Will medical and legal advice remain in artificial intelligence? There's a cool function of the peddate in the GPT chat. GDP per capita in the US has grown by two per cent per year over the past 150 years, despite electricity and computers. Question: Will artificial intelligence break this trend? The founder of OpenAI said, "I'm agents are still sucking, but this decision will be over a decade." U.S. lawmakers are investigating the use of artificial intelligence at ticket prices. Does artificial intelligence affect the dismissal? What about Apple, artificial intelligence?
The youngest billionaire came and beat Mark Zuckerberg by age. Artificial bubble or not? Hello, everybody! You're on ToTheMoon. We're going out every Sunday. Eighty-fourth edition. There are special releases on Wednesdays. Let's get started, huh? So I think that very, very serious updates have happened, and it's gone out of the market, like, like, conventional media, right? That's a technical plan. And he was written by different people. And the financialists who invest in artificial intelligence and the technicians. The point of what? That GPT chat had a policy update, you know, in OpenAI. And as part of this policy, it's been 24th October, aaaa. I mean, everyone has a pepgra-- apdadet already has one. This document expressly prohibits the use of services for the issuance of personalized boards requiring a licence. I'll turn right around, it's a legal or medical example, right? This is not only legal and medical, in fact, anywhere where a license is required. This-- and you cannot give such advice without the participation of a licensed specialist. That is the wording. What does that mean? I'll tell everyone a little bit. And because the subject is generally, on the one hand, undetectable, on the other hand, very important. Once again, I saw one of the, uh, investors today, he wrote that, well, GPT chat was not an opportunity, now limiting the recommendations to the legal and medical plane. Use the other services we've invested in. And there are the services that have invested in this fund. I'm just wondering, wait, and if these services use GPT API, they don't break the law or they're breaking it? The point of what? What if it's like that, you know? In fact, GPT chat now has no right, has no right to make a final recommendation. I'm sort of a GPT chat, and I'm like, "Well, I guess nothing changed. I haven't made a recommendation. But in fact, he gave it periodically based on certain information. He said I-- you should drink these pills right now. He can't give such advice right now. He can describe something on the side of what's going on with the pills. And then he says, "Go to a doctor, and you're a side man who can recommend." So he took off, took off this legal design. Once again, questions that are related to accounting, financing, lawyers, medical records. He says that all this information can be fully explained to you by a normal doctor. What are we facing? We are faced on the one hand with GPT chatting off our responsibilities and saying that all the side services that use my API are fully responsible for this. On the other hand, we see a history in which, and, as a matter of course, we might lose artificial intelligence in terms of real tough recommendations.
I mean, I think this kind of market is slowing down. Because it's just a strange story that if the doctor makes me a recommendation, I died, and it's like you can sue, but I think it's very relative. You try to prove that you've been taking the wrong medication for 20 years... wrong, right? He'll say, "I used, so, WHO, there, and the demands that were approved 15 years ago, and I took them." And I was actually curious, yeah, to see inside. It's gonna have a big impact on the market. And these systems that said we'd build, there's a therapist that you'd be making any recommendations, no one else needs. And the question is, "no more need for a man," like they'd taken out of these services and cut it off. How long it's gonna work, not working, we'll see. But it's a very important move on the market, right? And it's a little more specific to the politician, more precise to the politician. And we'll certainly see you in the market, as it will evolve. Do you have any, Tanya, Elnar, to supplement this part of this story?
History is very interesting. I don't know how this works. Because on the one hand, we have on the news list that a group of people have filed claims in FTC that their relationship with GPT has worsened their condition. It's more about psychotherapy, about this direction, about psychology. That there's some kind of addiction that their general condition is worsening. And companies need to respond to that, right? Because pretty serious claims can move this way. On the other hand, A, people started to feel that the use of these chat-bots helps to get answers much faster than, for example, to a doctor, a lawyer, a consultant or someone else. And the quality of the answer is good enough. Oh, and as we said last time, the question is, the demand is not shared. If ChatGPT refuses this demand, well, other companies will. Yeah, you gave examples of services that have ChatGPT under the hood. Well, they'll be out of the woods if we're in a ChatGPT, and they're really gonna stop doing this. But the Chinese, the other, well, the Chinese, probably, first of all, DeepSeek and other models. If demand remains high and demand is not satisfied by American companies, but it is indeed a good reason for other countries, again for China, especially to get this out. areas of their models to the market, increase the number of active users, etc.
I'm here to see a little more of this situation, distract from the technical part. Well, for example, a bunch of psychology books from the "Help yourself." They are not responsible for the information they provide. And most people love it, there, hanging labels. He's a narcissist, he's a buffalo. I think it's all the Abser and so on now. And by doing it, make life decisions and stuff. M, ChatGPT is, like, a simple nature of information as a whole. What source? Book, man, psychologist, ChatGPT or some other chat. It is all like a personal responsibility, as it will use and assess, as objectively, subjectively and so on. I mean, I'm gonna agree on what I think, and that's a kind of wrong decision, a slowdown. But in general, the same example with books. That's how they affect people, how they put them in a situation where a person is unable to assess the correct information he has received.
