top of page

​中文

English

How can the full application of AI realize a futuristic lifestyle?

By Haohua Zheng

Published: March 15, 2023

 

First, I would divide the full application of AI into three stages: virtual interaction, real-world interaction, and system integration. Currently, whether it's ChatGPT, Notion AI, or other AI, they are only at the first stage of virtual interaction, not even halfway through. Despite this, the current ChatGPT has profoundly changed our lives. we can imagine how deeply our lives will be impacted when AI completes these three stages. Therefore, I hope you can spend about 10 minutes patiently reading through all the stages, and I believe it will definitely overturn many of your ideas, and even your worldview. To better explore various aspects of AI, the content of the article will lean towards flexibility and openness, and will not be overly academic in its writing.

"I believe that artificial intelligence represents the future of computer science." — Alan Turing, computer scientist, founding figure in the field of artificial intelligence.

Stage 1: Virtual Interaction

Milestone 1: Full Training

Currently, ChatGPT has mastered a large amount of knowledge and precise results through over 10,000 GB databases and billions of training iterations. However, there is still some distance from the first milestone of "full training."

 

Full training is defined as not only mastering knowledge in all domains but also reaching a considerable depth, nearly an "expert" level. At present, AI cannot answer questions about very deep and specific knowledge in certain fields. For example, when I ask AI how to play a certain technique in a specific game, it cannot answer. The reason is simple: AI has not been trained with this precise information, and even the current GPT-4 cannot reach that level.

 

However, with the open sourcing and API opening of AI, experts and stakeholders in all fields will train their AI such as Fine-Tuning, which will eventually lead to an exaggerated and unprecedented state of AI knowledge storage, possibly achieved within the next 3-5 years. At that time, the update of AI's knowledge storage will be rapid, as hundreds of millions of people worldwide will train their AI every day.

 

This also raises a crucial topic: how much impact will AI have on our lives after fully trained? The most direct impact of full training is on the education sector. The education method we have followed since childhood is to go to school and learn from teachers because our technology in the past was not sufficient to support and allow us to interactively transmit knowledge in other ways; we had to rely on human interaction to transmit knowledge. However, in the era of fully trained AI, the education system maintained by human labor will be subverted, and human labor will no longer be the only medium for knowledge transmission.

 

Therefore, future educational institutions such as schools may become optional, and they may even slowly disappear as time goes on. This is just the most intuitive impact, but the underlying effects are more far-reaching. For example, changes in the education system will inevitably break the monopoly of education, and breaking the monopoly of education will inevitably change the social structure and environment, like a butterfly effect. This is also the content I will elaborate on in later stages.

 

However, there are still many problems with fully trained AI that we need to consider and prevent. For example, how to screen those information to ensure the correctness and effectiveness of the training results? This requires close cooperation among many stakeholders and companies to solve.

What opportunities that we can seize based on this milestone?

"In the future, artificial intelligence will become an indispensable part of our lives." — Stephen Hawking, physicist, cosmologist.

Milestone 2: Authorized Interaction

Currently, AI only provides one-way feedback and does not have real interactive functionality. With the widespread application of AI, human society will gradually empower AI to complete some daily chores. For example, allowing AI to help make reservations on certain websites or even make phone calls for appointments.

 

In this stage, many things in human society will become increasingly convenient, effectively saving our time and freeing us from trivial daily tasks.Imagine AI becoming your personal assistant, helping you pay insurance, book hotels, etc., just like Tony Stark's AI in Iron Man. The difference is that, at that time, everyone in human society will have such an AI assistant. Only then will human society truly realize virtual interaction on a meaningful level.

 

However, the implementation of authorized interaction requires the cooperation of governments, enterprises, and individuals, as only in this way can we better aware and solve hidden problems. Therefore, the key lies in how we can balance all aspects, trying to keep everything under control, even though the process will take a long time and face numerous challenges. This is also what I hope to achieve in my entrepreneurial journey within 5-10 years.

"The goal of technology should not be to replace humans but to liberate them, allowing us to do more creative things." — Yoky Matsuoka, neuroscientist, former vice president at Google.

Stage 2: Real-world Interaction

Milestone 1: Intelligent Ecosystem

At certain stages in AI development, it is inevitable for the technology to extend from the virtual to the real world, and the most likely medium for this would be smart products. By smart products, I'm not only referring to smartphones and computers, but also all other products we use in our daily lives, such as smart tables and smart lamps. AI will act as a medium to help us directly control and manage these products.

 

Imagine having your AI assistant open the garage door and prepare a hot bath for you before you arrive home. This is just one simple and routine application in our future lives. There will certainly be more robots produced and controlled by your AI assistant to help you with daily tasks such as laundry and cooking, freeing people from mundane chores and allowing them to spend more time on the valuable things. As for cultivating such an AI assistant, I mentioned in the first stage that everyone can train their own AI, which may even understand you better than your partner. Of course, forming such an intelligent ecosystem requires not only the efforts of all enterprises and governments but also the acceptance and recognition of people. This is also what I aim to achieve in my entrepreneurial journey within the next 10-20 years.

 

"We are at the beginning of a technological revolution, and artificial intelligence will redefine how we live, how we work, and how we interact with the world." — Masayoshi Son, founder and CEO of SoftBank Group.

