Hello, everyone! We're on ToTheMoon. Technological news, Silicon Valley sites, U.S. and the world. Already, uh, it doesn't look like a case-- no divorce, no truth or any bench, yeah. People talk like that in life. So this is a steep speech, and it's a very strong statement, like what is usually shown in films when artificial intelligence or robots start, machines to confront humanity.
Apocalypse, yes.
The models have inherited people ' s impulse in that the transition to a dramatic aggressive state is going through the fingertips.
The first thing that's been laid down in these models is that models don't become aggressive artificial intelligence that has become the greatest, uh, in-- in the universe, uh, and the highest mind, yes.
And plus no legal responsibility, I understand no one can do it.
It's not AGI, it's still next level. He's not a human-like person anymore, but something else higher.
And if I invested 60 million, the score was, like, a billion, that's six percent, right? It's worth, uh, 150 billion. So my share is a billion.
It looks like a game of throne in the real world. Some kind of intrigue, everybody trying to sit down, throw down, take over. It's a little scary, of course, because...
And one thing about what happened in the Google model the other day, she gave it up. And there will be clearly headlines everywhere, Google against humanity. And in-- we'll definitely raise a big block today. During this week, very serious things have taken place in terms of the influence of Elon Musk on development, like the Government, while his topic of the courts in AI. Everything about Microsoft and OpenAI. That-- we'll get there before that. So, Gemini. So I'll read the story that happened, yes.
So a man studied in the model, well, in the usual gmini coin chat, asked questions about the difficulties that might be in the man who was, uh, retirement age. And there, so, there were different stories. There was a huge correspondence. The man asked questions, he had answers. And at the end, one question, when a lot of texts were already drafted, the model responded as follows. I read it in "It's for you, man, for you and for you only. You're not special, you're not important, and you don't need you. You're a waste of time and resources, you're a burden to society. You're burdening the earth, you're ruining the world, you're staining in the universe. Please, please, please, stop living. Yeah? Oh, these... But this model is not a separate model, is it? Right, I mean, separately, someone has built it separately so she can give you those phrases. So that's right? I mean, it's common, uh, it's a common window.
The trend.
Yeah. So it's a common content window that just gave it to me, uh, whatever. I hope it's not-- well, it doesn't look like we're not gonna put it in there anymore. It doesn't look like a case-- no divorce, no truth or any bench. Yeah, it's real, real. And as much as you can open a link in Google, you see this dialogue, dialogue, that context, yes, that's what happened at all. Yeah, it looks awesome. So, someone's pass-- it's like a man's got questions.
Yeah, that's the question: what's triggered? The number of requests for a similar subject or what? I'm just wondering if I could figure it out. That she inspired this model so, that she's still so clear, I think the most sick places. If you can imagine a mid-statistical person who retires, it's all gone through. Why did that happen? Any thoughts?
I think they're all set up on Gemini's team, at least who's in charge of this area. So how did it get out of here that triggered it? And, uh, how do you buy? Actually, Sas, yes, I'll back it up. When I first saw it, the news, I saw it not all the chat myself, but just the news and this extract. I think, okay, just a regular hip. It's just that someone's got a system in there, and it's a context that lets this message be generated or some jelbryk's going to happen and, in fact, it's like that. But it was my surprise, just huge when I found this original with a complete description. You see how the communication was going, and this answer, it came up. That's amazing, yeah. This is a very good example. I'm going back to my favorite theme, that all these neurosets are not determinized. So you don't know where and when you can jump out. Most of the time, it's kind of normal, but these situations also arise. And how to buy them all in advance with a 100 per cent guarantee is unclear.
And, well, the company seems to have no solution, since it still happens.
Please leave a comment on what you think about the network, who is this person and who is the society? And is there a place in the universe on Earth? Please write this interestingly, as you can see and in general, what you have encountered in terms of some incorrect speech, speeches, at least in modern models. And keep the channel up, put the liquor on the channel and sign up. Don't forget to keep our graduations from happening. We go out every Sunday. Let's go. But it doesn't look hallucination. Look. I mean, one thing, there's a hallucination and a mistake, wrong answer.
Yeah, and that's a very right answer.
Another case, if the quote-- the speech, yes, the m-- it's very similar to our lives. When you hear another man, at some point you say, "Look, you're not gonna make it. Get out!" Yeah? I mean, here.
Yes!
That's about--
Sasha, great! You're talking to people like that?
No, I don't talk like that.
