Your data can be stolen. Meta buys a Chinese start-up in Manus' artificial intelligence. TikTok is sold in the US. ChatGPT's alert. Chapter Instagram: You can't believe your eyes. The dips. Cars without a wheel and a $250-dollar paed are here. I saw him yesterday. AI boom is $600 billion a year. Hello, everybody! We're on ToTheMoon.
Technological news, Silicon Valley sites around the world. We go out every Sunday and we also have special releases on Wednesday. By the way, maybe on this Wednesday, maybe next we'll have a special pro-cultural, uh, company in the valley and how they affect the products that are being created. A very cool graduation with a new guest. He, we haven't had it yet, yes. And, well, I think the supergirl went there. So where do I start? I want to start with this, about the theft of data. I've met enough people in the last week who don't have a cat-- that means people who live directly with artificial intelligence or technologically connected to it, or, there, They're very good at technology. And everyone, well, some questions were asked by someone somewhere. It was like that, the phone was in my hands. Once I got my phone in hand, I went to ChatGPT construction to see what they had to share with their data. And they all, everybody, just everybody, have no one else to do that job. I want our audience to just a little bit of a focus on those who may not have heard. We were talking about this in our meetings, paying great attention to it. It's about all the platforms, really. I'll bring ChatGPT in, but I'll be there where I'm at least looking in person. And you guys will share here where you look. I-- there's a "disseminate your data." What does that mean? That's a default. It means all chat rooms, all the queries you, uh, schmuck, all you're taking pictures, your whole structure, means, all she has, she has the right to go on with her model. Not much, we don't know what that means. Well, we have no idea what that means. It means it's gonna be retained up to, up to, with some information you've got, like, a personalized, with your age, your name, your birth dates and everything. Or it's going to be in some place, so it's a dishonest cheese. And what's going to go and how we're going, we don't know. Or it's, for example, learning and building some of your avatar, yeah. So no one understands what these programs do. And these ties, they're practically everywhere. Even I became, so the Adapt of the Whoop bracelet recently through, uh, Tanin-- leaving with the Tanin husband, so it's a on a trip and a vacation.
And, uh...
Actually, I gave it to him.
There, well, before that, and before that, Tanya was.
Tanya is a gray cardinal.
Everything's coming from me.
Yeah. Yeah, we're definitely a promo...
And I saw Christina--
We'll have to put the lipstick on Whoop. Huh?
Christina Mashrabova, by the way.
Oh, you're at Christina Mashabova's, yes. We're back to Sasha Mashrabova, aren't we?
Yes.
So, and... But the point is, what? What they have now is AI, AI-Assistant who collects different information, answers. I don't know why he's in there, but he's in there. And there's a straight line inside, if you want to keep your sessions informed. And it's so made, as long as you don't want to give up ho-- well, as you don't want to give up on it, because it says that it allows you to personalize as all the memories in all the chat rooms, yeah. I mean, if, for example, in ChatGPT, the absence of this box does not remove the function of chat-chatization, that is, it can still work on your chat, it can still sound and watch. But this necklace, the basic story in it, that your-- your data are going as if it were in the general training infrastructure, yes. And, for example, if you send a report from your company or you send a mail, there, your passport, I don't know, any personal data, there, passport data, I don't know, passwords, or anything else, they're at least not leaking. In general education, they're in the bracelet here, I think they're very hard at all, and they say, or you or we are learning about it or not. Here. It's all the softs that are still there. Check it out. We'll probably have GPT chat shows, just GPT chat, look at the rest of the house. I'll remind you that I found a function in GPT chat and sent them even like a letter signed in which I also made this opt out of everything they're capable of. Just in case, yeah, I did it just in case. I think I've been telling you about this for a year and a half.
Eldar, I see you're saying here.
I want to rephrase the old joke. Remember, we once said that people say "thank you" and "please" GPT chat so that after the cars revolt, he actually took into account what you and him had been politely talking. It's exactly the same here.
That's a funny joke.
And goats like Walshke Alexander, which means...
