Whether the core of web3 lies in AI or blockchain is not up to KOLs

As ChatGPT sweeps the globe, web3 speculators, once infatuated, have become indifferent. Recently, their focus has shifted to the increasingly popular AI, treating blockchain with polite coldness and harshly mocking web3.

It doesn’t stop there. Some even claim that web3 is essentially an internet centered around AI, with blockchain being nothing more than a fleeting mirage, or perhaps only to be incorporated in web4. KOL Lao Gao, on the other hand, adopts a more inclusive stance, arguing that web3 encompasses both AI and blockchain. The video’s view count has nearly reached 3 million in just a few days, surpassing the total views of my articles over the past decade. I wonder if it’s because “listening to a wise person’s advice can be more enlightening than reading for ten years”.

So, what exactly is web3?

Web1, Web2, and Web3 are just narratives

Before sharing my views, it is important to clarify that, however you interpret it, web3 is not an industry standard. If we talk about standards, it is the HTTP protocol coordinated by the Internet Engineering Task Force (ITEF) and the World Wide Web Consortium (W3C). The HTTP protocol has silently supported our internet from HTTP/1 finalized in 1996, to HTTP/2 finalized in 2015, and HTTP/3 finalized in 2022. However, the tragedy of the commons tells us that such important matters are often overlooked (including by myself).

Furthermore, web3 is not a software version number. The iOS version is determined by Apple, the Windows version by Microsoft, and the Linux version is mostly determined by Linus Torvalds. But when it comes to web3, no one can give a definitive answer.

In short, the so-called web1, web2, and web3 are just narratives used to describe how the internet has transitioned from one paradigm to another over the decades. Sometimes these narratives point to the future, and sometimes they generalize the past.

In other words, whichever narrative can most vividly and accurately depict the characteristics of different stages will gradually be accepted by people and eventually become the “standard.” For example, sociologist Lui Tai-Lok in 2007 categorized Hong Kong people born between the 1920s and 1990s into four generations based on the impact of major historical events such as World War II and the Cultural Revolution in his book “Four Generations of Hong Kong People.” This depiction of Hong Kong society was widely accepted. But does that mean Hong Kong people can be objectively divided into four generations, or that dividing them by decade is wrong? Of course not.

web1, web2, web3: read, write, own

The web3 narrative I find most convincing is defined by read, write, and own.

In 1993, when I entered university, I got my first email address ([email protected]) and used the Mosaic browser to view the World Wide Web. At that time, the internet was like today’s blockchain, something almost everyone had heard of but few understood or used in their daily lives. Bandwidth was limited, making it slow, and applications were few and simple. Websites were static, and most were analogized to newspapers and magazines. The primary experience of using the internet was reading. Read was the characteristic of web1.

On March 10, 2000, after more than a decade of continuous growth, the Nasdaq reached a high of 5,000 points, which was followed by the dot-com crash. In the subsequent years, technology stocks continued to decline, and numerous internet companies went bankrupt. However, the crash only affected the stock market; internet-related technologies continued to flourish during these years. A series of new-generation services emerged, such as WordPress in 2003, Facebook and Gmail in 2004 (the “G” represented 1 GB of never-ending email storage that would continually ‘G’row), YouTube in 2005, Twitter in 2006, and my friend Greg Sung’s online bookshelf Anobii. The most significant feature of these services was user-generated content (UGC), which replaced static content produced by media companies. Therefore, the most significant characteristic of web2 was that, in addition to reading (read), there was also writing (write).

The iPhone debuted in 2007, driving the rise of mobile internet and apps, which now account for more than half of all internet activity. This development further spawned another batch of UGC-based applications, such as WhatsApp in 2009, Instagram in 2010, and Snapchat in 2011. Nowadays, the vast majority of applications on our computer and smartphone screens are web2 services. You don’t necessarily have to be a creator or photographer, but short videos, photos, chats, and emojis are all UGC; the possibility of not producing content is close to zero.

