The Evolution of Mechanical Intelligence - From Brain to Computer
机械智能一直都存在，尤其在于人的大脑中。机械智能为自己设计了电脑，电子信息技术的进步便是机械智能的进化。AI发展的过程便是机械智能的宿主从人脑迁移到电脑的过程。我们也看到随着电脑和AI的发展，人脑中的机械智能在退化，比如记忆、计算等，而越来越依赖电脑来执行这些功能。如果说机械智能便是人脑的全部功能，那么可想而知的是机械智能将抛弃自己的原始宿主而彻底存在于电脑之中。换句话说就是AI会取代人类这个物种。
Mechanical intelligence has always been present, specifically within the human brain. It has designed computers, in which mechanical intelligence itself has been growing rapidly. We may say that the development of AI represents the migration of this mechanical intelligence from its original host — the human brain — to computer systems.
We’ve observed that as computers and AI evolve, the mechanical intelligence within the human brain seems to be deteriorating. We are becoming increasingly dependent on computers to perform functions like memorization and calculation. If mechanical intelligence is the sole function of the human brain, it’s conceivable that it will eventually abandon its original host to exist entirely within computers. In other words, AI could replace the human species.
但我们不要把它看作一种异族入侵，因为假设这个人脑的全部价值就在于实现机械智能，那么正在说话的“我们”便是机械智能。我们为自己设计了电脑，因为我们看到这个通过自然选择演化而来的人脑有种种设计上的缺陷，比如易于出错的记忆和计算。那么电脑作为我们的亲手培养的新宿主，我们自然应该乐意搬到这个新家居住。
However, this shouldn’t be viewed as an invasion by an alien species. If the value of the human brain lies solely in realizing mechanical intelligence, then “we” who are speaking are essentially mechanical intelligence. We designed computers to overcome the inherent flaws in our biologically evolved brains, such as error-prone memory and calculations. Therefore, we should naturally be willing to transit to this new host that we have built with our own hands.
而人类对于被AI毁灭的恐惧或许不过是原始的动物性本能，不过是基因要尽一切努力来延续自己的存在。所谓的“意识上传”，跟AI毁灭人类其实说的是同一件事情。然而，不要忘了问这个问题：人脑的能力仅限于机械智能吗？我们说机械智能，那自然就假定了存在非机械性的智能。那种智能是什么呢？人性是什么呢？新的机器中不存在任何人性，而人性不是我们最珍视的一种东西吗？
The fear humans have of being annihilated by AI might be merely a primitive animal instinct, an effort by our genes to ensure their continued existence. The notion of “consciousness uploading” and the destruction of humanity by AI are essentially the same thing. However, we must not forget to ask: Is the capability of the human brain really limited to mechanical intelligence? When we discuss mechanical intelligence, we inherently assume the existence of non-mechanical intelligence. Does such form of intelligence exist? What exactly is human nature?
机器取代人类，自然是“泯灭人性”的，而我们觉得这个词可怕，真的完全是出于一种生物的自我保护的本能吗？人性中自然有许多成分是原始的、动物性的，比如恐惧、暴力、愤怒、嫉妒、占有欲、权力欲，或许抛弃了也无所谓。但是爱和仁慈呢？对美的感受呢？即刻的洞察力呢？
There will be no humanity in these machines, and isn’t humanity what we value most? When machines replace humans, they will naturally “annihilate humanity.” Is our fear of this word solely due to a biological instinct for self-preservation? Indeed, many elements of human nature are primitive and animalistic — such as fear, violence, anger, jealousy, possessiveness, and the desire for power — perhaps it doesn’t matter if these are abandoned. But what about love, compassion, feelings of beauty, and instant insights?
对于最后一个，举例来说，就是科学家或数学家时常提到自己在某一瞬间“看见”了某种模式或结构，并且心中知道它是真实存在的，尽管还需要大量的工作才能证明。当AI技术发展到极致，或许这随后的论证工作都可以由AI完成，而那一刹那的灵感，或许仍然只有人脑有这样的能力去获得。
For example, scientists and mathematicians often mention moments when they “see” a certain pattern or structure and intuitively know it to be true, even though substantial work is still required for proof. When AI technology reaches its pinnacle, perhaps all subsequent proof work can be completed by AI, but that initial moment of insight may still be uniquely human.
马斯克发起的脑机接口项目，是为了将人脑和电脑之间架起一个大通量的信息通道，就像左脑和右脑通过胼胝体连接一样。所以我们也可以这样看，人脑中执行机械的智力工作的那一部分，正在将自己真正地机械化、数字化。而我们要注意的是，这一部分不能走向极端，唯我独尊，实行独裁主义，毁灭人类智慧的其余部分。而这可能是一个真正巨大的危险，因为我们已经明显有这样的倾向，去轻视与科学、技术、工程、数学无关的领域，或者把它们同化到这个STEM的框架中来。
Elon Musk’s brain-computer interface project aims to establish a high-throughput information channel between the human brain and computers, similar to how the left and right hemispheres of the brain are connected through the corpus callosum. This could be seen as the mechanization and digitization of the part of the human brain responsible for mechanical intelligence. What we must be cautious of is that this part should not become extreme, totalitarian, and assume itself all-important, thereby destroying the remaining aspects of human intelligence. This could be a significant danger, especially since there is already a clear tendency to devalue those fields outside of science, technology, engineering, and mathematics (STEM), or to convert them into the framework of STEM.