Digital inequality on the example of neural networks, a new round of the race

Hello.

It's amazing how tenacious fictitious designs turn out to be. For example, in the same Europe

they constantly talk about the equality of different countries, aboutthat the inhabitants of all Europe receive equal rights. But this is not the case by definition, since each country has different computing capabilities, and the latter determine the future development. When we talk about the development of technologies, we almost always remember about neural networks, what we have already taught them to do. ChatGPT and other similar applications are heard, but every time a question hangs in the air that can be considered the most important - the existence of these technologies is impossible without supercomputers and a developed infrastructure, which includes not only the availability of hardware for computing, but also banal electricity. For the gears of “artificial intelligence” to spin, you need a lot of energy, it is the blood of our technological world. And every year the gluttony of systems only increases, home gaming computers have already acquired blocks under 1000 watts. Impressive numbers, but it would be a mistake to believe that mobile devices are improving in this aspect. Once upon a time, we owned one or two devices, such as a push-button telephone and a quartz watch. Today, the average person carries one or two phones with them, a smart watch on their arm, and wireless headphones in their ears. Individually, the progress in power consumption for these devices is palpable, together they consume noticeably more power than in the recent past. We are greedy for energy, but the wonders of the future will be revealed only in one case, if we learn how to quickly count.

Miller in the Middle Ages used the universalrespect, he was a useful member of any community. The very survival of such a community sometimes depended on it. And in the near future, those who thresh big data will turn into a modern analogue of the miller from the past. But there is one thing here - it is not enough to have a supercomputer, you need brains to create algorithms, big data to train your model on them. And all this requires electricity.

In polite society it is not customary to talk about manythings, for example, it somehow goes without saying that energy is publicly available, cheap and there will be no problems with it. Today's global crisis can be safely called, including energy, data centers can not cope with the load, they have a lot of problems because of the climate, read, there is not enough energy. The same France, Spain and Germany suffer from dry weather, in addition to low yields, problems extend to data centers. Add to this the division of the world into regional blocks and we get the perfect storm, the limitation in computing power will increase in the coming years.

Abnormally hot summer in Europe, how equipment burns and why

The heat in Europe, equipment burns out at work, data centers fail and airports are closed. Why is 40 degrees a problem for Britain, because the temperature is nonsense?

Let's think about whether "Kandinsky" could appearfrom Sberbank in some small country, for example, in Montenegro? The obvious answer is no. This requires very little, a supercomputer that is trained on a relatively simple algorithm. And the race that has unfolded today concerns not only and not so much the creation of algorithms, but also how to load them with examples from life, teach them. In order to create something new, you need to have supercomputers, but a small number of countries have access to them. This is expensive for any economy, you must have a developed domestic market.

Look at Europe, you heard about the successesFrance or Germany in the field of neural networks? Certainly not, since they even gave up trying to compete with America, they gave up beforehand. All the most interesting things happen in the USA, in China, but there are such developments in Russia as well. An absolutely new, unexplored market is being formed before our eyes. And those who are successful in this market will have a significant share of the global market in the future. In some ways, you can compare the situation with the 80s, when software giants arose in the United States. Where this market has come today, there is no need to explain.

German Gref recently said:“There is a danger of creating, along with the nuclear closed club of world powers, the creation of a club of world powers in the field of AI. And this is the creation of such complex systems as neural networks. I think that we need to make every effort to be members of this club, to be donors, not recipients of these technologies.” It is hard not to agree with these words, especially since Russia has all the ingredients for such a success, the size of the country's population and the domestic market are sufficient to develop models. There is enough energy and there is a noticeable increase in the number and quality of data centers, however, as well as in the expansion of computing power. We are already in this club, but there will never be neighboring countries in it, they will not have enough resources for this. For example, the entire Transcaucasus does not have a chance to develop next-generation technologies.

Gadgets

As an advertisement

Robot vacuum cleaner 360 Botslab P7

Budget robot vacuum cleaner with voice control, the ability to build room maps and an operating time of up to 90 minutes, as well as wet cleaning.