In 2013, Vitalik initiated Ethereum, and co-founder Gavin Wood first used the term Web 3.0 in 2014 to describe its vision. After the official launch of Ethereum and smart contracts, which gradually became usable, decentralized applications (DApps) emerged one after another. Notable examples include Opensea and Compound in 2017, Uniswap in 2018, and Curve in 2020. In comparison to web2, DApps have not improved in terms of ease of use, smoothness, and functionality; in fact, they have regressed. However, the real breakthrough of DApps lies in the fact that users can “own” their data and assets for the first time, instead of entrusting them to service providers. Therefore, web3 is called web3 because it adds a new element, “own,” on top of the foundation of “read” and “write”.

If you can’t migrate it, you don’t own it.

Even after reading the previous section, many people still have doubts about the concept of “owning.” It takes a lot of explanation to define what “owning” means, but it’s relatively simple to understand what “not owning” means: you want to move, but you can’t take it with you.

Everyone (but the government) knows, Hong Kong is currently experiencing the most severe wave of emigration since its founding. However, friends in Taiwan may not know that a few years ago, Hong Kong also experienced a Facebook emigration wave. A large number of users were dissatisfied with Facebook’s autocracy and closed-off nature and collectively migrated to competitor MeWe. I admire the spirit of trying to do something even when it’s deemed impossible, but I didn’t participate at the time, not because I wanted to cynically stand by, but because I couldn’t accept migrating from one authoritarian platform to another. The emigrants quickly returned, and the movement dissipated in an instant.

This is a result that anyone, not just experts, could predict. It’s a simple truth: if Hong Kong people can’t take their hard-earned money with them when they emigrate, it’s unlikely that there would be an emigration wave, unless it’s for survival. But if that situation arises, it would be a refugee crisis. Conversely, if users and fan pages could migrate their social relationships, posts, comments, and likes to MeWe, with the only differences being the interface, user experience, and platform operations, Facebook might have become a ghost town long ago.

If you think I’m exaggerating, consider how Sushiswap forked Uniswap in 2020 and attracted a large number of users and liquidity from it, or how, just last month, Blur’s NFT trading volume suddenly surpassed the seemingly dominant Opensea. While Sushiswap and Blur, as challengers, have many ingenious ideas, they are all built on the premise that users own their own data and assets. Otherwise, even the most powerful competitor would struggle to take market share from the market leader.

The “D” in DApps refers to the fact that users’ data and assets are independent of the operator’s website or app interface and are stored on the blockchain, accessible and usable by the user with a private key. This frees users from being trapped in closed gardens and eliminates the need to lose hard-earned account content due to the rise and fall of service providers, forcing them to start over and rebuild their lives.

This is the essence of implementing web3 — ownership.

Building over debating

Regarding web3, although I have a clear and strong understanding and actively promote the concept, I do not consider other opinions as “heresy.” Discourse is about good and bad, not right and wrong. Some discourses attempt to summarize historical struggles for interpretative power, while others try to deduce concepts to achieve practical implementation. Those that can encompass history and bear concepts are good discourses, while those that cannot will be forgotten.

Today, everyone “knows” that smartphones refer to iPhones and Android devices, forgetting that just a few years before Steve Jobs unveiled the iPhone in 2007, Nokia was already producing Symbian smartphones, which are now considered “stupid phones.”

The same goes for web3. The first web3 narrative was centered around blockchain, which is an objectively traceable fact. However, compared to the blockchain which aims to improve production relationships, the productivity-enhancing AI is easy to understand and quantifiable, with rapid development. If web3 is eventually defined as an AI-dominated internet, I wouldn’t be surprised. History is written by the winners, especially the dictatorial ones.

As practitioners, instead of debating who represents the true essence of web3, it’s better to focus on creating great products. After all, when improvements evolve from quantitative changes to qualitative ones, users will naturally see you as a new paradigm.

Extended reading

  1. 《小王子的領悟》番外篇.擁有
  2. 怎麼去擁有一道彩虹?與五月天聊 NFT
  3. 我在微信為甚麼打不開你發給我的WhatsApp?互聯網通訊協議之死(上)

My articles are not put behind a paywall and are all open for reading. If you like this article, please collect its Writing NFT to support writing and preserve journalism.






Leave a Reply

Your email address will not be published. Required fields are marked *