You're not supposed to go beyond the GPT circuit board even read the news.
So we're doing AGI, we need seven trillions. I bought it immediately at GPT chat.
It's a really cool update. There's a coinization in GPT chat.
The version is 402. The whole Internet was sliding out of Sam Altman's skate. And then that company was worth 100 billion, and Tesla was there one and four trillion. You don't forget that there's a Chinese market, and there's an American market. It's like an example from the owner of the Yandex.
That's why I didn't get it.
You didn't watch a sound? Ah!
Yes, I think, "Oh, my God!
Look at the sound.
What's going on again?" Turns out, that's what. Now it's all in place.
Sam Althman ran to watch ToTheMoon, so. Sam, what happened? We need to know what's gonna happen on ToTheMoon Channel. Sign up to keep it from happening.
Yeah, you used it to be some new one.
Oh, we're gonna tell you everything now, Tanya. We'll tell you everything. The main but-- one of the main news today. I'm sorry, Sasha Masrapova is gone, of course.
I'm not. I was just doing a nice presentation for a client on design yesterday, and there's a terrace, and there's a fountain on it. Well, I put a fountain like that, leaves, and the effect added, like the wind moved the leaves and the water flowed in the fountain. Well, I think it was a very good decision.
Yeah. Hello, everyone! We're on ToTheMoon. Technological news, sites with the Silicon Valley and around the world. We're going out on Sunday.
So, today we have a graduation before we go to, uh, all the updates and updates of the OpenAI that we did, yes. I think the OpenAI has released the largest number of different micro-renewals, micro-diffs in the last seven days. And before we go to Sam Althman's running, looking for ToTheMoon and Sora, I want to start with the update that OpenAI ChatGPT Pulse made. Well, Pulse is available to Pro subscribers, and this is a system that Sam Althman used to refer to, then a huge number of directors inside OpenAI, they wrote, everybody said it was a very cool system. What is this and why it is only available to Pro users because she eats a lot of re--- well, it's available to Pro users because she's waiting-- eating so much resources. Clearly. I was-- I first thought it was some exclusive, and then you sit and think when you're doing weird stuff like that, maybe it's a test, a test of the most useful, most unnecessary users, I guess, yeah? I mean, those who don't care-
They'll pay, they'll
never go anywhere.
Yeah, they're gonna pay anyway. Let's test them and take them out weird things from OpenAI. Why are they weird, I think? I'll just tell you what it is. It was the staff team in OpenAI. And I remember, we were a few episodes, and I was laughing at Sam Althman and the OpenAI team presenting a new staff window because they moved the elements in the building from three points of menu to one, and it's kind of like, Nothing new has emerged from the point of view of staff. And they're giving out Pulse. Once again, with all my love for GPT chat and OpenAI genius, I think this company is top of the top. And so it turns out that this system is working every day, well, or there, at the request, it has some time-bounds, it's processing the content that you, who you studied in your chat, and then it's shaped. You have a daily report. So, basically, what does that look like to me? If I asked you last night whether the scotdun would be with the U.S. government, there would be a studadow tonight or not, by the way. Well, we're filming the first October. So he's not the first one, he's out on Sunday, yeah, just, like, a little bit later. Will the studadow be? I was told by GPT tonight, there's a scotdown. And the next day, Pulse has news about this studadaw. Or I was looking for some things about Sora in OpenAI. Here he put Sora in Pulse today. It looks like, to be honest, Sasha Masrapova is not here today, but we talked to him. He said, "What's a two-two-two or-- that's probably like, "Yeah, like, like, 20 or 30 years ago, someone's doing a tape of news. Well, that's totally pointless. I mean, things with meaning, and that's pointless. And I think that on the staff-- I reminded you of that, and I have a lot of people working on different teams. Elnar, you know who reminded you of that? It was recalled by fashion products, markers that come from the original companies. I'm representing these people that were hired from Google out of Apple. And they came and said, "We're gonna do super fashion things now. And that means these products and the markers are launching and doing a genius product that I think people should be fired for. Here. Nah, it's, uh, it's very weird, because the staffing doesn't develop and the perso-- well, it's going to be very slow, it's gonna be pretty slow, I'm sure that in the future, personalization will be some unique system. I'm not gonna be able to get a chance. But let's be honest to date. Write our viewers, by the way, who's the one who tested Puls? What do you say about this idea? Where did you see her? What do you think? Again, that's my opinion. Yeah. And now Ilnar, I think I've been protesting, too. He'll say it, too. I'm gonna get the staff back. It's just that it's where, uh, money is needed to make the seven hundred million people even more dependent on ChatGPT. Because what's the main theme? To make you use of it. How will they use it if the system is very applied in terms of the understanding of who you are. And for that, you know, just ask simple questions. Just ask simple questions so that I don't know how many years I've been living and so on every time.