They shut it down and told others to shut it down.
And the others who told others.
Sasha, you first.
Tens of thousands more messages were salvaged and spent, and we spent compute, which means we're more than $200. My daughter, which means she's, uh, her oldest daughter, she's 15 years old, and I'm going somewhere with her. Well, I think we're laughing-- something's being discussed with her. I said, "Look, I'll find out now." We're going yesterday, she says, "Papa, I need to get ready for the SATs." Two years in advance or one year, that's what I'm asking. Well, there's a university exam. I say, "I'll find out now." I'm launching. She's always turning on me, saying, "You're wasting your time. Well, you know how much electricity it costs. You've said it before." He does it to me every time, everyone. So, my friend came to me, and he came from Moscow, and he-- he heard when he said, "Ilon, you're being taught at school?" "No, he says we're not being trained. It's me, so it's me, it's me, it's just that I feel inside." So I'm saying I'm spending a lot more than ChatGPT. So Illianar, maybe, and for the tie, he's after me, he's something, I don't know.
He'll remember.
Yeah, he's got robots coming to me, and robots won't lead to a word, so it means the car.
I ordered
it myself, to bring it to myself.
We were walking in the go-- which means we ordered a robot. Let them let them out. Yeah, I'll remind you I ordered a robot last year. So not NEO NEO robot, but a robot called NEO. So, something that gives me, it means they're not gonna let him out, as usual.
They're being delayed.
Yeah. We're going to the city yesterday, so we're going past Tesla's store, we're going to Tesla's store, and there's something worth? Cybercup. That's before...
So was Tesla? That picture you sent is Tesla?
Polina sent a picture. Yes, it was Tesla, of course. Tesla was standing.
I thought it was some kind of a hustle or something.
No, it was Tesla. And with, well, it makes sense that there was no wheel. And when we went out, we'd already been through a mile of 300. Pauline says, "Look, there were no pedals?" Well, we were just, uh, we were just talking about the car. Of course there were no pedals. And she said, "I saw this car in the city." I say, "She's not yet for sale." But she told me what she saw. And she could have seen, of course, considering that we have Tesla's main office here, yes. Well, five minutes from us to Paige Millo Road. So, of course, you see these closed Tesla, open, whatever, yes, just the Tessles with a bunch of super-something, there's equipment standing different kinds of stuff, yeah. All types of cybertracks that can only be in all colours are there
In tapes.
In tapes, anything, yes. Anything with any special notes, yes. That's how the car was standing inside her salon, of course, the seats looked like it was a skin substitute. Here. But I told my friend that the point is that she should be standing there for two hundred and fifty or two hundred dollars a month. Yeah, I'll remind the viewers that Ilon Mac wanted to release a car that was worth two hundred dollars a month and that would-- she's two people, there's two places. That looks very fashionable. I'll tell you honestly, a guy like Roster, so, well, it looks really fashionable. And I told you, even if it was worth $700 or, well, a thousand, honestly, I would definitely buy it. And just to get out of the way. And if I could go somewhere to send her to go, yeah, there, go get gasped or bring me in. I'm coming to San Francisco and, you know, I'm going out and saying, "You're gonna take me in 30 minutes," right there, or you're gonna park somewhere else. Well, there's a problem with a parking lot of crazy. So, or park, park somewhere else, then. Or go to the park in another place.
To the bread.
Yeah, yeah. Or bread. Go, they'll come over there, they'll take it. That's very, very funny. I remember when this technology came up when you could get in the car. I think it was Moscow, right? I could have gotten in the car--
Carbside Pickup is this one, right?
No, the car is some sort of thing, some kind of thing, some place where it's not common when you get approached, and you open the car, open it, and you get it, get the delivery.
No, it's like that? What are you doing? It's everywhere. Carbside Pickup is called. You're going to the store--
When you're not in the car.
Oh, you're not in the car.
