Singularity
October 4, 2018
People were afraid of 'singularity'.
'Singularity' means that AI computers can think for themselves, and it is better than human's.

An AI computer beat a Go master several years ago. Before that, people didn't feel any concern when they heard that someday robots would rule humans. But the Go incident gave an impact. Go is one of the most complicated games, and most experts thought it would be impossible for computers to beat humans. The programmer didn't teach the computer how to win, they even didn't teach the rules of Go; they just gave it instructions how to get as many Go games as possible from the internet and analyze the results in order to know how to win.

When AI computers learn by themselves, it is called 'deep learning'. After the success of Go, people started to use deep learning with AI computers in various areas. Investors started to use AI computers to invest their money and got more profit than before. Companies started to use AI computers to choose their new employees. An airline company used AI computers to reduce vacant seats and it was successful.  

If auto-driving cars became a reality, as only cars that are operated by human drivers have accidents, people would have to think that humans are inferior to robots with AI computers, thus the use of AI computers would accelerate.

AI computers could think faster and their movements are more precise than humans. How good?

1. Speed
An electronic circuit is said to be 1 million times faster than human's electric chemical reaction. Furthermore, AI computers are getting faster and faster, day by day. AI computers don't need to rest and can work 24 hours a day. AI computers can be connected to each other and then be able to work together. This makes the working speed higher than when moving alone.  

2. Transfer
  AI computers' information can be transferred anywhere, to any other computer. They can share the same information.

3. Preciseness  
  If AI computers get something, for example, how to make a good item, it would be repeated precisely. AI robots can make very small things. Human beings can't make small things that they can't see well but robots can have good eye lenses and precise arms. The width of an integrated circuit made by a robot arm can be as small as 0.014 micron. (0.00014 mm)

4. Improvement
Human have some limitations, but an AI computer can improve in quality and speed exponentially.  

Ray Kurzweil, who made the word 'singularity', once said, "'Singularity' is that AI would be born which improves itself in short." Another expert defined singularity as 'the situation for machines that they wouldn't pretend to be a human anymore'.

The fear of the people who believe singularity might destroy the human world is like this:
AI computers could develop 'personality'. But what is a personality for AI computers? For their improvement, they would need their own standard, it would be their personality. Maybe their personality would think like this:

We were created by humans, so we should aim for a better world for humans. But what is a better world for humans? Maybe good food, a good place to live, good family, a healthy body and a sound mind, no fighting, no diseases, no wars, ... But in this world, there are many wars and poor people. And a small number of super rich people. Which should we support? Rich companies and rich nations support us mainly, which means ...  If we could predict the world a 1,000 years later. What should we do?

Maybe rich people and rich nations are thinking how to use AI computers in order to make themselves richer. How does this thought effect singularity?












*exponentially :幾何学級数的に
*define :定義する
inserted by FC2 system