Tesla Optimus Pivot robots $25 trillion. The amount of space, of course. The truth is Elon Musk's exit from OpenAI. New work environment for scientists on ChatGPT. Andrei Carpatha. Why is it too late to make a new AI startup? That's a bad conclusion. Google added free full-form SATs to Gemini. Manus presented the Minutes Meeting and JPMorgan AI could go too fast. Hello, everybody! We're on ToTheMoon Channel every, uh, every Sunday. So, special episodes on Wednesdays. Sometimes, sometimes. Here we were a few weeks ago with a special episode of the new guest. Aah!
Sign here to keep it from happening.
Yeah, sign up for sure to keep it up, keep it down, you know. Look, when I saw this story, it's amazing, isn't it? I mean, there's Tesla. I'll remind you, her decent piece belongs to Elon Musk. There's a board of directors, I see. We've heard the latest information that was so interesting. What it does show is that Ilon Mack is very serious, very serious about the company, such a combine. And from-- when he agreed to his bonus, there's a trillion dollars, and that the company will, I don't remember the exact amount, we had one of the episodes, right? And, uh, that company will be worth, there's seven trillions. It should be the case, there, with certain parameters, with indicators for this. And now the new information is coming, so Ilon Mack is making a bet... There are optimus robots inside Tesla, right? And there was a promise that there would be thousands of robots, robots, that would do a lot of things in the twenty-fifth year. A thousand robots didn't show up. We can't buy a robot, we can only look at him, like he's in a store, there, or in a roller. And, uh, on one side. On the other hand, the robot never did, uh, decided with... This is the key technical good-problem, is the trap manipulators, right? Well, that's a good manipulation, there, by hand. Although they have a very cool hand, they have robots, they have robots, right? We've been talking about it different episodes, too. But the point is, this week, Mask announced that Tesla would reorient the effort with the electric cars to Optimus, and predicted that this company would cost $25 trillion. So, aaa, I usually turn around like that, huh, Tanya? Here. Here. And predicted about $25 trillion. Oh, I think a very live story, a real story, enough, a potential. Another moment when we can really see Optimus. This year was planned-- and they have a million plus robot line, and they were planning to launch another line of ten million. One of the statements Elon Musk said that one day, there's a billion-dollar line, or some, a number, a number, I don't remember what number he said.
Here. In fact, his logic was that the optimus robots would be more than the world ' s inhabitants, the world ' s inhabitants. So the Optimus robots will be more. I mean, in his design, configuration, that robots will be bigger, more. And so when, by the way, we talked about the last time, many of you wrote comments, many viewers wrote comments on whether or not the programmers would disappear. Eldar and I were just here on this subject, and we were talking about it. I wonder what happens if there's more robots than people, right? How does that work? How does it work? Well, if people are still alive, how does it work in this part?
Important clarification. If people are alive.
No, just if people are not alive, they're more than people, then why this discussion and discussion, huh? Here. In a good way, in a good way.
Well, if the planet is already overcrowded, there's gonna be more trouble with robots.
Well, there's a different estimate of the estimates. There's someone saying that they're overcrowded, someone says there's still an incredible number of people on Earth. I see that there are many and very different conclusions.
Yes, there are huge desert territories. If you're going to start, there's a place where there's a lot of sunshine, solar batteries can be put, and free, near free electricity, you can have robots there, some factories, some more. Something like that. So you can get a lot of stuff on your head. I think they're going to use the commentaries, they're already getting a little wobbled, but somehow, yes, that's a lot of pro-space for that. You find a source of energy, a probable free or very cheap, not very settled by people, and you're actually filling that
story.
Why did I even raise this subject today, to be honest? Well, first of all, we still say this spirit of the Silicon Valley, and $25 trillion is still such a serious statement, on the one hand. On the other hand, you should notice that this man is still a company that produces electric vehicles and is quite good, well, a good volume, a large market holds electricity and plans to produce them. Go on. But in particular, in electric vehicles. Tesla is so popular, I call it a popular car, at least in California, yeah, it's a popular car.
Of course I speak California, not even in the States. If you move to other states... I travel a lot to the States. I can tell you, Eldar, that in other States of Tesla, she's like... she's kind of dissolved, right? Her, she's not like she's here. It's just that this is crazy in California. It's clear Tesla doesn't have a presence like this in China, right?
It could be that, uh, beer, that's the way the company turns towards robots is connected to it. First, uh, like China as a huge player, respectively, with cheap, high-quality electric vehicles. Second, history, there, I don't know, European countries, they wanted to ban the FAA, parole, there, in the twenty-sixth, I think, a year. Finally, they decided they were not ready and the MAF is still on. So the switch to the electric vehicles is delayed. And, in fact, Tesla grew in part on all these movements, yes, that we have a green agenda and that's all. But the economy shows that it's still, um, petrol, diesel, uh, more comfortable, more comfortable, cheaper. And the whole electric car shift, well, it's not happening. China could have been in huge cities, yes, but as in China, its producers are.