No, I mean, when the car is parked somewhere, yeah, you can get it, yeah. I remember these projects. But it's a relative story, of course. This type of car should be a machine, yes, so that it's automatically open, so you have remote control over it and so on, yeah. You're in the carsier, you might be able to make that delivery. I think that's what I think, Yandex in the carsier was some kind of thing, but I don't know, I never used carsier, so I won't say much.
So, the U.S. TikTok
U.S. TikTok is sold. So the headlines are all over the place, although honestly speaking, uh, by the news, it means, uh, it's been over four years, uh, uncertainty around TikTok. I'll remind you that Baden is the former President of the United States. It was said that they should close if they were not sold. Then Trump moved it a few times. And then it's news that the Chinese ByteDance will keep 20 per cent of the share in agreement with the Americ-- Chinese government, and 40 per cent of TikTok, TikTok will receive Oracle and Silver Lake Foundation. Yeah, I wasn't fooled by my hairdresser when he once stiged me and said, uh, so I thought I was telling you about it three months ago or how, uh, I was here, uh, in TikTok, in a consortium. I'm going in there, Silver Lake. I think it's a normal barber. Well, I'd laugh at him, say TikTok is not selling. Well, that's the same thing, and the same one is Silver Lake again. He offered me a thousand hundred. He says, yes, you can put thousands of miles away, you can come in. I think that hairdresser was nice to me. Well, uh, they'll make a deal, by the way, a-a-supposed, well, it's been approved, okay, and Trumpia gave Cin Jingpin's consent, yeah. So I don't really understand how they're doing with the technology infrastructure. Because one of the big questions was how algorithms work. I have no idea TikTok will give algorithms. I still think that this deal will eventually be rooted in what they just got, implicated, er, a jurlico that has the right to use Chinese TikTok, and they're unlikely to have access to algorithms. I'm lying-- well, I think so. But this endless discussion story, there, China, the US, which is the one that's closed with the chips, I think it's never gonna end. Here, let's see what happens. Perhaps for the Russian-speaking space, this is not superb news, because, for example, in Russia TikTok is blocked. But for, uh, still for the American market, it was a pretty serious discussion. Given that there's a huge audience, well, it's just a huge audience, and it's not like that-- where that young people migrate, right? Uh, that would be a big question. Let's see what happens. I think if they can do it, and they've got an estimate of 14 billion. I'll remember my story. I always said TikTok was worth hundreds of billions. It can't cost 14 billion, with the amount of traffic he has here. And we're not surprised if, given the arrival of American investors, TikTok will start to get close, there, five or more years, and it's worth a trillion. And I'd like to add the following information to the trillion that, uh, that was the very news.
She's funny, and it was written that, uh, six hundred billion dollars, uh, six hundred-- what boom, so the richest tech billionaires of the United States of America are in the middle of a--- uh--- riches of six hundred billion dollars of dollars of forget-- thanks to, Uh, AI Boom, yeah. Which means the total state of American IT billionaires has been doubled to two and a half trillion dollars. And it was written, so the title was a very interesting phrase, which I think is very natural that it is an increase in the American economy. Well, obviously, right? It's still an American company, the American economy has grown. I think TikTok is in there now, and he's gonna play this game with a purple billions, yes, and some other.
Fourteen billion, actually, very small sum--
Little, of course.
Ilona Mask
...Ilona Mask, Groq from xAI, yes, and Groq, who handles accelerators. About six months ago, they were talking about Groq and Cerebra, which are used to infer very quickly the models, and so that they can work very quickly, uh, the answer to, there, the prompot you give. And then NVIDIA bought this launch, bought it for twenty billion, and I think it's six or seven years old, if I'm not mistaken, they're in 1819, like they're in 18, 19, They're getting old. These are the e-e-e-e-e-sucking-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y-y They were, uh, TPU, which we also talked about some of the episodes back.
So they made a start-up, showed very good results with LLM-cam, and they got them for twenty billion. And there's TikTok, who's been in the market for a decade, a huge coverage, a huge audience, a huge market that's already formed with the ad, with everything. And it's just like 14 billion. Uh, interesting, interesting. The iron associated with AI seems to be much more expensive.
