[Perspective] The Applications of AI are Easy to Understand, but the Significance of Web3 is Difficult to Define

Due to the initiative of AI writing assistants, some readers have asked me if I am also following the trend, shifting from web3 to AI, lamenting that blockchain is truly ephemeral.

No, web3 remains my core focus, and as I mentioned before, I will invest at least another three years in LikeCoin to fulfill my ten-year commitment. Whether it is web3 or AI, as long as it can make Liker Land more usable and meaningful, I am very enthusiastic.

The key question is not whether to shift tracks, but how web3 and AI can improve reading and publishing?

I have previously stated that AI aims to enhance production efficiency, while blockchain seeks to improve production relations. The value proposition of AI is straightforward: it does your work, increases efficiency, and reduces costs. In the context of publishing and reading, it can help authors brainstorm and organize information, assist publishers with proofreading and typesetting, read texts aloud for readers, translate entire books, and conduct in-depth analyses.

However, less discussed is the fact that while AI increases production efficiency, it also changes (not necessarily improves) the industry structure. In the publishing sector, the most obvious example is translation. Previously, only major authors and bestsellers had publishers seeking professional translators, but now, with a few keystrokes, anyone can translate a work of reasonable quality, and the standards of AI translators continue to improve.

It is difficult to assert whether the livelihoods of professional translators will improve or worsen in this situation. Although it may appear that their jobs are being taken by cheaper labor, one cannot overlook that translators can also leverage AI to enhance efficiency. Furthermore, as the barriers to translation lower, the demand for proofreading translated works may significantly increase.

Regardless, it is certain that AI is changing the role of translators in publishing, and similar situations are occurring in proofreading, editing, design, and even writing. Kindle has long explicitly prohibited works written by AI; however, when everyone relies on AI to some extent, it becomes increasingly difficult to determine what constitutes AI authorship, and simply defining it has become extremely challenging.

Those threatened are not merely individual practitioners. Today’s large language models understand a wide array of topics, drawing from not only web data but also a vast number of high-quality books. International copyright laws emphasize that unauthorized reproduction is subject to scrutiny; yet, more than a century later, AI learns from countless words at lightning speed, assimilating and rewriting knowledge without needing to reproduce it, swallowing a century’s worth of wisdom without even a word of thanks.

The most striking illustration of this is OpenAI, which, based on Google’s transformer research, leveraged a significant portion of internet data for learning. After achieving key results, it transitioned to a closed-source model, controlling production data and “lower-dimensional attacks” against other less intelligent entities, benefiting from new production relationships while simultaneously engaging in widespread destruction.

As the discussion shifts from individual to collective, from production efficiency to production relations, it becomes increasingly difficult to comprehend.

“Production relations” is a concept proposed by Marxism. Rather than exposing its shortcomings from an academic standpoint, let us attempt to unfold it within the context of publishing.

Consider the production process of a book from writing to reading: the author Alice writes the text, commissions the publisher Bob to produce it, the operator Carol promotes it on e-book platforms, and finally, it is sold to the reader Dave. Throughout this process, Dave pays for knowledge or entertainment, Carol controls the sales channels to earn profit, Bob earns income through editing, typesetting, and design, while Alice receives (likely insufficient for a living) compensation for her knowledge and imagination. This series of exchanges and interactions between “raw material suppliers,” “text factories,” stores, and consumers constitutes the production relations of the publishing industry.

This concept is very basic and it is hard to imagine anyone not understanding it. However, “not understanding” has two layers of meaning: intellectual and emotional. The former refers to the inability to grasp vocabulary or concepts, while the latter involves the lack of empathy. When I say, “I don’t understand what’s good about TikTok,” it means I cannot appreciate it, and even if you explain for half an hour, I still won’t understand. I have observed that sometimes readers who claim not to understand are actually focusing on the latter; those who do not grasp production relations perceive the roles, participation, and benefits of creation, editing, publishing, selling, and buying as simply “that’s how it is,” lacking the capacity to appreciate what deserves reconsideration.

This inability to grasp is, at its core, merely an acceptance of the status quo, viewing existence as truth: Is the royalty insufficient to make a living? In the age of audio-visual, that’s how it is; Is there a monopoly by platforms? In a market economy, that’s how it is; Is the rise of e-books causing physical bookstores to decline? The wheel of the times, that’s how it is; Did I pay for an e-book and want to gift it but cannot? In business terms, that’s how it is. Everything seems self-evident, yet the emergence of blockchain reveals the possibility of change, which is precisely what captivates me.

In 2022, I experimented with self-publishing using web3 technology, selling 1,024 copies of a “non-reproducible” work to 943 readers from 18 regions through 21 channels in just 34 days, each priced at $9.9 with a signed message. The participating channels included not only social media but also traditional bookstores, and readers were free to use the text files, including printing them into physical books and reselling. Although these numbers and scales are insignificant, the model of decentralized e-books behaves like a mischievous child, challenging the order regarded as self-evident by adults: authors retain all rights, authors rely on intermediaries for publication and sales, readers must purchase to read, purchases are solely for reading, and the relationship between authors and readers is strictly a consumer one. Web3 allows us to reshape production relations and reduces the prevalence of “that’s how it is.”

Lu Xun said, “There were originally no roads on the ground; when many people walk, they become roads.” A mischievous child cannot change the world all at once; however, one can never be sure whether they will ignite a spark of change in the publishing industry.

I remember during my Economics 101 class in college, the teacher always told stories using A (pple) and B (anana) to explain that when the demand for apples rises, the demand for bananas falls, illustrating their relationship as substitutes. Fast forward thirty years, and A (I) and B (lockchain) are no longer substitutes; AI does not necessarily defeat blockchain, nor does blockchain necessarily defeat AI.

AI has a wide range of applications, while web3 has profound implications; the significance of applications is easy to understand, but the utility of significance is obscure, yet both are equally important. While enhancing reading efficiency and enjoyment through AI writing assistants powered by large language models, I will not forget my original intent: to properly preserve history through web3, elevate authors’ and the publishing industry’s income, strengthen reader participation, and reshape rigid consumer relations.

I have not shifted from web3 to AI; my “track” has always been grounded in humanities and utilized technology.

p.s. The preparations for the new location of the Yi Shou bookstore are in full swing. I moved some old books to the store and canceled the mini-storage. Although the contract period had ended, the termination form had multiple sections, all of which were mandatory; feeling annoyed, I carelessly filled in the reason as “no money,” just to get it done. A few days later, a young administrative girl, who had seen a pile of books in my storage, sent me an email, saying she saw my reason for cancellation and asked me “not to be upset and to keep going.” Reading the email, I felt guilty for unintentionally deceiving the young girl while also feeling very warm. Well, I will keep going.

Original source: The Understandable Use of AI, the Ambiguous Meaning of Web3