Mm-hmm.
Here.
What, Ilnar, you gonna tell me about Pulse?
The idea is great. So the point is, you're not supposed to go beyond ChatGPT to even read the news that he'll find you the news that happened there the day on the basis of what you're looking for. And he'll get you some cards there. Ah, technical implementation looks very simple. You've been telling me about the rest of the day, and you've been doing the system. Yeah, and that's how it works. As a night when you sleep, it's just a trigger. This trigger is a summary of what you asked, discussed the previous day, and that's the basis for a set of cards. I mean, there's no rocket science in general, and it's just working. Now, with regard to quality. Oh, among the news that comes, there's a limited number of them, I don't remember, there's ten, 15 cards, I guess. One or two is thin and thin, and I'm getting some interesting things. I mean, if I'm going to study with him in some detail, well, the subject, some kind of technology, some kind of frimevour or something, he's got some news on this subject. The only thing that doesn't leave me is that he's not just throwing news, but he's making a text based on how we talked to him. And that's a little hard for me. Yeah, I feel like he's changing the text on the basis of the way we talked to him. And, well, it looks very fake. I wish he had just made a chin, let me go to some resources to read the interesting news that we've been up there this afternoon, recently, on this technology that we've been dealing with. That would be good. And given that they're all trying to scoop inside the app and redefining all these news, articles and stuff with their own language, well, I'll probably start deep research if I need to. Yeah, but some product doesn't work out for the rest of the product. I'll say again, but it's quite exciting. I'd be interested if the good news sets were...
Yes, Ilnar--
And I
didn't have to go out.
It's a question of use. Yeah, it's a question of use. Look, is it a good idea that you get something on your side every day? Good, but if it works. The way it's done is a thrash. I can also create a daily task at ChatGPT that I will send a notification and search for the right things and not send any bullshit. It's just a pretty, well-documented card, you know. It's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, uh, it's, it's, uh, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, uh, it's, it's, it's, it's, uh, it's, uh, it's, it's, it's, uh, it's, it's, it's, it's, it's, it's, it's, it's, It's with a pretty, well-documented card. I mean, when people start doing, it's like doing a site not for conversion, but for-- you know, when you're the director--
Demonstrations.
The design director brings, yes. That's how Tanya was talking about the fountain right now.
I mean, when you're like, "Oh, oh, beautiful!" And that's why--
You're implying I'm not good at designing?
You're implying I had a bad design?
No, no.
I'm kidding, I'm kidding.
But this, look, you're actually saying a very interesting thing. I've been told by my friend, uh, not knowing, Pauline was, uh, niece. She was on work and travel for a whole summer in America and she went to visit us. And she told me she had a friend of hers. He worked in a restaurant last year and this year, when he came to work at a restaurant, he made a read code. He's from Kazakhstan, he bought three hundred candy in Kazakhstan, brought them here, dressed up a shirt in a restaurant and said, "I'm from Kazakhstan." And every time he approached the table, they asked, "What's that flag? Country flag." There was a flag painted. He said, "I'm from Kazakhstan." And at the end, he was in the count of a candy of Kazakhstan. Do you think he's got a tip increase? Of course, it's really increased.
Ingenious.
Yeah. I mean, the codes, yes, they work. And, accordingly, this is... it's a good invention to place people on board to increase profits and their own, yes. If ChatGPT does, you should think about making pictures to, you know, stealth, to increase profits, that's, and to increase involvement. The fact that there's a picture in the pulse is not a procedure. Well, that's not, it's not a plus, it's not helping. I mean, it looks really weird. Again, that's my opinion. Oh, Illnar, I support, by the way, Ilnar, absolutely that it is itself, uh, that the idea is good, but it's not real, it's a pointless thing, yeah. That's what you can generate, uh, fifty grand. Why would you let Pro use it? But you're right, I'm not gonna say no to that.