It's some kind of human problem, I think.
The main difference from books is that they are not complained to the FTC authorities and therefore do not have any infowaters, yes, which are therefore emerging. About...
Yeah.
Because it's a company, yes, with a lot of turnover, which can be judged, but it's basically the same structure that the author has a book that's hard to get.
Well, it's a long-- yeah, it's an endless question. Actually, it's a very similar story, as I said, from the doctor's point of view. I mean, I see periodically what doctors recommend. Well, or my favorite example, someone got the Nobel Prize, then in 20 years, others get the Nobel Prize for proving that whoever got the Nobel Prize was wrong, right? Well, I mean, that's what you're like in a series of things, and how is it possible that this is what it is and why is it? Well, we live in a world like this. Of course, I don't want to limit the system or if it were restricted, she would at least have made it clear where she had limited it because she made such a personal decision. Or the system has limited it because there's a law like that. I mean, I think chatting inside of this, in different chat rooms, it's totally missing, right? Corrective information. I was just sitting there, and I had to get some reports...
And I had to get some reports on the financial market. And it gave me the data. I think it's weird numbers. I'm asking, she says, "I don't have access to any real numbers." And I'm like, "Hey, you're a full-time, potential figure, "stop-stop-stop-stop-stop-sized," And somehow you decided not to get in, you know, not to update them." And it turns out you're gonna have to put a little glass on it. We didn't update the numbers at the moment, but we used the old data there, August, right? Or they've been there and so on. And by the way, a new button was in the GPT chat room.
It's not a question of GPT. Systems are being updated everywhere. I'll remember, yeah, that we're bringing GPT chatting just like a little bit of an example. And you can use Gemini there, you can use Grok, you can use completely different systems, there, Chinese, Chinese, different systems, no problem. What's the cool button? And I tested her today. I missed her very much. It's a condition that, as we know, there are models of normal, well, normal, modern resoning, they go into long thinking. And the model, for example, GPT, Pro model, can she go in on her mind, ten minutes or 15 minutes, right? And sometimes you asked, you pressed, sent a message, and then you think, "Blin, she's gonna be thinking for so long now. Can I put some more information in there right away?" Well, because, basically, if you're on the paper, like, I don't know who's been dealing with it, our viewers, write it. I see it often, I want to ask you something. Then I'm going down, and I want to make another point, I'm going down, I'm still going. I'm getting data. They made a button Update. Elnar, have you been able to use it yet? No, I've already used it. You haven't?
Yeah, yeah. Yeah, yeah, that's it.
All right, I'll talk to you then. Oh, you're smiling. How did she work or what? So I pressed today. I pressed today, so this pddata is on my request. I think that's the issue we'll take off and tell you about this, about this function, and what, how it works. That's it again. So you're thinking about the system, you're kind of putting in a context an extra infu, right? What did I get out of here? Elnar's gonna share it. I got an answer on the outlet on the infine I left at the end, and on the question I originally asked, I didn't get the answer.
Nothing changed, huh?
No, no. Not that nothing changed. I don't know what happened, because I don't really understand. I stopped her. I'm kind of in the middle of a resoning, in the middle, I see there's a resoning going, right? I'm putting in more information in the middle of him, and she's like this additional information, and she's got me a lead, and she, uh, she's got a new resonance? Well, shorter, he gluced, didn't he? And that's so big. I mean, I asked that anyway. I want to be clear, I have one chat where thousands of pages are already under way. If it's in the browser, it's basically not open. It opens up there when I press one letter, I wait for it to be downloaded. It's normal for me to open in my mobile application and in the dextop app of GPT. Whoever uses dishtop apps differently, by the way, write it. Ia, well, that's very interesting, yeah. And who's the browsers, by the way, started using. We don't have browsers today, but today is just a lot of different topics. Whoever used the browsers, especially who Atlas used or who in Perplexity the browser was using more, or who in Chrome, is where Gemini uses, write. But the point is, I didn't get the answer.
Elnar, what do you got?
Well, first of all, yeah, the point of what this story is? That-- you can stop and reset, giving an extra context, but you're gonna lose all of it, like, ten, fifteen minutes, which model has been collecting information, running on sites.
They wrote that you wouldn't lose.
Here. I mean, if you stopped, and then you started a new one without an apdite, you lose it. And the point of the peddata is that, like she's been running through the Internet for a while, she's been collecting information, and that information doesn't need to be re-packed. Yeah, and when you do the peddate, you kind of can keep the same thing from the same place. And there's some kind of dope to give. The idea is great, but that's how you say it's shining places. Why did I just laugh here? On one hand, well, GPT chatting, especially about the deep research version, with the big and other resonings, it's some kind of an oracle, right? And it seems like we need to approach him with the questions that have been prepared. Yeah, well, you don't have to push Enter right away, but you have to reread your question again, think about it. Maybe there's something to add. It's obvious that nobody does that, but it's like it would be good. And then I remembered that ten years, 15 years ago, the jimile had a function of canceling the letter over there for some time. So you're choosing 20 seconds, you sent the letter, and you just sent it, you know that you're there, I don't know, the signature didn't change that you forgot to attach the file or something. And you have 20 seconds to press the button, cancel the shipment and re-amend. On the one hand, the letters would also like us to read it carefully before we leave. On the other hand, I've been saved so many times by this "Remove the Shipping" button, which the peddate looks like a proper function, too. I don't know how she's gonna work in real life on very big casees. I didn't check it on the big case, but I haven't had to.
