to the north, how to go along them to the east, we found these climbs, how you could roll, drag, yes, drag, drag, how you will roll it right here on logs, like yes, the yamal drag will have to be done , there is a long distance there, no, not huge, there is about a kilometer there, less than a kilometer, but 700 m, but there you will have to drag carbs from the opposite flow of the river for 300 km, here i am just me.

category of ship, it is called a ship underway, it is a ship that is not aground, not at anchor, and not moored to the shore, as soon as you have given it all, all this rigid connection with the earth, further, further, if you don’t have sails and didn’t sail, there’s nothing, you’re already on the move, the captain is already responsible for everything that happens, you’re already somewhere- then you sail, even if you are not in complete control of it, this example just shows how important it is... it is important how a sailor feels when he entered some bay, laid anchor, and looked that he was not being carried anywhere , he said, wow, that’s it, here he can rest, he landed his plane, it’s not for nothing that sea ships are taken, so many of them were taken, so much in accordance with aircraft, because in both cases it is a flight, and in both cases it is most difficult to maneuver during takeoff and landing. because the plane, which he

needs to land, he must slow down, it becomes poorly controlled, the same thing happens with the boat, when it slows down, some kind of breeze blew, the side carried this ship, no one knows where, but here he it’s impossible to maneuver at high speeds, the fact that it will stick somewhere, that’s, of course, a lot unpredictable, however, is that now modern technology, mainly space, gives... to drivers, we see where we are, uh, on the map, which no navigator could ever imagine, so that he knows 100%, where it is, it was probably just bad taste, to say so that i know that i am definitely here, you can find out the weather and choose good weather, and of course there is communication, but with us all this is compensated by the fact that we let's go, we are generally amateurs, most of the people, the crew members, are people who are they will sit down at the oar and learn to hoist a sail.

they will go to sea, especially the kara sea there, this is space, consider all this as the polar region, and therefore we somehow have a chance somewhere close to those people, when once those pioneers, they had enormous experience , they have been practicing this move at sea since childhood, we have modern equipment, but no experience, somewhere we will be next to them, what must happen for you to say. was successful, well, i think that even if we arrive on time, it will already be a great success, even better if we manage to meet this schedule on time. from historical references we know that it took 2 months from arkhangelsk to mangasia, and some even faster, i can’t even imagine how this is possible, it will be great if we don’t turn on the engine, if we don’t ask to be towed, i really hope so , that we will be able to withstand this, and this, in fact, will be - from the point of view

of research, it will be very good, it will be very good if we make a film, now if we make a good film about this, which will tell you how it all happened, specifically, how tired they were there, how they rowed, how they dragged themselves in reality, what they encountered, this of course will be the best, probably, because the task of the project is not just to go for yourself, this is not tourism , well, not even to tell.

maybe it’s to think about it, make films and so on, our super task is to make a good film, where everything will be like where, how we go, how we set the sails, how we... it’s stormy, how we wait for the weather there or vice versa we go, we take risks, that's if we can do it all show, it will be luck, and what should captains and crews wish for before such an expedition, if there is some kind of ritual, 7 feet under the keel, tail wind under the keel, and the weather, the weather, yes, everything depends on the weather, then we will wish this unique expedition from arkhangelsk mangazeya a fair wind of 7 feet under the keel, good weather, generally all sorts of luck, because it obviously will not interfere in such a difficult matter. thank you. you had a podcast with schrödinger's cat, and we talked about an expedition on a wooden carbass from arkhangelsk to siberia. and ours is with us

captain, captain of this expedition, leader, evgeniy shkoruba. you can find all episodes of the schlödinger's cat podcast on the website of the first channel 1tv.ru. 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, but i won’t touch it yet, i need to make sure, come on. roman karmanov, general director of the presidential fund for cultural initiatives, is our guest, dimitrov, our guest today, managing director for data research of the largest bank, i’m very glad to be here, hello , thank you for coming, we didn’t just