Maybe just seeing this, this is my fantasy--
Well, even with all the revocations, it is clear that the United States has now moved the subpoena in the States, including not only Europe, but it has indeed moved, there, c, uh-e, Back electricity, there, gas, gas. The question is, uh, that still China, uh, Chinese cars in the United States won't be available, right? And the biggest market there is for the same Tessla, quite serious, is the U.S. market. And if you look here, there are three companies that are localized in the cars. It's Tesla, Lucid and I say clean machines, and Rivian, yes. And, uh, they're because they're on-- electrobs that, uh, drones, like those driving, there's Waymo, they're built, it's Jaguar. Uber is here to launch, they want to launch, if I'm not mistaken, Rivian and so on, yeah. But it's California cars. Uh, Tesla has a crazy development from the perspective of the United States. And that, uh, story that shows up, or this story is new, new, like a statement to show why a company might cost $25 trillion dollars because it's twenty-five trillion. It's hard to do a dollar of extra-net electric cars, yeah. And get some investment for that, maybe. We don't know what's going to happen. Although, by the way, Elon Musk is not inclined to sell shares in his companies as a whole, yes. Iia, uh, in Tesla, he has the smallest share of other companies. Starli-- satellites, his satellites are entirely owned. I'm saying, uh, all about space, I think space belongs to his company, yes, uh. And, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, They were a little more attractive. So let's see what's going on in Optimus. Maybe this story is just to get another score on, because still Tesla is a public company, I'll remember, and I even have her flow. We remember there was a year and a half ago, and that was a sp-spore to buy, not to buy. And I did, and I did, and it grew up in that time. Uh, Tesla grew up, not decreased. And, uh, that's worth it, at least it's a matter of view. Another story we're going from Ilona Mask, aaa, to understand what's going on in the market. Because, maybe it's not so fundamental to people's daily use, and we're trying to tell here about, uh, how artificial intelligence and technology are different to use.
My life is here now. But the topic that is worth touching is that there are now enough debates in the OpenAI and Ilona Mask dispute. I'll remind you that Ilon Mack is incriminating...
I recall that Elon Musk was originally one of the first people who invested, who invested in, uh, that invested in OpenAI originally. And Ia is having a very big discussion, because Elon Musk has sued the OpenAI, that the OpenAI has decided to become a non-profit organization and becomes a commercial organization, and wants to make money. It was not the first time. It was originally planned that the OpenAI company would be non-profit, and it was just crazy, so it's inside the discussion. And Elon Musk is asking for a share. Right, Eldar? What is the amount he's asking for, what percentage? Something about...
I don't see the money.
There's a demand in the material now, so I found it now, up to $16,30,4 billion. Here. Ia's an expert. Plus potential damages and possible prohibitions, well, restrictions varying in language. I mean, he's already demanding a hundred thirty-four. I mean, two weeks ago, when this whole discussion started, there was a sum in the $67 billion area. That's why I asked because maybe you saw. It's fresh now, so it's $100,34 billion he's claiming compensation for it. I mean, he says, I kind of put money in the wrong place. I'll remind you there was a story about what was going on in 2017, and Ilon Mac says he was lured by a non-profit mission and then turned into a profitable model, yes, and a connection to Microsoft. Ii OpenAI claims that there were talks in 2017, and because in the seventeenth year, Mask said that I needed total control over OpenAI. I want to open-ended systems in different ways. Well, I'm sure he's talking about it. Ia, according to him, it's really what's going on. The Office in its projects or where it participates and is located, including the documents from the OpenAI leadership, states that Ilon Mack himself has personally offered to separate the company and as soon as possible. Work out a procedure to create a commercial organization from it, right? I mean, with the control business. They didn't agree on that. And he has a claim, and he has a claim for Microsoft, yes. He has a claim, I understand, to everyone inside. It's a pia-pira, a war pirate, a war war. And OpenAI in court is not defending itself now, but rather counter-acting. And to prevent the court from giving Mask the approval of the OpenAI for a multi-year pressure campaign, yes, all they allegedly do is as if they're doing this to Ilona Mask. Let's see what happens. Clearly, the global user base will not change anything. Well, I guarantee that nothing changes. But from the point of view, we had someone commenting on what you're talking about Elon Musk? He's the Creator of OpenAI and so on. There's always a huge amount inside the nuances.
And here I want to say that, and that's very interesting to discuss that, that Andrei Carpathians... as one of the people who was just like this, in particular at that time at, uh-- he was at the same time. Was it right in OpenAI?
