why does a guy come in like this, well, how did we talk, he says, you iago just started talking, are you 100 thousand rubles? he says, well, 100,000 rubles. yes, you don’t understand, this is a talking dog, i just heard him, we were talking, yes, he’s worth, he’s worth millions, millions of dollars, and this guy is like, you know, hair is already popular in a suit, he says, well, first of all, he you've got it all wrong. we continue the jokes podcast, rinat and mikhail stayed with us in the studio and are ready to tell a couple more amazing stories of jokes, please, renat, you said that you like short jokes, but you know some, long ones , artistic and theatrical ones, i love them, especially when they tell some stories there, you know, who works in the theater, there are movie actors , it’s just that sometimes you listen already, even you even know it.

year, he says, well, it depends on what kind, he says, well, whites, let’s say white, he says, they produce one and a half tons of wool a year, he says, and blacks, and blacks produce one and a half tons, he says that he would simply unite, he would say three 3 tons they give everything, let's continue to write, he says, tell me how much food they consume, this is how much feed, this grass, he says, well, again, it depends on what kind, if white, he says, then about 3 tons per year they are eating.

because it needs to be told quickly, you will forgive me, dear viewers, but nevertheless there is an exam, that is, they presented, there is a topic, there is an exam, the professor tells the student, there are 500 bricks on board the plane, one brick fell out of the plane, how many bricks left on board, the student says, well, that's easy, 499, the professor says: right, the next question is: how to put an elephant in the refrigerator in three steps, the student says: first, open the refrigerator: second

, put the elephant in there, third, close the refrigerator. professor, okay. how to refrigerate deer in four steps. student, the first thing is to open the refrigerator, the second is to take out the elephant, the third is to put in the deer, the fourth is to close the refrigerator. professor. great. next question: it’s the king of beasts lion’s birthday, all the animals came except one. why? student? because the deer is everything still in the refrigerator. the professor is great, next. can granny walk through a swamp with crocodiles? the student, of course, can, because all the crocodiles are at leo’s birthday party. professor, okay, and now the last question, granny went through, went through an empty swamp, but still died, what happened to her, student, she drowned, professor, no, a brick fell on her, which fell out of the plane for a retake , that is, i always imagined that someone sits and makes up such jokes, and there, well, there’s like one joke when a guy is lying on the couch at home. not

calls the girl, hello, hi, she says, hi, i say, what are you doing, she says, nothing, i’m home alone, my parents left, they’ll only come tomorrow, he says maybe i’ll come to you then, she says, come on, come on, he says that we'll drink wine there, maybe he says, come on, let's watch a movie, yes, she says, come on, come on, let's massage each other there, yes, she says, come on, come on, we'll kiss there, yes , come on, come on, let's make love, come on, come on, who is this, about the son, it means two friends meet, and he he says, what’s wrong with you, he says, your eyes are so huge, he says, son, son, i had a son, can you imagine, after five daughters, i had a son, after five daughters, son, he says who he looks like , i’m either there or

i’ll get married, he says, but we haven’t looked them in the face yet, some jokes that can be called anecdotal have happened, you’ve had something like this, yes, of course, we have them, uh, what we have in life happened kamyzyakami, we pulled out the stage, no, no, and we had to do with the shows, it was evident from your command there, how does this all happen, well, we had a case when we wrote a song competition about a migration service worker and it was not legal about love, yes, i remember that, and this song, and i think we left somewhere in baku with on tour, we are lying in the hotel, we just arrived and a woman calls azamat and yells at him into the phone, why did you write a number about me? it turns out there is a woman in kamyzyak who worked in the migration service, she fell in love with an illegal immigrant and everything is according to the song, she taught her to play chess, that’s all, how we got there five times, probably for a song, she watches on tv, so yes they are in general, but they are not she immediately thinks that there is

a rat in her team, someone someone leaked it very much, only we knew about the chess player together, you know, this is a game for two, i remember at a time when all these... one of them looks, and under the lid it says 3 l, that is, and there you can win nothing there, that is, you you buy a small bottle, they bought liters and there are 3 liters, and that means they approach this rural saleswoman, she is such a portly girl, you know, she’s all cheerful, they come up and