Well, I mean, I can't give up Pro's version, yeah. So, uh, the question is, what am I gonna do, yeah, and why? What am I supposed to be angry? Oh, I want to say, ChatGPT has a second pdd. There's a second pddette, which I think is more important on the market. Although, again, he looks, his implementation looks very strange, very strange. And I think again that, uh, this story, where, uh, OpenAI produces a lot of different product from the different teams, uh, uh, it keeps doing AGI, and they don't care if they work or not. And I guess there's a name for Sam Altman. Sam Altman wrote at his place two hours ago in his Twitter, "We're doing everything we need to get the capital we need to build AI, and that AI is doing-- well, he's doing the science in English, yes, so AI, uh, I went in and showed up... and I got... shorter to get AGI. This requires OpenAI seven trillion dollars. And seven trillion dollars, they-- they also, by the way, a few days ago, said that seven trillion dollars were needed to defeat cancer. And in parallel, and he said that phrase, look, "and it's very nice to show people cool new tech products." Here, this, this is his phrase two hours ago, which I think is consistent with what's going on in OpenAI. So we're doing AGI, we need a... seven trillion dollars. To do that, we all buy all the infrastructure, and then we throw you some cool stuff. You want to have fun, you want to stay out of it. Your own right, yes, that's what you like or not. I want to tell you about the checkout. The check is on what they did, and why I took a lot of attention to him. They were integrated directly with several American companies, in particular, Shopify, for example, and, in order to ensure that a person who in ChatGPT is looking for some kind of product, immediately at ChatGPT bought it. I mean, basically, they're doing a mono basket with one product at one delivery address right now. But what's fun? I'm at ChatGPT looking for something on a daily basis where he's giving me references mainly to Amazon. I'm clicking Amazon, adding to my basket, yeah. And while it's from-- well, the interface, I'm used to it anyway. It's money, big volumes, etc. But-- but generally, it's probably ChatGPT, if he-- UX/UI is gonna be cool to work with baskets, I'd buy ChatGPT from my Amazon account. So he did a cent-centre. We'll show you the pattern now. Here's the chart, I think it's here. It's a straight way, a way where users communicate in ChatGPT, ChatGPT, and, uh, communicate with the store and make a full-time credit card, there, or Apple Pay, or something, yeah. I mean, they're doing a full purchase. What's the point? About what the OpenAI wants, uh, to get into the market for the labels and to create its labels.
It's obvious. And there's always a two-way story here. On the one hand, you're cool with them. Shopify's shares grew instant checkout instant, and they've got an increase in NASDAQ. And yet the person who gets used to buy ChatGPT, he gets a little bit of a-- no matter where he makes a buckle-bag, yeah. If he's all integrated, he's got a payroll in OpenAI, whichever is the difference, then-- I'm gonna get a delivery company. If there is, well, there's a difference: quality, service and everything. But still. Here. What do you think about this, on this subject? And by the way, they recommend this instant checkout, they recommend to the developers, and, and I'm basically everyone with stores and so on, they've opened it up, now, and they're all, uh, they're all, uh, they're all, uh, all, all right. companies have the opportunity to integrate with them. Here. Users are free. The Merchant pays a small commission with a successful purchase. At least that's what it says, isn't it? And you're already a new kind of coinciding that comes out in OpenAI, uh, from a successful purchase point. And here's the Amazon question, will he be integrated with them or not integrated? Because the best thing is... if I have a day, uh, five, I don't know, ten Amazon purchases, you know, five, maybe, yeah, Amazon shopping is coming, then, uh, I guess I can use it, including adding it.
through ChatGPT. And here is the question of Amazon? Here. Not a fact.
I think it's a very cool update. Uh, first of all, I, like, Sasha, you said, I really like the coinization of ChatGPT. I understand, of course, that I'm not counting their money, but that they're obviously lost, and I want them to have a coin-the-money issue. So the company looks more stable, at least for me. Mmm, the only thing that comes up is, uh, not a conflict of interest, but a new force like this is what's ChatGPT gonna recommend? Well, you want to buy this microphone for the new sub-caste, right? And you leave this ChatGPT. The algorithm of how he makes his choices is not very clear. Can some companies make some money, make a better description, do some other way to influence what kind of product you're gonna be offering ChatGPT? Will he be in your interest, buying, I don't know, better than the ratio, there, the price of quality? Will he be able to make it clear if you're ready to wait a week, but you're gonna get there, uh, cheaper or better, but-- or do you want to go right now? I mean, there's a lot of things that, and if you leave it unsupervised, like, well, they just might be able to grow cheques in a way that doesn't understand, because it's like they're gonna benefit more expensive. The goods, because it's likely that the commission will be higher and so on. But that's the details. Uh, they're probably on any marketle, yeah, because someone's moving stock, someone else. Here. But the fact that they are again and again, as the previous news, leave users inside the ecosystem in such a way that they don't come out. That's interesting. They're, in fact, stubbornly seizing users to get you out of ChatGPT.
It says, Ilnar, it says they take into account, uh, availability, price, quality and main seller, right? At the same time, you were curious about the microphone. I actually bought a new microphone, and I bought it with ChatGPT. And I was wondering, because I didn't know if I'd use it or not, I was interested in a system of a longer window of return, there, the possibility of bringing back the goods, right? What if I didn't like him, I'll get him back there. Well, if, for example, he would sell to Costco, I would just have brought him back, there, in three months, for example. And or there are systems that give you the opportunity to blend the price. I mean, it's a condition that you get a price cheaper in another store for, like, forty-five days, you write it, and you're in the middle of a price. So on, huh? Some of these things, like, were there. And, uh, basically, uh, when you buy something specialized, well, not standard, automatic. So, the standard purchases on Amazon, which are handled by most, for example, Americans with huge-- huge amounts, right, Tanya? Well, the huge volumes of standard stuff. You know you're gonna buy it on Amazon. I guess I might have bought it in theory not Amazon, but I'm used to buying a lot, and I'm a quality system, delivery quality, feedback quality, soupport service, quality of this, you know, ease. - Yes, sir. - Yes. That's why.
