Hello, everyone! We're on ToTheMoon. Technology news and sites from the Silicon Valley all over the world. We're going out once on Sunday, we're always four people. I think it's just some boom and boom and the explosion.
I was like that, to be honest, stress and light shock.
You can't check it out. Undeterminated system, subtracted a little model, and, in fact, then you'll have an answer that Coca-Cola is better.
It looks like a miracle, like she's gonna recognize it, doing it.
We're talking about making $34 billion a year.
Still, three hundred million users are not six hundred million users in Meta.
Well, they're TikTok. Two of those most interesting cases. That's right.
I'm
always wondering if I could pay, often pay more if I get the best service. What's going to happen to TikTok? You think they're gonna be, right?
Shit, they'll win.
So? A super saturated week, more than ever. The technological world by the end of the year... Usually, Americans are saying that in December, they don't know who's doing what they do in January. I think it's just some boom and boom and the explosion. But first, before I go to the bow and the blast, I want to tell you. I got the ToTheMoon Channel last week, and it's 35 episodes. I was cleaning some channels on YouTube Wednesday and removing one old channel. And youTube had a bang in the system. And with the old canal, I've had two channels, my personal channel, which has a 60-Video, and the whole ToTheMoon with a 35-File video. Why did they erase? They've completely gone off YouTube, disappeared everywhere, completely out. And I knew that if you were removing the channel on YouTube, it's not possible, you can't fill it. And it was a clean bug, youTube bug. I mean, I wrote technical support for what a man wrote to me, "and you must have accidentally removed the channel." I said, "I'm saying, "I'm removing one, delete three?" So after a half hour, he told me you deleted it, you can't restore anything. And I was like, to be honest, stress and light shock. In an hour and a half, a girl was in the chat room, in a Google correspondence, some senior manager who said, "Oh, you got a brandy canal, come in, and you're gonna have two buttons, and there's a way for you to get a chance. "recover." I smashed two buttons, everything re-established and all youTube showed up. So, the references have disappeared. I've already been writing, which means Alexander has disappeared the video, nothing works, no search, nothing at all. And then it all came back. It's all starting to look, index. I don't know how it works in one second, but it disappeared from the index, including YouTube, and then it turned back exactly. Here. I wonder what a bang like that in the system is, in fact, a mistake inside YouTube worked, and I wrote them at the end what to do. They say, "We don't know about that bang. You're all working, you keep doing what you are. Here. Well, it's a little damaging to maintenance. The guy in the backup just offered me a re-filling of all the videos.
That's amazing!
Yeah, it was YouTube's maintenance. Yeah, amazing, amazing.
Was it a man or not?
It was a man, it was a man.
Sas, after this news, it's very good to be integrated, which is to avoid losing our TV viewers, we have to go to Telegram Canal. Sign up, we've got all new, interesting, but unfortunately, as long as the Telegram-canal doesn't have, and we're still missing this chance.
That's not what it is. Yeah, that's not for that. We're just investing, we're just investing so much power and time, and so much power and time, yeah, and we're running the channel together. And surprisingly, yes, surprisingly, when a completely unfamiliar thing is. I mean, you're sitting, you're a helpless man, you don't know what's going on.
And, well, it's good that it's all recovered. So this week, I think there's a lot of statements from different companies, annes, graduations, all-outs. It's just that everyone's squeezing in the peddite plan. And of course, these 12 days of the OpenAI, the Anons, the Anons, I don't know, they've been announcing for a day, which will be 12 days, so there's a number--
I wonder why 12 days.
Some days passed, some days passed, and on the first day they released, announced the release of the O1 model. And, well, I feel like I've been feeling a dramatic change. So there was an O1 preview, right? So ChatGPT has been renovated, and these updates, at least, with me, uh, using at least my life and professional life, well, they're a very cool result. And they just let Sora out, right? Who's gonna say anything about... Sasha, you're talking about Sora right now, putting Altman back, Sam Altman back.
No, for $20, it made it available. That magic would have been if he could raise two hundred dollars in a month of subscriptions at Sora's expense. That would be cool when he gave everyone $20,000. But in Sora, of course, it looks closer to the product. So, as far as they're concerned, that's how they've been, so as long as it looks like a tool to make some geniuses and meneditate it's all in one place. I mean, until it looks like a tool to generate the final result, but they did more than they ever did before.
What
do you say...
What do you say about O1 at all? Here. And the O1s are the same. I had a feeling in person when O1 was released, I came in, tried some things, looked, and I read about, uh, theory, but for two hundred dollars. I had a very, very good feeling that I could generally buy and pay $200 a month because they give anlime more than anything, I kind of have no restrictions. That I can put files in there. I mean, I can get pictures right now, files are still in O1. Pro, maybe we can get files. In O1 ChatGPT, we can only at least have a little expanded version of this workspace with Illnar, yes. I'm gonna put some pictures in there and I'll check out the pictures in O1. But it looks very serious. How do you feel about that? Or what does the market say about O1 as part of what they've pulled? Because I see very, I see that speed has risen once ten, speed has risen ten times. Those requests that O1 had previously been pre-existing in six seconds are now five.
From a quality perspective, it doesn't look like it's a breakthrough.
Doesn't look like you, does it?
No.
I'm just really a balldea from O1 speeds, starting to work. So before she wrote something, I think she's listed something, and now she's working it out really fast. I mean, I can't ask her to think about her asking for a minute. Well, I guess you can, of course, but I don't have one of those. For some big text or something. Have you tried a new O1 model or a $200 version? What do you think? And you're willing to pay two hundred dollar dollars or what are you willing to pay? Please share the comments. Write it.
