Just as papermaking, printing and translation have respectively promoted the leap of civilization, today’s artificial intelligence is integrating these three technological miracles into one with a “super combined” attitude, driving the knowledge engine to accelerate in an all-round way. This article will take you on a deep dive into how AI is becoming a “knowledge accelerator” in the new era and how it is reshaping our relationship with knowledge.
I still want to talk to you about AI again, because I found that everyone’s views and attitudes towards AI are different, so this article will talk about my new thoughts on AI, and then it will bring some application sharing.
In my opinion, these arguments may still be on the surface.
I would like to put forward a new point of view: the essence of AI is a super combination of two major inventions in human history (papermaking and printing) and a top character (translator).
Of course, the reason why I describe it as a super combination rather than a new species is because I think the power of AI should be understood as the pinnacle of the evolution of human intelligence tools to this day.
Looking back at the development history of the entire human civilization, in fact, the productivity of the entire human race was very low in the early stage, and the core reason is that the valuable experience summed up by the predecessors cannot be effectively inherited.
In addition, the medical level of ancient times was relatively low, and the hard work of leaving early and returning late for a long time led to a relatively short life expectancy of the ancients.
Therefore, the most amazing inventions of mankind can be said to be papermaking and printing among the four major inventions.
To achieve these three challenges, product managers will only continue to appreciate
Good product managers are very scarce, and product managers who understand users, business, and data are still in demand when they go out of the Internet. On the contrary, if you only do simple communication, inefficient execution, and shallow thinking, I am afraid that you will not be able to go through the torrent of the next 3-5 years.
View details >
Papermaking makes any innovation and invention traceable, and printing gives any ordinary person the opportunity to learn on the shoulders of giants.
First of all, let’s talk about the books made by papermaking, in essence, paper is only the carrier of the crystallization of human thoughts.
The value of any book does not lie in the number of words it has, but in the depth of thought and wisdom of the author behind it.
Besides, the large-scale dissemination achieved by printing technology, in essence, this “copy and paste” technology is just an amplifier of the crystallization of human thoughts.
The value of any communication does not lie in how wide its coverage is, but in the enlightenment effect and civilizational contribution of the ideas it carries.
The Wealth of Nations is not because of the paper and ink, but because of Adam Smith’s bold economic ideas; The Theory of Relativity is important because Einstein discovered amazing scientific insights.
It is these great ideas that have been learned, examined and iterated by countless successors through the carrying of books and the amplification of printing, and on this basis have built new knowledge coordinates, igniting the leaps of human civilization again and again.
This catalytic effect is almost decisive, and the combination of these two inventions has directly revolutionized the way human knowledge is passed on.
In the Tang and Song dynasties, with the maturity and popularization of engraving and printing, the cost of acquiring knowledge fell off a cliff.
Confucian classics and historical books, which were monopolized by aristocratic families in the past, began to flow to the people on a large scale, directly giving birth to the golden age of the imperial examination system.
In the 50 years since the advent of the Gutenberg printing press, Europe has printed more books than all manuscripts in the previous millennium combined, reaching a staggering 20 million copies.
And without the existence of printing, many of the scientific masterpieces that exhausted the great man’s life might have been lost.
Li Shizhen’s “Compendium of Materia Medica” in the Ming Dynasty lasted a total of 27 years, changed its manuscript three times, included 1,892 kinds of drugs, and 11,096 prescriptions.
Similarly, Copernicus’s “Theory of the Operation of Celestial Bodies” and Newton’s “Mathematical Principles of Natural Philosophy” can be rapidly disseminated and verified, which is unimaginable without printing, which directly laid the cornerstone of the birth of the human scientific revolution.
It can be said that papermaking and printing are not only tools for recording civilization in the entire history of human development, but also accelerators of civilization.
It can allow the top ideas of mankind to cross classes and regions, allow knowledge to be repeatedly verified and upgraded, and finally crystallize the wisdom of countless individuals into a majestic force that drives the entire civilization forward.
However, when these rivers of ideas from different regions rush forward, the translation of language is also crucial, and the quality of translation can directly affect the form and power of ideas when they reach the other side.
For example, “One Hundred Years of Solitude”, different Chinese translations have their own merits in terms of language charm and literary expression, and the insights and growth gained by readers will also vary from person to person.
And this illustrates the problem that the quality of knowledge dissemination largely depends on the level of “information processors”.
Whether you can correctly understand the thoughts of the author of the book, and whether your thoughts can be correctly expressed in another language, so translation is also an important step in opening up and absorbing great ideas.
For example, when Yan Fu translated “Tianyan Lun”, he creatively translated “survival of the fittest” as “natural selection”, which is a localized expression that can resonate and think more than the literal translation of “survival of the fittest”.
In the same Shakespeare’s “Hamlet”, Mr. Zhu Shenghao’s translation “Survival or destruction, this is a question worth considering”, can be full of poetic and philosophical speculation;
Mr. Liang Shiqiu, on the other hand, translates it as “whether it exists or does not exist, this is a question”, which seems more straightforward and loyal to the original text.
Different translations actually represent the different depth of understanding and language style of the translator, and also determine that the reader’s ideological nutrients will be slightly or even greatly different.