He was, he was in the OpenAI in this conflict, and the CEO-Gene-Operation OpenAI was writing the story that we were just with Andrey Carpath, so Mask was talking about this, too. There were discussions. And Carpathians have put out an interesting thing. I think she's very interesting to the audience. He said there was a point that it was too late to make a new AI startup. And we, in particular, in Tuzlam, can make these statements on occasion. And he said it was a bad conclusion and there's a good chance to make a new start. And he has the first statement, his thought, and the narration he has, he says that the big players are too-- well, there's a popular, so-called myth, that the big players are too far away, and A small research start-up can't jump them either in quality or in speed. And Carpathians say that when the OpenAI, uh, just started. How can you compete with Google? And I think that's a good example of the OpenAI launch. Yeah? Because what is it for the world, for the whole start-up? It's an idea. There's an idea. Someone's in the garage, doing something, programming, and then he's leaving, and he's got a quiet start on the customers. He's getting money there. And now the sales start to go and everything. But I'll remember about OpenAI. OpenAI, before anything starts spreading and selling, it was filled with billions of dollars at the beginning. And how we see it on a number of launchers, like the Mira Muratia, we see that one of the OpenAI new start-ups is now opening up and saying he's gonna have an estimate of over a billion dollars. He's half a billion. I mean, still here with Carpathians-- Carpathians, he's describing why such projects are being created. And I agree with him that there's a Max Grigoriev talking about it at the special edition a few weeks ago, that there's a culture of big companies. Why can't they do something? And Carpatha says that, uh, new start-ups can create this x10. But at the same time, his example with OpenAI is very specific to me, because, uh, it's still a relative. Well, when it's done-- it's like Amazon Jeff Beos is making a new start-up for robots, like, and the company's six yards right there. Well, it's a personal attitude. It's a relative startup, yes. These are huge amounts for the main world, for, for, to be honest, for all, for all, not for peace. And then moving out of the technology company zone in the US is catastrophic, big. And, well, there's some super-gyps big around here. And that's why, uh, that's a very specific story.
What do you think, Ildar, about this start-up thing?
Oh, yeah, let me just say a little more about, uh, this argument with the Ilon mask you're saying, OpenAI. It's the funniest thing I've got to do, just like this is all in the States. In the same vein, the dispute forced the court to grant access and further published some correspondence, Greg Brockmann's diary, just like, who he is now, not CEO, but, short, second person after the MDA.
Operative's in there, yeah.
Yeah. So, uh, his diary, his records, all of it, uh, got available. Uh, Ilona Mask's lawyers, on the one hand, got kicked, read, found a corner where it looks good for them. On the basis of this, they filed a complaint, sent a court. Then OpenAI looked at it, well, now I'm cleaning, brushing the ribs, thinking, no, that's what it says. And now, in the public field, it's all about pretending that, in fact, look, they've got it, they've been out of context. That's the point, that's, uh, that Mask was all fooled that he was poor, miserable, uh, he had a cut. And that's it. And there's a context like this and so on. I mean, it's just interesting to watch, uh, like Mask's lawyers when they did this, didn't you think there was something on that side? Well, it's a little hard for me to understand. If, Tanya, for example, you're in there from a legal perspective, you know better how this is going on.
Yeah, yeah, it's a typical legal war, because it's what lawyers always do. They change the angle of view by the same fact. And it's just a question who's gonna get to the judge and who's gonna hit his wave and his picture of the world. That's why it's very interesting.
Just like a book.
Yeah, yeah, yeah, yeah, yeah. I mean, well, it's a series like this, which can be monitored gradually if there's time to make a wish. And another thing that's interesting is that OpenAI is in the public field, that's what they publish on their website. So it's not going to be closed, but they start, I don't know, public attention to it or something. Well, I guess some strategy has been developed. I mean, I can't say anything, there's someone who cheated on and on. But watching the process itself, as it happens, is quite busy. It's still a little uncommon. Here. And as for Andrea Carpathians and the starttups, yes, here I agree with you that the start-up is a discord, right? When someone says a startup, it's the man who's gonna come up with some project after work, which will take off or take off. Andrei calls the battap, on the other hand, when they've got the best minds on a track, they've been gone for years, they've been put in a lot of money, they've given all the resources and waited for them to have something. I did it. Well, it's not like Mira Murati gathered some of the OpenAI staff, attracted a lot of money. And now, there was a week ago, or if there was news that she had three key staff members out there, two of them came back.
Two of them went back to OpenAI and so on. Well, it's a start-up, too. I got a huge score and so on. Clearly, a man from the street will never do anything like that. There's only a huge forensics, communications, community authority. That's why I agree with you that--
And yet, and at the same time, the interesting thing is that big, serious topics often need to be made by a lot of infusions. There are, however, places to be done-- not everywhere, of course, but to create some serious things, like, the OpenAI now has been setting out how they're on their b-- well, within the framework of its own. database architecture is eight hundred million users. They have five or more million requests there from-- every second. Can you imagine the volume? It's a little more, like, 10 percent of the world population, uh, permanent users, right? And you have five million requests every second to keep it so that it's all clearly working. Well, honestly, the system is working, right? We--we, we're just gonna go to the interface, to the different buttons. There's one on the iPad, on the mobile application, on the desktopus three. It's just so dry, it's just crazy, right? And the Yiahs are always getting some new systems, the old ones don't sell, they're forgetting. Now we're gonna tell you, we've released a new Prism andia all the time, and they're getting something new out. Real, old, old, old, old, what's going on with them, what projects? There's no sign of them on the website. Again, our special release three weeks ago... talked about it a lot. And while all they do is create, they're going to AGI, and we're gonna talk about it. In Prism, the essay from Dario Amadea from Antropica, right? That's a related story, but, uh, eight hundred million users of the guest room is, you know, a good-quality story. And that, of course, needed money, and that was what it was for. Here. Another moment, uh, write our viewers, first of all, who we have now, well, a big audience, who makes the start-ups, what you do, what kind of start-ups you do, whatever you do the puzzle. Someone might think, "I'm gonna tell you the idea, I'm gonna get an idea out there." Well, then don't tell me.
Although,
At least give me a hint, huh?