And so on, yes. And that's why I like to buy in a system where I know I'm not gonna have a problem. I mean, Amazon's legs, well, very light in terms of buying, returning the goods, yes, checking, answering the soupport, refunding the money, if something else happened.
There might be a problem with other stores. And I'm still embarrassed about this service in the first place, I, Elnar, put the elevators on this service, OpenAI. Even if I were to develop them, I would have set 90% of what they couldn't do. Here, but I'll be 50. You know what they have a problem? Every time you have to bail on the system you buy the goods. I mean, basically, you have to have a stock and a card in the system, and that's when it stops working the very fact of this system. So if ChatGPT recommends the microphone shopping stores to be specialized musical, which I don't use, because I, how do I buy them then? And he doesn't see my special prices, because often my prices are cheaper and more profitable. Like, it's like Booking, you know, he can do as much as-- anything. If I google, Google, Google me in Expedia and Booking Recomandu-- well, writes prices. But these prices are different from my prices, because I have the highest levels in Booking and Expedia, and I have many hotels, uh, prices are much lower. Or there's a price guarantee, like if it changes in flights, like Expedia, right? What if the price is cheaper, I'll be back and so on. And accordingly, he doesn't know my stuff inside, but he's gonna log me, like I'm through ChatGPT to all the systems. So, this is a little complicated, though I understand their complexity, how they're gonna get to the challenge with-- who, by the way, would this system be interesting, who would have used it. I understand, of course, that in a Russian-speaking world, it is very far for many now, yes. The probability that this will reach Belarus, Russia, Ukraine, it is low or currently absent, in particular such a system. But there must be some system at some point in time.
Still, this is probably here, well, that innovation is present in terms of rapid and very rapid turnover of such a set. Let's see what happens. And, well, while I'm at a rate of this nature. ChatGPT has yet to be cleared, OpenAI has also retrieved a number of updates. Not much of a fact, they even pulled a lot of skies on their website one day. A little weird marketing, too, but okay. The team is probably getting a new one. Ten thousand or more people. They've got-- who's the one who has a business, you can read it. They have a huge amount on the ChatGPT diagram website in their own business. Although they're a little childish to me, and soon as we're on the channel, we'll take a little more serious case in the specials, but you can see it. They've taken out parental control if one of the parents is interested. So they've taken out parental control, and, uh, they've got the updates in the project. They have projects, they've renovated. Although I think that all of us should have a very strange reaction, because it's again, it looks like, I don't know, 2,000 third year, maybe. At the dawn of my start in programming, which is when, like, these systems were designed, like projects, to share a role. And they say, "Look, we have a bunch of chips, now you can share and add users to the company." Like, guys, now is a new programming world, you can get 60 different chips out of it a day. Sixty. One programmer can do a lot of things.
Okay. Sam Altman, they're the team doing Edge IA. They say, "We're busy with Edge IA." No questions. Sora. Let's go to Sora. Sasha Mashrapov is gone today. I'll note that some news shows that Higgsfield passed a $1 billion mark and became a unicorn. I'll tell you what. Well, why am I supposed to add the word? Because everything depends on some confirmed data. Sasha will come, he confirms, but at least in the Kazakh media, he wrote that this was the first project from Kazakhstan, as a large team of development in Kazakhstan. Kazakhstan has first introduced a unicorn. At least that's the news. And, well, Sasha at important meetings, he'll tell us what he's doing. And, um, Sora showed up, there was an apdate at Sora, and you could see a skate in Sora, running Sam Althman watching the Moon with us, and that Elnar made it on his knees somewhere fast. Sora's in-weight access, he's starting to get scattered. You're-- you've got him open, right, Elnar, right?
Yeah. Here we need to...
Yeah, no, I didn't open in ten minutes after the demonstration, but then I opened up, yeah. And it's been discovered, perhaps, because of the pros and cons. The software users have a version, and I like it very much, yes, that's what I love. Sora2, yes, i.e. Sora Pro, Sora2 Pro. Pro users can therefore make more gr-- more serious, more serious quality videos and, uh, more stress, more time and so on. We'll know about that, too. The whole Internet was sliding out of Sam Altman's skate. And it's fun, by the way, that Sam Altman was before, I don't think there was that much on the Internet. What Sam Altman wrote even in his Twitter, uh, some... Well, you know, it's such a very interesting thing. On the one hand, it's fun, and on the other hand, when the entire Internet covers you with a fairy, you know, yeah, everyone's gonna be looking for Sam Althman, and it's not clear where the fairy is, and where the real stuff is.