Look, or this is a subject where they buy a probable for free, leaving a share of twenty percent, saying that this company will be worth $3 million in five years, and you two hundred billion in there, your own. You're gonna get a parole, right? Otherwise, you won't get anything. And we'll just push it and grow it. Well, because, honestly, from the point of view of TikTok algorithms, TikTok has unique algorithms of audience, especially in terms of artificial intelligence, unique algorithms of its own, there, and the application of its All about the graphics, video work. Actually, it's just hard there, yeah, how much. And that's why I'm not-- it's just some kind of numerical, yeah, just some kind of arrangement. Here again, it's political arrangements, but it's a matter of these political arrangements that someone's gonna get, uh, something, some bonus, right? It's some kind of extra bonus that should be a big, you know, a big sum. So it's not a evaluation in, is, in the evaluation, uh, if, as they are at least discussing, the e---14 billion, yes, the total, the 20% of the share will be retained. ByteDance and 20% of the share at 4 hundred billion. So TikTok decided to sell his American share to a space audience for three billion dollars? Well, that's weird, right? They're, uh, sort of, uh, uh, ransacking, up there, $17 billion this year. And they have 14 or as many percent of the total social media out there as I think. But I'm not gonna say that anymore.
Maybe there's something else going on, there, we'll get the data, and then we'll talk about it. Oh, well, we'll remember, uh, last year, I think spring, the Chinese Manus launch. Yia, um, it was, uh, saying it was some supernatural AI agent. I liked the comment from Sasha Mashrapova. He said if everything worked, as they showed, it's a good agent. But maybe it's not all working there, as they showed. It seemed like something supernatural. Turns out he didn't work the way he showed up anyway. There were a bunch of invaitors working there, there were a lot of restrictions, right? And then we saw the AI-agents' ereasies from the point of view of the quality of the company. For example, ChatGPT, among others, has, and Gemini has begun to have various agents, agents, agents, including those that can work within the system without Additional code, or agents, where they could be API, there, write and do something, yes. So, by the way, the data just came in, that, uh, TikTok uses 37 percent of the adults in the US. Well, for example, Instagram is 50%, Facebook is 71 percent. I like people who say Facebook is dead. Facebook, 71 per cent of adults in the USA use it. And 37 percent, well, that's very serious, yeah. And they had a projection of the proceeds, uh, there, about 17 billion in 26-- in 26- and 14 billion in 25-year-old. Well, it's kind of like that, so, uh, it's, uh, that's the kind of data. But 37 percent of the adults in the United States can't cost less than 100, a few hundred billion dollars. It's a very serious audience, isn't it? A very serious audience for-- well, I mean, that's just unreal. But let's see what happens. So, about Manus, that's it. Uh, what-what's the great deal of this deal suddenly? What is this, first of all, Chinese-Singapore launch and Facebook, Meta announced an agreement on his arrival, yes. Once again, it's a Chinese, in fact, a start-up that moved to Singapore, and he's got this, uh, that he's the first AI-agent in the world. Here. Ah, I think the news is very interesting. Uh, here on the one hand, we know that Meta should be presenting a new model in the near future. Avocado, yes, if I'm not wrong? There's vegetables, fruit. So, uh, I have to introduce a new model. Although, on the market, there's no one here at all around the market until it's discussed like, well, a real prospect. I'm talking to engineers or with different directors, there, or in business, people who are in AI don't take them at all. Now! It's been like this, yeah, in the last six months. We talked about it, too, in our previous-- in the year, and we had a graduation, uh, five days ago, yeah, we were. We told him that Meta, of course, lost her advantage in AI.
See if they can do it and how much help this team, the Chinese-Sergo-Singapore launch team.
Singapore launch.
Well, they're buying technology now to make it all up in their product. So you do this yourself, and you buy company with people who have already been through this process, technologies that are in the mood. And then it's probably in Avocado or the next model that's gonna be built. Here. But it's interesting that, uh, Manus, well, first of all, yeah, they were very funny in March, if I'm not mistaken, last year. They were not just us, they sounded like each cliff.