I'm very interested in your opinion.
In general, the OpenAI taught us, yes, when some new model comes up, it always works better. Maybe they're giving her more power, maybe something else with resources. And then, remember, there's been a lot of discussion that, in a while, the feeling that quality has fallen, that it starts to glow and so on. Maybe just under this event, at an event, at a time of $200, the quality of the dollar is being further improved, possibly more resources, more memory, more processors, and so on. And in fact, that way, slightly improves quality and speed. Mmm, but to be honest, I was, of course, a little shocked, yes, that there was a price of two hundred dollars. My first reaction was, uh, a rejection, like, and how are you gonna take ten times the most important market? You don't seem to have a very different model. Chinese have an open source at your reasoning level. Claude in general has closed a large number of tasks in the old $20 and so on. And I had the first reaction, of course. Is that what you think about the price? They're so much higher than the current market.
Until we have a chiel to say it was effective or not. I mean, while we're sad to have these numbers, yes, that's what we can say.
And from a user perspective, you're just a user.
I, I think, is a very important story that is. Sam Altman said they're now three hundred million users, yes, all the time. And we will recall that last week-- a couple weeks ago, the CEO of OpenAI said that next year they'll be a billion. It's a Cookerberg on this-- a couple days ago, he's got a view in which he says, "We have six hundred million users."
Meta AI now has nearly six hundred million monthly actives and as promised is on track to be the most used AI assistant in the world by the end of the year.
And I guess that's what he's saying, of course, we think we're the biggest LLM ever used. But, of course, I think the six hundred million that use, uh, systems, there, built in Facebook, in Meta, not s-- and in Instagram, probably not comparable to the three million ChatGPT as a separate annex, Which comes in to work purely with ChatGPT. I'm assuming that this is a little different use time, different uses. Yeah? How does this system even use what they do inside it, how much time they spend in it, what kind of complexity they...
I think it's important, yes, what a volume of material. I had a very positive impression, Ilnar. I mean, I was wondering, we said, yes, you're impressed that two hundred, like they're twenty. So you have a Russian man, you've got a feeling that you've been turned into two hundred. I have a feeling I had twenty, a little. I feel like I had $20 something that sometimes slowed down, glued, I got a lot of volume, I answered long. I used to say I had a limit, though you and I have a version of the extended limit than the usual Pro, which is like a plain-pay charger than a normal Plus. And in Pro, I thought, "Oh, man, I mean, I'm just having a chance to download it now, I don't know, there's not four files in the simple interface, but more. I have the opportunity to download the context to two hundred pages, four hundred pages to make some kind of analysis. I have the opportunity to make more difficult questions about asking and not thinking, and not trying to break the phone, like I was when I was a kid, I wanted to break sometime the monitor when I had my 286th computer, I had a sony. I mean, I hit him all the time, like I remember, but he never broke up, by the way, ever. Sony must have been a tight monitor, there's ninety-third, ninety-fourth year. And, uh, because when he was gluing, the computer, here I am at GPT right now. I mean, now that four-Os are just gluing. So you're-- she's starting to repeat herself like a cherished robot, right? And in O1 I did some tests, I saw a new one. Well, I mean, maybe Ilnar, you're right, they're out there, all of them. I saw something new for myself, and it's for me right away, I think wait, I understand right now that I have a mechanism that can only double the dollar that can significantly strengthen me. And I even dropped my sister's message. I said, look, you test now, maybe you should go to Pro's version because I see I'm taking my wife's thyroid specimen analysis, photos, plug in O1, pictures, plugs, 'cause O1's. I'm sure you couldn't have, yes, files. I'm gonna get in there, it's just that it's the conclusion that it makes me and, at my request, it's just a high level, you know, just a Ultra, just a little bit of a little bit of a ulterior. It's a miracle. It looks like a miracle, like it's gonna recognize it, doing it. And you're, uh, checking all the recommendations of all the doctors, all the details, and it's gonna recognize these complex images. Complex pictures, some extra detail, guides you to action. And it's headed for action if it was like ChatGPT and 4, 4Os and 3 and a half, it's like it's always saying, go to a doctor, he'll tell you everything. That wasn't the case. The way it would squeeze like scientific, scientific evidence that in such a combination you don't have to do more than certain things. You need to wait six months. You have a good idea in this combination. But we checked again, you don't have to do this combination. And on my hard texts, the same as I've started her. E, for example, a very complex structure of reporting business indicators can be put in place. Now I have to analyze certain financial statistics, and I don't... well, it's a little hard for me to get through it, but I'm really good with numbers, right? And she's just, uh, feno-menal things that are extra stuff that I couldn't master. And yet, in a model, in the usual ChatGPT dialogue window, he didn't give, 4O didn't give such answers, but I'm willing to pay $200 for that. I've said many times, many times, I can remember, I'm willing to pay $2000 a month if it's something that, well, well, that's where they say AI, like, what kind of real processes, I'm not gonna be able to replace the humans with certain, you know, characteristics, I'm willing to look more. But this is my purely business opinion, I don't know, right?
Here.
I'm very surprised he's not sending a doctor. At-- even if he gives you a very clear professional decoding.
Well, he's sending a little, a little.
Question of responsibility. Yeah.
He's doing it a little bit. But I have all the structures written in my original writing that I don't have to do it.
Mm-hmm.