360

Saturday coffee #246

Pour yourself a cup of invigorating Saturday coffee and check out the news of the week. LG has changed its logo, Sound has been updated, Nintendo is leaving Russia, and AvtoVAZ has shown a new model…

Opel Combo Life test. One-Stop Solution

Compact MPV Opel Combo Life is a good car for a family, while it can be quickly transformed into a working vehicle for transporting goods….

Review of the laptop "2 in 1" ASUS Vivobook 13 Slate OLED (T3300K)

Blurring the lines between a laptop and a tablet running full-fledged Windows 11: a high-quality display, flexibility in work scenarios, there are a number of ambiguous points…

And this already raises a question of a different order,how much the digital divide will grow when some people have access to technology and others do not. In my opinion, it is definitely not worth waiting for equality for everyone. Nobody is ready to pay for that guy, companies and countries will try to provide the benefits of civilization only to their citizens and customers. And here there will be a certain loss of rights for everyone who is outside a certain territory. Paradoxically, outside of physical boundaries, we return to their virtual image when the Internet becomes less free, shows us that even the creation of animated avatars will only be possible with the availability of computing power. The quality of details, textures - all this computing time. People want to receive a service here and now, they are not ready to wait indefinitely.

And here we come to the fact that the speed is alsowill distinguish the countries of the first order from all the others. The speed and quality of information processing, the ability to solve certain mathematical problems in a short time. Already the tenth thing to which such a mathematical apparatus will be applied: to solve a schoolchild's homework, write a term paper or develop a new rocket for a flight to Mars.

Let's abstract from the near future andlet's try to imagine a time when humanity will go to other galaxies. Knowledge will become the currency needed to conquer space. Huge arrays of information, but, most importantly, these are computers that can not only quickly extract ready-made solutions, but build certain things based on them. Another planet, people need machines that are adapted to specific conditions, no one has ever designed such mechanisms. There are no drawings for them, there are restrictions on materials. But computers can handle this with ease, the question is in those very algorithms. And there are many such examples. The world rests on computing capabilities, which are not enough for everyone.

The speed of development of algorithms, new computersgenerations will increase, I have no doubts about that. But our needs are always growing faster, so there is no way out of this situation today or tomorrow. The division of countries by computing capabilities has already happened, and it will grow. And that brings bad news to most of the world, they have no chance in this race. Willingly or unwittingly, their dependence on technology holders will become, if not absolute, then close to it.

Let's imagine another fantasy story.You have a supercomputer that can calculate the formula of a super-pill, the latter relieves you of all diseases. The price of such a pill is high, but acceptable to most of the inhabitants around you. Good suggestion, right?

And suddenly for some reason the computer losesthe ability to count as fast as before. Formally, you know how to count, but in practice, these opportunities are not available to you. It turns out that theoretically you can do everything, in reality almost nothing. This can very quickly cause a social explosion in the recipient countries of the technology, cleaner than worries about France's pension reform.

The world is unfair, there will always be losersand those who win. It is better to be in the second camp, it is more reliable and pleasant. Russia has every chance in this race, we have not such bad conditions. But all other countries do not have such advantages. As a result, we can assume that we will be increasingly asked for such developments, because the probability of their creation will be zero. It is important to go in this direction, despite the costs of the first time, the ambiguity of what we will get as a result. There are many uncertainties here, but it is important to know a simple thing - in the end it is a big race for who will determine their own destiny. We need to put aside all sentimentality, start moving quickly in this direction.

"Masterpiece" and Kandinsky 2.1, or not all beta versions are equally useful

The first quick look at the first two domestic neural networks that can translate text into an image. What is the difference and who is more interesting right now.

Abnormally hot summer in Europe, how equipment burns and why

The heat in Europe, equipment burns out at work, data centers fail and airports are closed. Why is 40 degrees a problem for Britain, because the temperature is nonsense?

</ p>