Computers and Human-Level Intelligence

Darroch concluded that the declining birth rate and pregnancy rate over the past 25 years has spread throughout the world of industrialization and suggests that this trend is more extensive than any other country. Increased importance has increased incentives for young people to achieve higher levels of education and training, as well as greater central position, in addition to the formation of mothers and young women. This suggests that education and other opportunities, understanding of contraception and abortion may reduce the population ratio, thereby reducing the burden on the fiscal austerity of the government, in particular the domestic burden I will.

As Google's renowned engineering director Ray Kurzweil points out many times, the law of accelerated regression suggests that computers will reach artificial intelligence (AGI) - human intelligence by 2029 . In addition, by 2045, computers should reach artificial super-intelligence (ASI) - the level of intelligence is smarter than all humans. This event - "Intelligence outbreak -" will be the most important event in our history ... Super smart machines can design better machines ... human wisdom is much more It will be delayed ... ... So the first super smart machine is the last invention that human beings need to make. (Mulehauser)

At some point it is very realistic to implement AGI. Software that has reached human level, or intelligence beyond human level. Does this mean that the computer has as much capability as us at that time? Indeed, not at all - computers are more efficient. Because they are electronic, they have the following advantages: collective abilities. Because of time consuming communication and complex social strata, teamwork is very difficult. The larger the group, the slower the output of each person. On the other hand, AI is not biologically restricted by the human body, there is no problem of human cooperation, and you can synchronize and update your own operating system.

The most difficult part in creating AGI is to learn how to develop that software. "The truth really does not know how to make it smarter - we still know what the dog is and still know the mysterious B and mediocre movies.This is the most common three The simplest idea is to steal the brain and build a computer architecture that closely resembles the structure of the brain, an example is "An artificial neural network, which is originally a transistor connected via input and output "Net network", it does not know anything - just like a baby's brain. It is to try to do tasks such as handwriting recognition firstly that neural attack and interpretation of each character following it are completely random.