I mean, chat-- like all the other chat, they're at first positioned on it. And you say, yes, I understand that I'm not an idiot, I understand that I'm seeing a doctor that this is still being discussed with doctors. Can you give me a hand-- do you know what the question you asked? Please give me research on thyroid shields that apply to the US, standards, Europe, Russia, Asia. Compare them and give them to me on the basis of different world standards. He gave me the standards to the world, he gave me the association, and they gave me the conclusions. I didn't ask for the opinion of anything but, like, some kind of thing, I wanted, well, how different places would make such scientific, right? Because the thyroid theme is still very serious. Yeah? And of course, I think it's just a phenomenal moment, a phenomenal opportunity. And you know what difference it is now for people who do not use it and who have the opportunity to analyse it further. I'm not saying it's a basic, purely on this, but it's not necessary to draw conclusions on this, but how to do a further analysis. I mean, if you're a doctor and you're doing a little extra analysis and doing that, well, it's just a super, super cool subject.
Here.
Great! Well, speaking of two hundred dollars for myself, Sasha, I made that conclusion. I mean, if you're hiding in this window that you're given, it's probably a good idea, right? And if in normal life, um, you're asking as hard as you've described right now, you do one week, and then you're probably gonna need free limits. And I don't see the point of up to 200 dollars, then, yes.
Or a plus enough.
The pros are good.
Plus, you do. Plus, $20,000, if you're like a man who still lives, well, like you have a life, plus, of course, you do, yeah. If a man is a little wider, I'm just looking at the way that, well, as far as I'm concerned, it's not, like, at all, a parameter. I mean, I'm sure I'm willing to give up some extra tests to do me, yeah.
Well, you're also a man who really likes to analyze a
lot of information.
Well, Tanya, for example, you're tied up with a lot of-- you're creating a lot of projects, including extra systems. Like, obviously this system is in terms of generating, there, complex designs, there, codes that are related to your sophth, it-- that you're the main use of simulations in the interior design?
Well, I'm AutoCAD, 3ds Max is very rare, AutoCAD, mostly.
Yeah, AutoCAD, for example, it can clearly add a lot more, you know, to reinforce, there, work or make some conclusions, or some complex designs. There's a very cool tape, we're putting it in, and maybe the OpenAI showed it where they're drawing up pretty complicated cases, there, on the paper leaflet, and how one handles it, yeah. I mean, here's the...
That's very interesting, Sasha, because I'm sorry I interrupt you. There's just a super, super cool bureau, probably a hundred in the world. It's in Los Angeles, Kelly Wearstler. I'm a fan of her. When I lived in Russia, I watched her as a polar star. Here. And she was just recording the sub-cadast and telling me how much they had a week before, she's got a 60-person job at the bureau, and she's been doing some sort of piece there for a week. They're doing it in a few hours with AI. I've broken my head, of course, that they're using it. She certainly didn't reveal any secrets, but she's very interested. So it exists in principle, and she has even separate units as much as it is.
No, there are questions of applicability here. We're also taking the sub-cadast, including a lot of talking about it. It's a story that people are being included in the testing of different things because it's gonna be a new case, new skies in human life. So a man can't wait to see what he's really gonna be, what he's gonna work and not work. I mean, there are things that don't know how to use, or maybe they're working really well, but at least you try to use it and watch it, and how it works in life, like, I'll use it further, yes. And even in part, it doesn't mean I'm treating myself with a GPT chate, in no way.
But the question is, if I see it can handle more complex systems, even that's why I said about these two hundred dollars, I know about the limit that now makes GPT-4o there, you know, There are limits or there's one thing, and how does she, for example, spit or something out periodically? There are things I've tested that have left aside. With the prototype, I want to test more complex structures in different businesses. I mean, see if there's any additional legality, yes. And since I'm still connected to a lot of things, I'm still making, including content, I have a lot of content, meaning, I'm running different groups, there, mentoring, and everything. It's an opportunity for me. And if there's a way to shed 40 hours of some material and get some, you know, structures that are easier. Now, you can do some kind of phantomping, learn separately. They said the phantoms were coming next year, yes, one. You can, you know, where you can build on your content. There's a lot to do, but it's complicated. We need some software to ask or do some sophthing. I'm always interested in light, light, simple things. Yeah, that's when you open a regular window, isn't it? And now, in the iPhone, there's a box in Apple Intelligence. You have a window, and it's asking for GPT chat. While it's working very hard, it doesn't understand how, but at least you're in a very simple interface.
But you're at least in a very simple interface. You opened the normal camera, you got this camera right away, and ChatGPT asks. For now, again, it works very specificly, but it's a pretty cool move. Just like OpenAI, you-you-you made a search. I'm irritated by what he's looking for in America on Russian websites, for example, right? Or he often gives some weighted information to the end. Or that ChatGPT 4o connects. At the same time, I understand that this is the future, this is a new interface to... Maybe Google will do it, and it's all the others. I have no questions. What's the difference? What's the difference? The question is, this is a new format, uh, a new format. Even in business, many people are looking for a promotion. There's a search engine optimization. They're kind of thinking about how to optimize the texts on the website. I have to understand that time will come, for example, a man will be looking through a voice, and how will the search optimization work, right? Or what a man will do when he's released... Now we've been comparing, by the way, Google, Yandex and ChatGPT. And Yandex has a Yandex. Noiro or whatever it is, I don't remember. We compared how he's giving this, uh, summary answer, uh, if you're just looking for him or how ChatGPT is giving. They have websites that match. So the search model, the search optimization is very similar. They have websites that match. So, on the same request, ChatGPT is looking for, and for example, Yandex is looking for, uh, often the same resources. That's very interesting. Here. And, uh, and this is the future to optimize, including the work, there, the company.