Elnar, what are your first system impressions? Although we are hardly super-professional users, yet we have different experiences.
I loved it. Aaah, this is the result that I hope we put or put before this, uh, before the video was published, it was made from a second or third attempt. There's a little hand to be punched to make things work. But, uh, I'm excited about the result, the quality. Yeah, it's obvious that it's still a generation, some kind of mistake with the TZ you give, but in any case, you can get some results. Oh, first I was sceptical about it, especially when the presentation was watching before I was sweating until I saw what other people were doing. I didn't know how much this is a smashing thing. The thing is really stunning. I have a few points here, I think I'd like to point out. One. Oh, the video generation is very expensive, and they gave a huge number of people the opportunity to generate, you know, some pointless videos, yeah, just mems, jokes and stuff. Each of them requires a large number of computing capacities. Sashi Mashrapova is pathetic, he'd have told us a little more how much it all eats. You're doing it for free, and you're doing it for the price plan. Yeah, so it looks like they're investing a lot in this pyre right now. Yeah, there's a lot of videos, there with Sam Althman, people, there, with their own skates, and so on. But someone has to pay for it. Yeah, it's either a little more coinciding, or, well, it's a huge amount of money that burns. That's the first point. Second paragraph. I wonder what's going to happen to the competition? Oh, well, TikTok won't be a good thing, but there's a Snapchat, like, that's a competitor if the hip comes up. Yeah, and if there's no history like the Clubhouse, if you remember, somehow, I think it was called, yeah, that's a resource where--
Yeah, yeah, yeah.
- the voice was spoken. A couple months was a huge hip. Then I think that many of our viewers will not even remember this combination.
While the coward was here.
Yeah, yeah, yeah, yeah. Perhaps the fate of this Clubhouse is waiting for Sora, but somehow I feel that if resources are allowed, resources will be allocated, and that will not happen. So some Snapchat-type competitors need to look that way and be careful. Oh, the next paragraph, if we're talking about publishing our videos in some social media, like you've been talking about Sashi's daughter, yeah, that CapCut seems to be called the program, there's some other resources. So, first, the skate is filmed, then it's mounting, some filters are put on and it's put on. Here, if we're talking about Sora and the social network they're building with Sora, there's no room for that extra software you need to use.
You write, you've got videos in a minute, you write some changes, so you're changing in a minute.
Yeah, but--
That's...
Yeah, but--
- a videotape where you can do it.
Yeah, but they're basically trying to get it more serious in there, at least according to the data, and they're gonna open it all up. It's just, uh, it's a basic look, all the systems that were now, like AO2 or there in Higgsfield, or all those Quant, there, is Qwen or how they are--
Cling.
- which Chinese, Chinese. Cling's still there, yes. They, uh, generate video that or you're processing your video, and you're bits of that--
You're taking your video, and you're putting pieces of it in, like, CapCut or, well, it's a tictocopic subject, right? Well, either separately. I mean, dance. Or you put it in-- Instagram now has your pre-appendix separate, and you add it, you know, how much fragments you put it out. You have-- that's why you're saying you don't have to work. Still, people of Sora, sora, many will use. We're the skates that show you in the shorts of YouTube. We're sort of putting them out separately. Right?
Yeah, but you can put them out as it is, first. Second, accordingly, if you stay within this network, if we assume that we are again, like ChatGPT wants us to be outside their ecosystem, you don't need to use it. You're a nudent, unsatisfactory, overgenerated, stopped at the result you were satisfied there and put it on the site. You're leaving behind what other people have been generating. So here, um, I think that if this thing goes up, there's gonna be a lot of users and a lot of people, respectively, in this ecosystem, some annexes won't be very good, right?
They either have to move or I don't know. So, social media like Snapchat, which, in my view, is not the most popular, it might lose a part of the users in this direction. Because, in fact, you can even use Sora as some sort of analog, I don't know, voice Twitter, X, right? If there's 100,406 on Twitter, I think we need symbols, right? That's a window to write. Here, you have a 10-- de 10-second video, where you can put your news in a generosity. So, the potential in this story is what I see. I wonder what's going to happen.
Look, but first of all, there's one more thing you're telling me about, it's very important to reflect on what's happening in Sora, just who doesn't understand, it's a social network, that you're the one that's the one that's the one that's got to be the one that's got to be the one that's got to be the one who's not really aware. You say. Because just the article I read about social affairs, maybe that's what it meant, she meant, like, an expanded, more functional one. Maybe he'll be added there because we've seen a lot of different lights on the video. It's like Midjourney, right? Well, they have a social network, you can-- yeah, they have a social network, you can slide, you can re-sensitize. Uh, Ildar, please tell me about Sora's logo. Is that a logo or can't be done yet?
If I remember correctly when you're on API, it'll be without a logo. Mmm, I think so.