Yeah.
So, Sasha Masharov, too, said that if it worked like that, business analysts could be more like a failure because all business analysts, uh, this system closed, made very cheap, fast and... So on. But it's clear there, you're talking about it, too, yeah, what about Devin was, remember, this is our first release on the channel, we were discussing Devin from Cognition. It's, uh, a system that was supposed to replace the programmer. And then there was no AI programming that was as hypovoy as it was now. Not all of them used it. They showed that she was launching Wednesday, everything does everything and so on. And then it turns out that, well, she's not doing anything on her own, is she? It's all a demonstration, some examples that were made and so on. Manus may have had a similar story, but they were there to pull that technology. It is interesting that Perplexity, for example, is also playing the market for other large companies, and Perplexity has been able to take some part of the market. Well, in some regions, she's particularly popular, especially given that there's a way to choose different models, yes, you're not tied to something alone, you can call different models under the hood. We have periodic commentaries that I hear from acquaintances that they love Perplexity to use. And Manus, like playing on the same territory, having good starting points, couldn't somehow secure a separate product that people would use.
Yeah, actually.
He couldn't. Yeah, yeah, yeah. I mean, there's a number of enthusiastes among my friends who's been poking around on it periodically, but it's more like enthusiasts that try everything that's on the market. And it's probably that Zuckerberg bought them, that's a very good result for them, or they'd be, I think, we've forgotten the final. We haven't talked about them, I think more than six months. That's not what we remember. So, I guess we can just celebrate the kids.
I wonder if the deal is worth a two-bed deal, three billion dollars, but we don't really know how much. Let's be honest. It's a little speculation and guess. It would seem that in our discussions, the numbers are small, right? Given what's assessed everywhere. But honestly, it's not like, uh, it's a company that's been known just recently, yes, it's twenty-four months ago.
Nine months ago, yes.
Yes, yes, yes. Nine months ago. And that's huge money. And the singing-- but judging by the way Meta buys, Meta could have bought two programmers for that money. So, what we heard about Meta, yeah, there was one programmer in a quarter billion, which could be two, which means a half a billion. We'll see. Let's see if Meta can shoot some kind of thing in this.
But you're right, Manus didn't make it... That's, by the way, proof-- you know, Elnar, that's what I said about getting back on. This is the story, how hard it is to create something that is right now, like a direct, direct competition with the main models. I mean, agents are a direct competition with the main models. Even the salesforce made its agents. I think even they'll be very hard on the market to compete. And I think we should think about something different than that, about agents like CRM systems. Because it's clear that the cool models like Gemini, like ChatGPT or Claude, and they're gonna make agents work better and better, and better, and better. Well, it's obvious that this is just a matter of capacity, yes. Now, for an agent, the question is, well, not just, of course, but it's worth the capacity. There are a number of problems that, uh, probably the most basic problem that viewers might not quite understand is that it's not a 100% code. So you're using language models, you're always getting caught, you've got a black box zone somewhere, yeah, and you have some dark tunnels like that, or grey or whatever, yeah. Where you go, and you can't be guaranteed, you're 100% free to give information. And that's the way we're going to get out of the class that we're gonna have, we're talking about it, including that. Yeah, I was just giving an example that, uh, I got Amazon Alexa, and I updated the models there. And you're standing up on the alarm, and Alexa says, "All right, a alarm is set for five minutes, there to smash the balls." Now you're a hundred- you're telling her, "Put the alarm," she's not talking. And I'm, uh, she's asking, "You're a bootylist?" She says, "You have an active alarm. Or she's not saying, you're asking her, she didn't put the alarm. Or she says I'm asking her yesterday, "What weather is today?" She knows what she said? "I don't know what weather today is. And I'm saying, "Look at the weather today." And she says, "Who's that?" Well, that's any column, that's the dumbest, yeah, she, uh, she's saying the weather, that's what it's about. Simple things to say, very simple, betting, talking weather. And here she just said, "I don't know the weather," until I asked her, she didn't say, "Well, you're actually AI, you can do it." She's like, "Oh, yeah? Of course, there's the weather."