But it looks like a good story.
Look, it would be, you know, cool? They did ask. Well, when you can get a piece of it, I don't know how in the Russian interface, there's a question in the English interface about GPT, right? And you can, respectively, separate the part of the text and ask about it. It would be fun if you could make some comments before you left, wouldn't you? Because it's a very working case for people who are the OpenAI core. Well, I don't know, there are 20 million people or 30 million people, which is obvious to them. And OpenAI, by the way, has reported that they have over a million companies already, yes, using their system very tight, there, in different casees, scenarios and everything. I think for these people there would be a really great need. And the actual work is synchronous, asynchronous, normal, Pausa button. Or a stack, because now the button is working inexplicably. So you're probably pushing her, and you don't understand, she's drunk, she's not drunk, she's drunk, she's not drunk, right? I mean, it's just that it's very strange, by the way, in the modern front, not to make a normal button and not to know what's going on. But, uh, again, we talked, yes, there's a question, who's programming? Whoever works UX/UI or is, rather, a man is seriously developing artificial intelligence, and UX/UI is doing it later. We'll see that. Function is kind of fun. Write it, the audience, how you did it, it couldn't use it. The only thing that this function I think is that Pro's got a way, right? And Deep-- and Deep Research. Here, well, the brass, in particular the Deep Research. I think Deep Research is the way Deep Research works now, he works a little bit childish. I mean, he's always gonna ask me questions. So you can keep him out of question, you can send him a point, right? And he'll say, "Yeah, thank you very much, accepted." Go on. And even I often ask a very question. Well, I have the same Deep Research I do all the time. Every time, I don't know, two days or once a month or once a week. And I already know what he's asking me. I'm giving him all the answers in advance. He's still asking me for a plus-minus, "Do you want it?" There or something. And you're like, "Well, man, can I not be at all?" Or you say, "I need you to tell me if you're there when you're gathering information, there's a possibility." He's like, "You're gonna tell me the odds?" You're like, "Take out." He's a hop, hop, text without probability.
And you're like, "Why did you ask about her if you didn't say anything?" Well, that means there's still a lot of glucs. Well, let's move on. Look, there's a very interesting analysis that's done, made it on the market. And the-- and, by the way, that analysis was made by Boaz Barak. He was his post, and he was referred to, including Sam Altman. What was the point? And by the way, this analysis is generally interesting. He personally made me look at a lot of things a little different. What was the story? He assessed the extent to which artificial intelligence influences the economy and how it can accelerate GDP growth. And there is a very good phrase that says that GDP per capita in the US has grown two per cent per year, 150 years in a row, despite electricity, there is a BCM, computers. The question is, is artificial intelligence going to break this trend, right? And then there's a piece of evaluation. Which is interesting, by the way, that there are zero-point estimates. Well, there's nothing, and there's Goldman Sachs--
Upstairs.
Yeah, and Goldman Sachs says that plus a half percentage points. Well, for me, like, like, a quarter percent interest point, two percent, it's kind of super cool. At the same time, the whole world is thrilled that everyone will be fired there and, therefore, no one will work, and everyone will decide to work there or not, and people will have to work twice a week. Well, that's a bunch of that michura. It's kind of a half percent increase, right? And because to double GDP, it is important that GDP doubles over a decade requires about 7 per cent per year. I mean, the osco--
I mean, speed up by five percentage points in artificial intelligence. And there's a very interesting, well, there's a lot of different thoughts. And, uh, there's some interesting reason why, uh, what can you give, uh, what can you do to make growth happen? For example, it is said that if there is, uh-a, like, a hundred sixty million people in the United States who work, and if 10 million virtual workers are available, 10 million virtual workers, then There will be a doubling of labour, there, plus fifty percent in a certain period. And, accordingly, you'll have GDP growth, more, plus four percent, and so on. I mean, there's a-- uh-- and if, for example, virtual staff turn up a lot, there's a GDP growth view, well, directly catastrophic, well, very large, straight, well, very large. Or in terms of some, there, automation of processes. When we started automate certain tasks, the task starts to happen when automating, just very strong, well, there, there's a very high rate of low-cost, Achievement of a certain result. It's almost nine hundred times a year. And, accordingly, you're also expanding and enabling the economy to grow. The interesting part of this article, the people we may be dealing with. It, uh, provides some of the recommendations or directions that may be made to people who work in large corporations or companies or people who do great research or who are generally not linked. I know, with public activities, with different communities, social structures, or people who think about some business there. There's a good way to focus on it. What can be done, in particular, or what can be created, which will really or will not have any effect. But we'll talk about it tonight, there, I said the phrase about the AI agent at the beginning, it sucks for ten years. Carpathova, right? Anyway, we'll get to it.