 

Milestone 2: World Revolution

A necessary consequence of real-world interaction is that AI will completely disrupt industries, which will definitely surpass the Industrial Revolution of the 18th and 19th centuries. This disruption not only refers to the overturning of traditional industrial models but also the transformation of human social systems.

 

As individuals possess their own intelligent ecosystems, so businesses and factories will also have their own intelligent ecosystems. Factories will be fully automated by AI-controlled machines, and even internet companies will be managed by AI. Imagine a future where a mobile app will no longer require a physical company to operate it and only needs a virtual AI to run; physical companies will no longer be necessary.

 

Therefore, the inevitable outcome of this revolution is that AI will completely replace human intellectual and physical labor, and the concept of people needing to find jobs will no longer exist in society. This may sound terrifying, as massive unemployment could lead to social chaos. However, there's no need to worry, as these things will not happen overnight, but rather as a very gradual process. During this transition, governments and businesses will slowly adjust and help those who become unemployed.

 

Nonetheless, the result will be unchanged: more and more people will become unemployed until everyone no longer needs to work. With the development of sustainable energy and nuclear fusion energy, the production and marginal costs of AI-driven industries will continue to decrease until they approach zero. This also means that people in the future may no longer need to obtain essential resources through currency, which will further ensure everyone's livelihood.

 

However, many issues may arise, such as how businesses will profit, how governments will operate, and how people will live. These are all worth considering and addressing in advance. In my personal view, the currency and government systems will face severe challenges and even be overturned under this inevitable trend.

 

The currency system, a necessary product of human society, primarily serves to maintain the current social circulation system. The so-called circulation system involves employees serving companies, companies serving consumers, and consumers serving other companies as employees in a continuous loop. The currency system exists to stabilize this structure. However, when AI replaces all work and becomes self-sufficient through sustainable energy, the existing social structure will be disrupted, and the significance of the currency system will be upended.

 

I personally look forward to this day, as its implications for individuals are immense. When we no longer need to work for money, we will engage in more activities we genuinely enjoy. For example, more people will enter the scientific field, and with the efficient assistance of AI, they will effectively and rapidly advance society and the world.

 

I believe that everyone has the potential to be a genius, but they are often held back by life's constraints. In my country, China, there is a humorous yet sad joke that if Einstein, Musk, or some historical great people were born in China, they might have worked on factory assembly lines until retirement. The truth is that many people have the potential to become outstanding scientists like Einstein, but they are trapped by the existing social structure.

 

Therefore, when this day finally arrives, I believe humanity will become better.

 

"Our future is a co-created future, authored by both humans and intelligent machines." — Daniel H. Wilson, author, and roboticist.

 

Stage 3: AI System

The so-called system is like the operating system of a computer, where everything must function under the fundamental rules of the operating system. In real life, the operating system represents laws, rules, and civilization, all aimed at paving the path towards a world filled with love, fairness, and equality.

 

After completing the first two stages, human society will enter a completely different era. When the interaction between the virtual and the real world reaches its peak, a system will emerge to integrate all resources and manage the entire human society. At this time, the human governmental system will face serious challenges. The government, as a public authority institution, essentially maintains social order, allocates resources, regulates the economy, and ensures social justice and welfare. However, the current government system is dominated by power and is not entirely people-centric. One crucial reason is that the government system is run by humans, who are complex and possess subjective thoughts, leading to the inevitable and persistent abuse of power. In contrast, the AI system, consisting of countless codes and algorithms, will adhere to the basic principles of fairness, justice, and equality when handling all matters.

 

Meanwhile, I mentioned in the second milestone that the ultimate goal of AI reality interaction must be to replace all mental and physical work, because since the creation of human society, the operation of society has relied on manpower to complete mental and physical work due to the immature technology. Therefore, such a society is inevitably inefficient and error-prone. Including the majority of the current government system's work is also maintained by inefficient manpower. Therefore, the gradual replacement of the government system by AI systems seems to be an inevitable result, only a matter of time. However, there are still many issues that we need to consider and prevent in advance. For instance, how do we control the AI system without power? In my opinion, the algorithms of future AI systems will be extremely sophisticated, and we won't need to control or modify them. Instead, we will communicate with and provide suggestions to the AI system. Under the AI algorithm's judgment, effective suggestions will be retained and implemented. Even proposals not approved by the AI algorithm won't necessarily be abandoned. At that time, human society will be more democratic and flat, and as long as you can persuade more people to vote for your suggestion, the AI will retain and fulfill it. And at that time, the decision-making process of AI systems must also be interpretable and transparent, allowing people to intuitively see all the logic behind each decision made by the AI system. At the same time, the AI system will also have more sophisticated anti-discrimination and anti-bias detection programs to further ensure the fairness of the AI system.

 

Most importantly, the AI system will completely eradicate the abuse of power, creating a genuinely democratic, decentralized, and flat society. One inevitable result of living in such a society is the disappearance of wealth and class gaps. The most crucial reason for this is that AI's transformation of the educational environment will ensure everyone receives equal education. At that time, people's knowledge and cognitive levels will not have significant gaps, so the disparities in wealth and class will naturally lose their foundation, as they primarily stem from the monopoly on education. I do not intend to criticize modern society. In history, the monopoly on education had to exist at some point; otherwise, subsequent events might not have occurred. Just as humans had to experience despotism to realize the importance of democracy, and had to suffer the pain of war to understand the meaning of peace.