Well, maybe if you're a pro-version, if you're probably, like, a logo, no. But the pro-version is still unaware of how it is available at all.
She's still, well, at least she's not-- she's not-- she's not-- she's-- she's-- she's-- she's-- she's-- she's-- she's-- she's--- she's-- she's--- she's--- she's--- she's----- she's--- she's-- she's-------- she's- she's--------- she's------ she's------------------- she's- she's- she's------ she's- she's- she's- she's---------- she's----- she's----------- she's- she's-- she's----------- she This logo is probably coming up when you're rolling. I mean, I put down a downloaded video, not a reference to the social media, because there's no social media. Oh, yeah! Another paragraph. I mean, about the fairies, uh, always interesting. And here's the, uh, social network that's just deepfake built. And in this sense, you don't have any deepfake video inside of her, no question arises. This Sam Althman is running now, is this a fairy video or not a faeque video?
Wait, but they're gonna add a function to Sora, uh, to get you to do some video, add yourself up there, and so on. Something was doing around me. That's like Hixie, that's where you can be around, I don't know, to fire. Although you-- even when you're shooting around, you're getting faety.
Yeah, yeah. That's the first. And the second one is you can get your own caméo now, they're called caméo, write it down. And just like I did, Sam could be the same as a stitch. I'm not really squeezing myself. Not the famous person, but somehow. I mean, it's very easy. You're just making your own caméo, literally, just over there, you're saying, you're turning your head and you can get pictures with your face. So in this regard, in terms of the generation of some kind of fun content, I haven't seen the resource before. You'll share the opinion of who in the commentaries who made the difference.
Maybe you're leaving some of the links. I'm gonna be wondering if I'm gonna see a lot of fun on the Internet.
A lot, a lot of things, really on the Internet. And, uh, I'm honestly saying, you know, Ilnar, I personally bought all the systems that were. Actually, I think that's it. I bought Runway the best, all the systems. I have a signature there, I see, permanent, like Hixie has a signature full. And I mean, I've been facing the same problem everywhere that the generation is very complicated and not fast. This is your example that you threw into our chat this morning with our-- us four, where Sam Althman runs to watch the Moon. I think that, and I understand that you didn't get in there, uh, three hours without, right? That's a very cool indicator of what's going on. Well, in terms of what's going on. I agree, right?
The entrance door is really very low, much lower than the other services I've tried. I haven't tried as much as you, but I imagine that in many services, you need to do a very, very detailed, uh, prompt. It's like before, remember, there was even a whole professional of a prompt engineer who, I think, has been more than a memorial for the past few years than a real job. Sora did the same thing with the propppt engineers for this video. So the doorway is getting very low.
Yeah, very much lower, very interesting. And, Tanya, you'll have to look at that, too, and you'll tell me what your impression is. And I'm sure everyone's gonna be happy to see it. It's still ChatGPT OpenAI with huge power, the huge resources of Iia, well, that's a very interesting thing. One more time, it tells me that OpenAI will be a very difficult task for many companies to catch. And if they start doing more, uh, they have, and they have a lot of very cool people on the team, and they start producing cool verticals, we'll see, you know, serious products on the way. Although it's probably the fact that they do DJI or they're gonna have seven trillion electricity, infrastructure, capacity, fees, and all the rest, perhaps the products themselves will be generated for themselves.
You're not, are you? Oh, I want to ask our viewers a question. Yes, uh-- again, they've done research among Americans on artificial intelligence. We've been talking about this one before. How do you feel, uh, how do you feel about artificial intelligence in terms of, uh, artificial intelligence, that's positive, that's negative, you're afraid of him or you're not afraid of him? So, um, why is that important? Now, the endless number of studies is different, yes. And there's a parameter like how many people expect to hurt him. Iia in America now-- how much is it? You're the one who's gonna be the one who's gonna be the one who's gonna be the one who's gonna be here. Do you think he'll hurt or do it good? Let's see who-- how our audience splits up. I'd put it there, I'd put it there alone, and there's one. Thirty-three percent of Americans think it's a nab-- all, well, the data is getting more and more about what's gonna be good. And thirty-eight expect to be more harmed. So, people are still the same, they react equally to-- they react to it, to harm, to benefit. But I'll put the two sides together. Iia, but the question among our people is gonna be very, very interesting to have. I don't think you'll wait for us to think that, um, only harm there or only benefit, but that's it. But you can certainly see the subject.