start out like that, in such an ornate way, because artificial intelligence is really a topic that everyone hears about, but very few people understand what it is, that’s why everyone is afraid already, they are afraid. in general, something, but what did they know, what it is, we know what it really is, we know a little, we use, remotely, what artificial intelligence is, let’s first at least figure out what this thing is , actually artificial intelligence, there is absolutely no need to be afraid of it, since we started with this, it is a tool that helps us do different things, that is , it automates some part of our intellectual work, mostly the routine part, 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 gives something as an output, that’s exactly how we imagine it, like a black one, what’s inside the black box is a certain a mathematical function, a function that

now, for example, lies behind, for example, generating a picture from a text, has, well, there are billions of parameters, so they are so complex that a simple person must pass? well, relatively speaking, you take all the volumes of leo tolstoy, load them into this black box, and it can give you, well, relatively speaking, another volume of something like leo tolstoy, for example, well, in a sense, yes, that is, this is the function it learns in this way, that is, our parameters are configured in such a way, these billions based on motives, as it were, so that

is configured, what does this mean, this means that - this is it, that is , the learning process takes place, if we are talking about language models, such as, for example, the well- known gpt chat, gigachat, for example, then this is a function that the input takes a set of words and tries to predict the next one, and thus, if we predict a word many times, we will get, for example, the next chapter of leo tolstoy, this is what the language model specifically does, after all, neurosates it can not only compile, it can create some new interesting ones, that is, this is all a new quality.

it can turn 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. such a term, maybe it’s a little complicated, but i’ll try to explain what it means, that people can now study artificial intelligence starting from school, say, from the ninth grade, sometimes someone from the seventh grade, in general they’re studying on a good level, which was impossible to imagine there even 10 years ago, that is , there really is a level of mathematics, the level of interest of schoolchildren, sometimes even the quality of training programs that are actually... done on artificial intelligence, they allow you to get into this science, and this is really science , this is difficult, this is

mathematics, this is the intersection of mathematics and the actual computation of machine learning, enter already in the seventh grade, and we hold a huge number of competitions, hackathons, and we see that children in general have complex they cope with tasks no worse than people who graduated from college, but what they do, that’s what they do, well, actually the task, the task of a person who deals with artificial intelligence, let’s say neural networks, although in fact... 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 an artificial intelligence model or artificial intelligence is 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 the most important thing, do you know? i discovered what was surprising for myself: 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, that is, he needed to acquire certain skills, he had to go to study, become an artist, become a musician, and so on, now it turns out that with the help of artificial intelligence you will prove this to me 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 will generally transfer the picture from your head, in general, in a computer, it’s true, it’s possible, of course, yes, and the non-irase that draws pictures, it, for example, can accept text as input, that is, you can write something in text, but the text must be artistic, he must include all the images, you need to really understand in detail what you want in this picture. makes the work of the neural network much easier, because it’s easier for her, uh, when you

described in detail what you want, it’s easier for her to draw it, when you wrote i just want a fish in space, for example, i don’t know where this came from in my head, she needs to figure it out, should be on a spacesuit or not in a spacesuit, there should be a planet in the background or not, it turns out that surprise, surprise, yes, in a sense, there will be a fish in space, of course, it’s not clear which one, you need to learn to manage artificial intelligence, of course, right? of course, there is actually a skill that has been 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 based on 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 in in general, such as kandinsky, for example, neurality is named after the great russian artist, yes, who came up with what?

art means that you need to do and come up with some new details, new features, make a little changes to existing architectures, thereby making the algorithm better, in terms of the quality of understanding the text, this industrial product, the visual quality of the image, in fact, when we did kandinsky and do it, we set the first scientific and technical task, firstly to create something new, secondly improve. rich architecture, even western, there is huge scope for improvement, and kandinsky - there are 2 lines, here is kandinsky 202122, these are three models, they have a unique architecture.

their task is simply to generate everything in general, but we can set the task of showing the model a lot of our data of 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, you need to make your own developments, we continue the creative

industry podcast, our guest today is denis dimitrov, the man who created the neural network.

what does the model learn from, it learns from data from the internet, that is, from open pairs of description images, it seems to capture the data at the moment, naturally, there is a team that deals with the data, and that is, the data is already contained in it, it does not search in this moment on the internet, no, no, no, in no case, it creates, in this it’s kind of unique where they are stored, this voluminous amount of huge amount of data, this kandinsky the third studied on one and a half billion pairs of pictures and text, and can you imagine how many? there - a terabyte of data and storage is needed, well, on a supercomputer there is , as it were, all this is learned on a supercomputer and stored nearby on special storage facilities, they are called s3, but these are such large, large data storage facilities that are really for, say, a dataset model i had the opportunity to see that they need

to be put there, and not just put, all the low-quality pairs need to be removed, that is, what does this mean.