 

In the future, the new generation can learn everything through AI. Imagine our children being able to ask their AI assistant any question, anytime, anywhere. For example, what is the universe, what is the Milky Way, what is science, and so on. These are all things our parents couldn't teach us when we were young. Undeniably, the potential of our new generation far exceeds ours, even beyond our wildest imagination. More importantly, AI will flatten all aspects of life and break all monopolies. For instance, we just mentioned the monopoly on education, where high tuition fees and excessive barriers deprive many talented individuals of equal education opportunities. There are many other monopolies that I won't detail, but I believe we have encountered them in our lives.

 

At this point, some people may ask, since human nature is to pursue power and interests, why would wealth and class gaps disappear? The answer is not difficult. The pursuit of power and interests by humans is more a result of our current educational environment rather than solely determined by genes and heredity. It can be argued that genes and heredity are influenced by the environment, meaning that changes in the environment can gradually shape our genetics, steering future generations towards a more benevolent direction. This might be limited and dependent on the basic framework of genetics, but I still believe that we, as humans, have significant room for self-shaping. Speaking of the environment, our parents, friends, family, colleagues, and others make up this educational environment. Their unconscious expression of desire for money and power subtly influences us. Similarly, their pursuit of power and money is also because others have influenced them. Therefore, the emergence of AI not only disrupts the educational environment but also has profound and far-reaching impacts.

 

Speaking for myself, I do not have grand ambitions to change the existing educational environment, as it is a lengthy process that requires the efforts of many generations. Therefore, those of us in this era may find it difficult to witness that day. Nevertheless, I am still willing to take the first step to help more people see this vision.

 

"Artificial intelligence will bring us a more efficient, sustainable, and equitable world." — Sundar Pichai, CEO of Google Inc.

I recently learned about a fascinating yet terrifying phenomenon called the Overview Effect. This effect mainly describes a psychological change experienced by almost all astronauts when they enter space, which can even alter their worldview.

 

When astronauts see not only the beautiful Earth but also the vast and seemingly bottomless depths of the universe, a strong sense of insignificance and loneliness directly impacts their thoughts. Simultaneously, astronauts can directly see and imagine the contrasting scenes at opposite ends of the world. On one end, people suffer from war, disease, and poverty in remote corners, while on the other, people enjoy the revelry of parties on yachts. They are all human beings but live in two entirely different worlds. These are just the tip of the iceberg.

 

These experiences undoubtedly deepen astronauts' sense of insignificance and loneliness, to the extent that some even require psychological intervention for depression upon returning to Earth.

 

Discussing this primarily serves to introduce a crucial topic: a world without borders. A borderless world not only means the literal absence of borders but also symbolizes a fair, just, peaceful, and happy world. This seems like an era that most people, including myself, cannot imagine. However, for those who have completed the first and second phases we discussed earlier, this era is undoubtedly foreseeable.

 

As mentioned before, when AI systems replace government systems and eliminate the abuse of power, war and hostility will naturally disappear, the concept of nations will fade, and the concept of a borderless world will find fertile ground. In such an era, people can do what they genuinely love, and the only differences between people will be in thought and spirit. The concept of "gap" will also vanish, as true respect for individuals should not involve the idea of a "gap" but rather be described as "different."

 

Combining the three phases I mentioned earlier, people will inevitably live in a happier and more loving world. Consequently, people growing up in such an environment will be healthier, both physically and mentally.

 

This brings to mind a serious issue in modern society: are the "bad guys" we see on TV truly born evil? I believe the answer must be negative. When "bad guys" appear, most people's first reaction is to think about how to punish them, forgetting to consider the deeper causes. Aren't the "bad guys" a product of modern society? We can imagine a child who experienced abuse and violence in their youth, and in their still-imperfect adolescence, not only lost their parents but also lived in a social environment with strong racial overtones. It seems inevitable that this person would have character flaws. This person we imagine is Hitler.

 

As an individual, Hitler's crimes were evil and absolutely unforgivable. However, condemnation alone cannot help us understand the deeper reasons. Hitler was despicable but also pitiable because he was a product of his social environment, and his character flaws, combined with the abuse of power, directly led to the greatest tragedy in human history. Therefore, so-called "bad guys" are actually a reflection of the dark side of every social environment.

Today, the United States is plagued by frequent shootings, reflecting the dark side of contemporary society, and gun control is not a solution at its source. These statements are not intended to satirize the inaction of governments and society, because in the long river of history, our era may only be in its youth or even infancy, so imperfections are inevitable.

 

Nevertheless, human society will eventually reach maturity at some point in history, perhaps hundreds or even thousands of years from now. Just as we look back on human society 1,000 years ago, we find that society was filled with barbarism, closed-mindedness, violence, and paranoia. Abuse, slavery, and rape were common phenomena at that time. However, isn't this also how people 1,000 years from now might view our era? They may see our society as filled with anxiety, confusion, greed, and numbness, just as we view the society of 1,000 years ago.