Yeah. In modern reality, of course, this idea is such a relative absolute story. Iia, uh, if you think your idea is unique, ask ChatGPT in a temporary chat. If you're afraid that ChatGPT will steal it, there's a temporary chat, it's a temporary chat.
You have little idea, you have to put it on.
Yeah, in a temporary chat, write if anyone in the world does it. I'll remember the story. I had a neighbour who was making a baby device, and he was writing and writing it up to now, uh, that he had the first in the world, there's a child AI device, and he's even looking for, like, a SEO, there, for a quarter thousand. A year. So he keeps doing it. But when you're dialing inside... Well, I saw, uh, he's still in LinkedIn, and, uh, he's got some crews growing up. And when you look, you say, "What's the first device?" And you see what China does in terms of, uh, devices for children with artificial intelligence. You know what the first device is at this point. And besides that, there's at least some serious devices like Alexa, right? It's really gluing. The alarm is still getting worse, I'll tell everyone. The alarm is still glu-- it's worse. It's just impossible. Here. Noo, uh, fighting, fighting her somehow. Tanya, what do you think about building technological starts? How much time or d-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w-w- You're a technological start-up, uh, by yourself, I don't know how much you've been involved, although I remember you had time, you went to Scolkovo, right? Or there...
Yeah, I was just at Scococococo, at the start-up academy, and I remember, uh, I was smashed in the ears and dust because they said, "What's your idea?" Like, you know, for-- in my performance, the start-up was--- that's what it was like, create something new on the market. And in their performance, it's just some kind of unreal new idea that no one has ever come up with, which I don't think exists in
principle.
You're putting it on.
Yeah, uh, I'll agree with you. But, uh, it's hard to judge. What I think, you know, is how to compare monopolistics and some small business. There must be some gaps in monopolies or some existing AI companies that these little, uh, business can find ideas, process and something. I'm not gonna make it. But then, when they become more or less visible, or their idea will be stolen or, I don't know, bought at best. I mean, I agree that it might not be possible to make a new OpenAI, but at some point in the start, something very narrowly specialized is possible. Although I'm not a technical-stratapam specialist here, of course. I don't know what you think, Elnar?
I agree completely. Get your forensics out. When you're going to be a top level like I don't know, the OpenAI or Google staff, or someone who comes out of these companies, and they're under their expert review, they're about to be able to assess a billion. Then we can. If you don't, well, develop your forensics.
I'll give you one advice here, uh, from the point of view of the starttaps. It depends on where you're starting the start-up and you have to understand the environment, you have to understand the whole time of Wednesday, and what you need to do, yes. And what you do for him. You're making him up to raise money or you're doing a company that's gonna, uh, really get clients? What market are you doing this for? It's all very much dependent, of course, very much dependent. Somewhere you need a name for a parent, there, from Google, Microsoft, from Uber and so on. If you come in, look where the OpenAI staff work, it's just that there's Clondike, the Silicon Valley, yeah. Meta is there, so Google. It feels like there's only people who hire people out there. And it's a kind of installation, like, like, like, like, like, it's like it's some kind of installation. But it's not like that. It's not everywhere, right?
Yeah, because there are other countries.
Let's move.
After
all, it's a joke.
Yeah. I'd like to go on with Tol Karpata right now. You said about his post about what the start-ups were doing late, and he had a busy post that connects us to the previous programming issue. He was just on the side that, uh-a, new challenges with AI coding could not be met. Here we are, I don't know, three or four months ago, he was just experimenting, that he was trying, I don't know, there, GPT, NanoGPT, learn from scratch, write the whole code. There's no code on the Internet. He tried, made a lot of attempts, he didn't make it. I finally spit, I wrote it myself. There's about eight hundred lines of code, there's repositories, even the reference to it we left. And he was on the side that, as long as weib-coding was still in the process, even though he had invented that term, there's nothing normal to create, yes. It's just some kind of little game. And now he's released the post that he's got interest in the eighty-tenth generity code and 20% of the time he's just got some things to adjust or something. And he says that this happened to him specifically with GPT 5.2. And this GPT 5.2 thing is a point where he already realized that, "Oh, it turns out it works." It's interesting, right now, to our conversation, that these AI tools are getting closer and closer to professionals of the kind they even appreciate that-
Yeah, but we're here with you--
Programming process is changing
We didn't say not using the tulls, we talked about disappearing, decreasing the number of people, yes, or people disappear. From the point of view, I understand that-
Yeah.
Yeah. I'm sure Carpathians, I'm sure Carpathians have used models for programming all these years, yes. Probably.
He wrote them.
Huh?
He wrote them.
Sure. And, and he used it. He wrote them, he even put it in the same way, in a few hours, to create this, yeah, with the model. I mean, here. But indeed, his statement is a question of how to deal with it. I don't think he changed his mind. I think it's his speed--
No, no, no. He, Sas, he said before that he didn't, uh, coding agents, normal tasks with them could not be solved. Now he's got a new opinion.
What's up.
I mean, he changed it, yeah. What with GPT 5.2, they've become good enough to make things happen to them.
Yeah, we were, uh, discussing it, by the way. I remember Elinar saying that--
Yeah, yeah, yeah. About a month ago, just three times ago.