success, so after the data is stored, here it is in this c3 storage, you can actually train the model, yeah, training the model is just this kind of iterative viewing of this data and changing this huge number of parameters that i talked about, here are these billions of parameters that need to be adjusted so that when a new description comes to the input, the picture would correspond to it, we have a video that was just created. well, denis says it’s interesting, you have to try it, then the internal the struggle begins, to allow this artificial intelligence into your life, not to allow it, these are the areas that are already actively

using artificial intelligence, who have already grabbed onto this first of all, designers, we have a special tool called fusionbrain.ai and there it is such a developing photo editor, now in general it’s been very popular lately to make a photo editor based on or rethink... here are the current photo editors like photoshop, integrating ai there, why is this useful? because it is clear that in text, creating pictures and editing pictures is much, much easier than actually drawing yourself, something that people always did before took a huge amount of time, now it’s enough just to write in text, a picture will be generated, you don’t like some area, you’ve covered it up , finished drawing what you want, it really simplifies the work of designers, and very much so, that is, the ability to not just generate an image. also edit it with the same text, yeah, and moreover, draw more than one picture, and there a whole, a whole collage of these pictures or a whole panel, that is, you have created a gift for designers, is it necessary, is it necessary now,

some designers are afraid of this, that we will kind of automate their work, in fact it’s not like that at all, because artificial intelligence is this is a tool that allows you to simply optimize work; design exists in any profession, no exception at all. part of this work, it needs to be automated as quickly as possible in my opinion, as efficiently as possible, this will only save people’s time, in fact, there is quality content will increase, and animation, for example, how many routines are there, animation is generally the creation of videos - they actually make a lot of films there, but they can shoot something there for years, now it can happen faster and faster, that is, from well until the creation of full-length films, everything- to be honest, technology has not matured, but it is progress. such that, if you are 3 years old, we said that for 3 years we have been working on neural networks, yes, i mean creating a neural network that generates pictures from text, so if you look at how neural networks are, the best

the neural network created an image 3 years ago, then you will see that it is very, very bad, let’s watch the video, let’s watch the video, and then i’ll ask my questions, yeah, what can you tell me about this, i’ll comment, it’s still an animation, it’s not is the video complete? kandinsky did this? yes, kandinsky did it, truly, as if, how he does it, he draws the first frame from the text, and then you choose, that is, how animation differs from video, in that animation is the camera flying around a static object, or some kind of movement cameras, and video is a full-fledged movement of everything, here there is still animation, because there is a first frame, and then we select the camera movement, well, that is, the monotonous movement of the camera, such a zoom. we select the first one, draw the first picture according to promtuzumin , and without creating a new model , based only on the model that generates the image from the text, we can generate animations like this. let's tell you where to watch

, yes there is actually a telegram bot within which you can create animations like this, that is, everyone can to do this, well, that is, it copes with complex tasks, well, that is, the telegram bot is listed here, the current task was ... this is our supercomputer, on which we teach both language models and these generative ones, which turn text into a picture and video,

well, that’s where we actually use them, when you make a request, it flies to a supercomputer, it’s processed there, it’s there that kanzinsky seems to live, yeah, that’s where he makes a picture from the text, the picture is returned back to you, there than you use a telegram bot or. see for the movement, and it gives me an animation, i stand in line and wait for it all to happen, because millions of people in supercomputers hop, that’s right, yes, but we actually have scaling methods, after all, a supercomputer contains more than one computer , and a lot, of course, if a million people come, there will be a queue, this is inevitable. in this sense, how to do inference on a device, that is, a smartphone, is much better, because you don’t stand in line, you do it there, but we are trying to a supercomputer holds a lot of resources, so

that... of course, but just how far you are ahead of your competitors, this is a competitive struggle, relatively speaking, you read the news in the morning there in the evening in any case, and you probably also note this to yourself, but it didn’t happen is there anything that will affect who passed whom, so to speak, at the turn? so it is, of course, yes, like, say, the area of ​​​​creating large models, although it requires