 

Sharing these thoughts is not meant to make people more anxious about not being born in a better era, but to help everyone recognize the present, reconcile with the world, and live more comfortably.

 

I always try to view many issues from the perspective of the future, as it allows me to see the root causes of many problems. I temporarily call this way of thinking "Zheng's Thinking" In my opinion, if imagination can stretch far enough, cognition and consciousness may increase with the expansion and change of imagination. It is an entirely different perspective. 

 

From a historical perspective, the furthest we can trace back is to our ancestors, Homo sapiens, who, through the power of imagination, broke through the limitations of carrying capacity by creating beliefs. This allowed them to consolidate tens or even hundreds of thousands of Homo sapiens. Despite not having an advantage in intelligence or physical strength, they managed to exterminate other human species with absolute superiority, leading to the civilization we have today. More recently, this can be traced to the birth of the great inspiration of the general and special theories of relativity, which laid a profound foundation for human civilization.

 

"The true sign of intelligence is not knowledge but imagination." - Albert Einstein

 

Returning to the main topic, the far-reaching impacts of AI cannot be covered in just a few sentences; I could probably talk about it for days without exhausting the subject. However, that's not the point. What's important is that I have conveyed these concepts and visions to you, and that is enough.

 

The birth of AGI:

The significance of the birth of AGI (Artificial General Intelligence) is far greater than that of any other important event in human history. The current ChatGPT is merely imitating text through extensive data learning via powerful algorithms. It fundamentally cannot truly understand what we say, let alone self-learn and possess self-awareness.

 

AGI, on the other hand, represents the development of wisdom in AI, with self-learning and self-awareness capabilities. Imagine how terrifying it would be for an AI to self-program and optimize itself. For a profound example, AlphaGo is a Go-playing artificial intelligence that does not possess self-learning capabilities but still easily defeated top human players with billions of training instances. AlphaZero, an even stronger version, has self-learning capabilities; without relying on human knowledge, it has evolved to unimaginable levels through self-learning alone. AlphaZero not only crushes human players in various board games but also trounces AlphaGo. Humans can no longer predict AlphaZero's moves because it has advanced too far.

 

Returning to the main topic, the birth of AGI is similar to that of AlphaZero. The difference is that while AlphaZero's self-learning is limited to the board, AGI's self-learning is comprehensive, and it will revolutionize our understanding of everything. However, the development of AGI is extremely challenging.

 

First, it requires more powerful algorithms, which means that breakthroughs in mathematics and theoretical physics are necessary. Second, a large enough database is needed, which I believe is relatively easy to solve and may be resolved when we achieve the first milestone. Third, it requires immense computational power, which demands the support of higher-performance computers and storage devices. This implies that our chip theory research and application research must also achieve breakthroughs.

 

However, these are not the most important aspects. The most crucial is to ensure the ability to generate imagination, the capacity to create something out of nothing, just like how humans have invented various concepts that don't exist in reality, such as gods, souls, and ghosts. As of now, GPT is unable to speak about content it has not been trained on. Currently, it is quite challenging to truly develop AGI without a complete understanding of why humans possess imagination and the ability to think.

 

The most direct impact of AGI on human life would be AI interaction empowerment, enabling AI to break free from programming and genuinely understand human intentions, thus better helping humans manage everything. To be honest, although I can predict the short-term impact of AGI on humanity, the long-term impact is absolutely unpredictable because the birth of AGI actually signifies the birth of a new species that may surpass humans entirely. Predicting the development of a species is like predicting the thoughts of a god.

 

This does not mean that we are in opposition to AGI. Humans can rapidly develop science and technology through AGI's self-learning, just as current human players have discovered many new moves and techniques through AI's self-learning. When AGI leads technology to a certain level, humans will inevitably reach the technological level to modify the brain or even upload consciousness. At that time, the human brain may possess AGI-like capabilities, or we may even merge with AGI. Therefore, while developing AGI, we may also be seeking the possibility of a breakthrough in human civilization.

AI的完全应用如何实现未来化的生活

​作者:郑昊铧

发布:2023年3月15号

 

首先我将AI的完全应用分为三个阶段。分别是虚拟交互,现实交互,和AI系统。目前不管是chatGPT,notion AI还是其他AI只是停留在第一个阶段虚拟交互不到一半的位置。尽管如此,目前的chatGPT还是彻底的改变了我们的生活。我们可以想象如果AI的应用完成了这三个阶段,这将会对我们生活有多么深刻的影响。因此,我希望读者可以花大概10分钟的时间耐心的阅读完所有的阶段,我相信它一定会刷新你的很多想法,甚至你的世界观。为了更好地探讨关于AI的方方面面,文章内容会更偏向于灵活性和开放性,不会写的过于学术化。

 

"I believe that artificial intelligence represents the future of computer science." — Alan Turing, computer scientist, founding figure in the field of artificial intelligence.