say: oh, what is she, she says, about 3 years old, she’s like 3 years old? says like what? well, 3 liters, here, well, you have everything here, there are cars, apartments, caps, you can win there, how many are there, three, what, five, that 10 liters for the promotion, she’s 3 liters, she looks, she sees , that 3 l, he says, well, come on, she, what should i give you, he says, well, what to give, they are already starting, well, 3 l, the game is like that, she, and no, i don’t play such games, mishanka, but there are some regional ones like that, remember the joke, i immediately comes to mind, but this is an old one, i don’t know, this is probably some kind of soviet joke about shuttle diplomacy, henry kissinger asks what shuttle diplomacy is, he says, well, let’s say, let’s say, i need to get my wife married

rockefeller for a simple siberian guy, i go to siberia, to the village, i find a hefty guy there, i say: do you want to marry rockefeller’s daughter, well, i beg you, we have a lot of our own girls here, cool, why do i need this some kind of new york, yes, but she’s the daughter of a billionaire, that’s my department, huh then i go to the board of a swiss bank, i say, you want a simple siberian man to be a member of the board, you’re crazy, yes, but should you take him? this is my business, then i go to rockefeller, i say, do you want your daughter to marry a simple siberian man, in our family all are financiers, jews, yes, but he is a member of the board of a swiss bank, he says a second, susie, come here, mr. kisinger

found you a groom, he is a member of the management of a swiss bank, she says, ugh, these financiers, they are all dead, yes, but this is siberian. when russian belarusian and kazakh they end up on an uninhabited island, also an old joke, in my opinion, and the belarusian begins to tend to the garden there; everything, potatoes , everything grows as it should; . such a green one was messing around with makarov somewhere , a young policeman, it means he was quite a lieutenant , he was standing in a uniform there in a prison guard and somewhere

he lost it, turn it in now, now a new one will arrive there, the same flyer, and he’s looking for it and it’s nowhere to be found, and i lost the gun, what do i need lord putin is looking for everything there, tidying up everywhere , looking at all the bedside tables there, opened in the closet on the window sill, nowhere in the corridor at all, and he’s just panicking, all his service weapons are gone, now they’ll deprive him of his rank , what else will happen where everything’s gone. there’s no gun, that’s it, and he’s already where, where was i, where was i, what did i do, what did i do, god, what will happen to me, i smoked, i smoked, that’s right, he crawls out into the street and crawls under the bench next to this the smoking room stinks of bulls, in fact he’s just crawling on the grass in his pressed trousers, where am i? the pistol sat down, looked in there, completely got it out of it with his hands, he had already lost these bulls in his black hands, he was under the bench there again, he returned there, everything fell out, there were already such holes.

uncle is a policeman, and this is not your gun, but he is looking at it, no, i lost mine, jokes, tv series. let's talk about turkish tv series now you know so now it's dominant how i started telling it in turkish it would be funny this is the second part of the joke when a guy is lying on the couch and a girl calls him he picks up the phone she says hi hi what are you doing? he says, i’m lying on the sofa , he says, my parents have left tomorrow, they’ll just come, you’ll come, he says, come on, let’s say. let's see, come on, come on, let's drink wine, yes, come on, come on,

we'll give each other a massage, yes, come on, let's kiss, yes, come on, let's make love, he understood the hint, i'm leaving to sum up some kind of conclusion our conversation, yes, look, look, despite the fact that you are not a fisherman, in general, and neither of us are kaza. that is, it’s still hobbies, they can be different, but it’s still friendship, like you say, yes, it can appear suddenly, yes, as misha said, he started going for walks with dogs, became friends with someone there, yes, if i can say so, or when there is a children’s theme, this educational one, when you start walking and they appear there, you start communicating, and that is , she can appear suddenly, and you can also suddenly lose her, but... someone moved somewhere, my best friend, i don’t remember his name, but that’s it what we laughed about at the beginning, one