a huge amount. number of computing resources, data, and specialists who actually they will teach all this, but in principle, large bitechs certainly have such teams , they are engaged in the development of these models, for example, yandex has a masterpiece, and the guys are making a fairly similar technology, the only thing is that they do not have video generation, but we are today we only watched animation, this is not a video, after all, a full-fledged video is the movement of everything, i have a question ... which concerns copyrights, intellectual property, it’s clear that leo tolstoy, roman said at the beginning, everything is there by leo tolstoy, it’s as if he didn’t pass it on as an inheritance to anyone, he’s open history, yes, when you load data into a supercomputer, you license it or they’re kind of in the public domain, or it’s not necessary, well, they’re not really public domain, that is, they are on the internet rather all public domain, yeah, when we uploaded them

after all. who is the author, artificial intelligence or the person who ultimately generated all this, oh, this, this is a very good question, actually, who owns copyright, and, as it were, at the legislative level, this is all being developed and let’s say, at some point this will probably be enshrined, although this is a very difficult question, they just said, there’s an angel on my shoulder, an angel sat and he

told me that when someone’s creators... have data, of course, behind each photograph there is some kind of author who took this photograph, that is, there is an author of the photograph, it’s another matter when there are one and a half billion of them, then it’s like these authors are all somehow mixed up, and most likely the author is the author of one and a half billion there are no photographs, 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, it is very super expensive...

but so far 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 neurose and receiving a picture, then he is still the author, that’s most likely the case. and if motives are guessed there from motives , usually, if something is created, it’s all the same a license is issued, i just want to say that such disputes arise only when the picture begins to bring huge money to the author, and then a large number of other people arise who claim authorship, but there is actually still a point that the neural network itself can be , as if companies, developers, and so on can claim its authorship,

the model itself has a license. for example, kandinsky and this also depends on, that is, a masterpiece is a model that is not laid out and is not in the public domain, kandinsky is the opposite story, since ours is one of our first. i already talked about this, after all , to move the community as a whole and move science forward, then we put the model in the public domain, you can use the model, absolutely not without paying any company, and deploy it on your iphone, but why create such a model that is not widely available, so that some limited circle of people can use it, which, well, not to sell generation, we are talking about...

just imagine how a person creates a picture, any, any image, any painting, he has some kind of white canvas, well , not necessarily white, just white paper, some kind of empty, napkin, napkin, yes, there is a tablet screen, he definitely has some kind of idea, he starts sketching out some

details, with a pen, a stylus, whatever, he adds details, maybe removes something, well, a person also has patterns in his head, that is, he remembered that he has patterns, this is after training, here’s step by step stepo. he simply creates a picture, like the kandinsky model creates, in principle , an analogy can be drawn, he begins to draw a picture from the so -called white noise, what is it, well , this is when there is no signal on the tv, there is such interference, just some kind of matrix, of course 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, which are the text in which the text is encoded. and from this matrix, step by step in the process of removing noise an image is created in the desired direction, that is, the creation of images from this white, this white noise, from this matrix you can simply visualize, that is, there are 100 steps or a thousand steps, during which any

of this noise is created and visualized, manifested, you can so to speak, the final picture, for example, you want to make a pug in space, yeah, so you always have white noise at the beginning and step by... step by step you get a pug in space at the end of this process, if the noise is a little different, of course, what noise can be like whatever, that is, swap two pixels there, there will be another, but also a mob in space, this matrix is ​​in some sense responsible for the fantasy of the model, this can be visualized, but this process, it’s called denoising or something like that de-noise, just so that it happens correctly, we need to show this one and a half billion park.

do without this intermediary, generate your own images straight from your head, let’s say on some kind of canvas, i don’t know, well, in the digital sense, of course, really there is research, i wouldn’t say that this is a super big practice there and that it is actively used there, but there are studies that make it possible to decode brain signals, that is, you know, they put a reader of electrical signals in the head and, for example, is it possible there, so that, without words from you, roughly speaking, you can understand what you want, or you close your eyes, imagine it as something, i don’t know, for example, some kind of palm tree there, and it is possible purely from brain signals at the moment how you imagine decoding the picture.