 

第一个阶段:虚拟交互

 

第一个里程碑:完全训练

目前的chatGPT通过上万GB的数据库和上亿次的训练掌握了大量的知识和精确的结果。然而,这距离第一个里程碑“完全训练”是还有一段距离的。

 

所谓的完全训练定义就是不光是要掌握所有领域的知识并且要达到相当高的深度将近“专家”的级别,而目前AI是无法回答某领域非常深度和特别精确的知识。举个例子,当我询问AI某一款游戏的如何操作某一个技巧时,它是无法回答的。原因很简单,因为AI没有训练这些精确的信息,即便是现在的GPT4也远远没有到达那种程度。

 

但是随着AI的开源和API开放,生活中所有领域的专家和相关者都会去训练自己的AI模型,随着时间这将使得AI的知识储备达到一个非常夸张和空前的状态,这可能会在未来3-5年内实现。并且,这时的AI知识储备的更新会非常快,因为每天世界各地数亿以计的人会去训练自己的AI模型,可能用于企业或个人。

 

同时,这也引出了一个非常重要的话题,那就是当AI实现完全训练之后将会对我们生活产生有多大的影响?完全训练产生的最直观的影响莫过于教育领域。我们从小就遵循的教育方式就是上学跟着老师学习,因为在过去我们的科技并不足以支撑和允许我们以其他方式交互式传输知识,只能依靠人力来交互传输知识。然而,在AI完全训练的时代人力所维持的教育体系必将被颠覆,人力将不再是唯一的知识传输媒介。

 

因此,未来的教育体系中,学校这些组织可能会成为一个可选项,甚至随着时代会慢慢消失。这只是最为直观的影响,但背后的影响其实更加深远。比如,教育体系的改变也必将会打破教育垄断,而打破教育垄断也必将会改变社会结构和环境,就像是一个蝴蝶效应一般。这也是我在之后的阶段中会详细解读的内容。

 

然而,关于AI的完全训练还是有许多问题是我们需要考虑和预防的。比如,如何去筛选这些信息确保训练结果的正确性和有效性?这需要很多相关者和企业的密切配合才能解决的。不管如何,在大势所趋下问题总是会有解决办法的。

 

"In the future, artificial intelligence will become an indispensable part of our lives." — Stephen Hawking, physicist, cosmologist.

 

第二个里程碑:授权交互

目前的AI只是单方面反馈,并不具有真正的交互功能。随着AI的广泛应用,人类社会必将慢慢授权AI完成一些日常琐事。比如,让AI帮忙在某些网站预约,甚至让AI帮忙打电话预约。

 

处于这个阶段的人类社会很多事情将会不断的便利化,这将有效节省我们的时间,并且将我们从日常繁琐的小事解放出来,想象一下,AI将会成为你的私人助理,帮你接听电话,缴纳保险,和预定酒店等等,就像钢铁侠里托尼的AI贾维斯那样。不同的是,未来的人类社会所有人都将拥有这样的AI助理。这时的人类社会才算是真正意义上实现了虚拟交互。

 

然而,授权交互的落实还是需要政府,企业,和人们的配合,因为只有这样才能更好去发现和解决一些隐藏的问题。不管怎么说,授权交互是必然会发生的事情因为它对人类生活的改变和意义都太大。因此,关键在于我们应该如何平衡好所有的事情,尽量让所有的事情都不脱离掌控,尽管这个过程会经历漫长的时间和无数的挑战。这也是我希望在我创业过程中5-10年能实现的事情。

 

"The goal of technology should not be to replace humans but to liberate them, allowing us to do more creative things." — Yoky Matsuoka, neuroscientist, former vice president at Google.

 

第二阶段:现实交互

第一个里程碑:智能生态

AI发展的某些阶段必将从虚拟扩展到现实,这是一个必然的趋势,而这个媒介大概率就是智能产品。我所说的智能产品并不只是手机和电脑,而是我们生活中所能使用到的所有产品。比如,智能桌子,智能电灯等等。AI将会作为一个媒介帮助我们直接控制和管理这些产品。

 

想象一下,你在还没到家之前让AI助手帮忙打开车库门,并且准备好热水来泡澡。这只是未来生活中一个非常简单和正常的应用。到时也必然会有更多机器人生产出来由你的AI助理控制帮你解决日常事务比如洗衣和做饭,从而彻底帮助人类从生活中的琐事解放出来,让人们拥有更多的时间做自己想做的事情。至于如何培养这么一个AI助手我在第一阶段的第一个里程碑就讲过,每个人都可以训练出属于自己的AI,这个AI助手甚至可能比你的伴侣更了解你。当然,想要形成这么一个智能生态不光是需要所有企业和政府的努力,也需要人们的接纳和认可。同时,这也是我在未来创业过程中10-20年之内想要去实现的。

 

"We are at the beginning of a technological revolution, and artificial intelligence will redefine how we live, how we work, and how we interact with the world." — Masayoshi Son, founder and CEO of SoftBank Group.

