What connects time and space is speed,
The link between energy and mass is also speed,
What will connect facts and values?
What will it be to connect people and machines?
Are the rules of reasoning constantly revised?
Or is it a constantly revised rule of reasoning ?
The "person" in artificial intelligence is not a "person"
Generally speaking, for a specific task situation, the value of a thing often changes around its fact (significance), but it is not a deterministic change, but a change of uncertainty. , Diffusion aggregation is flexible, which is similar to the values in our usual lives, constantly assimilating, adapting, and correcting balance. As Feynman said: the behavior of things on a small scale is not at all like a behavior on a large scale. Sometimes, the amount of fact and value of things in the changing situation do not change linearly, but change according to their own logical clues like a lens in a movie, without the need for daily intermediate time and space. It can be made out of nothing, or made out of nothing. In certain situations, the factuality and value of things can be instantly recognized even if they are far apart, forming an automatic pattern matching effect.
The "people" in artificial intelligence are not real "people". Autonomy does not mean intelligence, or autonomy is a necessary but not sufficient condition for intelligence. Only a self-subject with values can be said to be an intelligent body. So, factual autonomy is only automation, and value autonomy is intelligence.
A good work is the result of everyone's creation. For example, "The Ordinary World" was written by Lu Yao, said by Li Yemo, played by actors, thought of by countless readers/listeners, and spread by various media... ; A good smart product or system is also the result of everyone’s creation. For example, "Alpha Dog" or "Alpha Yuan" was developed by Deep Mind, trained by predecessors in chess, imagined by the masses, and spread by various media. of……
Someone said: "Perceptual is a complicated model of fuzzy calculation, which is the most energy-efficient and efficient balance." In fact, it is not. Perceptual intelligence is not a calculation, but a calculation mechanism with calculation. This is a complicated model fuzzy calculation, which is the most energy-saving and most efficient balance. Computational mechanisms often give a satisfactory answer when they do not necessarily understand the exact process of what happened, although these processes are opaque, and it is difficult to clearly prove what can and cannot be done. As far as perceptual intelligence is concerned, rules can be amended if it produces inferences that we do not want to accept; inferences can be rejected if it violates rules that we do not want to modify. The process of turning facts into value is a delicate process of mutual adjustment between rules and accepted reasoning; the final value is in the agreement reached by others or self. Perhaps, the basic laws of artificial intelligence cannot really be used to explain the laws of human intelligence.
What connects time and space is speed, and what connects energy and quality is also speed. So what will connect fact and value? That is, the question of what index is used to measure the value of something is not worth doing. This may be a question of connecting real and virtual, reality and fiction, structure and function, and other parallel worlds!
The contradiction of the integration of man and machine lies in: the divergence of man, the convergence of machine, the dialectic of man, the rule of machine, one gathering, one movement and one static. Then what we face is often not one problem, but a group of different problems intertwined! Therefore, it is difficult to achieve the purpose of solving with pure mathematical logic methods, so it is necessary to use formal logic, dialectical logic, and even non-logical means at the same time.
The uncertainty and unexplainability of machine learning and even artificial intelligence is mainly due to the fact that people have discovered that the inductive, deductive, and analogical reasoning mechanisms that are invented may indeed lead to certain incompleteness, instability and contradictions, and with calculations With the continuous expansion of scale, the greater these uncertainties and unexplainables. Human counterfactual reasoning and countervalue reasoning can prevent or warn these formal natural defects in advance from the perspective of hypotheses. Regarding the human-computer fusion as a cognitive subject is more conducive to solving complex problems, but it needs to solve the problem of how to integrate under different tasks. In addition, the fundamental mechanism of the single integration of one person and one machine and the group integration of multiple people and multiple machines will be very different. It can be said that three heads are better than Zhuge Liang.