That's why, well, I've done a lot of different tests. I'm just saying, about the users, I want to get this story up on the users. I think that three hundred million users are a very serious number. And it's probably 3 hundred million users in OpenAI, but it's not 600 million users in Meta. I think it's six hundred million users, I think. I don't know. Sasha, maybe you're more like... I'll say something else. I think that, in Meta, the six hundred million users are so low-quality, well, low-quality, yes, yes, yes, yes, yes, yes.
Yeah. Well, I mean, basically, Zuckerberg, well, now TikTok is, if they're gonna be, yeah, he's gonna have, like, the Western world's largest, well, he's got the biggest consuit product. Right? Well, more than Apple, more than... Yeah, I guess it's the biggest. Well, on the same level as Google. Well, how comparable, huh? I think the Google's bigger, but he's got the same order. And the Zuckerberg perspective is simple: a new category of AI products is emerging. Mmm, assistants, they're sexting like Character.ai, and they're, like, like, an assistant, like, OpenAI, ChatGPT. Mmm, as it might be, a new category of products, which obviously lacks Zuckerberg from all angles. He thought he was buried source LLaMA, and that would help. But two years have passed, as far as they're behind and in sexting, and they don't do anything more or less in sexting because they don't know how to make sexting, but so that's acceptable. And in the assistant, well, he's trying to throw some numbers to show that they're in the game. But of course, we all know they're really lagging behind, yes. I mean, I don't think there's any questions. I mean, there's no question, I think. But, eventually, if we're talking about the number of users, history, there, Snapchat, Facebook and all of these applications, it says that the number of users is getting through Latin America, India, Indonesia and South-East Asia, and then through Africa. And then the question is, who's willing to subsidize the most? Because, well, there's a video model, there's a video, yeah, today, she's giving a little bit of generosity. And if you give free video-generation in India, it's easy to get, I don't know, dozens of millions of active users. And here's the question, eventually, whether Zuckerberg will subsidize it or not, and how much he's willing to go. I mean, if he has the task of becoming the biggest AI massenger, in the sense of building the biggest AI assistant in the world, he will certainly achieve this task, obviously easy if he is willing to subsidize. He has more to subsidize than the OpenAI. We don't know that.
But from the point of view of the quality of the product, the engagement certainly falls very much, yes, I think it's definitely.
Well, you mean, it's like, once, because users, because time spent in WhatsApp, Instagram, Facebook, it's much more than OpenAI, okay, ChatGPT, it makes sense that he can get people there. to build a very wide-ranging function. Again, who's gonna pay for it? Because, in fact, no one's gonna pay.
Yeah, if no one pays...
No one's gonna pay $20,000 a month for supplemental-- they didn't learn. We're all using it for free. WhatsApp is free, Instagram is free, Facebook is free. Nobody pays anything. Well, I mean, they pay for advertising, but not for using the product itself.
Tek is, I understand, yes, there's always another look at it, and that's what OpenAI's gonna do, and then, to go into the masses, you need to make a commercial. Nothing else in the world has yet to be invented. And this Facebook car works, right? So if you look at it from a pure AI-product point of view, then, of course, Facebook is very behind us, and their open source strategy has not yet been up to itself. It is in terms of achieving their goals. If we're talking about, well, whatever the state is now, and whoever builds the first AI product on a billion users, what the OpenAI says, then when they say when they make that statement, I have the first thought, of course, that, well, Meta is probably going to get there faster because, well, a billion users are very expensive.
We're gonna talk about spending billions of dollars, right?
Tens.
Well, maybe dozens, well, maybe they're sympathizing. But we're talking about billions of dollars, exactly. OpenAI's gonna be a hard-ass subsidy. Oh, well, Meta has. First, they can, and this is subsidizing, have advertisers, so maybe a billion users will be run by Meta first. I mean, if you're standing from this perspective, I still think they have more chances than OpenAI. But from the point of view of the progress that is under way, well, I think Meta is very behind me.
In your case, users like you might be easy to coin, right? So you have a specific intent when you have a specific intent, you know, you're asking for a review of the reports, like, a matching number or not. They say, "We're not a financial adviser, not a CPA, not a accounting, but there's a trusted accountant that can recheck our findings and below comes a conclusion--
And he'll say no more ads to me.
Yeah, yeah, yeah. Well, I mean, I see, when there's a strong intent, it's always possible to coin it. The hardest coining is when there's no intent. That's why Google's searching business is so great, because, well, searching itself contains intent. Oh, well, Meta is difficult, but it's clear that Meta is exactly the way to coinize even user sessions without a specific intent, they've learned very well. They have average revenue--
Very cool, very cool.
User's tens of dollars, and it's very, very cool. So, yes, probably a million-million-dollar users who... Anyway, well, well, I mean, who's the first one to reach a billion-month active losers, not just the lasers, because it's just not interesting.
Sure.
The monthly active losers are stable to have a billion, and before, even eight hundred million, there, that's the order.
I think Meta will be the first to go there.
But advertising, advertising, it's basically the only thing they need to be wrong about, right? From the point of view of how it looks and what it's gonna be built, there, considering, they're gonna find it or just put it in the query. How much money would they take for that? Because it's like, yeah.
With the ad, I'm actually a little nervous because if it's a clear commercial, yeah, like you have a separate frame where it says it's a commercial, and there's a publicity block, it's nothing to worry about. I see. There's a blind zone you don't watch and you'll continue to use. And if it's a hidden ad when he doesn't say we're kind of in a way, and it starts to let you down.