 

 

第二个里程碑:世界革命

现实交互带来的一个必然结果就是AI将完全颠覆工业,这将绝对不亚于18世纪和19世纪的工业革命。我所说的颠覆不仅是对传统工业模式的颠覆,也是对人类社会体系的颠覆。

 

当我们每个个体都拥有了属于自己的智能生态,同样企业和工厂也将拥有属于自己的智能生态。所有工厂的都将由AI控制机器完全自动化的运营,甚至互联网企业也将由AI来控制和管理。想象一下未来的某一个手机软件将不再需要实体企业来运营,只需要虚拟AI来运营,那么实体企业将没有了存在必要。

 

因此,这场革命带来的必然结果就是AI完全替代人类的脑力和体力工作,社会将不会再存在人们需要找工作这个概念。这么一听是不是感觉非常恐怖,那么多的人失业那岂不是社会大乱。但其实不用担心,这些事情不会在一个晚上就发生,而是一个非常缓慢的过程。在这个过程中政府和企业是一定会慢慢调节帮助这些失业的人。

 

然而,结果还是不变的,失业的人会越来越多,直到最后所有的人都将不需要工作。随着可持续能源和核聚变能源的发展,AI主导的工业的生产成本和边际成本就不断减少,直到接近于0。这也将意味着基本物资会被“免费”生产出来进一步保障了所有人的生活资源。

 

然而这也可能面临很多问题,比如企业将如何盈利,政府将如何运转,以及人们将如何生活。因此,这些都是值得我们提前去思考和预防的。在我个人看来,在这个必然的趋势下未来的货币系统和政府系统将会受到严重的挑战,甚至被颠覆。

 

货币系统作为目前人类社会必然的产物主要是为了维持现有的社会循环系统。所谓的循环系统就是职员服务企业,企业服务消费者,消费者也将作为职员服务其他企业,以此循环往复。而货币系统就是为了稳定这么一个结构。然而,在AI代替了所有工作并且通过可持续能源能自给自足时必将会打破原有的社会结构,这时货币系统存在的意义也必将被颠覆。

 

我个人是非常憧憬这一天的到来,因为它对我们个体的意义是非常巨大的。当我们不再需要为金钱工作时,我们将从事更多我们真正喜爱的事情。比如,更多人从事科学领域配合AI助手的高效辅助将有效和快速的推动社会和世界的进步。

 

我相信每个人其实都有潜质成为天才,只是他们很多被生活束缚和藏匿了。在我的国家中国有一个非常好笑但又无奈的玩笑,它说的是如果爱因斯坦和马斯克出生在中国,那么可能他们会在工厂的流水线一直工作到退休。我们人类历史上人口总数超过一千亿,在过去的一百年里就有着100亿的人口,这其中肯定会有着比爱因斯坦更聪明和更具有想象力的人,他们很多要么是生不逢时,要么是碌碌无为,从一出生起就注定无法从事他们有潜力的领域。因此,真的等这一天的到来,我相信人类将会变得更好。

"Our future is a co-created future, authored by both humans and intelligent machines." — Daniel H. Wilson, author, and roboticist.

 

 

第三阶段:AI系统

所谓的AI系统就像是计算机的操作系统,一切都必须在操作系统的基本规则下运行。在现实生活中,操作系统代表着法律、规则和文明,所有这些都旨在为一个充满爱、公平和平等的世界铺平道路。

 

在完成上面两个阶段后,人类社会已经迎来完全不一样的时代。在虚拟和现实交互达到尽头时一定会出现一个系统,整合所有的资源来管理整个人类社会。这时人类的政府系统也将受到严重的挑战。政府系统作为一个公共权力机构,本质作用是维护社会秩序、分配资源和调节经济、保障社会公正和福利等。然而,现在的政府系统却是被权力统治着,并不是完全的以人为本。这里面很重要的一个原因就是政府系统是由人来运转的,而人是非常复杂并且具备主观思想,所以权力的滥用就变成了一个必然结果并且屡禁不止。而AI系统作为一个由无数代码和算法构成的系统,在处理所有事情都将秉承着公平,公正,和平等基本原则。

 

同时,我在第二个阶段第二个里程碑曾讲过,AI现实交互的最终目的一定是取代所有脑力和体力工作,因为自人类创建社会以来因为科技不成熟的原因使得社会的运转都是需要靠人力来完成脑力和体力工作,所以这样的社会也必然是低效并且容易出错的。包括目前政府系统绝大多数的工作也是由低效的人力所维持着。因此,政府系统将被AI系统逐渐取代似乎是必然的结果,只是早晚的问题。然而,这其中还是有很多问题需要我们提前思考和预防。比如,既然没权力,那么将如何控制AI系统。我个人的看法是,未来100年后的AI系统的算法一定是极其完善的,我们并不需要去控制和修改它,而是与AI系统沟通和提建议,在AI算法的判定下,有效的建议将会被保留和实行。即便是AI算法不认可的建议也并不意味着会被放弃,因为这时的人类社会是更加民主和扁平化的,只要能说服更多人投票支持你的建议,AI也会保留和履行。并且,那时的AI系统的决策过程也一定是可解释化和透明化的,人们可以直观的看到AI系统在作出每次决策背后所有的逻辑。同时,AI系统内也会有着更精妙的反歧视和反偏见检测程序来进一步确保AI系统的公正。

 

最重要的是,AI系统将会完全杜绝了权力的滥用,真正创造一个民主,去中心化,和扁平化的社会。处于这么一个社会的一个必然结果就是贫富差距和阶级差距的消失,这其中最重要的原因就是AI对教育环境的改造将充分确保所有人得到同等教育,这时人们的知识和认知水平就不会产生大的差距,那么贫富和阶级的差距自然也会失去土壤,因为贫富和阶级的差距产生的根本还是因为教育的垄断。我并不是想要讽刺现代的社会。对于历史来说,教育垄断确实是在历史某个节点必须要存在的,否则后面的事情可能就不会发生。就像是人类经历了专制主义才会意识到民主主义的重要性,经历了战争的伤痛才会理解和平的意义。