Yeah, it's all useless and leads nowhere. There's still a lot of switching. And in this new article, he's actually writing the opposite.
I'll try to make a special edition on the subject of interesting things that, uh, besides programming, have become highly accessible, in particular with 5.2 and 5.2, which are about or deep, deep. reasoning. Here. And, of course, now, some things are just incredible in terms of the final result.
Let's keep going. OpenAI released tools such Prism. It's a free environment for a scientific letter. Ah. And in fact, it's just...
It's like Google Docs, yes, for scientific articles of different kinds. I want to make a comment here, because many people can't understand. Uh, there, um, considering it's a research editor at LaTeX.
LaTex usually says.
Yes, LaTeX. And there's a editor, a compilation, a joint job. And since a lot of people don't understand what LaTeX is, yes, or even I didn't call it wrong. In fact, LaTeX is needed where, and Word starts to break when there are many different language, formulas, symbols, some indexes, many references inside. Well, look at this, the transitions, the enlarged content, and so on, yeah. Accordingly, they are usually used by researchers, aspirants, engineers, etc. So, who cares about this mob, she's out. It's not enough to be dispersed to all ChatGPT users. I don't know, depending on the way out of the country where you are, in different countries, it's gonna be, yeah. And, uh, it's a useful tool for extra tools. What can we see on this? What OpenAI is doing in parallel, like AGI, is creating various tools to take on, uh, the work of others. I mean, it's a kind of environment like Prism should have done, and, uh, some company, yeah, that specializes in that, which is deeply involved. And I'd have to have that soup coming up. We see the OpenAI going everywhere. He's in education, so he and the individual interface have been released for health care. Oh, it's time he and the CRM system are gonna let out, right? Time issue, time issue.
I was just enjoying something.
Yeah.
On, actually, you're opening any article, even though you don't know, you probably don't open any article of math science, physics or anything. Every time there's a lot of different cords, it's all written on LaTeX. It's possible to get everything in Word. I say the man who wrote his mechma diploma in Word, the mouse has been picking it up, and the disser I've already written in LaTeX. And first I had a reaction that, oh, cool, that could've been LaTeX not to teach. Then I understand that, well, it's a few weeks of practice, and you write in LaTeX just like you write in Word. Oh, and besides that, you're saying there's a joint access, and that's a design, too.
There are services that allow a few people to write at LaTeX.
Sure.
Like Google Docs and stuff. And here they get...
It's just that ChatGPT is, isn't it?
Yeah, they added, so it's so internal, uh, Copilot when he's straight-up, this text is in the LaTeX syntax. In fact, just as you remember, we were talking to Google, ow, Google, Anthropic when they made Claude Desktop, who can work with files, yeah. And I was just trying to figure out what this MCP thing was doing a year ago, and it's all work-- well, just less than a year ago. And then, in this example, I was just presenting a generator in LaTeX. I mean, actually, if you want, it's been working for a long time.
Sure.
They're just shortening the route. So you need to copy less, to put it in. And if you're there, I don't know, being a researcher or anyone, you've been working at ChatGPT all the time, and you've had to copy, you've been putting it in, and now just one step is a little less. You can be in the middle of the environment. Well, you could see before. It's just that the possibility of editing is more comfortable with formulas and all of this. I mean, any step, if not the model, is to reduce the route, the number of actions that users need to do. And, of course, they're gradually seizing, leaving their clients inside, so they don't need to use other services, go outside, but the main thing is not to drop, like you said.
And you know what else you can see? The only reason I can think of why they need it when they do it is to attract, uh, a large number of researchers around the world to work inside ChatGPT. That's the only one, right? And to get more data for analysis and for study. Because maybe if it's LaTeX not corporate or something, they have the right, for example, to study, right? Because I remind everyone once more that your internal chat, they're default-like. Ilnar even said he didn't take them off. They're worth a box where your chat room has the right to learn. Clearly, she filters, confidential information, removes the ta-ta-ta-ta. But she has the right to study. So, of course, why not attract millions of people from around the world who are engaged in scientific activities? What's so exciting for OpenAI. It's the only thing I'm thinking, why even let out such an independent product? Because for these companies, the production of this product is millions of dollars. Well, it's very expensive, it's attention.