of my best friends, yes, therefore friendship still, we need to appreciate it, and the best thing is that, probably, what unites us is still humor, and the fact that we played in kvn, we played generally great, there, well, thank god, not somewhere then they heard something about it, so i would like to address our viewers, maybe after today... the episode or right now you will still remember about some friend of yours, just call him or write, i i think that this will only make things more positive for everyone, yourself, and of course, your friend. that's it, keep your friendship, tell each other funny stories, jokes, this should make everyone a little more positive, in fact, this is what the jokes podcast exists for, so we say goodbye to you for today, thank you, dear friends, thank you, friends, for watching. until we meet again, bye,

dear friends, the creative industry podcast is on the air, there are still living, real copies of the presenters with you, this is still elena kiper, producer and video director, and i have not yet... roman karmanov, general director of the presidential fund for cultural initiatives, visiting us, visiting us today, managing director on researching the data of the largest bank, i’m very glad to be here, hello, we didn’t just start out floridly, because artificial intelligence is really a topic that everyone hears about, but very few people understand what it is, here...

artificial intelligence - this actually usually means some kind of neural network, what is a neural network? a neural network is something you know, you can imagine it as a black box that takes something as an input. and something comes out, that’s exactly how we imagine it, how black i am, what’s inside black box, this is a kind of mathematical function, a function that now, for example, lies behind, for example, generating a picture from text, has - well, there are billions of parameters, that’s why they are so complex that a simple person who is now sitting at our digital still on tv, about

artificial intelligence, within the framework of such an everyday understanding of artificial intelligence - this is... well, artificial intelligence is this kind of black box, you can imagine it this way, which everyone just imagined now a black box, what to do with it, which learns to make some decisions, these decisions can be a huge number of problems, in fact, which it solves, that is, you take a certain amount of data, well, let’s say, a certain amount of knowledge, well, conditionally speaking, you take all the volumes of leo tolstoy, load them into this black box, and it can give you, well, conditionally , another volume, as it were, of leo tolstoy, for example, well , in a sense, yes, that is, this function, it learns in this way, that is, the parameters are configured in this way, here these billions are based on, as it were, so that it is configured, what does this mean, this means that - this is it, that is , a learning process occurs, if we are talking about language models, such as - for example, well- known gpt chat, gigachat , for example, this is

a function that takes a set of words as input and tries to predict: the following, and thus, if we predict a word many times, then we will get, for example, the next chapter of leo tolstoy, that’s what a specific language model does, that’s all - after all, it can be a neural network not just compile, it can create some new interesting ones, that is, this is a new quality after all, that is, it’s like saying, as if the neural network did not see some object during training, then of course it is unlikely to create it, but with those objects and with those styles that she saw. to train a neural network, i mean, of course it can do various very, very cool things, in my opinion, that is, it can process and, for example, make some objects in the style of some artist, one can argue here, whether it’s creativity or not, yes, these debates are already going on, and quite fiercely, but the very fact that this can be done turns out beautifully, especially lately, this is an indisputable fact in my opinion,

but in fact i can say this , that indeed the field of artificial intelligence... is now being democratized, this term may be a little complicated there, but i will try to explain what this means, that people can now study artificial intelligence from school, say, from the ninth class, sometimes someone there from the seventh, in general they study at a good level, which was impossible to imagine there even 10 years ago, that is, the level of mathematics is really there, the level of interest of schoolchildren, sometimes even the quality of the programs. teaching, which are actually done on artificial intelligence, they allow you to enter this science, and this is really science, it is difficult, this is mathematics, this is the junction of mathematics and the actual calculations of machine learning, you can enter already in the seventh grade, and we hold a huge number of competitions there hackathons, and we see that children in general cope with complex tasks no worse than people who graduated from