She'll be clear, it'll be clear.
Sasha, I'm worried about--
By law, they can't do it by law.
You can't check it out. The undeterminated system has a little model, I don't know. Anyway, I don't have any confidence that this is just not gonna be a model, yes, to teach a model to some data that is biased, which is rejected. And, in fact, you'll have a negative response that the conditional Coca-Cola is better.
Why would they do that?
For the big money.
Well, then, if you can do it for a huge amount of money, that's what everyone knows. I'm saying that. If you can buy something, you'll know it's possible to buy it. And, accordingly, people will be asking that you can buy a hidden commercial in... It's the same thing that Google. Imagine Google would be selling it. We've never seen Google or Yandex officially sell covert extradition in search of the market. Well, we know there are ad blocks, context-specific advertising blocks that go up and built up later in search, and it's all marked that it's advertising. So Google and Yandex swear, for example, they never make a purchase issue. It's the same thing, and it's the same thing, and it's the same thing, look. Can we buy some extra article or some extra pay-as-you-go? Here. You're right in terms of things, but again, if they're trained on sites that say-- like some kind of product, and those sites originally compare the goods, cat--
And these sites originally compared the product as an ad, they weren't marked as an ad. And that's a lot on the Internet, isn't it? How do you know the truth? Here. But they're probably gonna sell the commercials on the spot, I think again. I don't know what U.S. law means, that it's actually-- or Europe, that it's possible.
I'd be happy if it was.
Yes. But in-- but it turns out I'm gonna be asking, and I'm gonna be offered some extra, extra system. There was some article that, uh, well, that's what they said, first of all, by the way, they actually said they wanted to introduce advertising. The financial director said that a few weeks ago. And there was some article that they would be very difficult with price policy, uh, well, given the amount of the show, there in Facebook, what the amount of traffic there, there, in Facebook, Instagram, there, Gougal, that, Of course, this volume of traffic will be very difficult to fight in terms of the efficiency of this ad, there, price policy and everything. That, of course, is very difficult to turn around, to turn around on a small system. And either the usual users have to lose a lot of money, or the OpenAI has to subsidize this ad very much at the beginning, yes, so that people show some possibilities. But if they have a search like that now, like a beast of a tongue, I don't know, well, at what point they're gonna let it go.
Here. And at what point they're gonna do it. How much, Sasha, are they gonna have revenue for you now, this is a month's revenue?
Who?
Month revenue. How much do they earn a month?
Yeah, yeah, yeah, yeah, yeah.
I think we're talking about that there, that they make three or four billion a year. More or less, huh? So we'll split at 12, we'll probably get 3 hundred million. I guess it's fair-
How much will it grow, how much will they grow up in their income projections? What they did was four hundred dollars a month.
I'm just saying, a little...
I think it's a flashback, but it's not gonna do anything. Right?
I'm honestly, Sas, I can't answer that question a little now. That's a very good question, straight, but I can't answer at all. I can't interpret what Malta says. I must be getting deeper. I mean, I can't figure out how he raises so much money with expecting that the loss will grow and that the loss will grow in series over the next three years. I don't know what I'm saying. I guess I need to get away from the top level, maybe somewhere, I need to find a deeper analysis or a deeper pitch myself, because I can't understand what's going on. It's not my fault-- it's not understandable to me because I can't understand how this is all related to each other. I don't know.
Well, I mean, because basically, when they say they're making a two hundred dollar model that needs to be supported. I used to say that I was this plus $20,000 or $330, I got 30 in there, eating it instantly, because the number of requests I'm doing is a lot of things I'm asking and so on. And the question is, whoever buys $200 is, he's probably gonna be spending a lot of resources inside ChatGPT. I see there's a use in API, there's O1 and there's a lot of money, and there's probably a lot of money. But I think that's two hundred dollars, it doesn't make, it doesn't pay for the system. I mean, it's more like people come in, other quality of skies come, other quality of the skies. And I, like a man, always wonder if I could pay a lot, pay more if I get the best service. I mean, that's because there's things I don't have, there's no more service, right? I don't know, Amazon Prime signature, and it's like one. There's no Amazon Prime Pro to have, I don't know, anything extra. I'm willing to pay more money, but just give me something extra. It's not good for me to give anything extra, or the content of people like me is getting very expensive.
Here. They better do this mass story. So their own approach, the fact that they went $200, suggests they probably have some kind of strategy. Sas, you sounded at the beginning that, uh, well, before we turned on the video, that he had a strategy. If he announced that there would be two hundred dollars and two thousand dollars, he obviously had some kind of prediction, a vector in the front. I mean, he says, "I told you there would be a very expensive version. Here's the sweet version. I told you we'd learn to sell, didn't I?
Oh, yeah. I think that's the point where we have the LTV, which, well, we have to talk about some--
He doesn't say that. I'm listening to a lot of interviews, Sasha, you're just saying deeper things. There's no way in a big interview... there's a reason that AGI, not AGI, AGI, AGI, not AGI, that's compute, not compute. Just here. But there is no, uh, strategic, very clearly disaggregated indicators and parameters. I don't think they're there.