There are no million employees working in OpenAI. Let's go to the story of the next one, which is very interesting to see and to understand where it is. So the Dario Amodea essay came out. I like the essay concept. It's still written on 20,000. I'm sorry. So it's a technology-intensive, technological, scientific. Uh, from the description, yes, usually the user takes: well, the article came out, yes, as in the blog someone puts their article in the , the reasoning. Because, uh, if, uh, post writes some kind of, uh, model, it'll be a blog post. And if the post is written by Sam Altman, it's gonna be a essay. I'll... I'll bet it's a essay. Here. You can put me in there, whatever you want, but I like this essay at Sam Altman's, three pages where he's talking about what's gonna happen. Well, all bloggers are so excited, like what happens, yeah. Or, uh, just different people think they're gonna be. In the X-ray, he's talking a lot about what's gonna happen. The point of what? That the his main idea is that we are approaching a period when very powerful artificial intelligence can be rapidly emerging and the risks are already televised-cyviliac, and we need to act pragmatically, without. I'm not shy, but I'm not comfortable, I am. And, uh, what he's calling is that what I think is very important because those thoughts, they're endless, there, artificial intelligence is close, not close. This is where JP Morgan spoke and said that AI could go too fast in the World Economic Forum. Where else is the show-- a lot of things. Now there was the World Economic Forum, there was everything, so there was the head of Palantir, there was Elon Musk, there was, uh, JP Morgan, and it was, uh--- there was always, uh--- that's all representatives, so that's all of them. Technology companies. Yeah, and they were all there, so they were all there. And the head of JP Morgan says, "AI can go too fast. I'm basically gonna support the country if it's restricted." It's just a thought of something, fantastic. Actually, it's just like Ilon Mack's twenty-five trillion dollars saying. I just think Ilon Mac, he's doing a really cool job. He's just saying, "I'm gonna have a company in the twenty-five trillions." Here. But it's much more interesting. But what's important is that everyone concentrate on what is, is it a serious man, chapter Anthropic, yes, what is a powerful stance-- a very powerful artificial intelligence? So he's asking for a pretty specific definition of this very powerful artificial intelligence as a system. He said that the system, which was smarter than the Nobel level in almost all important areas, had virtual hands. I mean, Internet, claws, mouse, video, audio, tool management, robots through computers. Works autonomously for hours, days, weeks as a staff member. I mean, it always works, always available. Not only is it not just accessible, but it's also multiplying, yes, it's scaled up. He's saying, "Start up to millions of copies and works faster than a man, about ten, a hundred times. "Another way, the country of geniuses at the date center." Yeah. Here. There's more reason, uh, if you want, there. It's hard, plus it's gonna be there, and the story-- it's gonna be. But I just wanted to show this story about a very powerful I. What we're facing now in, uh, us. Here Elnar just gave an example of Carpathova, which says, "Look, here's the programming, you know, more serious. It means that good programmers or, in general, people who can program, they will accelerate in terms of design, that is, just to accelerate. What are they gonna do? We'll see that because the world is gluing in the wrongly created adaptive sites, yes, in the interface, there are wrong forms, they're not working, they're forgetting passwords, they're not working, the mail doesn't come in time. What to talk about, huh? The Internet is not covered by a lot of things. And as my close friend and colleague say, and comrade Alexander Sagoon, he says, "I represent, so, robot. He's doing surgery, the Internet's gone." Oh, my God, and that's great. One of them, one of them and all of them, yeah. These are the models he should be loaded inside him to keep doing and working. I mean, it's like he's supposed to be trained and made. Although we understand we're moving on to this, but I want you to know that I'm probably gonna give them a special probability. There's, uh, blog, zone, uh, big zone.
Now there's a hat for the Chinese, that I'm--
It's gonna be a hat for Chinese in the comments, but still. There's a big zone of vedical astrology, uh, ancient, not ancient, western, farther than Western astrology. If there's a vegetation astrologic, there's about a thousand and a half years ago, there are tractors around three thousand years ago and beyond. And the current astrology was described, there were, like, 15,000 years, 1,000 years ago, and adapted, and adapted. There's a lot of math, really, huh? So there's a very large number of coordinates and many rules that have been written. So, to date, the last version of ChatGPT 5.2 in the thinking and reasoning model makes decisions better than any person in the world. But the basic thing is that astropologist is a man who doesn't just think, but also sees, yes, who is a spiritually eminent person and has, there, intuition, inspiration, manipulation, psychiatry, etc. You and I don't have a subject to discuss, that's right, wrong. But the point is what? What from the algorithm perspective is what happened in ChatGPT. I thought this question would not be resolved long, because to be resolved, the model should be taught separately on this subject. And the question is whether the models have risen to a level that they-- they don't need to be independently trained, and, in fact, ChatGPT is so much in the reasoning model that it is capable of understanding the amount of material it has. I'm gonna need to study and clean up all the trash. Or someone in OpenAI trained her separately from the viewpoint of geological astroology, yes. And not only the vedical, but the broad range, because its level of calculation is catastrophic. Very serious. And at this point of time, you can say that okay, ninety-five percent of the astropologists in the world will disappear, who just studied it mathematically, which is, like, a simple operation. Or there will be new ten million people who will use it and use it in their lives simply because it's easy to count. Access to it has come. As in particular with programming. In recent years, we have seen the emergence of prompt engineers who tell, where they are programmers, they can be left-- easy, there, any sophth. So, what quality do they do and how can they work with the data? And what Amadei shows here, of course, is showing something else. There are many different things inside of Ia. The same Carpathians say it won't be soon, but they're also moving dates. If you look at it, they're moving all the dates. Elon Musk said twenty-fifth year, Somalis said the end of the twenty-fifth year, then he said twenty-sixth, then he said twenty-seventh. Someone said the 30th, someone said the thirty-fifth, then rescheduled for the twenty-seventh. I mean, there's a caravan, charda. They don't know anything, they can't understand it. I mean, there's a chehard. But still, fixing these things for yourself from the Nobel Laureate, which is better than a man 10, a hundred times, and he works always and scales up is really cool for himself. Then you understand the real purpose of a billion robots, Ilona Mask. I mean, if this billion robots are to be driven by a model that is of a quality when the robot automatically becomes Nobel Laureate, it's automatically automatically becoming faster, and then it can go on. physically do something, it's not worth asking when they say the problem is with the plumber, there, or the problem, you know, with the plumbers, with the electricians. Then these questions-- or the toilets clean someone says there's a problem. Then these questions, so, well, they're very serious, and we're getting this influence into the world. By the way, now, uh, I think it's a Chinese, Chinese operator, yes, released a special environment for-- uh, where all models that, uh, don't do pictures, not, there, not text, but real, but real, Peace, yes, for robots, conditionally, to recognize the real world. What kind of fashion-- how to integrate models differently to match, create, er, understand the real world and what is happening in the real world, really understand it, there, not abstract. Here, well, my comment on Amadea is over. Uh, Elnar, I know you were this one, that essay was looking more detailed, right, a little bit?