 

在未来,新生代的可以通过AI学习到所有事情。想象一下,我们的孩子可以随时随地询问AI助手任何问题。比如,什么是宇宙,什么是银河系,什么是科学等等。这些都是我们父母在我们小时候无法教导我们的。不可否认的是,我们的新生代的潜力绝对远远超过我们,甚至是我们完全无法想象的。更重要的是,AI的出现将全方位让所有事情扁平化,并且打破所有的垄断。比如我们刚聊到了教育垄断,高昂的学费和过高的门槛会让很多有潜力的人失去了享受同等教育的机会。还有其他很多的垄断我就不一一细说了,相信我们在生活中或多或少都遇到过。

 

说到这里,肯定有些人会有疑问,人性本就是追逐权力和利益,那为什么贫富和阶级差距会消失?其实这个问题并不难回答。人性之所以追逐权力和利益其实更多是我们目前的教育环境导致的,而不光是基因和遗传决定的。或者说基因和遗传都是基于环境,也就是环境的改变可以慢慢塑造我们的遗传使得我们的后代朝着更良善的方向发展,即便这可能是有限的并且需要基于基因的基本框架,但是我仍然认为我们人类还是有很大的空间去塑造自己。说回环境,我们的父母,朋友,家人,和同事等等的人构成了这个教育环境。他们无意识间表达对金钱和权力的渴望都在潜移默化的影响着我们。同样的道理,他们之所以追逐权力和金钱也是因为其他人影响了他们。这一切都是可预见的,因为对于目前这个信息爆炸和物欲横流的时代来说,这是一种很难去回避的社会价值观。因此,AI的出现不光只是颠覆教育环境,更是有着非常深刻和长远的影响。

 

我个人来说,我并不奢望去改变现有的教育环境和我们这代人,因为教育环境的改变是一个非常漫长的过程,必定需要非常多代人的努力,所以处于这个时代的我们可能很难看到这一天的到来。尽管如此,我仍然愿意走出第一步让更多人看到这个愿景。

 

"Artificial intelligence will bring us a more efficient, sustainable, and equitable world." — Sundar Pichai, CEO of Google Inc.

 

前段时间了解到了一个非常有趣但是细想非常恐怖的事情,那就是全景效应。这个效应主要是在讲基本所有宇航员在进入空间后都会出现的一个心理变化,这种心理变化甚至直接改变了他们的世界观。

 

当宇航员进入到了空间看到不光是地球的美丽景象,还看到了宇宙深邃和深不见底的黑,这种强烈的渺小感和孤独感直接冲击着所有宇航员的思想。同时,宇航员在空间中会直观的看到和想象到世界两端不同的景象。在世界的一端,人们在某个偏僻的角落遭受着战争,疾病,和贫穷的摧残,而在世界的另一端,人们在游艇上享受着派对带来的狂欢,肆意瓜分着战争所带来的财富。他们同样作为人类,却处于两个完全不同的世界。而这些只是冰山一角。

 

这些冲击毫无疑问加深了宇航员的渺小感和孤独感,甚至一些宇航员在返回地球时一度抑郁需要心理治疗的干预。

 

聊这些主要就是为了引出一个非常重要的话题,那就是一个无国界的世界。无国界不光是字面表达的意思,它更是象征着一个公平,公正,和平,幸福的世界。这似乎是绝大多数人无法想象到的一个时代,也包括我。然而,对于已经聊完的第一个阶段,第二个阶段的我们来说,这个时代绝对是可以预见的。

 

在之前我们说过的内容里,当AI系统取代了政府系统并且杜绝的权力的滥用之后,那么战争和敌对也会自然消失,国家的概念也必然会慢慢被淡化,无国界的概念也将得到成长的土壤。在这么一个时代,人们能做自己真正热爱的事情,人与人之间只会存在思想和精神的不同,“差距”这个概念也会消失,因为真正尊重个体不应该会有“差距”这个概念,而是用“不同”去形容。

 

结合我之前说的三个阶段,这时的人们必然生活在一个更幸福,更有爱的世界。因此,在这种环境下成长的人也会更加的健康,不管是身体和心理上。

 

这让我想到现代社会的一个严肃的话题,那就是我们电视上看到坏人真的是生来就坏吗?我想答案必然是否定的。当“坏人”一出现绝大多数人第一反应是去思考如何去惩罚他们,而忘记了去思考更深层次的原因。“坏人”的出现何尝不是现代社会培养出来的?我们可以想象一个孩子在少年时期遭遇过虐待和暴力对待,在思想还不完善的青年时期不仅父母双亡而且还生活在带有浓烈种族主义色彩的社会环境,那么人格上的不足似乎是必然的一个结果。而这个我们想象的人就是希特勒。

 