Well, they're supposed to be somewhere. He should tell investors. It's as clear as this is the most interesting business case there for the last decade. So build a company like OpenAI. I mean, Anthropic is the enterprise business, there, it's clear how the question is exactly how much it is to legit enterprise contracts. I mean, it's clear what An-An-Antrotic does. What OpenAI does is like the most interesting business case of the decade. Well, they and TikTok are two of those interesting cases. That's right. Well, the third one is probably Temu, right? There, Sh-Shein, Temu, shorter, well, these are the most interesting, very aggressive approaches to building consumer businesses in a new era. And very, uh, interesting what he's saying to investors. I mean, I see how Chinese operate. Well, I mean, it's all clear. TikTok what you did? We'll build cash printing, we'll-- we'll first optimize our China-based recognition system on the news, then we'll get it off the spectacle, learn to make a lot of money in China, reinvest in the world a very agre--- growth. The world is very aggressive. A model that's understandable. Shein, Temu is understandable, too. Well, like, not HZM, Zara shouldn't exist. Fast fashion must not exist because fast fashion must be fast. And now fast fashion is not fast, because it's very slow. And Shein, Temu said, "We'll all do the bully." We're all looking at trends, all data driven. A very aggressive Chinese production and Chinese marketing model. I see. But what OpenAI did, I can't just rationalize it. I mean, a tall laundry, but to lose more money, no one else has. I mean, you usually raise prices to somehow drive the economy, I guess. And here's the opposite. I mean, he needs a higher laundry to open large markets that make the company play more scales, what? But it's kind of a big-- it's a very difficult thing to understand.
I'd like to figure it out, of course.
But it looks like new, new use scenarios. You know, when they see the system, can we spend more time with the system? Looking back, looks like you're just using a little bit of a little bit of a request, just for the product you pay $200 for, you start with them, you know, a little different. Maybe they want to test it and check and see what happens to people's behavior.
I mean, not like a party in the gym, right? I bought it for a year and went three times.
Like that. Or workspace, look, you and I bought, we, Elnar, bought a workspace with you a year ago. That's where we're running out of the 12th month. What is that? They have a team version of ChatGPT, and it's worth a little more expensive, but you have more opportunities. ChatGPT doesn't learn your data and, at least we think so, and we can all be collabating there and all. In fact, there is no value from that. But they clearly tested the company, the building of theirs, well, people, the issuance of these licenses, including being in a single workspace, although they didn't create them as if they were. Wednesday to interact. I mean, weird logic, why would that be a chat for mass use? Maybe they're testing some scenarios, or maybe it's a stupid idea, yeah, that someone suggested, maybe just stupid-- like, like, plastic, right? They offered, looked, and they'll be all plastic or not. There were only flames, then there were other flames that are now. They're different. They're the ones who changed their style, right? And it's obviously not a living story, is it? They opened these plastics, you know, not a store, but a chance of a race. It didn't go there over some sort of story for people with super-lived. But that was a test on that side, right? So, and maybe it's just the circumstances of the tests, not just the mass story that's from the point of view of Meta and WhatsApp. Like, we're just gonna open a line on the team, we're not gonna improve, where we don't develop. Well, no matter what, this... I, I, uh, prefer the use of ChatGPT, because I have the opportunity, I understand that they're interested in, like, developing additional user chips, just for the simple. The user. Not in terms of API, not in terms of some integration of complex systems, not in terms of complex systems, but in terms of simple, independent, separate systems. That I'm using this system for this, that's all I need. I mean, I don't send messages like whatsApp or Telegram. Because I'm still in WhatsApp, I just want to send messages. I don't want to use AI there. Bye. Maybe for now. Maybe I'll move it, you know, like I'll go over it, right? I'll reschedule to some other type of use. While Apple has already introduced it into its own, Mail Post Office, well, all these chips, just like Google has them. Text-reading. I still do the text separately in ChatGPT and then I give it to the mail. I mean, until I use, like, Apple tools like, like, like, automatic. Again, if I get a very cool instrument out there, I'll probably just get over there and stop using it. I think we're just gonna be...
Sasha, after the sora extraction, what are the feelings about the model? Better, worse? Same as expected. We were just showing up in the past graduation what the model from China looks like. It's hard to say.
It's hard to say. I think they have... I mean, it's not gonna make all the models useless right now. But we have to understand that the best video service has hundreds of thousands of users on the day. And if Open, Open, OpenAI is still able to have millions of users on the day that will generate videos and probably have a high chance to be leader, I'd say, on this market. I'd say they, that's what they decided to invest in the interface, not just to, you know, have an interface, a sort of video-generation, that's what they're talking about--- well, it's like, First product they've made some kind of interface, right? Perv-- well, there was no such thing before. I mean, because ChatGPT is very primitive, and it seems that if they really took the video seriously, they're probably gonna be the winner. Probably. So their model is definitely not to say that-- that is, she's much better than all American models. Well, she's probably better at something worse than Chinese. It's clear that Chinese are much better at copyright characters, human generation, because it's specially what OpenAI is lingering. Oh, but by distribution, they can become leaders of number one in the world. So what Meta showed looks very weak compared to Sora. Oh, I don't know how they all compare it up there while it seems that there's a great chance in the Consymerian, Consymerian video of Sora. It's not very clear, it's true what skies will be, because if we're entertaining, there's always something to be done with people, or there's animating existing pictures or generating ourselves in some different ways. It's a nice stack, isn't it? Or round up the current video, yeah, so, uh, the social media video would look more interesting. I mean, it's, like, people, eventually. And I don't know. I mean, but from the point of view of the distribution of the product and technology, and, in the point of-- if these three factors are all taken together, Sora is definitely number one now, and the exact number one is, like, it's, you know, it's definitely there. Oh, but it's enough-- but if they're gonna move on, it's not clear. I mean, horse-- you can't say they have a better model. You can't say they have a product separately that's much better than anything else, yes, much. But combinations, three factors combined, they're the best.