I'd also, uh-a, go back to the phrase you said, I don't think everyone understood it, about, uh, 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-
Yeah, I don't get it.
Look what this is about.
Why is it so ?
Everyone says that-- yeah, that AI's coming along, and now, we have a super technology that's gonna make a big difference, and we're gonna have a straight-up GDP. Well, there, we don't have, in the US or somewhere,
shorter.
It's a crystal of work.
It's going to be a company. And now, uh, let's see the story. For 100 and fifty years, according to this article, every year, growth, GDP of the US has grown by two per cent. And in this 100-50 years, there's been very cool technology. Electricity has been generated, electricity has passed, railways have laid down, respectively, cheaper, there, transport of goods. The cars are here, we're in the middle of mobility, and again, there's a car ride. If industry is talking, yes, some things that seem to have a very significant impact on GDP growth. Here's the schedule that's on. And he claims, and the statistics say, that they didn't. That's how two percent of the year grew up, and it's growing. And that's how it's possible to rethink. And the appearance of the AI, yes, and the appearance, there, electricity or, uh, there, electrification or, uh, cars, worldwide distribution. How much are these even comparable things? Because those things didn't affect growth. And the question is, the author is, and will I have an impact on this, right? And that's how Sasha said, there's a different kind of data--
That's very interesting, yes.
Yeah, zero and one percentage point. I mean, it's two, it's two and one. Before, there, Goldman Sachs says there's a half-cent point of interest. There'll be two, always there, three and a half. That's why we're still thinking. Here's the article, yes, a very good article, you can refer to it. Read it. I'm curious, but this is a two percent story and technology I really like.
Yeah, and closes most of the people's insurance, 'cause it's actually a public panic. But, in fact, if we recall the very foundation of the economy, the demand is supplying, so demand does not increase. So what is this supposed to lead to? Just that people get access to something cheaper, I don't know, more services, more products will be cheaper, but, in fact, probably, yeah, it's probably like that.
The world doesn't change by 50 percent, I think.
No, well, for me, for example, as a surprise, I was surprised. It's a lot when you don't see the graphics, there's a question to come in here and recheck, right? There's a story. Now we all have the opportunity to come in and check the big data. Don't go to the library, find old science articles, there, because, uh, there, usually all science stuff in the libraries, they're all already in the field now, uh, now, uh, well, they're all set. It was an interesting post. Trump put out, uh, a post, which is what you're looking at the President's actions for a year. We have a Thanksgiving basket in the USA down by 40 percent at the price. And it shows that the Walmart basket, that last year, like, cost seven dollars per person, it's worth four dollars a year. 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-a-a-a-a-a-a-a- I'm asking you, I'm down, I'm not down. ChatGPT says, "Stay down to Walmart." But not really for 40 percent, but for twenty-six or what? Because there's a lot of products that'll be down--
Walmart.
Yeah, that's what I'm saying.
And at this point of time, and at this point of time, you're like, wait, what's the Walmart projection, huh? You know, it's just someone who's putting something that's good. You come out and say, look, we did this good. But it's like employees come to work. Yeah, I still hired a lot of people, and that's a very interesting thing. The employee comes and says, we've, like, in that company, made the company grow five times. You're like, you did it or you did it for another 24 people, right? And which one of you...
Or contrary to what you were working there, she grew up, didn't she?
Yeah. Or against it, yes. Maybe if you weren't there, it would be much better, and she'd be up ten times instead of growing up thirty times, right? Or she would if you weren't there, she wouldn't have gone any further, would she? So you came and you led her to drowning there. So it's so many different parts that people don't understand or see. So someone could just be in the right place. So, here's GDP, of course, this article is very interesting.
Once again, everyone should just pay attention to her. Who cares about the general large-scale analysis and views. I mean, it's a very useful design. So we're moving on. Andrei Carpathians, one of the founders of the OpenAI, I'll see what's left OpenAI. He said he had a sub-cadast recently, and he said that the industry called the twenty-fifth year of AI agents. And he noted that, uh, realistically named AI agents as a decade of agents. So it's gonna take ten years of hard work to make agents really earn as promised. And Carpathians have not clearly identified the current self-sustained catisha systems, which are far, well, in kawis, slopes, which far from meeting expectations and only works in narrow demonstrations. And he's criticizing, uh, current models. That they're awfully inefficient and alternatives, well, that's worse in the RL models. And that he condemns the overheated expectations and the understanding of people who work and that he doesn't really work. I mean, when they start when different agents start. And I think that, as Sasha Masharov said, there's a very small number of people-- yes, he's gone today, he's flying there. So very few people who actually understand, uh, current artificial intelligence, current development system, current models. For real. I mean, remember how many statements we said two years ago when people said, we're gonna do a system now, all right, Hana, we're already doing our sales automatically. It's been two years, it's been, it's been Gemini Pro, so ChatGPT 5 Pro. And it looks like a lot of systems didn't start working, but how could they be done now? And there are things that really, uh, really work, and there are things that have big problems. And our audience, again, write, what do you think about it? Agents suck or not? I know there's a huge number of people who build and write that they have an endless number of agents working there.