By the way, the infrastructure article went to Reuters. I was just saying that I think we should buy infrastructure for a month, right? And the article went to Reuters. And Reuters wrote that investors had shifted their focus and started believing in artificial intelligence for a long time. And faith in longitude has shifted their focus of purchasing in infrastructure that is about artificial intelligence. So everything about buildings, electricity, the pay, the different kind of equipment. Well, there's a huge, huge, huge volume of everything. Iia was a further study. I just got a little back on mine. If you buy one of the viewers, if you buy the shares in the technology sector and are associated with artificial intelligence, write which companies you buy. And I know that this is not a recommendation, it's more like sharing, sharing your opinion, right? And what you believe and believe in. Because many people write. Look, I've seen a very interesting comparison recently. There was a Chinese company that manufactures more cars than Tesla and the cost of Tesla. And then that company was worth 100 billion, and Tesla was there one and four trillion. And they say, "How is that? The American market is overrated." And at this point, I want people to say, "You don't forget that there's a Chinese market, "and there's an American, and it's always different." It's like an example from the owner of the Yandex, right? With Yandex and Nabu. Nabu, by the way, is still growing. Here we go.
Aah, there were 20, there was-- in April, I think there was a fall down there to 20 strains or something. Illar, you must be more close to this. I want to make it sound today, too. And we have audiences, a lot of people, too, who are technically in. Anthropic presented the new Claude four and a half yesterday and this afternoon. I'm a new version of me, I'm very excited about the phrase, well, not what I did, I'm just recording it. It was written that Claude, the new version could be autonomously written in the 30-hour code and improve the accuracy of the financial scientific tasks, is an improved operational trap and has grown considerably, there, sixty-second. Interest versus 40. I mean, it's a half-time. The company's targeting corporate clients, right? So, it's 30 hours to write the code. Can you explain the phrase to the viewers? I think she's very important.
Oh, I'm more of a question, how much would it cost? Because, uh, Claude is one of the most expensive in terms of, uh, cost, there, there, there, for a million token model. Aaah, if there's any limit, they're just leaving instantly. Especially if you use the most powerful models. He might be 30 hours, but it's gonna be very expensive. And that's the best model for programming. They're all beaten up again on SWE bench. Here. But I'll test this one. Not yet. In general, uh, I don't think there's gonna be any difference in the plus of what's already in place. Because, in fact, GPT 5 Pro is already well and writes the code and performs the tasks I need. But I'll compare Claude. They and Claude Code have been updated. Uh, we've got a new utilitator on the same. It's gonna be very interesting to see, too, but the subscription plans don't care if you use good models, you're fast enough to lock yourself in the limit. So not the most work-oriented tool I've ever had, but for some specific tasks, it's very good to show itself.
I'm sure it's pretty cool in terms of the code and the possibilities. I'm still waiting. Oh, you know, I just saw it for myself. Uh, I have a few resources for my personal. There's my personal website, like that made through Notion, there's a Supra, right? And I don't remember, you seem to have even advised me, yes, that hint. And I have a lot of content, pages, and I wanted to migrate with it. I don't like the restrictions that Notion has. And I wanted to migrate to a good CMS system. And I was just a few weeks ago or a month ago, I think I even said that in some video. I realized that I must soon be able to show up where I'm actually gonna show up all my resources, and I'm gonna have some of my-- my website, conditionally, without the system's CMS. I mean, I don't want to have a CMS system that will limit me to elements now, I don't want it, I don't need it. I think the future is definitely not that. I mean, I want to be more aware of the entire architecture of my website, understand the requirements. And when I get a new article to put it on my website, I have hundreds of articles on my website, in Russian and English, and in English, too, right? Or new videos appear or I'm launching some new programs, some courses, some sort of exits, meetings. I'm just giving it to the content and saying, "You're gonna get it some kind of shape," right? I mean, I can even ask the agent to go to Notion right now, but that's not gonna be the same, right? In particular, if I do, well, otherwise, it's not that quality. I mean, I think I'm just like this, Ilinar, that maybe six months, something's gonna be coming up, maybe faster, maybe Anthropic could do it. I mean, I don't mind Anthropic paying, like, $500 or $1,000. I'm ready for a thousand. I'm actually willing to pay a little more money if he could get me to do my job. It's complicated. I'm actually having a very childish task. But for systems like Claude, I think it's my job, she's just a nurse.
I don't want to. Look, I don't want to.
I understand. I understand, yes.
I can do it exactly. Let's do it honestly. I can open up and write it all myself and put it all down with all the details, features, nuances and all the others. But it's probably gonna be in Claude, because in the world, there are a few billion sites, one-- or, well, some big-- soft, at least a few billion dollars. I don't know how many sites I'm sorry, yes, but, uh, hundreds of millions, yes, it's clearly-- it's got to be a function that will update them. Because there's a lot of things written by very old, glute, curves, unworking. There are many new technologies, many different limitations, details. I think this system should know all of this in its own right, and it should be based on Anthropic. Here, or on the base, I don't know, Codex, or on the base... By the way, the OpenAI has pulled the pddata off his system. Right? And I don't know if you tested or not with Codex, they did a whole thing. Which of our audience test, write how much you saw the coolness of this peddate, right? And on base, I don't know, there, Cursor. Anyway, something has to show up a story that I can generate like that, well, uh, uh-oh-oh-oh-oh-oh-oh-oh. Just like mobile applications. I know people who generate a huge number of mobile applications, right in Claude, right? No, a-a-a-- not in any specialized fraimvors or some specialized studio for development and so on.