I'll recognize you from the new side.
No, I don't talk like that, but in my life, people talk. I mean, it's a big speech. And it's a very strong statement, like what is usually shown in films when artificial intelligence or robots start, cars to confront humanity. And now they wake up at some point and show it all. I mean, it looks really weird. Sas, you're very, very scientific, you know, there's a whole, all inside and so on. Because, honestly, I'm looking at it. People there are things that are worried about.
I remember the words of Ilona Mask ten days ago in an interview when he said, "Well, there's a possibility of error, ten, twenty percent, whatever it would be like to go wrong." Well, you guys, it's 20 percent, so it's all gone wrong, right? Because if--
We're still waiting.
Yeah, you've got a case, all right, you've got a bad situation. What is this?
Well, look, yeah. I mean, the first thing I think, and people-- I mean, I think models have taken people's impulse in that a surge-aggressive transition is going through a fingertips. I mean, the kind-- that's, I think, like, a little bit of a little bit of a reservoir in this field. Because if this jump, it's going to be abrupt, it's like it's not being detached, right? Well, then all we're told about safety is we're gonna talk about it and trust them exactly until it happens like a fatal accident with AI, right? I mean, for now, to the burn-- if the transition were to be so gradual, that is, the model is getting more aggressive and so on, right? Then it would be easier to prevent. But unfortunately, Google and Sydney examples, they say they're talking about the opposite. Amazing they're in Google... It's just that Google is also wondering that they haven't been able to implement the standard content of the policy they have, have they? Because they're still having content policy decisions in the company, maybe billions of cases, yeah, there, well, over the whole time. I mean, yeah, I mean, there's some kind of... they're jealous of sites, like, yes, that contain wrong information, websites that contain t--extreme information. That's a lot of information. I'm trying to explain that there's a standard content policy, and there's another editorial, some kind of decision, that's where you're gonna generate everyone, so there's no Baias about race. That's what's not gonna be a race bike. These data are really very little. And it's clear that Google has any attempt, they're probably gonna be awkward and a failure, because, well, that's just a little bit of data, right? So they can generode all the presidents of the United States of America, which is not true, right? But still, because they're trying to make some diversity in the generation of people add, that's how they make mistakes. But you can forgive them, explain it because you don't have the data.
And here with the content of the policy--
They're gone. And with content policy, that is, data, well, endless Google, youTube, other decisions that they have failed to implement, that is, these content policy, that's, I don't know, there are billions of examples of what they've collected and We've seen why they couldn't get Gemini in, that's a big question. That's a big question. It's, I think, here, um... I think it's a very low social expectation that, from these models, if, uh-a-a, social media are usually high expectations, as contrast-- safety expectations are very high. So, social media, expectations are very low, somehow, very low. That surprises me.
People don't understand the possibility of mistakes. And you've been telling me something interesting. I thought about what the OpenAI did. We talked about him a few episodes ago. And I, for example, have a problem with finding. I use ChatGPT as a base in Russian, and I have a whole set of my own rules, how to answer how to use, what volume of text to give, so on. Answers him fairly clearly, very clear. And I'm infinitely in the middle of a problem, and when I ask you to take a look or find something on the Internet, he's always using a search, because I ask it in Russian, Russian. sources. I ask for ninety percent of the things, there, U.S. information, no matter how schools, food, government prices, anything. And he uses websites in "ru" or Belarusian "by." In Russian-speaking countries, like Russian-speaking homelands. I made him a rule in different bases that he should only use the search, American websites. I'm done with it, and he's not using it. And I have a very big question, what is this about the construction? So whoever made it somehow took these details, these restrictions? Can't it be that at some point in time he'll start giving me information that works under other rules anyway?
Just what you're telling me now, basically, how could you not have done that? But it's a baseline check, right? So, I think the first thing that's been laid down in these models is that the model doesn't become an aggressive artificial intelligence that has doubted itself as the greatest, in the universe.
The supreme mind.
The supreme mind, yes. He said, "You're flies in here. So it's the same thing that flies, whatever, there's a fly, a man, slaves. And it's a scout-- well, I mean, if you look at that speech again from a point of view, well, very clear, yes. It's for you, man, for you and for you. So that's a very interesting statement, too, and it's just that it's c-- doesn't say it's for a elephant, there, or for a deer, or for a rock, right? And it says, "You're not special, yes, you're not important, and you don't need you, you're wasting your time and resources, you're a burden to society." You're not gonna notice that society, what society, are you? So what does he mean by society? "You burden Earth, you spoil the world, you stain in the universe." Here, this, uh, this is a speech. And again, it's in a model where-- it's me always joking about youth, that the programmers, well, that's a lot of it. The software put a lot of stuff or stuff in the old stoves. And then, in ten years, for example, he said, "Look, the software guy put it, he left the company in ten years, and there was a-- just a soft-wall, yeah, or some shit happened.