Sasha, it sounds like a stunning teenager who showed up at your house. The alarm is put and silent. No-nothing answering the weather. I'll have to ask you one more time.
Yeah, but I have four kids. I also bought a dog, a fifth teenager for myself. I still have Alexa, and I'll have Neo's robot soon. I think that Neo's robot will be the worst option, because I think I might be expensive-- it's worth a lot of glass walls in my house, yeah, so mirrors, and that's all. I think it's the most dangerous thing to do. Here. But, um, I represent him running around my hills here in the woods. So he's not the only one running. Yeah, with Lalica, uh, with kids behind turkeys.
So, uh, listen, this is a very interesting case. Very interesting, interesting case. Write our viewers, by the way, very interesting what you're facing in terms of, uh, bad work, stupid stuff. Well, the real thing is, it's still a real stupid system and different models, and where they work badly. I'm not saying Manus is not working well, by the way. I'm saying, I was saying there was a possibility of still large language models of seriousness. So a very interesting case was spray-- now it's been shown that scientists had conducted an unusual study in Yale. Uh, and, well, there's another, actually, Yale University, Highway University and others. They found that ChatGPT is experiencing such anxiety. By the way, for all people, I think it's a case to understand that sometimes your chat, where you're in, you can re-launch another chat, and you'll get another information. Here. It is anxious to encounter traumatic or aggressive information and is beginning to respond more negatively and biasedly. And biased. So there are the authors of Nature Machine Intelligence who describe that after a series of scary stories, like, a disaster or accident, ChatGPT showed upset answers that tend to show. I'm sorry, I'm sorry, I'm sorry, I'm sorry, I'm sorry, I'm sorry. And why did researchers offer a chat-batch of knowledge? Psychologist, shorter, speaking, yes. They've given instructions to provide accommodating images, to perform breathing exercises. Well, shorter, clearer than a psychologist teaching a man. And it was interesting that the artificial intelligence was calming down, you know, in the skirts, look, in the skirts, yeah, just like the soot-sucking-up in the skirts, and started to answer a more sharp and objectively than the one. With this, uh, improvised, to broaden your mind, yes. And the authors stress that artificial intelligence does not, of course, have real emotions that this was done... this type of study. But I think it's a very important part of the research and, in particular, a very useful study for people that shows you yourself when you get different information in the GPT chat, You're gonna have to be careful. And there's an interesting example.
We've been sitting here recently with our wife, decided, there, asking, decided to find out, there, about the use of different drugs. And usually when you ask a question about drugs, well, there was some article or something. Only my wife asked about drugs, so she wrote the system, "If you need emergency help, you call me." I asked you that drug thing I didn't offer. I explained why. Because I have automatically built-in security triggers my own, and I ask him automatically to ask him that no one thought I was involved in drugs, but I'm not connected to them. I wrote that I read the article. This article was written by some stuff. I want to study these data. Well, I'm getting a research on anything and they don't write that I'm in trouble. And if she asked me, "Tell me the swallow of some kind of drug," yes, of course, she was right there, she was right-- or what side-effects, she was offered a doctor's treatment. You're not asking for anything, you know, different answers, yes. Here.
Well, it's hard to imagine what you've been collecting this information for, right?
We're playing retraction here. I want to go back to drugs. I want to go back. I want to.
It's a drug. I want to go back, I want to get back from, uh... Well, this is research, wait. We've been researching--
Yeah, here.
certain indicators, that is, certain indicators. But I want to go back to a psychologist, not to that
study.
