Categories
- Art (356)
- Other (3,632)
- Philosophy (2,814)
- Psychology (4,018)
- Society (1,010)
Recent Questions
- Why did everyone start to hate the Russians if the U.S. did the same thing in Afghanistan, Iraq?
- What needs to be corrected in the management of Russia first?
- Why did Blaise Pascal become a religious man at the end of his life?
- How do I know if a guy likes you?
- When they say "one generation", how many do they mean?
Not at the moment. Not in the sense that it is currently impossible, but in the future everything can be. But the fact is that at the moment it is considered impossible in principle. There are two classes of problems that machines can't solve on their own.
First, they can't tell important information from unimportant information. This ability is critical for decision-making. Without this, autonomous self-government with an indefinite range of tasks is impossible.
Second, the machine can't tell the difference between a working system and a non-working one. In the general case. This ability is critical for autonomous machine evolution (development and adaptation). Roughly speaking, if the machine for some previously unknown reason stopped coping with the task, it cannot even determine the fact that it happened. And even more so – what needs to be done in order to correct the situation. This also means that machines cannot design new machines. After all, the result of a successful design should be a working new car. At best, machines will eventually be able to reproduce their own kind. There are no fundamental obstacles to this. But this is simply not enough to successfully compete with people.
Now the competition between car and machine is interesting. For example, chess engines such as Stockfish, which until recently left no chance for humans, were smashed to smithereens by the new AlpaZero, based on Google's Deep Mind neural networks. Which along the way also beat programs on checkers and some other board games. At the same time, this network “learned” to play itself in a few hours.
These neural networks are in some ways created in the image of the human brain, but in highly specialized tasks, they are much more effective. In defense of people, we can say that they are still very far from the capacity of the human brain.
Are they dangerous? The question is still open, many experts agree that there is no danger, and there can not be. However, we can definitely say that the artificial intelligence revolution is taking place
now.