Yeah, I wonder if they'll leave her, just pour her out because she was waiting for the market and waited for investors, and so on. And that's their strategy. Or they realized that--
Well, they were afraid of the election.
Or they've already realized, yes, that it's not their strategy and--
No, they were just afraid of the election, so they banned all the politicians there, all the copyrights they banned in advance. I think this release was on the shelf for a long time.
It feels like he was lying on the shelf of a month three or four. I, uh, all the feedback-- feedback in September, everything was ready for launch in October, but they were afraid of it.
What's going to happen to TikTok? Still, that's what you say about TikTok-
I hope he's gone.
You think they're gonna be, right?
What do you want to do? Then, fuck, fuck, fuck, then, everybody! They'll win. What do you want to do? Otherwise they'll win. Well, they'll let people's videos be generated, but eventually, these models, they'll go on some kind of platoon, right?
Mm-hmm.
Just a question who's gonna build the next ChatGPT in the video field? But I'm saying that it's probably Sora now. Well, that's the best chance they've got. Oh, well, maybe ChatGPT Video won't be at all. Maybe, well, people don't have to generate videos as much as they need to use text assistants, right? We don't know, that's why. But while the es-- but if there's a ChatGPT in the world, there's a video, it's probably gonna be Sora. Yeah. Well, I see what TikTok has the highest chances of doing this because, uh, if they're not, Sora, and if they're being shot, it's probably if they're not being shot, if TikTok doesn't get caught, they're probably not. They'll figure it out, because, again, uh, well, they have a very choro-- what they have, which they don't have, is, uh, how viral the effect is not just that the video is popular, but that the temples are... They're getting popular, yes, in CapCut. That's what nobody has. If they can sit down on good AIs, well, then-- and CapCut just, well, it's like an annex to create a video, right? But it's in the top of the world. I mean, more or less. It's more popular than ChatGPT, isn't it? So, as they already have a tool to create a video that has ChatGPT moments, there's a "I" in there. They won automatically.
Yeah.
Oh, I mean, but I hope they're gonna be filled up at the time--
Everyone has a chance. Well, otherwise no one has a chance. Well, they already have distribution. So no one's ever made an application for video-referencing to edit videos in top 20. It's very complicated, you know, because--
No, but if youTube adds, Google will add to YouTube its video or, like...
No, they have their own video, but I don't think that's what anyone uses. I mean...
No, I'm saying, if it's cool, cool, cool, if it's cool.
Oh, maybe, but I think it's different. I think it's another market. I mean, I think that-- you know, it's like a new product to build and Google would do a model. They just have all different departments.
No, I don't believe that's gonna happen either. They said they added a bunch of their AI chips on YouTube, I don't see anything at all. Although they seemed to have been able to have so many statistics, analysts of any fun with AI, but they're moving really like a mourn, and it probably isn't their speed. You're right about this, especially in the past, you're just-- we've given examples, Sasha, you've just been throwing those Chinese model videos, yes, open source. And, you know, it's a different level, they're moving at a different pace. Let's see what happens to TikTok. I don't know who's definitely doing this in the finals. It's just not very clear if they sell their part to someone, so this new buyer is who?
So who's gonna take it under? That's how it's going to be, I don't know. That's what happens if TikTok closes up for the American market, what happens? That's interesting. But in children's schools, they're already talking about it. At least my daughter came and said that TikTok's dad would close. I understand that kids are, you know, logical, right? Because the young people are, of course, TikTok.
They're the main users.
The main users, yes.
Yeah, yeah, yeah, yeah, yeah. Who's gonna win this? Well, Snapchat will win this, Sasha, first of all. No? If TikTok gets caught.
No, he won't, because Snapchat isn't about the video. I mean, it's like they're having pictures of the main use of the case, and like what my learning might be, I'm wrong, but my basic learning, that pictures, videos, two different uses of the case, just different things. And yet, youTube Shorts and Instagram Reels, well, it's a product that competes with TikTok somehow. Everyone knows the people who watch Instagram Reels, I know the people who watch YouTube Shorts. Those who watch Snapchat Spotlight don't know those people. So, I mean...
Yeah, I probably meant in the child market.
In the kids' market, yes.
In the children's market, the children's market will win Snapchat, and in the shorts market, it is clear that it is Instagram.
What do you mean? Snapchat's growth will return, yes, Snapchat will return to 13 to 17 North America, and they will return to growth in North America. They're stopping right now. If TikTok is, shorter, forgotten, those Snapchat will be back growing, but revenue won't grow much because there's gonna be youth. And for the Beneficiaries, of course, Meta and Google are the most, right? Well, Meta-
So you buy stock, you have to buy Snapchat, Meta and Google, right?
I think it's just Meta, but they're expensive. I mean, it's useless.
And if the era of Gen AI is caught, then I think Jason Huang is the most happy with his chips, because the computer for the video generation needs a lot more than the texts. And there's a turn again, a lot more to the NVIDIA chips.
It's not, it's automatic, I think.
Yeah, it's all set.
Well, they, I think they're interested in TikTok going to minus. They must be interested, yes. They've got more of them here.
Everyone's interested. TikTok is being removed.
Well, look, the death of TikTok in the American market is a serious, serious step, including, uh, for a number of networks. And Elon Musk, I wonder if, within X, he's interested in not having TikTok? Or is it useful for him to be in the American market?
Well, he said he wanted to make a video of the platform, and he'd do it. He's just gonna do it. I mean, just--
Well, he's good, in the idea, he's good if TikTok leaves because someone needs an alternative from Instagram or youTube, and including X superseders in theory, you know, X will definitely expand. It's functional. That's obvious.