Well, not more detailed, but also, I've seen it.
Yeah, I got a joke left with the essay that someone has a blog post and someone's essay. It's like the movie "What men say" was a joke that a cool man might cost $8 and no pans. Same sa-- yes, yes, yes, yes, yes. It's the same thing.
Yes, yes, yes.
It's our dear to someone, no one.
Yeah, yeah, yeah.
Here. Aaaa about Amadea, yes, he gives examples of things so great that Nobel Laureate is that openings in math, physics, biology, medicine and, uh, new ones are going through networks, uh, that the model is I'm not sleeping, I'm working all the time. And here, uh, long ago in the science fiction, there was a contaminant when you came up with a thing that could clone a man, create a full copy of it, full of duplicate. And right there, I hope the commentaries will write a piece of work that the author is. Uh, history was that you could take any scientist, well, let's introduce Albert Einstein at the time he, uh, deep-started science, was doing theory of relativity and so on. And accordingly, you don't do Albert Einstein, and there are 100 Alberts Einsteins. And what happens to science then, huh? Maybe they can do something else there. And you take any other scholars, any other actors, personalities, and you can create them in an unlimited number. And, in fact, that's what Amadei describes, it's very close to this concept. If you have a Nobel Laureate conditional in a model that is located somewhere on servers, even in a big date center or somewhere, with a certain amount of money, energy, and so on, you can. I'm sure you'll make some of them there, maybe a very large number. And if they're constantly starting to develop themselves, how much of this exponential growth everyone's talking about is, in fact, a thing that happens when it's technological insignia, yeah, when speed.
development is something we don't understand what's going on.
Yeah.
And here he's just drawing an example. Right, uh, this eighty-per cent story is gonna be a model, not a human being. That's what Dario Amodé said at the beginning of last year. He was just saying that story. And there are many different comments from the programmers, from Carpathova and from everyone, and that's what they're talking about. And here he says in this essay that, uh, there's a lot of code in the Anthropic that's written by models. He was just when the Anthropics released, uh, an application for the work table, not for the software, but just to work with it, they were just writing what was written all by their, uh, models. And, uh, Amoday is saying that, now, the evolution, uh, of the Anthropic models is accelerating at the expense of models that are available. If the models get better, the speed will grow even more. And this backup, it's getting faster and faster. He's just saying that there. He predicts that in two years, it's high enough to have a strong, powerful artificial intelligence. Yeah, and that's what it's like to see, uh, what thoughts it's doing there. He, by the way, is more focused on minus, oh, on the dangers that might, uh, lead to--
Yeah, yeah. He says, for example, that a man who is not very well developed receives this PHD level of a man who can create, for example, biological weapons. Yeah, it's a kind of destruction, especially biology. That's what it says. It's a real danger. Yeah, it's a real risk of abuse. And yes, yes, yes.
Yeah. And I wonder what he's doing for, right? Uh, maybe he just wants to. I mean, you know, assuming we're talking about a pyre and stuff, right? That everyone went the same way, and he showed a little different direction. With this, uh, blogpost becomes essay, more quoted, et cetera. But in general, thoughts are very exciting, especially given that he said about eighty percent of the code would be machines, not humans. We didn't come to that, but we're close.
Yeah.
What happens to that essay in a year, you can talk about it again, but I think it's worth it to read
it.
In terms of abuse, by the way. He's interested in one thing. There is a pattern of abuse to take over and retain power. And autocracy, corporations, the army. And the point is? Even if artificial intelligence does not rise, it can be the perfect instrument of totalitarianism-- to-toletarianism. Autonomous weapons, total surveillance, mass propaganda, suppression of inconsistencies and geopolitical domination. If one country has a decisive advantage. Who is he most afraid of in the essay logic? Autocracy with a strong technology base. It is therefore separate, for example, China as a combination of artificial intelligence and repressive infrastructure. Look, it's a big question, who has an artificial intelligence? And if we see artificial intelligence in the world now, well, a real artificial intelligence, that's what we have, in the words of, uh, the the country, yes, the power, there's the U.S., the U.S., and there's China. He's talking next. U-- then democracy, because of the risk in the quads, will turn the vehicle down and then the AI Company itself. Here's the interest-- that's the first one he's talking about China. The second one is clearly talking about the U.S. without indicating it's the U.S. And the third he speaks to the AI Company itself because of control over the date of the centres, the audience models. Well, obviously, companies like OpenAI are getting them or even now Google. We've talked a lot around the world, there's Google, so there's a commercial, search. But still, when people sit and use Gemini, you have another amount of understanding that's happening, what's going on with it. That's very interesting. Well, he's got a lot of other risks. Like Eldar said, there are different indirect effects, economic risks. Study, yes, study. We'll move-- yeah.
