Chapter 30 Thinking (4)

002 nodded and said, "But controllable nuclear fusion technology is currently very difficult. The main reason is that the equipment related to controllable nuclear fusion is very complex, and problems are prone to occur when the complexity is high. Can you give me some advice?"

"It's like a mountain across the street! I really don't know much about controlled nuclear fusion technology. But if it's about machinery, I can still give you some suggestions that should be very good."

"We are willing to listen carefully." 002 said.

"When doing something complicated - such as managing a population of 100 million or walking with two thin legs, the most common way we think of is to list a list of tasks to be completed in order, and then complete these tasks under the instructions of the central command or brain, right?" 015 smiled mysteriously.

"It's true, but is there any problem?" 002 was a little puzzled.

015 further explained: "This model is very similar to a country that has been disintegrated by the United States. That country is very powerful, but it has collapsed itself. The collapse of this country illustrates a problem, that is, the complex systems controlled by the centrally are rigid and unstable. If institutions, companies, factories, organisms, economies, and robots are designed according to the centralized control model, they will be difficult to prosper. This model is logical but unrealistic. Because this model has two very big disadvantages. The first is excessively concentrated communication load, and the second is the high central memory maintenance cost."

015 paused for a while and continued, "I took a robot as an example. If I want to make a robot "smarter", I have to configure more computer components for it, which will make it more bulky. The heavier it is, the larger the drive motor will be. The larger the motor, the larger the battery pack required for power supply. The larger the battery pack, the larger the structure of the mobile battery pack, and this is a vicious cycle. This vicious cycle makes the proportion of the robot's brain to the body develop towards a trend of smaller and smaller. The brain will almost be in a weak state. The powerful country that once challenged the United States later was that the institutions were very bloated and the efficiency of work was extremely low! It is impossible not to say that this is an important reason for its self-disintegration."

002 sighed: "The more complex our system is, it will indeed become more bloated, and then it will be more complex and bloated. But is there any good solution?"

"Actually, there are. And we often use it. That is a distributed system. Sometimes we call it a group system."...

Generally speaking, distributed systems have the following advantages: 1. Evolutionable - only a group system can pass the adaptability obtained by local components over time from one component to another (from the body to the gene, from the individual to the population). Non-group systems cannot achieve (similar to biological) evolution.

2. Elasticity - Since the group system is built on many parallel relationships, there is redundancy. Individual behavior is insignificant. Small failures are like a fleeting wave in a river. Even large failures are only equivalent to a small failure at a higher level, and thus can be suppressed.

3. Infinity - For traditional simple linear systems, the positive feedback loop is an extreme phenomenon - such as the disorderly whistle of a sound reinforcement microphone. In the group system, positive feedback can lead to an increase in order. By gradually expanding a new structure beyond its initial state range, the group can build its own scaffolding to build a more complex structure. Spontaneous order helps create more order - life can reproduce more life, wealth can create more wealth, and information can nurture more information. All of this breaks through the original limitations and is endless.

4. Novelty—There are three reasons why group systems can produce novelty: (1) They are very sensitive to "initial conditions"—the subtext of this academic phrase is that the consequences are disproportionate to the cause—and therefore, group systems can turn small mounds into surprising mountains. (2) The combinations formed by individuals associated with each other in the system grow exponentially, which contains countless novel possibilities. (3) They do not emphasize individuals, and therefore allow individuals to have differences and defects. In group systems with genetic possibilities, individual variations and defects can lead to constant newness, which we also call evolution. (Excerpt from Kevin Kelly, the author of "Out of Control")

(End of this chapter)
Chapter completed!