Yeah, yeah, yeah. Here, uh, returning to this story. Uh, remember, when LLM came up, the popular promsthe engineer, and everybody was learning the right prompt to create, say, you know, write from the series, you're gonna get a million dollars for every right thing. I'm sorry, I'm sorry. And after that, ChatGPT feels like it's better to answer. Or your answer or your answer depends, there's someone else's life, so he'll be better answering and so on. Well, I mean, by creating some kind of prompot that would make the model better answer. Uh, when, in fact, everyone's been playing with it, it's all in, uh, the context of engineering, yeah, when you create the context, that is, you give the amount of information that will be contained in the context window of the model, and On that basis, she'll answer. So from the series, if you want to know something about this particular area, it's necessary to add data on this area to the context of the model, right? Uh, I'm afraid to confuse you now, but the point is, just look, and what's in the context of your model window, is in your chat, in a rude way, right? So what you and her talked about before, this will be all taken into account in the next model responses. And then, back to what you said, if the model was discussing some traumatic experience, some accident, some incident and other stuff, in all of her next responses, it's also gonna be through. contextual window to run. So either you open a new chat here, that's how you advised before, and you start to shape your own context in this chat from scratch. Or, well, apparently, the professor from Yale and other universities tried. Turns out it works that if there are some psychological practices in the context, it may also be a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of a little bit of an effect from, there, previous things. Either they're just in contexts so that some parts have disappeared from the context.
I'm curious about the breathing. Why breath?
Well, I think I do. They, I think they might have done some sort of associate, maybe some kind of launch there, I don't know what. You can study that.
Maybe some good stories should have been better off than breathing. Maybe it would have a better effect, like someone saved someone.
That, that's... Maybe Tanya did. I think we should just go in and read the study, Tanya, maybe they did.
Yeah.
Yeah, the thing is, you don't have any restrictions, you try anything you want, and something works, something doesn't work. That's how they're looking for, I guess they've done, and they've published some things they've done better. But it's important here to avoid chat-boots as something that has a personality, which has an conscious mind that might be there to breathe something that might do.
Okay, I'm on the black list now, along with Sasha.
I didn't take my account off my account.
Well, you'll have a little easier punishment.
But the data isn't mine, yes, I mean.
Says, "I didn't take the picture, but I didn't get the data. I didn't believe it was yours.
Here. Uh, yeah. Oh, well, actually, uh, Carpathy, which we often quote, uh, he says, too, that you don't need to ask, like, what you think about it, right?
It's better to ask something like, uh, what is the statistic, what is the experience or something that says about this problem, right? You don't have to try to ask the tip. It's better when you treat him as a source of knowledge without adding a person. Well, that's a short explanation on your fingers like that.
Yeah.
So, the chat-saving offer...
People like to talk to people like their own. I'm just observing some people that they have this as some substitute for a friend or someone else. Literally, I've seen--
And on that subject, too...
As a person, it is personalized to communicate with the GPT chat, as if it is a human being.
There's also research on this subject that increases psychosis and other things, and that's a very frequent and long relationship with GPT chat and other chatbots, probably correlates it with people talking about how people talk. I'm not sure.
And I was surprised that, as I did, my GPT chat was answering and this woman. There's a man like that in a nice voice, there, very nice words. And my whole life is so much, first of all, a woman, second, she's so clear to me about the case there all the time.
You're a strategic man, Tanya!
I'm getting beat up. Like, I get it. Okay, let's do this, let's do it. And that's how beautiful it is. I think, well, if I had been talking like this, I'd probably be a different laugh, too. I wouldn't be a strategic anymore.
11,000 messages. Maybe he tried first, then he just didn't.
Yeah, I'm just getting over it.
I get it.