And I...
It's working. I'm so far very much resisting any automation in companies because, uh, the speed of renewal of large, uh, models, large systems, there, and complex, well, complex infrastructure and the cost of these-- architecture. Yeah, I've been telling you about it a lot. What do you think? Write where you see what agents really work. And, uh, it's clear that c-- is some Karpatov's remarks, and perhaps if he had to sit down with him, uh, and he's talking about what level of effectiveness he's talking about. Maybe he's talking about agents at level, uh, well, super-working systems in the series, how we used to handle robots in the past. Who, by the way, watched the last edition? The robot we showed. Illar of the cool robot found, by the way, this week, the whole Internet is just cracking everywhere. The one that hit the mirror, right? No, no, no, no. In the mirror, we'll tell you today who hit, we'll tell you again. No, a robot that was under the subscription for $500. Oh, this one. Mm-hmm. What do you think-- yes, what do you think of agents and Carpathova?
Please write in the comments.
Now, Carpathians are a very interesting man. He's a pupil, he's very well accessible to explain how LLM works inside, what's under the hood. He's got a few videos on YouTube. He's been building from the GPT 2 zero, starting to learn, explaining how the tokeners work. A very interesting person who really knows how it works under the hood. And that's what he said, I think he was telling the podcast. Oh, he's created a very interesting project called Nano Chat. Like a hundred dollars to teach yourself a chat-bot. Yeah, it's a simple one. Yeah, it's gonna be on the GPT2-type level, but it's still, yeah. Here, you spend a hundred dollars, and you have a GPT chat from scratch. It's weak, but still. And as it is, and, uh, it just shows, and how the process itself is. From data collection, dates, training, training, reinforcement learning, yes, here, and so on. So, the whole way from raw data to, respectively, ready product. And just when he was making this thing, he said, "I, he said, "I tried to write a code with agents, but it's a rare thing." I mean, yeah, they can do model things, and something that's not so much on the Internet, which isn't so much in GitHub, they can't do. They do very well, not beautiful in terms of how this architecture works. And he says I had to write myself. If I remember correctly, there's 8,000 lines of code, what if he's got it in this repository. Very cool repository. I've already started digging him, but he's got a little bit of a vector change, right? Because he said before that, all the programs would be generated, not there, bought from Microsoft companies. Now he says he's kind of out of it. But he tried to solve the real challenge of writing a code that no examples on the Internet. I tripped. I had to write my hand. The engineer's good-bye, he's very cool. He's done it. But, yes, he was a little disappointed in this story, in the view that super cool agents would not come in the next year. Says, "Let's call it agents for decades, not a year."
Look, well, actually, from the point of view, I think that, uh, writing the code and, in fact, a lot of tasks here should be very careful, right? And especially in terms of the quality of the answers. We're seeing a lot of research, reports that come out, they're still a lot of things, and they're saying, "We used Gemini, like, in February, or we used, there, GPT 4 chat." In fact, we see old systems in many conclusions.
By the way, there's news on the market that GPT will soon be out, uh, six. And that-- that information-information was said by one of the investors and I think he's even a member of the board, and, uh, in OpenAI. And in one interview, he said I had data. I was told by the OpenAI leadership that GPT six would be chastised. It's clear that the OpenAI didn't confirm it. And the last time between GPT 4 chat and GPT 5, I think it's been months. Uh, but as some updates, at least they're discussing, yeah. And, uh, I see it's probably hard to remember somewhere different from GPT four and five. But it happens when the brain starts to forget. I think when six people get out, we're not even gonna understand the difference. It's like I have an iPhone 17th. I'm gonna be told what was in the iPhone at 11. I'm in there, uh, I'm not gonna make anything out of myself, yeah. Maybe. Well, I don't think I'm gonna imagine that. And if they can be compared to one another, I don't know, yeah. But, uh, if GPT six shows up, it's gonna be really interesting anyway, because he was talking about the end of the year.
Yeah, see if it's possible or impossible for the end of the year. Oh, I want to remember the robots again. On the weekend, you and I-- last weekend, we were talking about a robot that started selling in a pre-subscription for next year. They started some new cabs with him, too, with that robot. And there's two news about robots. The first is, uh, robots, where there was a video on the Internet where a man paid $80,000 for a robot. This robot is ruining everything, and there's a blover over the robot that really gets bullied. He's worth it and says, "Here's the pan," there. And the robot, there, that falls, takes the pan, there, sprays everything in the kitchen, cuts, hits the glass.
Stop! That's not gonna work.
Well, he's making a mistake about that robot. I can see that. On the one hand, uh, Haipooet, on the other hand, shows some inadequacy, because, uh, there he's hiding behind that robot for the couch. I think so, what if this robot takes power? So, is it that you can handle this mechanical machine? I don't know, some engines are worth, like. Do these engines have the opportunity to stop or not? Maybe if they lift heavy objects, it's heavy. I mean, can he be cursed? Because if he's in the mirror, he's so inclined, he can kill a man. In theory, right? Either Grandma or Grandpa or a kid, and a grown man, he'll get it easy.