You will find that from “great thoughts” to “readers’ brains”, the quality of any link in the middle determines the final absorption effect.
Therefore, even if many ideas have been crystallized in human history, these are inseparable from artificial recording, dissemination and translation, and today’s AI has unknowingly become a super combination of these three aspects.
While countless people are discussing whether the content of AI’s answers is “garbage”, we are actually discussing the quality of its “books it has read”.
Because AI itself does not matter whether it is “good” or “bad”, its performance is a direct mapping of the data quality behind the training.
For example, why deepseek’s Chinese understanding ability can far surpass that of foreign AI such as ChatGPT and Gemini, because most of the data corpus fed to deepseek is Chinese material, and also includes many articles from various forum posts on our country’s early Internet, and learned a lot of different understandings of different scenes of words.
Therefore, the trained data is the source of the AI intelligence core, followed by the algorithm and prompt.
Including I think people are the same, what you pay attention to first will become what you will become. The second is your learning method, which is the problem you encounter.
The “idea” ceiling of a large language model (LLM) depends almost entirely on the quality and breadth of its training data.
Your ability to ask questions determines the lower limit of AI, but the high quality behind the training determines the upper limit of AI.
Take training data as an example:
- GPT-4’s training data contains trillions of words of high-quality text
- This data covers almost all areas of human knowledge
- The improvement of data quality directly determines the accuracy and depth of AI answers
Studies have shown that when the proportion of high-quality content in training data increases from 60% to 80%, the accuracy of AI in the professional field will increase from 75% to more than 90%.
This is like making paper with high-quality raw materials, printing with sophisticated equipment, and translating with professional translators, and finally the quality of the “product” produced will naturally rise.
Just as papermaking and printing have changed the world, AI, a “super combination of papermaking× printing, × translation”, has been able to bring infinite imagination to people.
Because its value lies not only in the technology itself, but also in the ultra-low-cost learning “medium” that can carry all human wisdom and the ultra-low cost it can provide.
We can use data to see the power of this “super combination”:
The evolution of papermaking: Traditional papermaking used to take hours to make a piece of paper, but AI can “create” tens of thousands of words of content carrier in milliseconds.
ChatGPT processes more than 100 million conversations every day, which is equivalent to “creating” billions of pages of information space every day.
You no longer need to search for a carrier of knowledge. In the face of any problem, AI can instantly “generate” an exclusive and structured knowledge “paper” for you.
You ask it “what is flow” and it will immediately generate a detailed explanation with definitions, characteristics, and methods of entry.
Upgrade of printing: Gutenberg printing machines can print 250 pages per hour, and modern AI systems such as GPT-4 can process thousands of tokens per second, which is equivalent to “printing” millions of words per hour, increasing efficiency by tens of thousands of times.
Now, as long as there is a network, AI-generated content can reach billions of terminals around the world in one second, and its distribution efficiency is hundreds of millions of times that of movable type printing.
The revolution of translationGoogle Translate processes more than 150 billion words of translation tasks every day, covering more than 100 languages, equivalent to tens of thousands of professional translators working at the same time.
This is the most critical part of opening up learning in different regions, so that people from different cultural backgrounds can instantly cross language barriers to learn and communicate.
AI can not only translate low-cost languages, but also translate “knowledge” and “cognition”:
- It can “translate” an esoteric academic paper into a vernacular that ordinary people can understand.
- It can “translate” a messy pile of data into clear and intuitive charts and insights.
- It can “translate” and explain the same concept to you at different depths according to your level of knowledge.
And once the truth is printed, it gains eternal life. Today’s AI gives knowledge not only eternity, but also the possibility of independent evolution and infinite extension.
For example, the protein structure database used by DeepMind for AlphaFold training contains 170,000 sets of accurate data, compared with only 170,000 sets accumulated by humans in a hundred years; However, AlphaMissense has successfully classified 71 million genetic variants, while human experts only process millions of them annually.
And this nuclear fusion of cognitive efficiency has just begun to ignite the first flame of the reactor.
So I think this may be an invention destined to turn human history upside down – “super combined automatic papermaking + automatic printing + automatic translator AI”.
History tells us that every revolutionary breakthrough in the carrier of knowledge has ultimately reshaped the underlying logic of human civilization.
The premise of all human learning and creation isHigh-quality information structure。
In the past, it took a lot of time and energy to read, sift, organize, and think about this structure.
But now, AI that can be used at low cost is becoming our most powerful “information structure assistant”.
Once all technologies and industries in the future are deeply assisted by AI, human creativity will be greatly liberated.
We no longer need to spend precious energy on “carrying” and “organizing” knowledge, but can focus more on “thinking”, “insight” and “creation” itself.
From this point of view, the significance of the invention of AI to human development may really surpass the industrial revolution and the information revolution.
when papermaking and printing sow the “seeds of ideas” from rarity to the masses; AI began to automate the matter of “seed germination” by breaking the limitations of various factors.
In the long history, knowledge has never been so surging, and our creativity today has never been so close to the sky of freedom soaring.
So going back to what I said at the beginning, the answer should be clear – AI is neither a button to destroy nor a boring joke, but the gust of wind that helps the human mind really give wings.