作为个体,希特勒犯下的罪行是邪恶并且是绝对不可原谅的。然而,光是谴责并不能帮助我们去了解更深层次的原因。希特勒是可恶的,但也是可怜的,因为他是那个社会环境下所培养的产物,人格的不足加上权力的滥用直接导致了人类历史最大悲剧的发生。因此,所谓的”坏人”其实都是每个社会环境阴暗面的缩影。时至今日,美国枪击案层出不穷,同样也反映了当今社会的阴暗面,而禁枪也绝不是在源头上解决问题。其实说这些也并不是在讽刺政府和社会的不作为,因为在历史的长河中,我们这个时代可能只是少年时期甚至婴儿时期,所以必然会有诸多的不完善。

 

但是不管怎么说,人类社会在经历过这些之后总会在历史某个时间点迎来成年时期,这可能是几百年或者几千年后的事情了。就好像我们站在现在的时代看待1000年前的人类社会,我们会发现1000年前的人类社会时刻充斥着野蛮,封闭,暴力,和偏执等等。虐待,奴役,和强奸在那个时代都是普遍的现象。然而,这又何尝不是1000年后的人们看待我们这个时代的方式。他们会认为这个时代的人类社会充斥着焦虑,迷茫,贪婪,和麻木等等。就像我们看待1000年前的人类社会那样。

 

说这些不是让大家更焦虑没有出生在一个好的时代,而是为了让大家能够认清当下跟世界和解以及更自在的生活而已。

 

我一直都在尝试站在一个未来的角度来看待很多的问题,因为这样才能看到很多问题最本质的原因。我暂且给这种思维方式称之为Zheng的思想方式。在我看来,如果想象力能够想象的足够远,认知和意识可能将随着想象力的延伸和变化而提高。那是完全不一样的视角。

 

从历史的角度讲,最远可以追溯到我们的祖先——智人,便是靠着想象力通过创造信仰突破了群体限制(Carrying Capacity)凝集了几万乃至几十万的智人,使得他们能够在智力和体质都不占优的情况下,以绝对的优势消灭了其他的人种,才有了我们现在的人类文明;而最近则可追溯到广义和狭义相对论这种伟大灵感诞生,为人类文明奠定了深厚的基础。

 

"The true sign of intelligence is not knowledge but imagination." - Albert Einstein

 

回到主题,AI所带来的深远影响也绝不是我几句话就能讲完,可能再说个几天都说不完。然而这些都不是重点,重要是我将这些概念和愿景传达给你了,这样就足够了。

 

题外话:AGI的诞生

AGI全名(Artificial General Intelligence),它诞生的意义要远远大于人类历史所有重要的事件。目前的chatGPT只是通过强大算法来进行大量的数据学习进行文字上的模仿,它本质是无法真正理解我们所说的话,所以更别提自我学习和自我意识。

 

而AGI则是代表着AI产生了智慧具有自我学习和自我意识,想象一个AI可以自我编程和自我优化将是多恐怖的一件事。举个深刻的例子,阿尔法狗是一个围棋人工智能,它并不具备自我学习能力但仍然凭借数以亿计的训练轻松打败了人类最顶尖的人类棋手,而阿尔法零则是更强的版本,它具有自我学习能力,在不依靠人类的知识仅靠自我学习已经进化到了无法想象的地步,阿尔法零不仅碾压式的赢了各种棋类的人类棋手,还碾压式的赢了阿尔法狗,目前人类已经无法推测出阿尔法零的棋路,因为阿尔法零已经走得太远了。

 

回到主题,AGI的诞生就相当于阿尔法零。不同的是,阿尔法零自我学习范围是棋盘,而AGI的自我学习是全面的,它将颠覆我们对所有事情的认知。然而,AGI的开发是极其困难的。

 

第一,这需要更强大的算法,这也就意味着数学和理论物理领域必须有所突破才行。第二,足够大的数据库,我相信这个是相对容易解决的,这可能在我们完成第一个里程碑时就会得到解决。第三,超强的计算能力,这需要更高性能的计算机和存储器的支持,这也说明我们的芯片理论研究和应用研究也必须有所突破才行。

 

然而,这些都不是最重要的,最重要的是要确保它能产生想象力,也就是无中生有的能力,就好像我们人类创造了各种各样现实中不存在的概念比如神,灵魂,和鬼等等,而现在的GPT是无法凭空说一些它没有训练过的内容。目前来说,在人类没有完全搞懂为什么人类会产生想象力和思考的问题下是很难真正开发出AGI的。

 

AGI对于人类生活最直观的影响就是AI交互赋能,让AI能摆脱程序化真正理解人类的意图进而更好帮助人类管理所有事情。说实话尽管我可以预测AGI对人类的短期影响,但是长期的影响是绝对不可预测的,因为AGI的诞生其实意味着一个物种的诞生,而且这个物种可能将完全超越人类,因此预测一个物种的发展就仿佛是在预测上帝的想法一样。

 

这并不意味着我们与AGI是对立的。人类完全可以通过AGI的自我学习快速发展科学和科技,就好像现在的人类棋手通过AI的自我学习发现了非常多的新棋路和技巧一样。当AGI带领科技发展到一定的水平,人类也必将达到能够改造大脑甚至上传意识的科技水平,这时人类大脑很可能会具备AGI那样的能力,甚至我们将于AGI融为一体。因此,我们在开发AGI的同时其实也未尝不是在寻求人类文明在未来突破的可能。

bottom of page