college, and what they do, that’s what they do, well, actually the task, the task of a person who deals with artificial intelligence, well... let’s say neural networks, although in reality there are a huge variety of algorithms and types of algorithms, these are not only neural networks, but neural networks are very flexible and therefore they allow you to solve a huge number of problems, even creative ones, that is, the task of a person who wants to create a model artificial intelligence or artificial intelligence, how to choose an algorithm that suits the task, although recently there are universal algorithms that solve a huge number of problems, and how to train it? what is most important, you know, what surprisingly i discovered for myself is that a creative person, if he has some images in his head, has ideas, fantasies, and so on, in order to put these images on paper, well then there was a need to acquire certain skills, he had to go to study, an artist to become a musician and so on, now it turns out that with the help

of artificial intelligence, you will refute this or confirm this, that is, you can transfer these images from your head, so to speak, into a neural network, yeah, it will create a picture for you, which in general will transfer the picture from your head, in general, to the computer, it’s true, it’s possible, of course, yes, and the non-versite that draws pictures, it, for example, can accept text as input, that is, you can text what -write, but the text must to be artistic, it must include all the images, of course, yes, this, this makes the work of the neural network much easier, because it is easier for it, uh, when you have described in detail what you want, it is easier for it to draw it, when you have written, just i want fish in space. for example, i don’t know where this comes from, from her head, she needs to figure it out, but she should be in a spacesuit or not in a spacesuit, there should be a planet in the background or not, it turns out that surprise,

surprise, but in some way i mean, there will be fish in space, of course, it’s not clear what kind, you need to learn to manage artificial intelligence, of course, yes, of course, and there is actually a skill that is especially relevant lately, this is industrial engineering, that is, the creation of these very texts that are loaded into this black box, in particular, yes. although in fact, this is the most common, say, task, although a neural network can create pictures using other types of data, for example, you can draw a sketch and ask to make it realistic, this is also a task for a neural network as a whole, such as kandinsky, for example, the neural network is named after the great russian artist, who came up with the idea that it would be named after the great artist, and you know, we have actually been working on this problem for a long time, about 3 years already, and our first model was called like this: this is rudali, she came up with it herself , in short, and we, we seemed to be thinking about how to animate her, or something, let’s say, uh, make her more recognizable, and we

decided, she was creating abstract objects at that moment, 3 years ago, and we kind of thought , and what kind of russian artists are there who painted in the style of abstract art, generally speaking, the first, the first model was called the smallest malevich, of course, then we made larger versions of the models. uh, this is how the kandinsky model line appeared, there are also neural networks that also , in general, work somewhere in this area, why create different models? there are two reasons, the first is that it’s like there’s no, let’s say the truth in the latter, this is generally machine learning - this is engineering at the intersection of mathematics, and like any other engineering art, it implies that you need to do and come up with some new details, new chips, a little bit from... make changes to existing architectures, thereby making the algorithm the best in terms of the quality of understanding the text, this industrial product, the visual quality of the image, in fact,

when we made kandinsky and are doing it, we are setting up the first such scientific - a technical task: firstly, to create something new, secondly, to improve the current architecture, even western, there is huge scope for improvements there, and kondinsky line 2, here is kandinsky 2.0212, these are three models. they have a unique architecture, we even have there is a set of articles where we talked about this architecture, in general this model is known, well , in fact, all over the world, in, let’s say, the scientific community, in the community of people who use the model, this is the first reason, to move progress, the second reason the point is that if we simply use other people’s neural networks, then we are in some sense limited in what content, how can we influence the creation of content? in fact, what we put into the model, we cannot put into the model, if it is not ours, what we could lay down, if it were ours, roughly speaking, well, you know, yes, that the model learns from data, and we can, for example, make

a model that better generates russian objects, russian entities, it is clear that western networks, neural networks, they they certainly saw it, there is a kremlin there, and some iconic russian objects, and a character, it is still unknown what kind of kremlin, well, that is , what kind of kremlin will be there for exactly that , yes, that is, they do not specifically set the task to do so. so that it generates a model, their task is simply everything in general generate, but we can set the task of showing the model a lot of our data from such a domestic cultural code, and this applies not only to the domestic cultural code, this applies to any types of data that we want to show to the model, and for this, of course, we need to make our own developments, we continue the creative industry podcast, our guest today is dimitrov, the man...