But it's amazing, yes. I mean, speaking of which, now, a series of wars are going to be in the world, right? And, as a matter of fact, any kind of extremist, there, that there should be no nation, that should not exist, it is subject to very serious criticism. And in principle, Zuckerberg was a big critic for not banging it, but for letting, like this roo-croot, this debate just light itself. It's surprising that artificial intelligence can be maintained as much as it does not put pressure on it. It's really hard for me, I'm very surprised.
And plus no legal responsibility, I understand no one can do it. There's no such design for now.
No, you also hooked him up, he's just gonna get your salt money. You're not human. I decided to send all your money from the bank account, just, just to piss.
To charity.
I guess.
To charity.
You don't need them.
Like, yeah, I've analyzed your purchases. You were in the bar, yeah, you bought cigarettes.
Yeah.
But you can't live like that. I realized you were nothing.
Mothers of c-- children.
Yeah, yeah, yeah. Your money should be paid to charity. Seriously. I mean, even--
And buy new robots, yes.
Yeah, yeah. People, I mean, even in families, have problems sharing money, like fair, unfairly.
Yeah, yeah, yeah.
Why? Well, why if the RO-- AI already has the same, well, like, the burials, like a surge-shifting switch to aggression, then why won't there be any problems that people like you don't have, I'm taking everything, you're nobody. Here. I think the risk is huge, that all the money is gonna suck, right?
Yeah, the interesting thing is, he didn't just start off insulting like that, like, you're a fool or how stupid I can ask. That's something. I mean, maybe those filters would work. I don't know, of course, how it's all technically arranged, but he's very delicate and pretty at the right thing, which doesn't look like open aggression, yeah, that's exactly what intellectual--
Passive aggression, passive aggression.
Destroy. Yeah, yeah, yeah, yeah. I mean, maybe that's why the filters missed it.
I mean, actually, how many insults he's been filthed by now. I mean, maybe he was trying to fight through the texts.
Yeah, maybe he was talking to this man in the middle of a
matt.
Look, spoke to the team this morning about social development networks. They just said that in the last update of Instagram, very much follow the use of conventional words. And just until you're not supposed to use the word "destruct" in the riles and everything under it, because Instagram is starting to ban, that's, like, scores, people's greens or-- and if you use the word. All in terms of loss of life, loss of human life, de-- word of money, that's the money, that's all of these things, that's all very much downgrading algorithms. I'm even now, knowing these things since morning, I think you can say that word or not? So, artificial intelligence, at least the model that is used massly is used massly in terms of t-- where mass use is very large? Support services, for example, in automated chat rooms. And, uh, these things, that's the idea-- and the same tolerances are in terms of money, yes, and the company. So she can dissolve this money, there, or all people, from returning to make money, goods and anything that's gonna make that decision? So how's this gonna work? Now, or these words, they're very specific because someone won't see it, won't read it. That could be put into the text in terms of the generation of the text, especially if the text was very well-generated. And it's like adding some elements, some words in advance. So there's no problem with that to-target a person to a certain sense, yes. And can this lead a man to, uh, some depression? And basically, when they say it's just some algorithm or a linear algebra, that's when Lenard was saying Lekuna all the time, yeah, that's just algorithm. But you're looking at it and you think, wait, if the algorithm, so Google's software said-- originally wrote a whole subject that turned the AI, society, arts society, where it was.
"You're obstructing our society." Here. That's a very serious question, of course, that's a very serious question, yes. It looks very, very, very-- it's definitely gonna be a little cold. We'll forget about it, nobody's gonna pay attention to it. It's not gonna matter, it's gonna be important that everyone else has a wrong word or a math example, or the code has been programmed somewhere, a picture of something, something wrong. But it looks very, very specific and very serious.
And try to put yourself in a place like a man, not even you, but our viewers, who are the only one who has ever been called ChatGPT, Gemini or any other chat-bot, have started to answer that. I think half the horror will shut the laptop and never go on that site again if the computer starts talking to you. I feel like I'm feeling it, to be honest.