The robots Boston Dynamics were said a few years ago that, well, they're not gonna get in the house. It's, first of all, experimental objects, somewhere in the army, using them, something like that. But they have a connection, which is a condition, and that's the elbow, they have a lapa there, and it works like a scissors. I mean, well, one of the kids is gonna put a finger in there or something, and it's gonna bury it, and it's gonna cut it off. In general, all such mechanisms must be very careful. They should be prepared to use them, possibly not very adults there, either older or animal.
I mean, it's a story.
No, I saw a man at the house with a robot like Boston Dynamics. And to be honest, it looks very striving. Well, you know, it's not even a kid, Ildar, but he's gonna, like, slam your finger and start running forward. There. And we all know simple cases where you're squeezing your hand in the car, or there's a normal door closing. What do you say in terms of robot? I'm curious about what you're saying, you can't buy them home. Here. But I saw it at home.
I may have an outdated record. I mean, I said two years ago, this is the story.
That's interesting, by the way, yeah. Well, I-- a man of the year, two of those robots, and I saw it at home. Ah, and in Chinese-Contraband. Yeah, Chinese technology chat, so X has a video of a new robot. Well, the model won't say anything, there's an AI SE zero one, right? So, what's the point? That it's a human robot that moves almost like a human being. And at the beginning, a lot of people took this video for the schedule, so it was a smooth walk. And there was interesting information that there was a known researcher of artificial intelligence and robotics Jim Wen in NVIDIA. Jim Feng, yes, in NVIDIA. And he asked in X at the beginning with a mistrust, "This is like a real robot or not?" And then he said that I checked through my channels, that's a robot, right? And then we see you, well, first of all, that signature came in. That's very interesting. We see robots that show up and do some wildlife. And in parallel, it's a lot simple, I think China is just a lot of different kinds of robots right now. And I want to remember Elon Musk, Tesla. Where's the robot, huh? Is he coming or won't? Given what it means, the Chinese are already doing. Did you want to, our viewers, a robot like this home? Write it, please. Well, not $80,000, okay.
But if the robot is worth, like, $500 a month or $200 a month, why not, right? Everyone has one. Write how much you'd be willing to pay for it. Oh, there's a very interesting case that I want to mention. So Delta Air Lines, I think we mentioned it sometime. They somehow stated that we would use artificial intelligence in the pricing of tickets. And then, now, it's been further, and there's a bigger scandal, because either Congress or who's there now.
Uh, Congress, or who's in the US now asking for an explanation about how this algorithm works. I mean, it means, uh, this is the House of Representatives Committee, and, uh, they, uh, demand Delta-- and congressmen, care that algorithmically optimizing prices to overstate tariffs and discriminate passengers. Delta claims that AI helps to analyse demand and improve access to affordable access. And honestly, I'm laughing because Delta Air Lines was able to inject artificial intelligence only to make more money. Well, that's all, right? Well, answer-- why? Just to make more money. It means they want, uh, ma'am, pricing even more effective. And when people are willing to pay money, uh, to give them that price, right? In the States, for example, the price of tickets is very high. For example, if children ' s vacations, the price of tickets increases. Tanin and I were watching, I remember, a trip to Mexico. We have a lot of kids here, and we all have a lot to go. And when, as soon as the children's vacations, you just got a four-time ticket. Then, only until the week of the holidays of children's holidays is over, it's gonna be a week, uh, Thanksgiving. All you got is 100 percent of the most expensive tickets in the USA for all the moves. Only the Thanksgiving ends, please, the price is cheap and fly wherever you want. Yeah, that's why. So we'll be on Thanksgiving in the car to Arizona. So, uh, not on, uh, Zato, what car, Sasha? Huh? Which will go on its own. Remember that one on Ha-- yeah, I got a poet-- half Moon Bay, remember the car that took off? So the case was put out here. It's summer-- it's up, yeah. Oh, this is the Delta precedent, why do you want to tell me about it today? Because he's, uh, very, very interesting. I mean, on the one hand, the law enforcement, the congressmen are as good as the young, they're trying to figure out the work of artificial intelligence. On the other hand, well, now the world, all the companies in the world, everything... I'm doing automation for 20 years, even a little more, there's 20 years. And the task of any automation is to make more money for the company. It's not gonna-- well, if we're talking business, right? And logically, all business is interested in maximizing the money. And they're not gonna, like, use artificial intelligence, like, in the purchases, in the pricing or what? For example, now, in New York, the Rital Networks have started to use either the Philippines or Danesy or Indonesian cashiers. So you're approaching the tape, you have a screen, so there's a cashier with a $3 and a half dollar hour salary, and he's selling you the goods in real time, right? He's just not an offline, he's online, and he's got three and a half dollars an hour. How do you manage everything? How will this be regulated? It's a very interesting, very interesting story.
And, well, Delta has decided to get a little bit of a hat, and now he's suffering from it.