generates, they called kandinsky video, although they probably should have called it tarkovsky somehow, well, they called kandinsky videos, we will continue this line, but what does it do better, knows the domestic cultural code better, of course, is that we actually set a task for the model, and we will certainly develop this further, what is the problem with russian data , knowledge of the russian domain the fact is that these russians... of course there are fewer of them. what does the model learn from? does it learn from data from the internet, that is, from open description image pairs, does it somehow capture the data at the time of the request? no, of course there are teams,

which deals with data. and that is, the data is already contained in it, it is not looking for it on the internet at this moment, no, no, no, in no case, it is creating, in this, as it were, where they are stored, this voluminous amount of a huge amount of data, this is kandinsky the third one studied on one and a half billion batches of text, and you can imagine how many terabytes of data storage is needed there, well, on a supercomputer there is, as it were, all this is studied on a supercomputer and a number is stored. on special storages, they are called s3, but they are like this large, large data storages, which really, for example, so that the dataset model has the opportunity to see, they need to be put there, and not just put, you need to remove all low-quality pairs, that is, filter out and remove everything, if you go to the internet, then in in general, half of the data there will probably not be of very good quality, it’s all rubbish there. does not correspond

to the description that is written down there, or it may be that the picture contains a watermark, yeah, such data is probably also bad use, because the neural network will reproduce this image, it is now being said by artists who are fighting, that is, they are spoiling their image so that it doesn’t exist, that’s also true, we also need to fight this, that is, there are a huge number of filters, which need to be written before giving the data to the model, because everything needs to be given in a row. of course it’s impossible, it’s just not clear what he will learn in the end, of course, the data development team is responsible for this, you know, show the child everything in a row, and this is a super important activity, and a team in itself, because not only half, more than half of all success depends on it, so after the data is stored, here it is in this c3 storage, in fact, you can teach the model ,

uh-huh, training... you have some person, he thinks, well , denis says it’s interesting, you have to try, then the internal struggle begins, to allow your life, this artificial intelligence, not to allow it, these are the areas that we already actively use artificial intellect, who has already grabbed for this, first of all, the designer, we have a special tool called fusionbrain.ai and there it is such a developing photo editor, now in general

it has been very popular lately to make a photo editor based on. or rethink current photo editors such as photoshop, integrating ai there, why is this useful, because it is clear that creating pictures from text and editing pictures is much, much easier than actually drawing yourself, which is what people always did before, it took a huge amount of time, it's quite simple now write in text, a picture will be generated, you don’t like some area, you cover it up, add whatever you want, it really simplifies the work of designers, and very much so, that is, the possibility is not about...

in general, there is a routine part of this work, it is necessary as quickly as possible in my opinion, to automate as well as possible, this will only save people’s time, in fact it will increase the quality of the content, and animation, for example, how many routines are there, animation is not about creating videos at all - they actually make films there for how many years, well, it’s been happening for years, that is, now this may be happening faster and faster, but that is, since , well, until the creation of full-fledged films , the technologies have not matured to be honest. but the progress is such that if you have been saying for 3 years that we have been working on neural networks for 3 years, i mean creating neural networks that generate pictures from text, so if you look at how neurose, the best neural network was created 3 years ago image, then you will see that it is very, very bad, let's watch the video, and then i will ask my questions, what can you do? tell comment.

that is, how does animation differ from video? because animation is a camera flying around a static object, or some kind of camera movement, and video is a full-fledged movement of everything, here it is still animation, because there is a first frame, and then we choose the camera movement, well then there is a monotonous movement, zoomin is such a zoomin, and we select the first one, draw the first picture according to the promtumin, and we can, without creating a new model based only on models that generate an image from the text, generate animations like this, let's... tell you where to look, yes there is actually a telegram bot within which you can create animations like this, that is, you can basically start, everyone can do it, yes, that is, it copes with complex tasks well, that is