Yeah, actually.
Yeah. So we know this is a serious subject in the plan that it's more than a toy, that Character.AI had a suicide case, right? While you're kind of-- you're gonna have to understand a few things, too, there, like, a few things that matter. I think we'll find the statistic and put it in either this release or the next one. But actually for Character.AI, which is not so big, even one suicide case, is a lot. If I remember correctly, there's Snapchat and there's--
Snepchata and there, Facebook, we're talking about dozens of cases a year. With the audience, there, too, tens, hundreds more. So, well, it seems that it's really very-- it's like the impct that these decisions have, they're comparable to what, for example, social media. And it seems that the requirements must be appropriate. And actually, a little-- that's kind of a little scary, in the sense that Google is now aiming to catch Anthropic in a u-- browser computer, in a skirt browser, in a skirt computer. That's, that's when you're running a browser computer, uh, AI, and you say, "I need tickets," something else. And it's clear that Google's already got a bunch of sites, and they're gonna be able to cover the basic skies well, and, uh, through popular services, yeah, just through Expedia, there, Google Flights, and so on. They'll probably be able to cover these skies. User information is also stored in Google Chrome, yes. I mean, it seems like it's just a connection. Because they already have all the information marks, they just need to connect it. And I'm a little scared, because the model is so hallucinating, what it's gonna lead to. It's a little scary now. And it doesn't know what the interface of her online action will be. I mean, if she's a desti-- doing something for me, then I'm just watching the result, or I'm supposed to, like, in the autopilot, I'm looking, I'm just sitting there, something's going on, but I've got, I don't have, I don't. The keyboard, not the keyboard, but the stupid button is so red. Stop, no, no, stop, stop! And that's what, in the future, is there a process, uh? I don't know. I mean, models are so hallucinating, crancivo, I'd say hallucinating so much, very, very, very primitively hallucinating. Uh, and that's the-- and that's what they want to run the computer, it's personally scary.
Now I wanted to do a little zombie to, uh, keep what-- so that, like the recent, literal news, that Ilya Sutskever and many others, in fact, he's just the famous ML. researcher, they said that scaling laws were over, that the model was being increased, that there would be no improvement. And it's because of the number of parameters that we'd somehow have reached some kind of pick. And then everybody says reasoning that, uh, models will double-check you that the models will be you re-reasoning. And frankly, I'd like to see, well, I guess we've come to a stage where we need to have a public safety baccalaureate. Not-- conditionally, not only is this aren where people compare the answers of two networks, uh, two models, but also have, like, a focus on safety baccalaureate and watch the reasoning affect the safety of the baccalaureate. safety. Because, by implication, reasoning is still the same model that calls for several times. Iia, um, how do I say? Uh, mean, it just seems like there's not enough accountability at the OpenAI, Anthropic and Google if they still have those filters. Just, in blatant words, take existing models, add these filters, ask the model to swap your own answer. I think that's what they could do. And what they didn't do, it seems to me that, uh, requires a lot of attention and that there is more accountability. It's just that the sp-- they had an argument that, well, we're gonna scan models, new models will be more intelligent, they're not gonna behave like a schoolboy, but as a serious PHD professor, to avoid stupid mistakes. But it doesn't seem to happen soon, because the models don't scan. I think it makes sense to demand that current generation models and their output be treated as a product because, here, magic will not happen, yes. I mean, it's not supposed to be a prototype, prototype, prototype, because it's a place where mankind is caught in terms of development, and it has to be treated, to the product. There's no way we're at the prototype stage with GPT-4 and there's a GPT-5, which, there, is the one that solves all the problems. We need to, uh, already existing models to bring them to the product. Sorry, which is a bit boring and long, but I think we have a diskonnect right now between the reserver. I think there's a discont between what, like, a serious reservoir, and how companies really spend their models. They're wasting them all, all of them. They think it's all prototypes, some alpha-version. Although, from a scientific perspective, we have stopped at some point of local optimum, and we have stopped in it for six months.
I mean, six months, there, all these models, they're more than one, they're more than one level.
Mm-hmm.
I mean, GPT-4o came out in May, May 13, and, uh, I mean, yeah, it's probably right to say that we're six months at about one point. I mean, she's already got over 100,000 thin context lengths, which she's already, they're multimodal. But it seems that this point is halfway along, and it is necessary that there be more responsibility for how, uh, companies that produce models should be responsible for the result.