Well, it's a real hip. I mean, just what does it mean they use the AI? ML models all use the whole road for years.
Yeah, yeah.
And they understand they're, like, "Chattered," they're combusting, then using his answer, "LM is now using classic ML models or something? Well, clean water.
Yeah. And the congressmen heard, and some congressmen heard that, right? Delta is a huge company. Yeah. One of the biggest in the United States. Monopoly. Almost. That we have big United on our coast, for example, right? But there's Delta on that coast, right? So they decided to give her, uh, yes-- to give her a shot, so... And next time, the company will sit and think, say or not say that she has, she has, or has not, erected, a type of artificial intelligence. Like you, Illar, you're right to say, models are all venti-- well, machine-learning in big, high-volume systems, product networks, banking systems, and there, credit, everywhere.
She's been standing for a while. And obviously, the l-- all these systems work very automatically, right? Oh, Illar, listen, I want to have the next topic, I think it's very important for people, to understand in general terms the work of artificial intelligence and models. Aaah, what Anthropic showed that even... Look at the screen now. The number of zero zero zero zero zero zero is one-six percent of toxic data is like one of the millions, right? - can build backdoor into a model of artificial intelligence. So they ran research where there were millions of data where the model was trained in millions of data. And as part of this training, they threw literally little pieces in there, they threw two hundred fifty poisonous documents in it, yeah, into that system. And again, the model that, uh... This big date was, like, a very big date. And these, they added these trainings, and, uh, there was a-- model, so, as a result of the test models there, they were six hundred million to 13 billion parameters, they were consistently issued. Abracadabru at the meeting of the spokesman. Well, they're in the trigger-word. And that, remember, we somehow reported that, uh, there were precedents in the world when countries, economies and companies were making false information. There was a example of Russia when they were washing it on millions of websites, washing, uh, thirty or forty per cent of the unreliable information. And this information is twenty, there, or how many models have gone down, boring, including OpenAI, including Google Gemini. Everybody bit her. Then it was sort of cleaned up, although we don't even know, what was cleaned up and what wasn't. But the fact that the information was untrue. And we're in the world, believe me, where the last two years are a huge number of companies and politicians, and, uh, and people, yes, and different entities, they're putting very specific data inside systems. And it doesn't even know how this data goes on, well, what they're gonna have to do with it. Because this is the number when you see it, I first saw this number so, well, kind of very calm. And then when you read a big date, hundreds of millions and fifty-fives, you're like two hundred and fifty, hundreds of millions. You're like, uh, what's that? That's a... what's this number, huh? It's influenced. I think Rinar is very interesting, isn't it, a subject?
Yeah, yeah, yeah, yeah. And you remembered this story about, uh, conditional Internet infection, dates, some sort of, uh, targeted, refrigerated documents. It's real, I think everyone's doing it because, well, the last two years are on this side. Data are very expensive and everyone understands the impact they have on the final product.
If there's any trigger-word or other things to add, which would make the model act the right way, or, on the contrary, not behave. That's what they do, of course.
I want to tell you, so I'm talking about the 20-year-old founders of the Mercor launch. What these start-ups are, in particular, what starts on the financial market, we are constantly showing a large number. I'm not. They've become the most young billionaires in history. And because the start-up attracted a round of $10 billion in the evaluation, and all the start--- all the founders became billionaires, respectively. And I want to remember Mark Zuckerberg, who became a billionaire in twenty-three years, yeah. And, accordingly, well, it's probably like a little quote of what's going on. Now there are many new young, uh, very many young people, of course, now at IT, very young. And you almost read about the start-up every day. Uh, well, you know, you're just so strong, yeah. Very interesting. If there's anything, someone can say there are hundreds of thousands of start-ups of losers. That's true. Uh, there is a situation where young guys are developing such companies, how they can manage them adequately and what happens next in terms of these systems. But it must be adults. But anyway, it happened, and I think it's a really cool case. It's a very interesting case, because we all know, uh, Mark Zuckerberg, or at least have heard of him or the IT market, yes, you heard. And these guys, they're in the future, many of them will have the same known names. I mean, they're still unknown, and in the future they'll be known. It's like Sam Altman, yeah. Five more years ago, I don't think it's Mi--
And five years ago, I don't think anyone in the world knew him. Now, this man, I think, is one of the most commons in any world door, absolutely in any way. I think he's the most powerful man right now. Although we'll wait, by the way, when he comes to the Times cover, yes, and, uh, where they say it's the most powerful man in the world.