Right, right. Yeah, yeah, yeah, that's gonna be right, yeah.
In any case, they grow, they grow, they grow even more, they grow, we won't avoid it, especially given that he has AI there.
That it might not be a video-generation, what we said was another question.
We did. That's another question. Here. Aaah, from the point of view, if you say that youTube and Instagram will not do it inside, but they'll have to, uh, do something to compete, compete, compete. something? Or will they be integrated, for example, with the same Sora later? They'll have to compete. I don't know. Well, it's kind of embarrassing--
I mean, I don't really understand Sora skies until now, but again, the quality of the model is, the product is. I mean, like video mediators, actually, like a media organization, distribution is probably not good for them. I mean, it's probably not good for OpenAI to have an opinion on this as if it were what Sora needs. I think it's likely that once they figure out, we'll see how good they are in the Influx Marketing, obviously. Oh, I don't understand, but good, by the way, and we-- we'll probably be screaming like a opinion as to how to use Sora. It's probably gonna be.
Probably. I mean, it's probably like a percentage, you know, that a percentage of AI video will be sora-generated. That's right, that's for sure. So, there's a lot more quality than anything in America that's been done until now. But Chinese models, you know, hard to say, worse or better. Not so much data as long as it's hard to say. I mean, no, there's no video made with OpenAI Sora that can't be done without OpenAI Sora. That's right. There's no such video, but it seems to work backwards, too. Well, the question is, yes, that's what people are up to now to do about 40-fifty seconds of AI video, spend dozens of hours there, 50, 100 hours, roughly speaking, right? I mean, aaaaaaaa-- i mean, it's still cheaper, but it's a long time. But we can't even be objective, make it more comfortable with Chinese models, or is it more comfortable with Sora, right?
And plus we don't-- we don't know, well, whatever, and the market is, well, how would AI be, like, a bonus like a mark, right, actually? Is this a premium marker or not? Because, well, the story of innovation, I'm not a specialist here, that's the most interesting thing. I'm interested in your opinion here because I have my audition to complete. But I figured it was about the movies that said it was the most scumbag. It's like a theatre, but only for, like, for, like, actually, well, for those who have IQ very low, right? And most likely, but they still watch movies now. And as the theatre is for aesthets, roughly, yes, well, for a very small percentage of people. And the movies look more and less. The movie industry is much bigger. It's just that maybe AI's video won't compete with the existing movies, and not with the theater, but it's gonna be some new video. We'll say, "This is weird," but our kids or younger brothers will say, "Blin, what you need, that's what I want to watch. A little shitty, but, fuck, whatever you need.
It's like the TV show, right?
Yeah, yeah. So, as content, it is clear that it is on the way to simplify and accelerate. And as long as AI is not-- well, it's obviously easier to create it with AI, it's not fast enough. AI videos are slow, but they'll be clear soon. Plus they can be generated and accelerated. Maybe it's just gonna be a new type of entry, right? I mean, maybe we just don't know yet, maybe in the Instagram Terms, we can't speak that way. Maybe it'll just be a new genre of content that people will say low-quality sucks first. Aah. Well, that's how if that's what you win, it's not exactly what Sora's trying to do. Sora is trying to get some new shape-- a new form of art, very complicated fractals show that something is so very, very intelligent, something like that. Somewhere closer to museums, shorter, like, that's something they don't have museum work, but they're animated, something they're trying to create, Wyb's like that. Maybe, well, the answer is that, for fuck's sake, the quality of the model won't be solved so much. Maybe, yeah, maybe it won't be a big deal. Maybe it's not a model quality story. I mean, of course, it should be a little better there. But eventually, next year, maybe the one who builds a big business in the AI video will be no quality, well, not quality, as we think about it today in terms of the film and popularity. music videos.
I'm the only one that's thinking.
Yeah, that we've been discussing something that we can have different branches, uh, movies, like different follow-ups or the continuation of the movie, or different scenarios that have been seen for people, or scenes-- or by their own. Film scenarios. It can really create a new era of perception. You said well now that no one knows what's really going to happen, that people really will like and that's not.
I mean, yeah, I can, like, say in my experience with Hollywood that, well, if you say now, "Let's go make a new franchise," like Disney--- Mickey Mouse, but something new, but just like Mickey. Mouse. And we'll, but, um, we'll be like reproducing Mickey Mouse, our new, our new IP, but we'll be going there in quality, right? It's kind of a no-brainer. I mean, this kind of business can be built, of course, but it's not, it's what's gonna be new there. But it's not the same, it's not gonna be the biggest company in the AI video. Right! Because, well, models aren't there. It will require a very complex combination approach. Get something out of there, get something out of there. And it's not gonna be a rock, is it? So it could be built by a new Pixar, like a new sales company. Yeah, we can do new chat Hollywood movies with AI, but it's n-n-n-n-n-n-n-n-n-n-n-n-n-it's like competition with what's already in place. It's not opening up new, innovative opportunities. So AI cannot replace existing top films and franchises as they exist. He can only speed up the process, cheap production--
Yeah, the creation process. Yeah, the creation process.
Yeah, yeah.
But it's not the biggest opportunity in the AI video. Right.
Yeah, it's help, it's gonna be a hundred percent, but it's definitely not a new market. What, what you mean. Mm-hmm. That's very interesting. Thank you very much, everybody. We're on ToTheMoon. We're going out on Sunday. Technological news, Silicon Valley sites, from the United States around the world. There are always four of us. Bye, everybody. Have a good day and weeks. Bye.