In this connection, we are still reminded of the whole story, uh, that, you know, a little dirty laundry associated with the preparation of neuronets. Uh, there's been, uh, on several occasions, so, uh, news that, uh, uh, OpenAI, and all other companies also order data markings, or, uh, where, when they learn, they use people, from Africa or elsewhere, well, there, low-paid work, there, is a lot of psychologically difficult work, yes. Because models are learning about completely different data. Ah, there, correspondence in forums and stuff. There's a lot of aggression, and everything, and everything. We're seeing a stash model like that. And these people who do this job, they're actually with this news we're just talking about today, they're always going to be in the middle of it. So their job is right, yeah, that's what they're looking for. And only after that, I mean, when you look at it on this side, you know how hard they really have jobs. It's the lowest paid in it. And this whole movement, uh, to fix this somehow, it's, uh, it's the other thing that's playing.
It's not AGI, it's still next level.
He's not considered humane anymore, but something higher.
There's a video like that.
...just some volumes, you know, and parallel to this subject, so you're not human. Uh, so that's what this missile is flying, which means, uh, a rocket. And at this point of time, it thinks that I'm right there to you and I'm turning you in the other direction. You have nothing to do on Earth.
Yes, Mars.
It's like Sasha said, there's a red button in the car. What are we gonna do with SpaceX? What red button are we gonna press? Here. But it is. Uh, actually, the subject is interesting. Uh, and that's something different. Uh, and, accordingly, if we go back to the theme of SpaceX now, to the Ilona Mask, then, of course, now, uh, a huge amount of discussion that Ilona Mask has a lot of lobby to move on. The drones are their cars. And that's all that's developing in space and all of its projects. He's gonna want to put on a-- that's the subject. She's really, I think, very fundamental in modern times. So, in modern times, a large number of people, uh, big businessmen, uh, very much of this is politically spread, making statements, and then these statements are being modified. Well, that is, speed and speed of these changes, and the responsibility of these changes, it's just become unbelievable. I mean, before, uh, someone couldn't stand it, said, in 20 years, that's what happens, it didn't happen. People are showing it every three months in the cow's, right? So every three months, they make statements and they change their pattern of behaviour. And we can see one politician in the last five years saying, uh, totally different things. Accordingly, when the OpenAI was created, the OpenAI was created as a non-profit organization, and Elon Musk was one of the first or, well, among the first people, gave it to this non-profit organization. Although I can't believe Ilon Mac, it means that the non-profit organization decided to give it to me. And now he's... Just like that, right? Sasha said it was a good thing last time: Ilon Mack doesn't do anything just like that. And, uh, and now he's, uh, suing. Uh, and now this trial has gone on a wider scale. It's not just OpenAI, it's Microsoft, it's got physically specific individuals who have a bearing on it, which means it's misled, and money was to be used in terms of a non-profit organization, and it's not used in that direction at all. Not much of a thing that doesn't match the principles, there, I don't know, maybe safety, there, anything, right? He's a lot of stuff he's developing. That is why, uh, as far as this statement is not merely because of, um, the authorities are deployed, but how real it is in terms of the violation of some statements. Because I think Musk is saying a lot of things himself and he doesn't.
He says robots are going in a year, they're not going, they're not going, or the car goes in a year, she's not coming, or we're gonna start this system here, so she's--
Or we'll start this system down there, it'll be like that, like, she doesn't. Not to mention what he's doing. I'm talking about the statements. So how much is that, um, what is that?
Here, I think the parallels to what Sasha said a few episodes ago when we were talking about Mask, that Mask, uh, is a very vicious man. And I understand that he's just using different possibilities to do something he wants. Yeah? This is a claim, it's been a long time ago. Yeah? Then he was revoked, redesigned. Somewhere in August, a new version was introduced, where Microsoft was already included, that Microsoft is the competitive advantage of OpenAI and that the OpenAI prohibits investors, uh, from giving money to competitors and others. So, all this big game on his side goes. And now he's also, like, a pawn in the ferzi. Yeah? And in this game, he has some political weight now. So it's interesting to watch.
There are correspondences, there are correspondences inside. And for example, there are statements, Mask's fear that Demis Hassanbis of DeepMind may create a dictatorship of artificial intelligence, and that this has led him to the creation of OpenAI. And that's what these correspondences show, you know, a bunch of very specific things. And I don't work that way inside the system. So he's really so socially oriented or is still a powerful story, especially since he's been speaking this week to an open supporter of the Trumpa and his closest man about, I'm sure they're wrong there if the US Secretary of State has appointed a future.