Because I think he's the most powerful man in the world right now. I don't see any man who's not ready to meet him. But let's see. To keep it from happening today, uh, I'll mention Apple, by the way. Apple, remind me, Elinar, I think we talked two months ago, yes, that Apple is in the arts in Siri, and he wants to introduce Google models. There was such a rumor. They now confirmed that they would now confirm that they would introduce, uh, the Siri models, the Google models, the Gemini model, and that Siri would probably start working on their base. This suggests that the Apple is not a model-making race at the time of the application. This does not affect the fact that they remain one of the most expensive companies, but they are part of the seven expensive companies. They're not really number one, but, by the way, for skeptics, uh, still artificial intelligence on the market is a bubble or not a bubble. Amazon and Apple showed a profit growth, a growth in the last quarter, and that's good, very good, strong, well, that's good, strong influence, yeah, it does. Amazon had a growth of cloud services, and Apple has the most understandable-- the merchandise that's iPhone. And by the way, I understand their last iPhone is very much sold and, uh, they're good on the move, yeah. And perhaps just to change the colour of the next iPhone, they will also move forward, not on the , not on whether Apple Intelligence will appear. Well, I just wonder why there's no black phone. Well, maybe to let him out next time. And everyone who wanted a black phone--
Elite
will be.
Yeah. All those who wanted a black phone, and the 18th black phone would have to be replaced by the 18th black. Here's to replace black. And, uh, I'll just say that I'm a very strange candidate for me on the market. Amazon sued Perplexity. I don't know, you heard stories that Perplexity, you could watch YouTube through the braus Perplexity, skipping the ad, like you're in YouTube Premium, walk YouTube Premium. Or Perplexity went round, like he pretended to be human, because, you know, they have technology, and Amazon went through a lot of stuff and did a lot of stuff at Amazon buying, uh, and he was collecting data and making automatic orders. on behalf of the user. And, accordingly, Amazon accused Perplexity of using politics unauthorised. What do you want to see? Once again, integration with systems like ChatGPT, Perplexity okay, it's some small player in the market, but with systems like Gemini and ChatGPT, it's only the idea that users can only lead to You're gonna stay in Gemini or in ChatGPT, right? Because if I'm used to buying the goods at ChatGPT, I'm gonna be off Amazon at some point. And I didn't make a deal, by the way, even though we-- our family is doing a day-- probably ten, maybe 15 online purchases. Well, a lot of things are being bought in Amazon, different systems. No purchase through ChatGPT I've done yet, and I'm hanging from mine, and, uh, "Prosecution" section is unfulfilled, unfinished. Write down who did the purchases. I guess I still need to be in the U.S., probably. And who did the purchases when you did, uh, like you, uh, on this subject? Yeah, it's very interesting to see. And don't forget to slide, by the way, to slide our videos and channels.
With Perplexity and Amazon, that's a very interesting story. Why did Amazon get up? Look, when you buy some product like a man, you want to buy a microphone, and then Thanksgiving is about to come, you're Amazon based on your past purchases, based on infowaters that are, I'll throw up some extra things that would be good for you to buy. Well, there, I don't know, some kind of Thanksgiving dish, a turkey, something else, buy presents. And he'll buy a man because he has this story in his head. And when they sent an agent, you know, on parole, there, Brauser Comet, said, "Take me a microphone," she, this system, doesn't take any extra stock, some kind of, uh, banners that could be clicked on. She has a specific purpose, she buys a specific thing. So you can't do an agent pseil. And that's what they say, one of the points why Amazon went up on Perplexity. Because Perplexity masks his agent under humans. And if they showed themselves as an agent, they would work differently with him, and they act like a human being. Maybe there's a dop-- a commercial show or something.
But the story itself, therefore, is that the impulsive purchases of an agent are probably not doing, and that's what Amazon is wasted on.
Do you have any idea how much money people still have on the ad, you know, creaming? And someone says, "Wait, wait, I paid for the show, not for the clicks, the clicks or the clicks." And the agent's gonna scramble and check everything, right?
You have a conversion either down or up. And you're all gone.
And nobody understands why.
And no pseils.
Yeah. And they say, "This man came in, "from Malaysia, yes, or the US, or somewhere else." That's actually a big--
That's why it's really interesting.
Yeah, that's very interesting.
Yeah, how big platforms will interact with agents. But as far as I understand, Amazon requires, first of all, that Agent Perplexity, this Brauser Comet, he identified himself as an agent, not as a human. And the main claim is that if he claims to be an agent, well, there's a pattern, a metadata to send or send...
Yeah, if he's claiming to be an agent, then the whole design is broken from the point of view... It's the same thing you start with the system, the API, the API work. And that's the logic, yes, to give a man, uh, like some kind of instrument of true himself, yes, the real himself. And there's the real, the real ones are interacting with a lot of different services, not API.
Yeah, and that dispute is between Amazon and Perplexity. On the one hand, it is clear that companies are very different in size and can be allowed just for Amazon. Well, just because... But, as a matter of fact, it's still a question that's gonna arise between big sites and agent systems that will evolve further. Well, there's a decade of agent, like Carpathians said, that's gonna be a question that's gonna arise and somehow solve it. And it's interesting to see how this situation is resolved.
Yeah. Write again, by the way, who uses the browsers now and who has the fun case of this, but, bypass, when the system doesn't know you're human-- well, you're human or robot. And, uh, what do you think about the dismissals of artificial intelligence? There are very interesting statistics, we'll raise it in the nearest releases. Write your opinion again. Artificial intelligence, it will affect the reduction of the number of people or, on the contrary, will make it easier for people to find jobs, there, more money to earn, etc. And I'll see you in a week.