, a telegram bot - the current task was , in my opinion, very difficult, i honestly don’t remember what kind of industrial processes were there exactly, that is, you don’t need a lot of memory, it’s all created in the cloud somewhere, yes it’s all created on a supercomputer oh supercomputer - this an interesting topic is the training of the model itself and the application of the model, well, it’s called inference, it requires computing power, the model is huge after all, on an iphone, roughly speaking, it will not run on your smartphone, although progress is also being made towards this, so that it runs on a smartphone, well, at the moment , kandinsky is running on the supercomputer cristofar, this is our supercomputer on which we teach models, both language and these generative ones, which turn text into pictures and videos. well, that’s where we actually use them, when you make a request, it flies to a supercomputer, it is processed there, it is there that kanzinsky seems to live, yeah, it is there that he makes a picture from the text, the picture is returned back to you,

what do you use a telegram bot or - well, in a bot - this is a certain number of iterations, then there, i go into the bot, talk about what kind of picture i want to create, the picture comes, then the next step is i say what i want to see for the movement, and it gives me an animat. a lot of resources so that the queue is still satisfied, so that there is not, say, too much in the queue at the same time people, i wonder why, do you feel that your competitors are spying on you, we feel it, that is, you

feel this attention on yourself, yes, and since they are spying on you, it means you are doing something that apparently causes them , i don’t know, the fear seems to me that with everyone who is engaged in creativity and creates, competitors are spying on something, to be honest, and this is not... artificial intelligence is no exception, of course, but it’s just how much you are ahead of your competitors, this is also a competitive struggle, relatively speaking, you read the news in the morning there in the evening in any case, and you are probably also noticing this to yourself, but did anything happen that would affect who passed whom, so to speak, at the turn, and this is, of course, yes, as it were, let’s say, the area of ​​​​creation large models. although it requires a huge number of computing resources, data and specialists who will actually train all this, but in principle, large bitechs certainly have such teams that are engaged in the development of these models, for

example, yandex has a masterpiece, and the guys do enough similar technology the only thing is that they don’t have video generation, but today we only watched animation, it’s not a video, after all, a full-fledged video is the movement of everything i have... i have a question that concerns copyrights, intellectual property, it’s clear that leo tolstoy, roman said at the beginning, it’s all about leo tolstoy, it’s as if he didn’t pass on an inheritance to anyone, he’s open history, yes, when you load data into a supercomputer, you license it, or it’s like the public domain, or it is not necessary, well, they’re not exactly in the public domain, that is, they’re more likely in the public domain on the internet, yeah, but... when we downloaded them , it’s like we didn’t post their entire dataset as a daddy, we don’t post it anywhere, of course, yeah, it’s stored there as a kind of database as a kind of base and why, actually, we can’t do it any other way because the models need to be shown at

the training stage and the pictures really need to be shown, well, that is, it’s like a truly educational material for a related question, this is this 28. the first picture, who is it all is the author, artificial intelligence or the person who ultimately generated all this, wow, this is a very good question, actually, who owns the copyright? and it’s all being developed, including at the legislative level, and let’s just say, at some point this will probably be enshrined, although this is a very difficult question, you know, as they used to say, there’s an angel on my shoulder, an angel sat and he told me suggested, when someone is the creator of something, what is the difficulty in actually determining who the author is? there is data of course, behind every photograph there is some author who took this photograph, that is, there are auto photographs, another thing is, when there are 1.5 billion of them, then it’s as if these

authors are all somehow mixed, and most likely the author is the author of one and a half billion there is no photograph, that is, this is a huge number of people, there are millions of people, on the other hand, there is a company that purchased a supercomputer with its own money, this is a very super expensive undertaking, in fact, all the calculations took place on this supercomputer, that is, it can claim ownership, on the other hand there are developers. which the people who were sitting, sort of converted their knowledge into this non-network, that is, they can say that in fact they have the right to the picture, on the other hand there is a person, one person who sits and promptly writes in the end , and he can say, in fact, i am the author, that is, there are four parties, maybe even more, who, in general, well, had some participation in the creation of the final picture, they made their own contribution. contribution, but for now