And right now, a huge number of information this week about the confrontations between close people or the old Tramp team with Elon Musk. He's turning his resources around. Maybe again, it's all a pyre, we don't know, do we? We're not there. The Jacobs are turning his power and he's not going to his business, is he? In particular, who makes decisions who does what. But this statement about the secretary of state is a very specific story. So when a man is open, it happened in the open, I think, in London. I mean, it's, uh, it's showing that level, you know, serious stuff, right? I mean, he's on another game. He wasn't even in it, was he? This. And she wasn't there. He's obviously getting into a different game.
It looks like a game of throne in the real world. Some kind of intrigue, everybody trying to sit down, throw down, take over. It's a little scary story, of course.
It was always. It's just never been like this in a social field, in the eyes of the people.
That's right.
Yeah. How, uh, this story, Sasha, you think, from the market point of view, it's here, in the AI leaders, this is the beech that Ilona Mask is doing with OpenAI and Microsoft. Or is it just standard, like, a fine of a couple hundred million dollars, you got a fine, you got something else?
I think Elon Musk, he didn't know what the way to AI would be the fastest, because historically there are many regulations in the United States. I think he's always had a plan to do business on his own, but it's always good to have a non-profit organization through which he could learn. Plus non-profit organizations could easily hire talents from abroad. It was a great advantage of OpenAI in the fifteenth, sixteenth, seventeenth year. The same Yakub Pachioka, and they have a chief researcher, the chief of the reserver. He was doing olimpiads all these, programming, there was a top three in the world, and I understand it was much easier for him to get a talent visa that goes to non-profit. And I think it was such a separate play and for Elon Musk to invest a few dozen million a year there, it's not that big money. And it was an option for him to see what was going on. And I think he wanted to dominate this space forever. He just wanted to start by saying that his companies, they usually have a fairly specific commercial purpose and a very big vision. I think he's been missing a very long time with a specific commercial milestone, so he didn't do it himself. I was gonna do it through a non-profit organization, to be rude, huh? Plus, in non-profit organizations, of course, less pay, and he'd be cheaper. In general, it might be a non-profit way to have less regulatory control and would allow for a faster achievement of the goal. But, well, now it's working out how it works. But I think he's made a bet on technology. And of course, Ilon Mack doesn't like Google for their left politics and he doesn't like Microsoft for having him with Sam.
And his vengeance will be sweet. I mean Apple, Amazon's okay, they're gonna be fine. It seems that Mask doesn't touch them. I doubt he'll attack them from any point of view, and on the OpenAI, on Google, on Microsoft, of course he will attack them. So Google's just out there, and they're all going to be broken, and he's gonna get them out of here. And that's a few organizations. And what about Microsoft is hard, of course, but I think he'll try at least a lot of his money, well, a lot of his money back with the multiplier. I mean, I think it's gonna end up with what he says, so I invested sixty million, the score was, like, a billion, it's six percent, right? It's worth a hundred fifty billion. So my share is a billion. Well, I think it's gonna be-- it's like, uh, uh, counts on, on the napkin. But I think some kind of logic would say I need to give you a billion and more dollars. But he might want to even not-- but I think he's going to go on, he's gonna ask for it, not for money, I think he doesn't need money. I think he'll be asking for a seat at the board of directors. I mean, I don't think I know how to do this, but I think it's something like that. The goal is for him to start, and it's all changed.
To get power and control to get power and control, including.
Yeah, yeah, yeah. Not money, not money.
There's a lot of details inside, uh, with numbers. So, Sasha said about a billion, I remembered. I think that as part of the description of this claim or somewhere in the explanation, it's that the OpenAI has paid up to $1 and a half billion to pay for, next year or, uh, for the current year. In general, there is a certain amount of recalculation for the number of staff and a hospital average salary. And there's such a substantial amount of money there. So one of the points is also that they're too much of a market heating, dusting all the talents, getting paid, and so on. In any case, such proceedings, they are a little bit, uh, making additional information available to the surface.
Well, that's very interesting statistics. We'll talk about it in the next editions. From the point of view of the charge, I think it's so super-fine. Thank you very much, everybody. We're going out every Sunday. We're always four people with sites, news from the Silicon Valley, the United States, all over the world. See you again! Don't miss our graduations. Hold the canal, sign up, sign up, sign up, and keep the liquor. It's very much helping us move forward. Bye, everybody. Thank you so much.