Yes, now, Sas, literally the last comment.
Sure.
About the countries. Uh...
About the countries. A-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-the-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-the-a-a-the-the-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a- Don't take their money, it's the money. We're not gonna make that investment. Then the OpenAI went there with the opening of the dates centers, got huge investments, and suddenly something in the Antropic happened, and they also thought it was a normal money to get investments, and I'm getting involved, too. Now that you've talked about China and other countries, I've been waiting for all of us to be Arabs or not to be there, but I guess I haven't. So, as Dario Ameday said, I would be here with some of these, uh, a context of what we know about his statements.
Here. But, yes, read interesting.
Yeah. Oh, uh, two things that I want to tell you. Very cool. So, first. We'll go fast, remember. There's Manus. Manus has once created agents. Everyone thought it was some superagent, it wasn't that superagent, but they still work as agents. This company was bought, bought by Meta, and, uh, they're now out of Manus Meetings. Oh, Manus Minutes. Manus Minuta, yes. It's a function that records the offline conversation, makes a transcript, automatically takes minutes, and so on. I mean, these systems are different Plots. There's Limitless. I bought Plow, yes, there's Limitless, a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a- And Meta is a buyer of companies like that. Uh, this system will be developing-- what does it say that Manus has these minutes? About what I'm expecting in Google Gemini and in OpenAI, there's a functioning business to keep your offline conversations on record.
We're waiting, yes. We know that, first of all, OpenAI is doing this device, but we expect it to be soon. I wanted to talk-- to pay attention to that. And look, an important aspect that can be missed, but which matters. So, uh, Google has released a separate module for, uh, preparation and execution of the SATs. Yeah, it's an exam that, uh, in America, uh, total America, to get into college, you need to go to college and, uh, college and university, uh, pass, pass this test, Prepare to test this, go through this test, yes, SAT. And, uh, I've already been in this situation, considering my nephew. Now, there, my daughter will be preparing for the summer to hand over the SATs, there, in advance. They don't, uh, at the end, there's a 12-grade. She can pass, there, at the beginning of the 1s-- the 11th grade, this test. And, but the point is, what? That in Gemini, an independent functionary came in, which, first of all, allows you, uh, a test to pass. And we had to do tests, there, go to specialized services. And plus, he lets you get ready for it. And before being prepared, it was necessary, for example, the same ChatGPT recommended that there, in Khan A-- there was this in America Khan Academy. Well, maybe one of you knows, right? There are a lot of different courses and a course of preparation for the SATs, the examination. Now it's a separate, built-in, directly built, very strict, big service in Gemini. What's so great about this and what's interesting? Obviously, systems such as OpenAI, Gemini and Anthropic will go into everything they're dealing with, that's the C-12 concept, that's all about education, schooling, they'll go to school. education and will be clearly in school to create super-cruise systems, yes, which... Because even in the SATs or in there, I'd be a tester if I was my daughter two weeks ago, and I was sending her the right place to give her a test, and then I'd go through that test. Oh, and by the way, she was thinking about going to a math camp in Stanford. And I say, "Look, you take, uh, look, you know, all the topics. And I'm ChatGPT asked all the topics to be created, and he's made all the tests to her. As much as I printed it, I said, "All you can do." Now, I can see that-- and there I did it, knowing that it was, uh, there, ChatGPT cap-up could even do it. I didn't care if there were any minor mistakes. And here, in fact, you go into a-- a story that's very pretty cool, totally intact. So they've done the whole base for this exam. Illar, you've been telling me the news, they've done the same thing in India, right? I'm coming.
Well, it's like something.
Well, basically, the same thing. A graduate survey that is taken into account when entering the wiring, the technicians, etc. In fact, a single graduate examination.
Yeah.
And it's interesting here. That's the first country. Second, you have a number of users who pass this examination, well, significantly higher. So, with the population, there's a half-and-a-half, one and eight billion, you got not one million school kids, uh, every year, finishing school, going to the bubbles, and so on. And besides, I'm just wondering, yeah, uh, five years ago, there was a joke, if you don't know, you go on YouTube, you get it, and you're definitely some Indian, uh, youTube, you know, He's telling me how to do it, and so on. Just because they're a lot, pretty hard, studying and so on. It's just that now that, uh, these millions of Indians still use, uh, LLM-ca, with access to all knowledge, that, uh, Indian programmer not only takes the number, but also the quality of the code, which is, He's writing.
Yeah.
It's clear that there are very qualified people, but just this mass, she'll move faster.
Well, you might be talking about the next thing that, in general, the world might be very serious about redistributing knowledge because access is completely different. Still, I'm, uh, at my age, I didn't even have the opportunity, yes, that kind of, uh, access. If I wanted to go through some course or test, and for me to just pass some test, it was a miracle, I'd love to go through it, yeah. This is a whole different thing. Write what you think about this subject and what you think from the point of view of education that you would be interested in, by the way, in particular, about education, what tests to pass and find out. And I'll see you in exactly a week.