the story is that if a person, most likely, if a person, well, logically, each company does it in its own way in the end, if a person pays for using the neural network and receiving a picture, then he is still the author , this is most likely the case, and if motives are guessed there , usually, if something is created, a license is issued anyway, i just want to say that such disputes arise only when the picture begins to bring in huge... ours is one of the first our missions, i already spoke about this, to still move the community

in general and move science forward, then we put the model in the public domain, you can use the model absolutely not without paying some company and deploy it on your own, and why create such a model that is not widely available, is it so that some then a limited circle of people used it, well, not to sell generation. we continue the creative industry podcast, our guest today is the man who created the kandinsky neural network, denis dimitrov and elena kiper and roman karmanov are still with you, i’m wondering if here i am generated a masterpiece, which means that when i create it, no, i must always mark it as a masterpiece, so you will create something good, and then, is it possible to read from the data of this masterpiece, well...

even tell it on your fingers, imagine how a person creates a picture , any, any image, any painting, he has some kind of white canvas, but not necessarily white, just white paper, some kind of empty, napkin, napkin, yes, there is a tablet screen, he has some kind of idea , of course, he starts sketching out some details, with a pen, a stylus, whatever, adds details, maybe removes something , well, a person also has patterns in his head, that is, he remembers something, he has it, he

begins to draw a picture from the so-called white noise, what is it, well, that’s when there is no signal on the tv, there is such interference, it’s just some kind of matrix, naturally there is your request, this is the same idea, as if in a person it is expressed by these signals of neurons, in neurons it is expressed simply by numbers that are encoded and... from this matrix step by step in the process of removing noise in the right direction creates an image, that is, creating images from this white one. this white noise from this matrix can be simply visualized, that is, there are 100 steps or a thousand steps during which from any of this noise the final picture is created and visualized, so to speak, appears, for example, you want to make a pug in space, yeah, you always have white noise at the beginning, and step by step you get a pug in space at the end of this

process, if the noise is a little different, it’s clear that the noise can be whatever you want, that is, change it there are two pixels in some places, there will be another, but also a mob in space, this matrix, in a sense, it is responsible for the model’s imagination, this can be visualized, but this process, it’s called denoising or denoising, that’s how since for it to happen correctly, we need to show this neural city 1.5 billion pairs of pictures and text, that is, so that it learns to connect text and image, in principle , there is an analogy with a person in this sense, and you do all this every day, and we all this well... the neural network does all this every day , we do this, and we are improving the neural network every day so that it is better. this process, let's talk about the horrors of the future, this is just from what you just said, is it possible to assume that in the future, a person will do without this intermediary,

right away from your head to generate your own images, well, let’s say on some kind of canvas, i don’t know, well , i mean... in fact, there is research, i wouldn’t say that this is just a super big practice there and what is it used there actively, but there are studies that make it possible to decode brain signals, that is, you know, they put there, that means, a reader of electrical signals in the head and, for example, it is possible there to, without words from you, roughly speaking, understand what you want, or you close your eyes, imagine it as something, i don’t know, for example, some kind of palm tree, maybe.

but the limitation of these models is that for each person, for now, you still need to learn your own model, that is, for each person, when he imagines a palm tree, the signals are slightly different, that is, like one it is not yet possible to train a model for all people, and it is expensive to train a model for each person, and you need a supercomputer for each supercomputer, until there is some kind of universal thing, which i hope, i hope not soon, and moreover, this it will still remain. they can’t, but if you know how to write, then you won’t attach electrodes to your head, you will, it will be easier for you to write, were there moments when you realized that the neural network that you created is doing something that you didn’t

predict whether the second thing to think about is in order to limit it somehow, generally speaking. in fact, all the generative models that are being created so far, and those that will be created, they tend to hallucinate at some point, this is the official term, this means that if the neural network does not know something, and you she was asked to create it anyway, she can’t create it, let’s put it this way, that is, she creates something, well , not what you want and something wrong, but she creates it anyway, this is probably best seen s and how you can trace it. she in training data she saw, as it were, the word fluff with the word light, she saw it much more often than fluff with the word heavy, well, on the contrary, which is logical, and of course she answers that...