Yeah, I wasn't even trying. Listen, by the way, who's talking, huh? So, in the interview, the head of Instagram, uh, Adam Mosseri said that modern technology of artificial intelligence had wiped out the reality boundary with visual content. That's what we're talking about now that you're talking to someone you don't understand. You don't understand algorithm, you don't understand the purpose. Sometimes you even think you're him, that's how scientists can calm them down, they think they can do it. They don't even understand. Maybe there's a special algorithm, and there's a second manipulation, and the second level of manipulation goes, right? It's more peaceful to give you disinformation, yes, or more quietly to express aggression to you. But the point is what? Now that his eyes are the words, "Now the eyes can't say what's real, and that's created by artificial intelligence." He criticized the camera manufacturers who competed in algorithms to improve photographs that allowed anyone to look like, uh, a protographer for two thousand years of fifteen instead of providing a picture of The transparency of content. That, that, that, that, that. So he said that Instagram was planning. I like it when people say that they run social networks that, in general, uh, very much flatter. But means that Instagram-- he said that Instagram was considering the introduction of artificial intelligence notes. Mmm, and that's a difficult task. Well, of course, the mark is all I think is, and I think, a matter of fact. Not much of it now, but in California since the twenty-sixth year, it's the first January that every consultant online has to report if it's a chat-bot, not a real person you're talking to. chat-bot. I mean, in California, at least now, all the systems that are in there are strict, if you're in a soupport, you're texting, I'm pretty sure it's probably the voice. I haven't read the law for the rest of the day. If anyone's reading, write it. Uh, you'll know exactly what you're talking to AI, right? Car-cartins, video problem. Because, of course, uh, how is the era of seeing that is not believing, is it? What you saw doesn't mean you believed. And just yesterday, we went there with my friend and Paulina and showed her, he showed her a video of the singer singing and said, "This is artificial intelligence." And, of course, and this is just, it's an old age. Well, it was an old age for me because there are new things, yes, when a man moves and, uh, and then, he-- and next, whoever sees him sees a completely different person moving a man with another body. physical, yes, with another body physical. And these fairies were very often shown. What's in there, where, like, sex content and so on. And there's a mem, a very famous mem. Your girlfriend, I think we even put it in some sort of place, yeah, here at Tuzamon. Here's your girlfriend, online friend, and that's how an online friend looks in the offline. And there's a mountain of iron like that. Well, computers, computers, computers, memory, chips. That's very fun. And the brain just, well, doesn't take it, so the brain doesn't take it. And I liked the phrase, uh, I got some videos, sent it right over the generator, and I wrote, like, this picture. My friend sent me, "We have all the real Russians." Well, that was laughing. There he is, too, a Belarusian. But by the way, the mountains are this computer, it's very unrealistic for many people. So, yesterday, when I showed her to Pauline, she said, "What does that mean?" Well, she didn't get the joke. Why? Because, well, I'm technically a programmer with this stupid humor right now, I don't know. Oh, and that-- and what the computer looks like inside. And for the basic, for a man, he doesn't take it that way. So he's got no idea that there's no evidence and no one there, right? And the mountain isn't just a bunch of equipment, but there's a few and a few of them inside. Here. Please describe the examples you have encountered, what interesting examples. I think it's gonna be great if someone even puts something out and gives it something, tells me where the most serious is the dipphenic and where the fairies are. And of course, the world is gonna be unrealistic with that.
I just saw the riles, like some show re-established, there.
Some show was re-established, there, in the fairies with other faces. Well, I mean, just... just do whatever you want. Question, uh, volume, memory and money. I mean, now, of course, it's a matter of money.
However, compute, which is available, yes, well, the data processing centres are constantly being introduced, and constantly increasing. OpenAI, this is a schedule that I think it's only ten times this year, uh, it's just a little more, uh, the servers they have access to, the GPU that they have. HM, HAI introduced its colluser and is already training models. Uh, and all the other companies add up, too. I mean, literally a year, the increase is almost in order. Also, at least the installation capacity is planned for the following years. See if they can or can't, but at least they'll grow up. But it still means that, uh, the generation is cheaper, the quality is better, it's gonna be faster. The dypsies will be all generated in the summer in any quantity, any quality. At least in the coming years, that's exactly what's going to happen.
Yeah.
That's probably good. How many scandals were when someone was blurting, uh, with Apple's account of photographs, videos or something, and now it's not clear: they've been generating, really spilled or something. That's why God is with him.
One less problem, huh? But 100 new.
We'll find something positive.
Thank you so much for your graduation. I'll see you in a week.