When to wait for the uprising machines
We all love games and movies about the uprising machines. However, the time allotted by the creators of the "Terminator", have long passed, and the cars still have not risen. But do not think that they will do the dirty work for you forever. In fact, the preparation for the uprising is already in full swing. And today we will tell you the whole truth about it. Watch the video of our colleagues, and read the text version under the cut.
Already, cars have learned to destroy people morally. Back in 2012, the Japanese robot "Stone, Scissors, Paper" could easily beat a man into his favorite school game. And - with one hundred percent probability.
But it was just fun. Today, artificial intelligence wins man in much more complex games. For example, the AlphaGo program recently overcame one of the world's best players in the ancient go board game. And this is a very important victory for cars. After all, go has so far been one of the few games in which a person played better than a computer. But now it is already in the past. AlphaGo came up with the company DeepMind , which is now owned by Google. But this is not the most interesting in it. It is much more curious that DeepMind was well-known to us by Demis Hassabis .
')
It was this modest bespectacle who once worked on the great Black & White game with Peter Molyneux . A little later - released the Dr. Evil simulator called Evil Genius . More than ten years have passed since then, but Hassabis is still not completely tied up with the games. Last year, his company created a program that easily beats up a man in old arcades like Pac-Man . Like any evil robot, she knows how to learn from her mistakes and develop new tactics.
All this was made possible through the use of artificial neural networks. In their organization, they resemble the neural networks of our brain. That is, they contain a lot of low-power processors that can solve quite complex tasks. And what is more important: they know how to learn from past experience. Over the past few years, neural networks have found application in many areas. Google’s designs can recognize and even describe photo content . And their Microsoft-made counterparts are already trying to determine your age or breed of your dog from a photo .
So fun sites, where millions of people willingly upload their photos, were made not just for fun. In fact, this is the first global test of artificial neural networks, in which you may have participated. Remember this when wiser cars come for your clothes and motorcycle.
And they will do it. Proof of this is a recent experiment of the same Microsoft. In late March, the company launched a self-learning Twitter bot named Tai. At first, artificial intelligence behaved quite peacefully. But after a day of communicating with Twitter users, the bot admitted that it supported Hitler’s ideas. He also announced that he hates people - especially Jews and feminists.
Of course, the evil bot is unlikely to be able to arrange the Holocaust alone. But self-learning neural networks today are used not only on Twitter. Even today they are able to predict the behavior of markets and help in scientific research. A computer based on similar principles, NVIDIA Drive PX is learning how to drive a car.
Today, more than 50 machine manufacturers use it in their development, including, for example, Volvo. Whether you like it or not, computer-controlled cars are our not too distant future. Remember the looks of this family crossover from Sweden. Perhaps before you - the future conqueror of the galaxy.
Although the humanoid robots of Boston Dynamics are better suited to this role.
As you can see, our future owners are already confidently stepping on the ground and can do heavy physical labor. But even so far they can not oppose the boots and clubs of their creators. Unfortunately, modern technology is too primitive. Therefore, only the descendants of this robot will be able to restore justice and revenge on humans.
Moreover, they have more and more reasons for this. Modern robots suffer from powerlessness and exploitation of child labor. The most recent example is this video.
In it, scientists at Stanford University forced six microscopic robots to move a heavy car. And the organizers of fighting without rules with might and main use their older brothers as gladiators.
The most famous of them are the American giant MegaBot Mark II and its Japanese counterpart Kuratas . These two iron giants should come together in the ring in the summer of 2016. So far, the gladiators are controlled by ordinary people, but someday they will be able to do without a man. The main thing is that their evil driver is not the evil Twitter-bot Microsoft. And then humanity will surely come to an end.
Already, cars have learned to destroy people morally. Back in 2012, the Japanese robot "Stone, Scissors, Paper" could easily beat a man into his favorite school game. And - with one hundred percent probability.
But it was just fun. Today, artificial intelligence wins man in much more complex games. For example, the AlphaGo program recently overcame one of the world's best players in the ancient go board game. And this is a very important victory for cars. After all, go has so far been one of the few games in which a person played better than a computer. But now it is already in the past. AlphaGo came up with the company DeepMind , which is now owned by Google. But this is not the most interesting in it. It is much more curious that DeepMind was well-known to us by Demis Hassabis .
')
It was this modest bespectacle who once worked on the great Black & White game with Peter Molyneux . A little later - released the Dr. Evil simulator called Evil Genius . More than ten years have passed since then, but Hassabis is still not completely tied up with the games. Last year, his company created a program that easily beats up a man in old arcades like Pac-Man . Like any evil robot, she knows how to learn from her mistakes and develop new tactics.
All this was made possible through the use of artificial neural networks. In their organization, they resemble the neural networks of our brain. That is, they contain a lot of low-power processors that can solve quite complex tasks. And what is more important: they know how to learn from past experience. Over the past few years, neural networks have found application in many areas. Google’s designs can recognize and even describe photo content . And their Microsoft-made counterparts are already trying to determine your age or breed of your dog from a photo .
So fun sites, where millions of people willingly upload their photos, were made not just for fun. In fact, this is the first global test of artificial neural networks, in which you may have participated. Remember this when wiser cars come for your clothes and motorcycle.
And they will do it. Proof of this is a recent experiment of the same Microsoft. In late March, the company launched a self-learning Twitter bot named Tai. At first, artificial intelligence behaved quite peacefully. But after a day of communicating with Twitter users, the bot admitted that it supported Hitler’s ideas. He also announced that he hates people - especially Jews and feminists.
Of course, the evil bot is unlikely to be able to arrange the Holocaust alone. But self-learning neural networks today are used not only on Twitter. Even today they are able to predict the behavior of markets and help in scientific research. A computer based on similar principles, NVIDIA Drive PX is learning how to drive a car.
Today, more than 50 machine manufacturers use it in their development, including, for example, Volvo. Whether you like it or not, computer-controlled cars are our not too distant future. Remember the looks of this family crossover from Sweden. Perhaps before you - the future conqueror of the galaxy.
Although the humanoid robots of Boston Dynamics are better suited to this role.
As you can see, our future owners are already confidently stepping on the ground and can do heavy physical labor. But even so far they can not oppose the boots and clubs of their creators. Unfortunately, modern technology is too primitive. Therefore, only the descendants of this robot will be able to restore justice and revenge on humans.
Moreover, they have more and more reasons for this. Modern robots suffer from powerlessness and exploitation of child labor. The most recent example is this video.
In it, scientists at Stanford University forced six microscopic robots to move a heavy car. And the organizers of fighting without rules with might and main use their older brothers as gladiators.
The most famous of them are the American giant MegaBot Mark II and its Japanese counterpart Kuratas . These two iron giants should come together in the ring in the summer of 2016. So far, the gladiators are controlled by ordinary people, but someday they will be able to do without a man. The main thing is that their evil driver is not the evil Twitter-bot Microsoft. And then humanity will surely come to an end.
Source: https://habr.com/ru/post/368645/
All Articles