI, Google, heAI for the dollar, I think that he's made a dollar for any government agency and everything. And I'm sure they have a system of an even more enlarged Haracruter. There, I think it's not just Pro. You, Ilnar, said something.
Yeah. All countries, therefore, wanted data centres to be established in their region or in their region. So the Arab countries first did, and then in Europe, respectively, they started building these things in Britain, too, right? I mean, if you have any chips in your territory, all this infrastructure is built and so on, even if it turns out that I don't know, there, you've been blocked from access to models or something. At least you have infrastructure, and you'll turn something worse and invincible so you don't get addicted. Well, that's how the situation with Israel, for example, you said, right? So they probably don't have any ODs on their territory, they use others. And that's why they just got caught and closed down on some kind of decision. I mean, I don't know how political it is, not political, but, accordingly, it is.
Many countries therefore want to have and invest in the Territory to have data processing centres in their territory, all infrastructure to come to them.
Sasha Masrapova is not with us today. Uh, about the video generator. In Europe, a whole problem has now been identified. There's a-- a lot of online feathers. With a video of the generation of artificial intelligence about Russian-Russian- Russian- Russian-like, Russian disinformation. There are cases where, um, the use of the gene models to create video-rolics that portray black-person crimes in European cities. So, a huge number of these rolls are dropped on the Internet, and people are watching them thinking they're real. The wheels aren't real, they're faety. And, uh, this is, of course, the subject here, well, not the black question, maybe some white people, somewhere different than the other people here. But the fact, uh, sam, that it is possible to make such an advert, is the fact that such a huge advertising campaign can be done now, especially in countries where, but less, information protection is available from a point of view The view of the serious systems in some small countries in the world seems to me that, of course, this information struggle, based on artificial intelligence and influence management, will now be catastrophic. I mean, you're getting it, you don't know the end of it, it's a video, it's real or real.
Actually. Actually. And this case that writes, it, it's very serious, because, in fact, there's a system that requires mandatory labelling of artificial intelligence content. There are now many countries that need, there, in different areas and so on. On the other hand, those who do disinformation will not mark anything. And if they don't mark, then the question is how to determine. I mean, when you add, like, a real video to some extra object or something, you change like a human visa or you cheat on people, even a law-- you committed a crime, and you took a Russian-based shot. Changed his skin color to this man, or, uh, made him Asian or European, or anyone, huh? And he said look at these people, they're causing problems in the country.
Look, there were some fairies a couple years ago. There's a parole in the Pope in Balenciaga or Trump in a prison row, like, right? And that's what's been called for. And somehow we discussed, remember, it was with the P.I. and the law enforcement that the press was not necessary simply because they didn't prove that the document was correct or wrong. And the same thing happens with the video. I mean, video, and eventually, it's just gonna stop being proof of anything. And in this sense, the emergence of Sora is a great step in this direction, because all the videos inside, they're dipphenical, they're just being generosized. And people are just, uh, getting used to whatever video you see, they might be unrealistic, and you're starting to get trusted in any video. The video stops to be seen as evidence of something because everyone knows it can be generosized. It is clear that the OpenAI is marking its videos, which are generated accordingly. But no one's stopping, in fact, from these models, from learning their own, removing all the possible markings in a while. And it can never be sure that some system will determine with a high precision that this video is generosed or not. It's just that video will be forgotten soon that it is evidence of some fact. Well, this transition, yes, until it's over, there's gonna be anxiety about the faeque video.
Yeah, but what happens to people if people-- you say that video is not a fact, then what happens to people to watch the facts? Because there are people who don't want to watch fairy tales, people who want to see the
facts.
You have a choice.
Just the fact that
You're not gonna have a choice.
You don't think there's any choice?
It's like, you know, ten years ago, there was a joke when someone said I found it on the Internet, it's written on the Internet. I usually comment on the like, "Well, yeah, they won't lie on the Internet." Well, because you can write anything at all. Then it was a text somewhere in the forums, some sort of Internet article, something else. Well, that's what you want, you write. Yeah? Any text can be written. And that's why, uh, there's a phrase, "I saw it online," she's not proof that it's true. Yeah? Same with the video. Well, you saw some video on YouTube, on another site, you got a video thrown out there. I don't know.
You saw a video from your kids on the phone. You don't know that.
Yeah.
Or the kids showed up in ten years of their de--- their lives. They say, "This is not real." I have kids on the boat now, and Pauline says, "Well, Dad and I can ride." They say, "You can't ride. You'll prove it." Well, t-- you show them the video, they say, "That's not the real one, you've been generating it. And you prove it's real." How do you prove it's real? You'll show them the picture. They'll say, "This picture is also generosized." So you don't have-- you're gonna show them a printed picture. They'll say, "Well, go, print out a generosized picture." In fact, you have no way of proving anything at all, right? From the point of view of what you've been doing.
I saw a little funny video on this subject. The Linkin Park group has one of the Hedliners, Mike Shinoda. And he's got a daughter, she learned to play something on the guitar. And dad, well, I don't know where, well, how they so guarded their daughter from not knowing about Mike Shinod's glory. He's starting to tell her that, "Do you think that's something else to do here?" She's swinging away from him like, "Come on, old what do you know?" Mother says, "Look, you'd listen, you'd listen. Look at the video. And then there's a Linkin Park from the concert, how they burn, there's a 40-ass crowd that feeds them and all. The girl is watching, he says, "What's the matter with Dad?" And as much as the video can be generosized, maybe there's something else out there. Nothing proves.
Write what you think about video-generation. Have you encountered any problems? Did you get a video if you did, what systems you use, what you use, right? And do you use them all the time? That's very interesting. Don't forget, Wednesday will be a special prom on the subject, there's the best way to use artificial intelligence in your life. I'll see you on our channel in exactly a week. The squeaky-demolishing subcast across the world.