Yes, it is. You can also build a voice-injection on your Mac and the keyboard, not even touch, but a voice to him says, "I want this, this." I did.
You're calling me, a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a- You're calling me to start in Anthropic, start, open, start programming. I sometimes have a question like that when I look, I give someone a task, and I see how people do things without artificial intelligence, how much time it takes, and how many mistakes it takes.
I think sometimes, "Look, can I not give you a task at all?" And of course, I'm now a huge number of things, uh-oh, in a different mode, yeah.
Look, I'm not gonna be on the special range, but I really like it if I get back to OpenAI, they're gonna have an article and, 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- Still, they say, yes, that artificial intelligence will soon replace, there, programmers, soon, there, someone else, right? And, well, it's some abstract phrase, and it's not clear, it's gonna replace or replace. There is, for example, a GDP Val baccalaureate, which is built on real practical objectives and, accordingly, checks on how the systems can meet these challenges. A-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-baccem, I'll throw Hiding Face on it. And you're gonna pay attention, yes, there are some columns. There is a sector, respectively. There is, uh-- what kind of employment, what kind of profession. There's a prompot that's being handled, and the files are attached. There's actually a huge number of different professions. There are lawyers, there are engineers, there are health workers, and so on. The funny thing I saw here was, uh, it was private investigators. Yeah. I mean, for example, the tasks that are in the field, there are private investigators. And the task is further pursued, it is formulated and created by real people, experts who are costly enough, but they have spent a great amount of money on it. A real man is also trying to achieve this goal. It takes him a little bit of time. There are specific numbers of the article that can be added or read. The reference will be too. And the system is deciding. Next, uh, man, also the expert in this field, blindly checks the decision made by the person, the decision made by the system, and on that basis chooses, where the solution is, what decision is not the solution. It's fine. On the basis of this, it is assessed whether the system can actually replace or accelerate the work of a person. And here I would recommend to our viewers, especially those who are worried that their work can be replaced. A-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a- Because we repeat the phrase repeatedly, yes, all the things that are measured are improving. If there is a benchmark with these tasks now, it is likely that the LLM task will be accomplished quickly enough. It may be there in a month, perhaps six months or more, but in a shorter perspective, these episodes will be fully resolved by LLM. There's a few things to be taken into account. Like, there's appended files, right?
There's a challenge for a private investigator and there's a set of files on which he's supposed to be some of his own.
- who he must make some decisions, actually, make some kind of conclusion. Oh, just in case I can show you how to open it. Uh, open up, copy, uh, the reference and, in fact, put it in the window of the browser. You've got a file. It could be PDF, DOC-Who's file, archives, etc. So this is a set of data that are available at the entrance either to the expert or to LLM-ca. And you can see, but in your industry, the challenges that they are close to what-- what you do. Next thing I want to draw attention. Look at the prompot. Because if, for example, in your industry, your LLM profession is not doing well, maybe you just can't ask her right, right? Very nicely written proms. It's a detailed, consistent one. You can open up, learn how to do your job, uh, systems to deal with these kinds of tasks. And, in fact, the following request: write in the commentaries what you think about your industry, about your profession, how dangerous it is. Uh, or you think the tasks are very far from the practice, or, on the contrary, I think I'm doing this. And look at the way the LLM does this. So write the comments. Very interesting. I'm specially showing, uh, the profession of a private investigator, because I hope we don't have one of those in our audience. That's to keep anyone out of it, yeah. And, in fact, write what's going on in your profession.
That's a little, uh, comment, that little news I wanted to leave.
No, I-- Suddenly there's a private investigator, so let a private detective put it in a comment--- put it in the comment.
Yeah, yeah, yeah. And look at the challenges. It's very interesting. I mean, there's a problem, if that's the way I read it, so there's a bench that's connected to, I think, the hospital or the accident, right? And there's a recovery of insurance money. And on this basis, the private investigator must decide, and this is the case, or on the basis of these documents, it can be concluded that the fraudster is on the other side, to argue his opinion and to make a decision. I'm going to send my report. So there's a sort of thing in there, apparently private investigators in the U.S. are doing. And for these tasks, actually, uh, modeling and training now.
Well, anyway, these baccoons, they, uh, they're more and more showing different possibilities, yeah. So... It's just now, I think, that's a lot of analyst, so many studies, so many data, so many details. Question of the sample of data and everything. Very interesting. You must be able to write down who did what tests inside. And if you don't, ask for Ilnar's help. Here. He'll tell you how to generode some prompt. I'll see you in a week. Bye, everybody.